webdoc.sub.gwdg.dewebdoc.sub.gwdg.de/ebook/dissts/Bochum/Kroener2000.pdf · 2005-10-31 · Table of...

The caudolateral neostriatum of the avian forebrain

and its modulation by dopaminergic afferents

Inaugural-Dissertation

zur

Erlangung des Grades eines Doktors der Naturwissenschaften

in der

Fakultät für Psychologie

der

RUHR-UNIVERSITÄT BOCHUM

vorgelegt von:

Sven Kröner

März 2000

Gedruckt mit Genehmigung der Fakultät für Psychologie der Ruhr-Universität Bochum

Referent: Prof. Dr. Onur Güntürkün (Fakultät für Psychologie)Koreferent: PD Dr. Kurt Gottmann

Tag der mündlichen Prüfung: 29.05.2000

Table of contents

Introduction 1

Chapter 1 Afferent and efferent connections of the caudolateral

neostriatum in the pigeon (Columba livia): A retro- and

anterograde pathway tracing study.

23

Chapter 2 A polysensory pathway to the forebrain of the pigeon: the

ascending projections of the n. dorsolateralis posterior

thalami (DLP).

68

Chapter 3 Characterization of cell types in the chick neostriatum

caudolaterale: Electrophysiological and morphological

properties.

74

Chapter 4 The dopaminergic innervation of the avian telencephalon 102

Chapter 5 The dopaminergic innervation of the pigeon

telencephalon: Distribution of DARPP-32 and co-

occurrence with glutamate decarboxylase and tyrosine

hydroxylase.

152

Chapter 6 Effects of dopamine on the excitability of neurons in the

chick NCL.

175

General discussion 193

List of abbreviations 202

Literature 204

Acknowledgements 236

Curriculum vitae 237

Introduction 1

Introduction

One of the long-standing issues that memory research continues to confront is the dilemma of

localizing a complex function in an anatomically highly differentiated central nervous system.

Is memory a mental function that can be identified within an isolated neural structure or

specific groups of neurons? Traditionally, this question has been framed by psychological

inquiry. In the last century, clinical observations in patients with focal lesions begun to

establish a correspondence between specific symptoms and localized regions of injury, most

notably left hemisphere frontal sites with expression of speech (Broca, 1861) and left

hemisphere posterior temporal lobe sites with speech comprehension (Wernicke, 1874). These

and subsequent cases strengthened the localizationist view, and it was not unreasonable to

expect that other mental processes, like memory functions, could similarly be localized to

circumscribed areas of the cortex. The classical studies of Lashley (1950) demonstrated,

however, that the „engram“, as a single entity, does not exist, but that rather the behavioral

effects of brain lesions correlated with the size of the lesion and not their location, with

different areas making an essential contribution. Since the time of Lashley, it has become

increasingly clear that memory is divisible into different processes and takes heterogeneous

forms based on the type of memory (e.g., classical conditioning versus procedural), the

content of memory (e.g., episodic versus semantic), the temporal parameters of memory (short

term versus long term), and the level of processing (encoding or retrieval).

An important cognitive feature of „higher“1 organisms is their ability to temporarily structure

their behaviour and to actively „hold“ in mind information relevant for goal-attainment. Food-

hoarding birds for example, must remember a large number (up to 33.000) of caches in which

they hide their food (reviewed in Shettleworth, 1983). The information about the location of

the caches is certainly stored in long-term memory. In order to effectively retrieve the food,

however, the animal must actively remember both the places already visited (and update this

information with every new place visited), and must plan a route to the closest location that

contains food (c.f. Shettleworth, 1983).

The active retention of information in memory for the guidance of responses in forthcoming

situations is embedded in the concept of „working memory“ (Baddeley, 1992). Working

1 The term„higher“ is not used in the sense of „further developed“ with regard to a scala naturae, but to describethe complex abilities of the cognitive system. Therefore, the processes underlying „higher“ cognitive functionsshow some degree of abstraction and may be engaged in various behaviours. This might be reflected for examplein the ability of an organism to rapidly adapt to new situations.

Introduction 2

memory encompasses both storage and processing functions, and within cognitive psychology

it represents an extension of an earlier concept, short-term memory, a limited-capacity

temporary memory store, typified by the model proposed by Atkinson and Shiffrin (1968).

Among different species working memory probably takes different forms, and its functions

may range from the „simple“ active retention of goal-related information, to an import role in

(speech)-comprehension, thinking, and planning in primates and humans (e.g. Baddeley,

1992; Kimberg and Farah, 1993; Goldman-Rakic, 1996; Cohen et al., 1997). In order to

understand this sentence, for example, a reader must maintain the first half of the sentence in

working memory while reading the second half.

In mammals, working memory is closely related to the functions of the prefrontal cortex

(PFC) and its innervation by the mesocortical dopaminergic system (see below), and in recent

years considerable progress has been made in identifying the cellular elements that participate

in working memory in mammals (e.g. Goldman-Rakic, 1996 for review). Important aspects in

this quest relate to the specifity of the input-output relationships (i.e. the functional

heterogeneity of brain areas, and the laminar and synaptic organizations of these connections),

the distribution of receptors, and the physiological properties of the neurons involved.

In birds, a region in the posterior forebrain, called neostriatum caudolaterale (NCL), has

recently been shown to be crucially involved in working memory (Mogensen and Divac, 1982,

1993; Gagliardo et al., 1996, 1997; Güntürkün, 1997; Kalt et al., 1999), and other tasks which

in mammals target „frontal“ executive functions (see below), e.g. reversal learning (Hartmann

and Güntürkün, 1998), or response inhibition (Güntürkün, 1997; Aldavert et al., 1999). In

addition, the observation that the NCL shows a high dopamine (DA) content (Divac et al.,

1985) and a strong innervation by dopaminergic fibers (Waldmann and Güntürkün, 1993;

Divac et al. 1994), led to the notion that the NCL might be the avian equivalent of mammalian

PFC. In contrast to mammals, however, virtually nothing is known about the anatomical and

physiological properties of the circuitry that might subserve higher cognitive functions in non-

mammalian species.

The organization of the forebrain of birds and other amniotes (i.e. reptiles) is radically

different from that of mammals (see chapter 1 and 4), reflecting at least 250 million years of

divergent evolution. A comparison between birds and mammals may thus shed light on the

structural and physiological requirements of a proposed canonical circuit involved in working

memory, and more specifically, how the avian brain manages to solve these tasks in the

absence of the layered isocortex of mammals.

Introduction 3

In the studies presented here, I will describe some aspects of the anatomical and physiological

organization of the NCL and its modulation by dopaminergic afferents, to gain insights into

the neuronal circuitry that may underly working memory and other higher cognitive functions

in birds. It should be stressed, however, that within this view no structural homology between

the NCL and the PFC is assumed, but that rather a functional analogy between these brain

areas may exist with regard to some functions, but not necessarily others as well. The term

„homology“ not only implies similar functions of an organ in different species, but also

stipulates a common ancestor (Campbell and Hodos, 1970). Functional „analogies“ in contrast

may reflect parallel evolution, i.e. similar adaptive trends and corresponding selective

pressures (Rehkämper and Zilles, 1991).

For comparison with the results obtained here in birds, the following sections give a brief

overview over some of the functions of the PFC and its interaction with the mesocortical

dopaminergic system. It is sought to describe the basic neuronal circuitry that underlies these

functions (as far as it is known), and specifically its relationship to the layered structure of

mammalian neocortex.

The prefrontal cortex - connectivity and areal organization

The prefrontal cortex occupies the rostral pole of the isocortex of mammals. In various

mammalian species the PFC extends over increasingly larger areas of brain (e.g. primates >

dogs > cats > rats), and its relative growth is larger than that of the other associative cortices,

reflecting a phylogenetic development (Fuster, 1989).

The most commonly used hodological criterion to define the PFC in various species is the

projection of the thalamic nuc. mediodorsalis (MD) (Rose and Woolsey, 1948; Divac et al.,

1978; Goldman-Rakic and Porrino, 1985; Groenewegen et al., 1990). In primates this

projection is topographically organized and allows a rough anatomical division of the PFC

(figure 1). The dorsolateral PFC (Walkers areas 8, 9, 10, 45, 46; Walker, 1940) receives

afferents from the lateral part of the MD (pars parvocellularis), and the

orbitofrontal/ventromedial PFC (Walkers areas 11, 12, 13) is innervated by the medial part of

the MD (pars magnocellularis) (Goldman-Rakic and Porrino, 1985; Groenewegen, 1988;

Groenewegen et al., 1990; Barbas et al., 1991; see Fuster, 1989, for review). However, the

definition of the PFC by its afferents from MD remains a matter of debate, as the MD projects

Introduction 4

also to brain areas beyond the PFC, and on

the other hand the PFC receives substantial

innervation from other thalamic nuclei, e.g.

the anterior nuclei and the pulvinar

(Markowitsch and Pritzel, 1979; Goldman-

Rakic and Porrino, 1985; Groenewegen et

al., 1990; Barbas et al., 1991). Based on

criteria defined by Campbell and Hodos

(1970), Kolb (1984) proposed, that in

rodents the prelimbic and infralimbic

regions of the PFC might be functionally

analogous to the dorsolateral PFC in

primates (but see Preuss, 1995; Williams

and Goldman-Rakic, 1998).

In addition to thalamically relayed input,

direct subcortical and limbic afferents

reach the PFC from the brainstem

tegmentum, the pons, the hypothalamus,

and the amygdala (for reviews see Fuster, 1989; Groenewegen et al., 1990). In various

mammalian species the PFC has furthermore been shown to receive afferent fibers from the

hippocampal complex, the cingulate cortex, and from adjacent areas of the limbic cortex

(Goldman-Rakic et al., 1984; Reep, 1984; Pandya and Yeterian, 1990; Conde et al., 1995). In

the context of working memory, the connections of the PFC with the MD, the hippocampus,

the tegmentum, and the nuc. accumbens (see below) appear to be of particular importance, as

they form a unitary functional network, and lesions in theses regions result in impairments

similar to those observed after lesions of the PFC (e.g. Sakurai and Sugimoto, 1985; Floresco

et al., 1997; Seamans et al., 1998; see below).

Sensory input from other cortical areas reaches the PFC via a set of interconnected pathways

(for reviews: Jones and Powell, 1970; Pandya and Yeterian, 1990). The primary sensory area

of each modality projects first to an adjacent area in parietal, occipital, or temporal cortex

(figure 2). This is the beginning of a sequential order of cortical areas that make up a pathway

for that modality. Each area in the sequence projects not only to the next in line but also to a

discrete area of the frontal cortex, which in turn reciprocates by sending fibers back to the

Sulcus arcuatus

Sulcus principalis

Superior prefrontal convexity

Inferior prefrontal convexity

Orbitofrontalcortex

Inferior prefrontal convexity

B

A

Figure 1: Simplified scheme of the basic subdivisions of theprimate prefrontal cortex (PFC). A: lateral view. B: ventral view.Areas in the superior prefrontal convexity and the sulcusprincipalis are commonly designated as dorsolateral PFC. Theorbitofrontal cortex and areas in the inferior prefrontal convexityconstitute the ventromdedial PFC. The orbitofrontal cortex ispart of the limbic association cortex (Data from Rosenkilde,1979).

Introduction 5

projecting area. The fields that constitute the third link in each of the three pathways - namely,

parietal area 7 (somatic), temporal area 22 (auditory), and inferotemporal area 21 (visual) -

project to the prefrontal cortex and, in addition, to the cortex in the dephts of the superior

temporal sulcus, which constitutes another multimodal area. Although the termination fields

show considerable overlap, especially around the principal sulcus of primates (Area 46), a

strong topography of these afferents exists, resulting in a distinct pattern of connections for

each of the subdivisions of PFC (primates: Jones and Powell, 1970; Selemon and Goldman-

Rakic, 1988; Pandya and Yeterian, 1990; Romanski et al., 1999; rodents: Condé et al., 1995;

Reep et al., 1996).

The (reciprocal) connections of the PFC with the posterior parietal lobe2 can be considered a

prototypical example for the organization of cortico-cortical associational fibers. Both, the

sources of the afferents and their areas of termination are organized in modules. The largest

2 In the parasensory posterior cortex visual and somesthetic information are integrated for the representation ofthe spatial location of an object. The connection between area 7 in the parietal lobe and the PFC is part of anetwork responsible for voluntary movements, particularly eye-movements (Selemon and Goldman-Rakic, 1988),and the so-called „dorsal visual-stream“ conveys information about the „where“ of an object. In contrast, theventral visual stream originating in the temporal lobe, relays information necessary for object-recognition, the„what“ of an object (Wilson et al., 1993; Rao et al., 1997; see Romanski et al., 1999, for auditory processing) .

46

46

8

6

13

AA3

V1

V2AA1

VA1

VA2

VA3

V3

SA3

A1

SA1SA2

V4

AA2

86

S1

12

467b

7a

45

Figure 2: Simplified scheme of the neocortical sensory and associational afferents to the prefrontal cortex in primates. Theparasensory association areas send both hierachially organized indirect, and parallel projections to the PFC. In addition to theprojections depicted along the lateral surface of the brain, afferents from (para-)sensory areas reach the PFC via the cingulateand enthorinal cortex (Data from Jones and Powell, 1970; Pandya and Yeterian, 1990). Abbreviations: S, somatosensory; SA,somato-parasensory association areas; A, auditory; AA, auditory-parasensory association areas; V, visual; VA, visual-parasensory areas.

Introduction 6

number of cortico-cortical projection neurons can be found in layers III, V, and VI (Schwartz

and Goldman-Rakic, 1984; Selemon and Goldman-Rakic, 1988). The afferents from area 7 in

the parietal lobe terminate in a feedforward pattern mainly in layers I, IV and VI of the PFC,

while the feedback from the PFC terminates in layers I and VI, but spares layer IV (Selemon

and Goldman-Rakic, 1988). Within the gaps of the ipsilateral connections terminate callosal

fibers from the contralateral PFC or parietal cortex. These callosal connections originate from

homotopic functional modules and their termination areas show large overlap in the

contralateral hemisphere (Schwartz and Goldman-Rakic, 1984; Selemon and Goldman-Rakic,

1988; Berendse et al., 1992).

The overall pattern of cortical afferents and their coarse topography to the PFC can be

summarized such, that the less differentiated cortices, particularly structures of the limbic

system, innervate the ventromedial (orbitofrontal) PFC, while the „cognitive“, i.e. sensory

aspects of a stimulus are processed in the dorsolateral PFC (c.f. Fuster, 1989).

As already indicated, all of the above mentioned connections of the PFC are reciprocal, and

the „feedback“ from the PFC largely retains the topography of the original input (Selemon and

Goldman-Rakic, 1988; Sesack et al., 1989; Pandya and Yeterian, 1990). Of special

importance are furthermore the reciprocal connections of the PFC with areas in the premotor,

supplementary motor, and motor cortex (Brodmann´s areas 6 and 8; Brodmann, 1909), as via

these efferents the PFC initiates and controls the execution of movements. An exception from

the general rule of reciprocity are the projections of the PFC to the basal ganglia. The inputs

from the PFC to the dorsal- (caudate-putamen) and ventral striatum (nuc. accumbens) are

relayed back only indirectly via the globus pallidus, nuclei in the thalamus, and the substantia

nigra, respectively (Alexander et al., 1990; Berendse et al., 1992; Goldman-Rakic, 1995;

Parent and Hazrati, 1995). Depending on their precise location of termination, the striatal

efferents of the PFC may participate in motoric, associative and limbic aspects of behaviour

(Parent, 1990; Goldman-Rakic, 1995; Parent and Hazrati, 1995).

Microarchitecture of the prefrontal cortex

In primates, the clearest cytoarchitectonic feature of the PFC is the distinct granular layer IV,

that distinguishes the PFC from the remaining frontal cortex; and therefore the PFC has also

been dubbed „frontal granular cortex“ (Walker, 1940). In rodents, however, the granular layer

is not as clearly developed. Afferent fibers terminate in all six layers of the PFC, but the

projections of most brain regions show a distinct laminar pattern within the PFC (c.f. figure 4;

Introduction 7

and the example of area 7 above). Synaptic inputs from many other association cortical

regions converge in layer I-II, where the apical tuft of the layer III, and V-VI pyramidal

neurons extends (Sesack et al., 1989; Berendse et al., 1992; Condé et al. 1995). Inputs to the

soma or proximal/basal dendrites of pyramidal cells in layers III and V-VI, respectively, are

known to arise from neighboring reciprocally connected cells within the same cortical region

(Levitt et al. 1993; Kritzer and Goldman-Rakic, 1995; Pucak et al. 1996; Melchitzky et al.,

1998; Gonzalez-Burgos et al., 2000). This reciprocal interaction may give rise to a

reverberative ensemble of local neurons. Such local activity, confined within the PFC, may

underlie working memory processes, because the PFC must rely on sustained firing to „hold“

information in the absence of continuous presence of previously presented sensory cues (Amit

1995; Goldman-Rakic, 1996; Durstewitz et al., 1999). In analogy to the columns of primary

sensory cortex, clusters of neighbouring cells in the PFC are assumed to form functional

modules (Pucak et al. 1996; Melchitzky et al., 1998; Gonzalez-Burgos et al., 2000), which

might share the memory for a certain location (Rao et al., 1999), or display similar object

selectivity (Rainer et al., 1998b). Local interneurons that contain γ-aminobutyric-acid (GABA)

may play a pivotal role in the fine-tuning of reverberating activity within a cortical module

(Rao et al., 1999), and/or provide lateral inhibition between modules encoding different

features, e.g. opposite spatial loctions (Wilson et al., 1994; Rao et al., 1999). GABAergic

interneurons also appear to be a major target of the dopaminergic innervation (see below).

Behavioural neurophysiology of the prefrontal cortex

Lesion studies

Animals that have been subjected to extensive ablation of their PFC show severe impairments

in the planning and organization of complex behavioural sequences (Kimberg and Farah,

1993; Mogensen and Holm, 1994). The results from numerous lesion studies in mammals

indicate that the disturbance of 3 (interacting) basic functions of the PFC contribute to the

impairments observed in these „frontal animals“ (reviewed in Fuster, 1989; Dias et al., 1997).

1 Focussing of attention: The animal with prefrontal lesions has a fundamental difficulty in

supressing its attention to irrelevant stimuli and maintaining its attention on relevant

stimuli. This distractability is also responsible for hyperreaction to external stimuli, which

in turn accounts for a general motor hyperactivity (Fuster, 1989, for review).

2 Supression of competing tendencies: A loss of inhibition of false response tendencies can

be seen in tasks in which the preoperative incentive values of the discriminanda are

Introduction 8

reversed („reversal-learning“), or in tasks that demand the (successive) discrimination of

two stimuli of which one requires a given response, and the other no response at all

(Go/No-Go tasks). Especially lesions of the ventromedial PFC impair the animals ability to

supress interference from competing tendencies (Brutkowski, 1964; Sakurai and Sugimoto,

1985; Dias et al., 1997), typically leading to perseveration and a loss of response control

(Iversen and Mishkin, 1970).

3 Working memory: As outlined above, working memory describes the process of active

retention and/or manipulation of items in short-term memory, required for the guidance of

responses in forthcoming situations. Experimentally, impairments in working memory

functions can be assessed by variants of the „delayed-response“ or „delayed-alternation“

tasks, which were established in memory research by Jacobsen in the mid-thirties (e.g.

Jacobsen, 1935). Permanent or reversible lesions of the PFC (by cooling, electrical

stimulation or pharmacological blockade), result in severe impairments in the acqusition

and the performance of delay-tasks. The impairments become particulary obvious after

lesions of the dorsolateral PFC (around the sulcus principalis), or its homologue area in the

rat (Jacobsen, 1935; Sawaguchi and Goldman-Rakic, 1991; Kesner et al., 1996; Fuster,

1989, for review).

In addition to these impairments regarding attention and cognition, changes in the affective

and social behaviour are often observed in frontal animals and human patients (see below).

Cellular correlates of working memory

The polysensory character of the PFC indicated from the pattern of converging afferents, is

also reflected in the response characteristics of single neurons. Neurons in the PFC are able to

encode the behavioural significance of a stimulus across various modalities (Yajeya et al.,

1988; Watanabe, 1992). Subsets of prefrontal neurons in the area of the principal sulcus (i) are

activated phasically in the presence of a visual stimulus, (ii) are activated tonically during the

delay period over which the stimulus is kept on line, or (iii) show phasic reactivation in

relation to the initiation of a memory-guided response. Many prefrontal neurons respond in

more than one phase of the trial, i.e., during the cue, delay, and/or response periods, that

means their activity is related to the subfunctions of registration, memory, and motor control,

respectively (Fuster and Alexander, 1971; Fuster, 1973; Funahashi et al., 1989, 1993; Di

Pellegrino and Wise, 1991; Wilson et al., 1993; Funahashi and Kubota, 1994; Rao et al.,

1997, 1999; Asaad et al., 1998; Rainer et al., 1998a,b; Sawaguchi and Yamane, 1999; see

Introduction 9

Fuster, 1989, and Goldman-Rakic, 1996 for reviews). The sustained delay activity of neurons

in the mammalian PFC encodes information relevant to the current behavioural goal (Yajeya

et al., 1988; Funahshi et al., 1989; Di Pellegrino and Wise, 1993; Rao et al., 1997; Asaad et

al., 1998; Rainer et al., 1998a,b; Sawaguchi and Yamane, 1999). This delay activity has been

shown to be content specfic with individual neurons encoding e.g. the location of an object in

space (Funahashi et al., 1989; Fuster, 1973, Rao et al., 1997); the direction of a prior response

(Funahashi et al., 1993), or the identity of an object such as a face or a fruit (Quintana et al.,

1988; Wilson et al., 1993; Rao et al., 1997; Rainer et al., 1998b). In addition, the activity of

prefrontal neurons is often polarized, exhibiting excitatory responses to targets in the preferred

direction and inhibitory responses to targets at nonpreferred directions (Funahashi et al., 1989;

Rainer et al., 1998a; Sawaguchi and Yamane, 1999; Rao et al., 1999). Both a disruption of

delay activity as well as delay activity coding for the wrong stimulus or response are

correlated with subsequent behavioural errors (Fuster, 1973; Quintana et al., 1988; Funahashi

et al., 1989). In contrast to neurons in the temporal and parietal cortices that also show delay-

activity during a working memory task (Miller et al., 1996; Constantinidis and Steinmetz,

1996), networks in the PFC are able to sustain delay activity even in the pressence of

interfering inputs (Miller et al., 1996). It has been hypothesized that this unique feature of the

PFC is due to a „protective“ action of DA (Durstewitz et al., 1999).

Data from functional imaging studies in humans have spawned the discussion about the

functional segregation of the PFC. While all of these studies have demonstrated the

participation of the PFC in different task-specific working memory networks (e.g. Cohen et

al., 1997; Belger et al., 1998; D´Esposito et al., 1999; Haxby et al., 2000), it remains a matter

of debate, whether differential activation of dorsal and ventral areas within the PFC, reflects

content-specific or process-specific domains. One view proposes that the dorsal-ventral

subdivisions of the PFC might be concerned with spatial and object working memory,

respectively. Alternatively it has been suggested that the dorsal PFC is related to the

manipulation of objects in memory, while the ventral PFC merely subserves the active

retention of this information (Belger et al., 1998; D´Esposito et al., 1999; but see Haxby et al.,

2000).

Pathophysiology of the prefrontal cortex in humans

The clinical symptoms following lesions of the PFC in man, most likely also relate to the

basic functions of the PFC reviewed for animals above. In humans the deficits following

Introduction 10

frontal lesions are categorized with respect to symptoms pertaining to perception and motility,

cognition, and affect and emotion (Fuster, 1989).

Perception and Motility. The perceptive abilities of frontal patients are often impaired in as

much as they show increased distractability, or, quite contrary, sensory neglect as a specific

form of attention deficit. Other attentional or „sensory“ deficits regard to spatial orientation

and gaze control (in Fuster, 1989). The motoric deficits of frontal patients depend on the focus

of the lesion. Dorsolateral lesions greatly reduce spontaneous motor activity, leading to

hypokinesis which ranges from a certain „aspontaneity“ to severe mutism (Stuss and Benson,

1986). In contrast, lesions of the orbitofrontal PFC often result in hyperkinesis, and excessive

and aimless motility (in Fuster, 1989).

Cognitive Symptoms. The common cognitive motif following lesions of the PFC are deficits in

the temporal integration of stimuli and actions (Fuster, 1989). Lesions of the dorsolateral PFC

result in disturbances of short-term memory, and the prospective planning of sequences

(Milner and Petrides, 1984; Robbins, 1996). In patients, the ability to sequentially organize a

task is often tested with the „Tower of London“ (Robbins, 1996), while short-term memory

impairments can also be probed with variants of delay-tasks (Milner and Petrides, 1984;

Fuster, 1989). Lesions in the orbitofrontal PFC, however, result in the inability to suppress

interfering tendencies and result in perserverative behaviour (Fuster, 1989; c.f. also Dias et al.,

1997), which can be assessed with the „Wisconsin-Card-Sorting-Test“ (Robbins, 1996).

Affect and Emotion. Symptoms of disturbed affect and emotion can coarsly be divided into

two complexes. The so-called „apathetic syndrome“ is characterized by a low awareness, lack

of initiative and hypokinesis. Most specific, however, is the generalized blunting of affects

and emotional responses, therefore it has also been termed „pseudodepression“ (Stuss and

Benson, 1986). The counterpart of this apathy is the „euphoric syndrome“ which results from

orbitofrontal lesions, and which along with euphoria is characterized by distractibility,

hypermotility, and a disinhibited hunger and sexual drive (Fuster, 1989).

In addition to deficits resulting from lesions of the PFC, multiple lines of evidence suggest

that certain PFC regions may be a locus of dysfunction or structural pathology in such mental

illnesses as schizophrenia (reviewed in Yang et al., 1999), or obsessive-compulsive disorder

(reviewed in Fuster, 1989). The PFC also appears to be markedly affected in Alzheimer’s

disease (Hof et al., 1990; Terry et al., 1991). It has been reported for example that one major

correlate of cognitive deficiency in alzheimers disease is the synaptic loss in the PFC,

Introduction 11

contributing about 70% of the strength of the correlation with psychometric scores (Terry et

al., 1991).

The mesocortical dopamine system

The dopaminergic innervation of the

forebrain of mammals and birds is

constituted by a small number

(!30.000 - 40.000 neurons on each

side of the rat brain) of highly

collateralized neurons residing in the

ventral mesencephalon (e.g.

Björklund and Lindvall 1984; Fallon

and Loughlin, 1995; Williams and

Goldman-Rakic, 1998; c.f. chapter 4

for birds). These mesencephalic cell

groups are designated as A8, A9, and

A10, according to the nomenclature

of Dahlström and Fuxe (1964), and

they generally correspond to the DA

cells of the substantia nigra (SN, A9),

ventral tegmental area (VTA, A10),

and the retrorubral area (A8)3 (Porrino

and Goldman-Rakic 1982; Berger et

al., 1991; Williams and Goldman-

Rakic, 1998).

Functionally, different mesencephalic dopaminergic projections exist which have been

designated as mesostriatal-, mesolimbic-, or mesocortical DA system, respectively, based on

their main targets of innervation (figure 3; Björklund and Lindvall, 1984; Cooper et al., 1996).

The mesolimbic DA system projects to various limbic areas, including the amygdaloid

complex, hippocampus, septum, nucleus accumbens, and entorhinal cortex . The mesostriatal

DA system provides strong innervation to the striatal aspects of the basal ganglia (i.e. caudate-

3 Due to the similarity of their projection, DA cells in the retrorubral field are often grouped with the substantianigra, pars compacta (Björklund and Lindvall, 1984; Reiner et al., 1994)

Figure 3: Organization of the mesencephalic dopamine system inmammals (brain of a rat). A: The mesocortical- and mesolimbic DAsystem. B: The mesostriatal DA system. In primates the mesocorticaldopaminergic innervation is much expanded and extends over largeareas of the posterior cortices. See text for details.

A

locuscoeruleus

lat. nucleusparabrachialis

entorhinalcortex

amygdala

piriformcortex

ant. cingulatecortex

prefrontalcortex

lateralseptum

hippocampus

habenula

A10

A8A9

Bcaudate-putamen

bed nuc.-stria terminalis

tuberculumolfactorium

nuc. accumbens

perirhinalcortex

A10

A9A8

Introduction 12

putamen), and to a lesser extent to the globus pallidus and tuberculum olfactorium (Björklund

and Lindvall, 1984; Cooper et al., 1996).

The organization and extent of the cortical innervation by the mesocortical DA system varies

with the species studied (reviewed in Berger et al., 1991). In primates, tyrosine hydroxylase-

or dopamine-immunoreactive fibers, originating in a continuum of cells in A8-A10, are found

in the motor, premotor, supplementary motor area, parietal, temporal, and posterior cingulate

cortices (sensorimotor), in addition to prefrontal, anterior cingulate, insular, piriform,

perirhinal, and entorhinal cortices (association). This innervation shows a strong rostro-caudal

gradient (also seen in rodents), the innervation being densest in the frontal cortex, particularly

in areas 4 and 6 (Björklund and Lindvall, 1984; Berger et al., 1988, 1991; Goldman-Rakic et

al., 1992; Williams and Goldman-Rakic, 1993). The rodent mesocortical DA innervation is

mainly confined to the limbic cortices, including the prefrontal, anterior cingulate, insular,

piriform, perirhinal, and entorhinal cortices (Björklund and Lindvall 1984; Berger et al. 1991).

Input to the rodent PFC derives from largely separate populations of dopaminergic neurons

located in the SN and VTA, respectively (Swanson 1982; Björklund and Lindvall, 1984;

Fallon and Laughlin, 1995).

The laminar organization of the mesocortical DA innervation generally shows a bilaminar

pattern, but again with distinct species-differences between rodents and primates (c.f. figure

3). In the rodent PFC, afferents from the medial VTA provide dense DA input to the deep

layers V-VI. The considerably weaker DA innervation to the superficial layer I-III originates

in the medial SN and the lateral VTA (van Eden et al., 1987; Berger et al., 1991).

Furthermore, on the ground of developmental, morphological, and pharmacological

differences (reviewed in Berger et al., 1991), these distinct inputs have been divided into two

classes, with class 1 DA neurons innervating the deep cortical layers. There is also evidence

that class 1 and 2 neurons show functional diverences, as class 1 neurons are more susceptible

to the effects of stress (Deutch et al., 1991). It has been hypothesized that the two modes of

dopaminergic innervation in the avian forebrain might relate to these two classes of DA

neurons in rodents (Wynne and Güntürkün, 1995; c.f. chapter 4).

In the primate granular cortices, the superficial layers I-III are generally innervated even

denser than layers V-VI, whereas in the motor and anterior cingulate (agranular) cortices all

layers receive a similarly dense dopaminergic input (Berger et al., 1988, 1991; Berger and

Gaspar, 1994; Goldman-Rakic et al., 1992; Lewis et al., 1992; Williams and Goldman-Rakic,

1993; Krimer et al., 1997). In the PFC of rats and primates dopaminergic fibers make about 40

Introduction 13

- 90 % specialized synaptic contacts (!39 % in the sulcus principalis of primates, Smiley and

Goldman-Rakic, 1993; !56 % in the suprarhinal, and !93 % in the anteromedial PFC of rats,

Séguéla et al., 1988; but see Smiley and Goldman-Rakic, 1993 for technical considerations),

which are mostly of the symmetric type. The remaining axonal varicosities constitute non-

synaptic release sites and probably contribute to the actions of DA via volume transmission

(Smiley and Goldman-Rakic, 1993). The most common synaptic target of the DA terminals in

the PFC of rodents and primates appear to be the dendritic spines and shafts of pyramidal

neurons (van Eden et al., 1987; Goldman-Rakic et al., 1989; Smiley and Goldman-Rakic,

1993). Many of the postsynaptic spines innervated by DA terminals also receive unlabeled

asymmetric (putative excitatory) terminals. Through this „triadic” synaptic arrangement DA

might both, pre- and postsynaptically, modulate the excitatory afferents to pyramidal cells

(van Eden et al. 1987; Séguéla et al. 1988; Goldman-Rakic et al. 1989; Smiley and Goldman-

Rakic 1993; c.f. chapter 4 for the situation in birds). However, dopaminergic fibers in the PFC

have also been observed to terminate on smooth GABAergic, stellate cells (Goldman-Rakic et

al., 1989; Benes, et al., 1993; Smiley and Goldman-Rakic, 1993). In addition, the localization

of DA receptors in interneurons and the physiological actions of DA in the PFC strongly

implicate a dopaminergic modulation of GABAergic interneurons (see below).

Dopamine receptors in the PFC

Dopamine receptors belong to the superfamily of seven transmembrane spanning G protein-

coupled receptors. In recent years, molecular cloning studies have led to a revision of the

traditional classification of dopamine receptors into D1 and D2 subtypes simply based on their

effects on the adenylyl-cyclase system (Kebabian and Calne, 1979). So far at least 5 different

DA receptors (D1 - D5) have been identified. The pharmacological properties in response to

„classical“ D1 and D2 receptor ligands, respectively, distinguish a D1-like family (D1 and

D5), and a D2-like family (D2, D3, D4) of DA receptors (Seeman and van Tol, 1993; Cooper

et al., 1996). In accordance with the classical view, receptors of the D1 family are coupled

positively to the adenylyl-cyclase, while D2 receptors inhibit adenylyl-cyclase activity

(Kebabian and Calne, 1979; Cooper et al., 1996; Robinson and Caron, 1997).

Insight into the distinct pharmacological profiles and the differential distribution of the DA

receptor subtypes are of high clinical relevance for the development of new drug treatments.

For example, the therapeutic effects of neuroleptics are mediated via receptors of the D2

family. The blockade of D2 receptors by „typical“ neuroleptics such as haloperidol, however,

Introduction 14

often results in side-effects on the extrapyramidal motoric system, leading to tardive

dyskinesia or parkinson-like symptoms. This is related to the fact that D2 receptors are most

abundant in the neostriatum. „Atypical“ antipsychotics such as clozapine, on the other hand,

have only a low affinity - only about 10 % of the classical neuroleptics - for the D2 receptor,

but a much higher affinity for D3 or D4 receptors, which are highly abundant in the nuc.

accumbens and the PFC. The latter structures and their dopaminergic innervation appear to be

critically involved in both the positive and negative symptoms of schizophrenia (Seeman and

van Tol, 1993).

The best studied effector protein modulated following the activation of DA receptors in

neurons is DARPP-32 (c.f. chapter 5). DARPP-32 is a Dopamine- and cAMP-Regulated

PhosphoProtein of molecular weigth 32,000 that serves as a „third messenger" in the

intracellular cascade that follows D1-receptor stimulation (Hemmings et al., 1987a, 1987b,

1995; Fienberg et al., 1998). Via activation of the adenylyl cyclase, D1-receptor action

stimulates cAMP synthesis, which through protein kinase A (PKA) leads to phosphorylation

of DARPP-32. Phosphorylated DARPP-32 in turn inhibits protein phosphatase-1. In striatal

neurons protein phosphatase-1 controls the state of phosphorylation of a wide array of

phosphproteins, including neurotransmitter receptors (e.g. for α-amino-3-hydroxy-5-

methylisoxazole-4-propionic acid [AMPA] and N-methyl-D-aspartate [NMDA]), ion channels

(e.g. Ca++ channels and voltage dependent Na+ channels), ion pumps (e.g. the Na+-K+-

ATPase), and transcription factors (Hemmings et al., 1995; Nishi et al., 1997; Fienberg et al.,

1998). Conversely, DA, acting on D2-like receptors, through both inhibition of PKA and

activation of calcium/calmodulin-dependent protein phosphatase (protein phosphatase

2B/calcineurin), may cause the dephosphorylation of DARPP-32 (Nishi et al., 1997), enabling

a possible interplay of D1 and D2-like receptors.

The laminar distribution of DA receptors in the frontal cortex largely parallels the bilaminar

pattern of the innervation by dopaminergic fibers. In rodent frontal cortex there is

considerable expression of mRNA for the D1 receptors in deep layers V-VI. A comparatively

lower expression of D2 receptor mRNA is distributed in superficial layers I-III and in deep

layer V (Gaspar et al. 1995; see Yang et al., 1999 for review). In general, in the frontal

cortices of rats and primates, the density of D1 receptors has been estimated to be about five-

to tenfold higher than that of D2 receptors (Lidow et al., 1991; Joyce et al., 1993). In rodent

frontal cortex there is also evidence for a considerable number of D4 receptors, but not for D3

and D5 receptors (Joyce et al., 1993; Ariano et al. 1997, see Yang et al., 1999 for review). In

Introduction 15

primate PFC expression of mRNA for all 5 DA receptor subtypes has been found (Lidow et

al., 1998). The largest numbers of receptor proteins were found for the D1, D4 and D5

subtypes in pyramidal neurons of superficial layers II-III and deep layers V-VI, with layer V

neurons showing a stronger expression of D4 and D5 receptors (Lidow et al. 1991, 1998).

Immunoreactivity and mRNA expression for D1, D2 and D4 receptors has also been found in

GABAergic interneurons of rat and primate frontal cortex (Mrzljak et al., 1996; Ariano et al.,

1997; LeMoin and Gaspar, 1998; Muly et al., 1998).

contralateral PFCipsilateral PFC

entorhinal cortexhippocampus

amygdalamidline thalamus

intracortical associational

hippocampusintracortical associational

MDhippocampus

midline thalamus


entorhinal cortex

contralateral PFC

hippocampus

amygdala


contralateral PFC

hippocampusmidline thalamus

ventromedial thalamus


rodents primates

SN (A 9)

VTA (A 10)

1

2

4

5/6

3

Layers AfferentsDopamine

column 1 column 2

dendrite

spine

sensoryafferent

D1D2?

Glu

D1/D2?

DAGlu

DAvaricosity

Figure 4: A simplified scheme of the functional architecture of the prefrontal cortex and the innervation from themesocortical dopamine system. On the left, the laminar distribution of known afferents to the PFC, and the pattern ofdopaminergic fibers in rodents and primates is shown. On the right, the neuronal elements of a proposed canonical circuitare shown. Neurons in the PFC (shown are pyramidal cells - green - in layers III and V/VI and interneurons - blue) appearto be arranged in columns, or „stripes“, which are believed to belong to distinct functional modules. The axon collaterals ofnon-pyramidal interneurons (yellow) can either provide „fine tuning“ between the pyramidal neurons of a functional module(not shown), or may exert strong lateral inhibition onto neurons of a distinct functional column/module. Reverberatingactivity between neurons of a certain lamina and/or a functional module could provide a means for the „holding on-line“ ofa certain information. Insert: Dopamine released from varicosities of dopaminergic fibers may act directly at receptorswithin a synapse of the spine, or may diffuse over longer distance to activate extrasynaptic receptors, e.g. at other sites of thedendrite, or on afferents presynaptically to the neuron (forming a synaptic triad). See text for details.

Introduction 16

Physiology of the mesocortical dopaminergic system

Midbrain dopaminergic neurons exhibit a burst of action potentials immediately following

unexpected salient events, which include sudden novel stimuli (Schultz and Romo, 1990;

Ljungberg et al., 1992), and primary rewards (Ljungberg et al., 1992; Schultz et al., 1995).

During the learning-phase of a task, reward-related activation of DA neurons first occurs in

response to the (unexpected) primary reward (the UCS). If the UCS becomes predictable,

activation of DA neurons „shifts“ to the occurence of the conditioned stimulus (CS)

(Ljungberg et al., 1992), and finally to the occurence of a cue that signals the CS (Schultz et

al., 1993; Schultz, 1998 for review). Schultz and coworkers (Schultz et al., 1995; Schultz,

1998; c.f. also Montague et al., 1996) have argued that the activity of dopaminergic neurons

provides an essential signal for reinforcment learning. This model is based on the observation

that dopamine neurons report rewards relative to their prediction rather than signaling primary

rewards unconditionally. The dopamine response is positive (activation) when primary

rewards occur without being predicted. The response is nil when rewards occur as predicted.

The response is negative (depression) when predicted rewards are omitted. (Ljungberg et al.,

1992; Schultz et al., 1993; Schultz, 1998 for review). The dopamine reward signal is

supplemented by reward-processing neurons in the striatum, amygdala, and frontal cortex

(Apicella et al., 1991, 1992; Schultz et al., 1992; Schultz, 1998), and these regions provide

direct and indirect feedback to the mesencephalic DA neurons via a cortico-striato-tegmental

loop (Alexander et al., 1990; Goldman-Rakic, 1995). However, Redgrave and coworkers

(Redgrave et al., 1999) have recently proposed another functional interpretation of the activity

of DA midbrain neurons. In their view DA release is responsible for reallocating attentional

and behavioural resources, in a sense of „switching“ attention to new or salient stimuli.

In accordance with both models, dopaminergic midbrain neurons are activated at the onset of

working memory tasks (Schultz et al., 1993) and DA levels in the PFC increase during delay-

task performance (Watanabe et al., 1997). In contrast, the behavioural impairments following

lesion of the mesencephalic dopamine neurons or the blockade of dopaminergic transmission

in the PFC are very similar to those observed after lesions of the PFC itself (Brozoski et al.,

1979; Simon et al., 1980; Sokolowski and Salamone, 1994; Sokolowski et al., 1994).

Destruction of dopaminergic neurons in area A 10 (Simon et al., 1980) or of dopaminergic

fibers within the PFC by 6-hydroxydopamine (6-OHDA) (Brozoski et al., 1979) result in

deficits in delay-tasks, and an impairment in response control (Sokolowski and Salamone,

1994). The effect of DA on performance in delay-tasks is mediated by the D1 receptor, and an

Introduction 17

optimal level of stimulation by DA appears to be critical for proper functioning (Sawaguchi et

al., 1990b; Sawaguchi and Goldman-Rakic, 1991, 1994; Williams and Goldman-Rakic, 1995;

Murphy et al., 1996; Sawaguchi, 1997; Zahrt et al., 1997; Arnsten and Goldman-Rakic, 1998;

Seamans et al., 1998).

The cellular bases of the effects of DA in the cortex, however, remain enigmatic. In vivo

extracellular single-unit recording studies have shown that iontophoretically applied DA either

increased or decreased spontaneous neuronal firing in the neocortex (Bradshaw et al. 1985;

Sesack and Bunney 1989; Yang and Mogenson, 1990). Dopamine applied iontophoretically or

released by VTA stimulation, suppressed spontaneous, as well as presumed glutamate-

mediated mediodorsal thalamic-evoked firing in the rat PFC in-vivo (Ferron et al. 1984;

Sesack and Bunney 1989; Godbout et al. 1991; Pirot et al. 1992). In contrast, neuronal firing

induced by iontophoretic application of acetylcholine or NMDA is enhanced by very low

doses of iontophoretically applied DA (Yang and Mogenson 1990; Cepeda et al. 1992).

These findings are in good agreement with the long proposed neuromodulatory action of DA.

By definition, a neuromodulator, in contrast to the classical neurotransmitters, is not inhibitory

or excitatory per se, but when applied in conjunction with other substances it has the ability to

alter their actions. The modulatory nature of DA also appears to extend to a voltage- (or

„activity-“) dependence of its effects, in that DA has inhibitory effects at hyperpolarized

membrane potentials, but enhancing effects at depolarized potentials (see below).

Another important methodological difference that hampers comparison across different in-

vivo studies may result from the use of anesthetics that block N-methyl-D-aspartate (NMDA)

conductances and thus one of the major excitatory effects of DA. Both in the striatum and

PFC, DA appears to differentially modulate AMPA and NMDA receptors. Application of DA

reduces AMPA currents, but enhances NMDA conductances (Cepeda et al., 1992, 1998;

Durstewitz et al., 2000; but see Law-Tho et al., 1994). This in turn is partly due to the

different voltage dependence of these currents. Activation of NMDA receptors requires prior

depolarization of the membrane, whereas AMPA receptors show no voltage-dependent

activation. Similarly in vitro, the effects of DA on striatal neurons, acting via the D1 receptor,

are inhibitory when the membrane potential is hyperpolarized (which resembles the so called

„down“-state of striatal neurons observed in-vivo), but are excitatory when the cell is in the

depolarized „up“-state (Hernández et al., 1997). There is evidence that in the cortex DA has

comparable enhancing effects particularly on highly activated neurons (Sawaguchi et al.,

1988, 1990a; Durstewitz et al., 2000). Although DA inhibits spontaneous activity within the

Introduction 18

PFC of anesthetized animals (see above), it enhances task-related single-unit activity more

than background activity in the behaving animal (Sawaguchi et al., 1988, 1990a).

In the cortex, this „content-specific“ enhancement of neuronal activity results from a complex

interplay of ionic conductances that govern the action-potential threshold of pyramidal

neurons and regulate the transmission of synaptic input through the apical dendrite (where

most of the cell´s input arrives).

In the cortex DA, via its action at the D1 receptor, appears to shift the threshold of the

persistent Na+ current (INaP) towards more hyperpolarized potentials (Yang and Seamans

1996; Durstewitz et al., 2000; but see Geijo-Barrientos and Pastore, 1995). Activation of INaP

or its prolonged activation will reduce the threshold for generating an action potential, and

render the cell more susceptible for excitatory inputs (Schwindt, 1992; Yang and Seamans,

1996). Dopamine might also act on a slowly inactivating K+ current (IKS) that usually delays

spike generation and lengthens inter-spike intervals, which in turn suppresses repetitive firing

(Schwindt et al. 1988; Kitai and Surmeier, 1993; Yang and Seamans, 1996). A reduction of

IKS shortens inter-spike-intervals and reduces accomodation of firing (Yang and Seamans,

1996; Nisenbaum et al., 1998). The concurrent actions of DA on both INaP and IKS might

reduce spike treshold and contribute to an increased firing rate observed in vivo (Sawaguchi et

al., 1988; 1990a,b) and in vitro (Cepeda et al., 1992; Yang and Seamans, 1996).

In addition, DA has been shown to excert a differential effect on high-voltage-activated

(HVA) Ca++ currents. Dopamine augments the activation of L-type Ca++ channels, which

functionally serve to enhance the amplitude and duration of excitatory synaptic inputs

(Seamans et al., 1997), thus further increasing the neurons excitability (Hernández-López et

al. 1997; Cépeda et al. 1998). Finally, in both cortical and striatal neurons, DA reduces the

amplitude of isolated N-type Ca++ currents, which has been interpreted as a „filtering“-

mechanisms that permits only strong synaptic inputs (e.g. behaviourally salient stimuli) to

reach the soma (Surmeier et al., 1995;Yang and Seamans, 1996). The properties of these

conductances modulated by DA are described in more detail in the discussion of chapter 6.

In the mammalian PFC, DA also appears to modulate the activity of GABAergic interneurons

and spontaneous GABA release. Dopamine directly depolarizes interneurons in the PFC and

enhances spontaneous and evoked IPSPs recorded in pyramidal neurons (Penit-Soria et al.,

1987; Godbout et al., 1991; Rétaux et al., 1991; Pirot et al., 1992; Zhou and Hablitz, 1999;

Durstewitz et al., 2000). Accordingly, part of the suppressive effects of DA on spontaneous

Introduction 19

firing of pyramidal PFC neurons observed in vivo might be due to the action of DA on

GABAergic interneurons.

Aims of the study

As outlined above, the experiments in this thesis try to unravel the anatomical and

physiological properties of a proposed canonical circuit involved in working memory and

other executive functions in a non-mammalian species. The NCL of pigeons (Columba livia)

and chicks (Gallus gallus domesticus) serves as a model structure in this quest as behavioural

studies have shown its involvement in working memory (e.g. Mogensen and Divac, 1982,

1993; Kalt et al., 1999), reversal learning (Hartmann and Güntürkün, 1998), and response

inhibition (Güntürkün, 1997; Aldavert et al., 1999). Furthermore, the strong dopaminergic

innervation (e.g. Waldmann and Güntürkün, 1993) and recent evidence from a behavioural

study investigating the effects of pharmacological blockade of dopaminergic transmission

(Güntürkün and Durstwitz, in press), support the involvement of DA in the executive

functions of the NCL.

The studies undertaken here to uncover the cellular bases of these behaviours adress two main

topics.

1. The functional organization of the NCL.

2. The influence of the dopaminergic system on identified cell types within the NCL.

The functional organization of the NCL

The term working memory refers to the control of reactions depending on past experiences

and the ongoing stream of sensory information that is currently being experienced (c.f.

Baddeley, 1992; Goldman-Rakic, 1995). As exemplified for the PFC above, this „memory for

action“ requires (i) the retrieval of learned responses from long-term memory (possibly from

memory stores in the hippocampus or posterior cortices), (ii) the simultaneous

integration/evaluation of sensory information about the environment (provided by the

posterior cortices and the thalamus) for the selection of the appropiate response, and (iii) the

initiation of the response. Thus, a clear prerequisite for a structure in non-mammalian species

that might subserve working memory is the ability to integrate all available sensory

Introduction 20

information and to exert influence over motor and limbic structures. While the available

behavioural data (e.g. Mogensen and Divac, 1982, 1993; Hartmann and Güntürkün, 1998)

strongly indicate that such sensorimotor integration takes place in the NCL, the anatomical or

electrophysiological basis for these computations is unclear. The experiments described in

chapter 1 have thus examined the areal termination pattern of afferents from other

telencephalic structures to the NCL, by means of anterograde and retrograde pathway tracing

techniques. In chapter 2, these findings are complemented by the anatomical description of

thalamic input from a multimodal nucleus in the dorsal thalamus. In addition, the organization

of efferents from the NCL to (para-)sensory structures, as well as limbic and motoric areas in

the avian forebrain, through which the NCL might modulate its own input and initiate motoric

responses, respectively, are characterized in chapter 1. Taken together, these anatomical

results reveal the nature of polymodal processing within NCL and possible functional

subdivisions of the NCL.

The projection from the mediodorsal thalamus to the PFC serves as a hallmark for the

definition of the PFC (Rose and Woolsey, 1948; Divac et al., 1978; Goldman-Rakic and

Porrino, 1985; Groenewegen et al., 1990), and lesions of the MD result in deficits in delay-

tasks comparable to those following lesions of the PFC (Sakurai and Sugimoto, 1985). It has

been suggested that in birds the projection from the multimodal nuc. dorsolateralis posterior

(DLP) to the NCL might form a functional equivalent to the MD-PFC connection (Güntürkün,

1997). In chapter 2 the nature of short-latency visual evoked potentials in the NCL is

examined by extracellular electrophysiological recordings in-vivo, and the effects of lesions to

the DLP on visual responses in the NCL are studied. Together with the anatomical data also

presented in chapter 2 (see above), the results demonstrate direct thalamic modulation of NCL

activity.

To understand the neuronal circuitry that underlies higher cognitive functions in vertebrates, a

detailed knowledge about the celltypes and their possible function within the proposed

canonical circuit is required. It is axiomatic in neuroscience that the informational output of

most neurons is defined by their temporal patterns of action potentials (e.g. McCormick et al.,

1985). This notion is implicit in all contemporary models of the function of the brain, making

it essential to understand how each neuron transforms its input into output. Intracellular or

patch-clamp electrophysiological recordings give the opportunity to determine the intrinsic

electrophysiological properties of neurons by measuring the membrane´s response to

hyperpolarizing and depolarizing current pulses. The intrinsic membrane properties of

Introduction 21

neurons in the mammalian and avian forebrain are not homogeneous, but instead produce

several categories of transform characteristics that can be used to categorize cells into certain

types (e.g. McCormick et al., 1985; Larkman and Mason, 1990; Kubota and Taniguchi, 1998).

In chapter 3 the principal cell types of the chick NCL are characterized both

electrophysiologically and morphologically in vitro, using whole-cell patch-clamp recordings

in combination with intracellular staining. The distinct firing patterns and axonal connections

of the principal neurons can be interpreted with respect to the integrative role they play within

the circuitry of the NCL, thus unraveling the „building blocks“ of mental computation within

NCL.

The influence of the dopaminergic system on identified cell types within the NCL.

The dopaminergic innervation in the avian telencephalon takes two different, and possibly

discrete, forms. The first, which is also predominant in the mammalian cortex (e.g. Williams

and Goldman-Rakic, 1993), is characterized by dopaminergic fibers that travel in close

vicinity along the somata and dendrites of target-neurons while forming „en-passant“' a large

number of bouton-like axonal swellings. The second one, called the „basket“'-type, is a

peculiar arrangement of dopaminergic fibers in the avian telencephalon. In this case, single

fibers densely coil around the somata and intial dendrites of postsynaptic targets, enwrapping

them in basket-like structures (Wynne and Güntürkün, 1995; c.f. chapter 4). It can be

speculated that this particularly strong dopaminergic innervation also differs functionally from

the more diffuse „en-passant“ innervation. The relative frequency of these modes of

innervation differs among various brain areas (Wynne and Güntürkün, 1995; chapter 4), but

the NCL is characterized by a strong innervation from both types of dopaminergic fiber

systems, thus in principle enabling a direct comparison of the two modes of DA innervation.

Chapter 4 provides a review about the dopaminergic innervation of the avian forebrain and

discusses anatomical and behavioural findings with regard to the functional organization of

the avian brain, and some general implications of these findings for sensorimotor integration

and learning. Chapter 5 presents original data also reviewed in chapter 4. This study

attempted to identify the cell type that receives dense basket-type DA innervation by

colocalizing the dopaminergic fibers4 with postsynaptic markers of neuroactive substances

4 In fact, immunoreactivity against tyrosine-hydroxylase (TH), which is the rate limiting enzyme in the synthesisof all catecholamines, is used for the demonstration of presumed dopaminergic fibers, as there is a goodcorrespondence between TH immunoreactivity and a direct marker of DA (c.f. chapter 4), comparable to thesituation in the prefrontal cortex (Williams and Goldman-Rakic, 1993).

Introduction 22

(i.e. glutamate-decarboxylase, as a marker of GABAergic neurons) or receptor proteins (i.e.

DARPP-32 as an indicator for D1 receptors). Specifically it was investigated whether

GABAergic interneurons are a main target of dopaminergic fibers, and if so, whether the

effects of DA might be mediated via the DA D1 receptor as is the case in the PFC (Krimer et

al., 1997; Muly et al., 1998).

Finally, the effects of DA on membrane excitability are tested directly in chick NCL neurons

in-vitro. In chapter 6, patch-clamp recordings from neurons of all cell types are made (c.f.

chapter 3) and changes in the firing patterns in response to DA application are analyzed. The

observed changes in action-potential firing provide insight into the mechanisms and substrate-

specifity (i.e. cell types and receptors involved) by which DA might enable sustained activity

during a delay task. Furthermore, double-labeling of intracellularly filled neurons with

dopaminergic fibers is used to identify classes of neuron that receive basket-type DA

innervation. The responses of neurons receiving basket-innervation to application of DA are

compared to the responses of neurons that receive only en-passant innervation of

dopaminergic fibres. Taken together, this will elucidate whether the basket-type innervation

selctively innervates a distint class of neurons, and/or whether the effects of DA excerted

through this strong somatic innervation differs qualitatively or quantitatively from the

dendritic en-passant innervation.

Taken together, the results show the structural and physiological bases of polymodal sensory

integration and motor control in the avian telencephalon, which appear to underlie higher

cognitive functions. Furthermore, the functional organization of the mesotelencephalic

dopaminergic system and dopaminoceptive elements in the caudal forebrain are studied,

which appear to be critically involved in cognitive processing and motor control, both in birds

and mammals.

Chapter 1: Connections of the avian neostriatum caudolaterale 2323

Afferent and efferent connections of the caudolateral neostriatum in the

pigeon (Columba livia): A retro- and anterograde pathway tracing study.

Sven Kröner and Onur Güntürkün

ABSTRACT

The avian caudolateral neostriatum (NCL) was first identified on the basis of its dense dopaminergic

innervation. This fact and data from lesion studies have led to the notion that NCL might be the avian equivalent

of prefrontal cortex (PFC). A key feature of the PFC is the ability to integrate information from all modalities

needed for the generation of motor plans. By using antero- and retrograde pathway tracing techniques we

investigated the organization of sensory afferents to the NCL and the connections with limbic and

somatomotor centers in the basal ganglia and archistriatum. Data from all tracing experiments were compared

with the distribution of tyrosine-hydroxylase (TH) immunoreactive fibers, serving as a marker of dopaminergic

innervation. The results show that NCL is reciprocally connected with the secondary sensory areas of all

modalities and with at least two parasensory areas. Retrograde tracing also demonstrated further afferents from

the deep layers of the Wulst and from the frontolateral neostriatum as well as the sources of thalamic input.

Efferents of NCL project onto parts of the avian basal ganglia considered to serve somatomotor or limbic

functions. Projections to the archistriatum are mainly directed to the somatomotor part of the intermediate

archistriatum. In addition, cells in caudal NCL were found to be connected with the ventral and posterior

archistriatum, which are considered avian equivalents of mammalian amygdala. All afferents and projection

neurons were confined to the plexus of densest TH innervation. Our results show that the NCL is positioned to

amalgamate information from all modalities and to exert control over limbic and somatomotor areas. This

organization might comprise the neural basis for such complex behaviours as working memory or spatial

orientation.

Journal of Comparative Neurology 407:228-260 (1999)

INTRODUCTION

The neostriatum caudolaterale (NCL) of the pigeon has been compared with the mammalian

prefrontal cortex (PFC), due to its dense dopaminergic innervation (Divac et al., 1985, 1994;

Divac and Mogensen, 1985; Waldmann and Güntürkün, 1993; Wynne and Güntürkün, 1995),

and the behavioural deficits in working memory (Mogensen and Divac, 1982, 1993, Gagliardo

et al., 1996, 1997, Güntürkün, 1997), reversal learning (Hartmann and Güntürkün, 1998), and


go/no-go tasks (Güntürkün, 1997) that follow its ablation. Working memory tasks selectively

target the control of reactions depending on past experiences and the ongoing stream of sensory

information that is currently being experienced, which in mammals is a key feature of the PFC

(Goldman-Rakic, 1987; Fuster, 1989). A clear prerequisite for a structure in non-mammalian

species that can subserve such functions and may thus be regarded as an analogue of the PFC is

the ability to integrate all available sensory information and to exert influence over motor and

limbic structures.

A number of previous studies have demonstrated afferent input to the lateral neostriatum from

visual (Ritchie, 1979; Shimizu et al., 1995;), auditory (Wild et al., 1993) and somatosensory

(Shimizu et al., 1995) areas, as well as a projection from the multimodal thalamic nucleus

dorsolateralis posterior (DLP); (Waldmann and Güntürkün, 1993; Leutgeb et al., 1996). Based

on cytoarchitectonic and hodological data, Rehkämper and Zilles (1991) have proposed that the

complete posterior neostriatum, including its caudolateral aspect, might represent an area of

multimodal integration. Furthermore, in a recent retrograde labeling study focused on the NCL

(Leutgeb et al, 1996) it has been shown that the NCL is reached by afferents from all major

secondary sensory areas within the pigeon brain. Until now, however, a detailed study of the

termination pattern of these afferents and a possible functional segregation within the pigeon´s

NCL is still lacking. Retrograde tracing data from a very recent study in chicks (Metzger et al.,

1998) revealed distinct and non-overlapping termination zones for the trigeminal, tectofugal,

and auditory systems within the NCL. Since retrograde tracing in pigeons (Leutgeb et al., 1996)

suggested a substantially different organization of the NCL with largely overlapping sensory

compartments, this report describes the afferent and efferent connections of the NCL as studied

by retro- and anterograde pathway tracing and compares it with the distribution of tyrosine

hydroxylase-immunoreactive fibers.

MATERIALS AND METHODS

Neuroanatomical pathway tracing experiments

Fifty-seven adult pigeons (Columba livia) from local stock provided the data presented here. Treatment of

animals conformed to NIH guidelines and specifications of the German Animal Welfare Act. Accordingly,

prior to surgery, the animals were deeply anaesthetized with 0.33 - 0.4 ml Equithesin per 100 g body weight.

Animals received pressure-injections of either the sensitive antero- and retrograde tracer Cholera Toxin,

subunit b (CTb, 1% in A.dest.; List Labs, Campbell, CA) or biotynilated dextran amines as anterograde tracer

(BDA, 10,000 molecular weight form, 10% in sodium phosphate buffer, pH 7.3; Molecular Probes, Leiden,

The Netherlands). Tracer was delivered through glass micropipettes (tip diameter 15-20 µm) attached to a


nanoliter-injector (World Precision Instruments, Sarasota, FL). Since previous studies (Leutgeb et al., 1996;

Metzger et al., 1998) and our own data have indicated that the majority of connections of the NCL are

restricted to the ipsilateral hemisphere, most animals received bilateral injections to minimize the number of

pigeons used. Stereotaxic coordinates for the injections were determined by using the atlas of Karten and

Hodos (1967).

TABLE 1: Experimental Parameters1

Tracer

BDA CTbInjection site Number of

animalsNumber of

cases L R L R

NCL 7 11 − − 6 5Nd 3 4 − − 2 2field L 3 6 1 − 2 3Ep 6 10 − 1 5 4NFT 3 5 − − 2 3rostral HA 3 5 1 − 2 2caudal HA 4 6 1 1 2 2NIMm 4 6 1 − 3 2NIMl 6 10 − − 5 5Ai and Av 6 10 − 3 2 5Ap 3 5 − 1 2 2LPO 5 8 − − 4 4PA 4 6 − 1 3 2

1 R, Right side; L, Left side; for other abbreviations, see list

The afferent sources of the NCL were determined by injections of CTb (30 - 54 nl) into medial and lateral

parts of the NCL along its rostro-caudal extent, as well as into the directly adjacent neostriatum dorsale (Nd,

Bonke et al., 1979; Wild et al., 1993; Wild, 1994). We included Nd in our analysis because the pattern of

catecholaminergic innervation suggests that Nd is continuous with NCL (figure 1) and might constitute its

auditory subcomponent. To study the areal pattern of afferent input to the NCL, injections of CTb (9 - 80 nl)

and BDA (40 - 100 nl) were made into most of the telencephalic regions which by means of retrograde

transport had been shown to constitute afferent sources of the NCL. Efferent connections of the NCL to motor

and limbic structures were studied by placing injections into the somatomotor and limbic parts of the

archistriatum (Zeier and Karten, 1971) as well as into parts of the avian basal ganglia that are comparable to

parts of mammalian striatum.

Immunohistochemical procedures

Survival times were 3 days for CTb and 4-6 days for BDA. Fifteen minutes prior to perfusion the animals were

injected with 1,000 IU heparin and deeply anaesthetized with 0.4 - 0.5 ml Equithesin per 100 g body weight.

Pigeons were then perfused transcardially with 200 ml 0.9% saline (40°C) followed by 1,000 ml of 4%

paraformaldehyde in 0.12 M phosphate buffer (PB; 4°C, pH 7.4). After perfusion, brains were dissected and

postfixed in the same fixative to which 30% sucrose was added, and then transferred to 30% sucrose in


phosphate buffer containing 0.9% NaCl (PBS; pH 7.4) for approximately 18 hours at 4°C. Brains were cut in

frontal slices of 40 µm on a freezing microtome and collected in PBS containing 0.01% NaN3 as a

preservative. Representative sections were then processed for the avidin-biotin-conjugate technique (ABC).

Immunohistochemical labeling for Cholera Toxin b: Endogeneous peroxidases were blocked by

preincubating slices in a solution of 0.5% H2O2. Slices were washed and for immunohistochemistry of CTb

free-floating sections were incubated overnight at 4°C in anti-CTb from goat (Jackson, West Grove, PA;

1:20,000) in PBS containing 0.3% Triton X-100 (Sigma, Deisenhofen, Germany). The following steps were

carried out at room temperature, separated by three washes in PBS of 10 minutes each. After washing, slices

were first incubated for 1 hour in biotinylated donkey anti-goat (Jackson,1:500 in 0.3% Triton X-100 PBS) for

1 hour and then in the avidin-biotin complex (ABC Elite, Vector Labs, Burlingame, CA; 1:100 in PBS with

0.3% Triton X-100) for 1 hour. Washes in PBS were followed by two additional washes in 0.12 M acetate

buffer (pH 6). Staining was achieved by the 3,3'-diaminobenzidine (DAB) technique with heavy-metal

amplification (modified from Adams, 1981) by adding H8N2NiO8S2 (2.5 g/ 100 ml), NH4Cl and CoCl2 (both

40 mg/ 100 ml). After 15 minutes of preincubation the reaction was catalyzed with a solution of 0.5% H2O2.

The reaction was stopped by rinsing the tissue in 0.12 M acetate buffer and PBS. Slices were then mounted,

dehydrated and coverslipped.

Labeling for BDA: Visualization of BDA was identical to that of CTb, except that primary and secondary

antibodies could be omitted and slices were directly incubated in ABC, followed by the staining procedure

described above. Selected series labeled for BDA or CTb were counterstained with cresyl violet.

Immunohistochemical labeling for tyrosine hydroxylase: To compare the distribution of afferent fibers and

retrogradely labeled neurons within NCL with the distribution of putative dopaminergic fibers, several sections

were processed for tyrosine hydroxylase (TH). Therefore, additional sections stained for CTb or BDA were

either double-labeled with TH, or occasionally, sections adjacent to these were single-labeled for TH. The

same basic procedure as for immunohistochemistry of CTb was applied for labeling of TH: After blocking of

endogeneous peroxidases and thorough washing, free floating sections were incubated overnight at 4°C in

monoclonal mouse anti-TH (Boehringer, Mannheim, Germany) diluted 1:200 in PBS containing 0.3 %Triton

X. Slices were then incubated at room temperature in biotinylated rabbit anti-mouse (Chemicon, Temecula,

CA; 1:200 in 0.3% Triton PBS) for 1 hour and finally in ABC for 1 hour. Staining was achieved using the DAB-

technique as described for CTb. In those cases that were double-labeled, enhancement of the DAB staining by

nickel ammonium sulfate was ommitted, resulting in a light brown signal for TH which could be clearly

distinguished from the black reaction product of the tracers.

RESULTS

Retrograde tracing of afferent sources of the NCL

Our injections of CTb into medial and lateral parts of NCL generally replicated the results of an

earlier study (Leutgeb et al., 1996). All injections were confined to NCL as defined by TH and

DA-like immunoreactivity (Waldmann and Güntürkün, 1993; Wynne and Güntürkün, 1995;


present study), and extended from the caudal tip of the telencephalon at approximately A 4.25 to

rostral A 7.00.

Telencephalic afferents to NCL

Areas that were found to project to the NCL included the medial aspect of the hyperstriatum

accessorium (HA) throughout its visual (Karten et al., 1973; Shimizu et al., 1995) and

somatosensory part (Delius and Benetto, 1972, Wild, 1987b, Funke, 1989; but see Deng and

Wang, 1992, 1993), the ectostriatal belt (Ep) surrounding the ectostriatum dorsally and laterally

(Karten and Hodos, 1970) and the adjoining lateral neostriatum, the anterior neostriatum

overlying the nucleus basalis (neostriatum fronto-trigeminale, NFT, of Wild et al., 1985), and

field L1 and L3 of the auditory field L complex. A large number of cells were labeled in the

medial part of the intermediate neostriatum (NIM of Veenman et al., 1995b) and the overlying

intermediate and medial parts of the hyperstriatum ventrale (HV); part of these neurons

probably correspond to the previously described parasensory area in the intermediate

NCL

ApH

AvAi

AiddAidv

NCm

NCL NCL

Ap

NCm

ApH

Av

NCL

Aidd

Ai

TPO

BNST

ApH

PPPA

Hv

Tn

NCL

ApH

Aidd

PA

Hv

AvAi

Tn

A 5.75A 4.25 A 5.00

A 6.75 A 7.25

Figure 1: Camera lucida drawing of tyrosine hydroxylase (TH) immunoreactive fibers in the caudal telencephalon of the pigeon.Staining for TH produces a pattern very similar to that seen with an antibody against dopamine which has been previously usedto define the neostriatum caudolaterale (NCL); (Waldmann and Güntürkün, 1993; Wynne and Güntürkün, 1995; Metzger et al.,1996). For abbreviations see list.


neostriatum, situated between the rostral pole of field L and the caudomedial border of the

visual ectostriatum, that receives multimodal input via the DLP (Gamlin and Cohen, 1986;

Wild, 1987a, 1994; Funke, 1989; Korzeniewska and Güntürkün, 1990). These cells were

predominantly labeled if injections were made into more rostral and medial (including Nd)

parts of the NCL.

After injections into caudal NCL, a distinct cell group occupying the remaining medialmost

aspect of the NIM between approximately A 8.50 and A 12.50 was labeled. Cells within this

area also extended across the lamina hyperstriatica (LH) into the intermediate and medial parts

of the HV. These two cell groups roughly followed the cytoarchitectonic borders of the regions

Ne 9 or Ne 8 and Ne 3, respectively, as described by Rehkämper et al., (1985). For the reason

of simplicity we will refer to them as the medial part of the NIM (NIMm) and the lateral part of

the NIM (NIMl), respectively.

In the caudomedial neostriatum (NCm) scattered cells were found dorsal to the lamina

medullaris dorsalis (LMD). Another band of cells stretched along the border between the

hyperstriatum dorsale (HD) and the dorsal division of the ventral hyperstriatum (HVdv). These

cells could be found in most cases, irrespective of the injection site within NCL. In cases that

were centered more medially or caudally within NCL this band of cells also extended into a

small triangular shaped zone within the vallecula. Finally, in several cases a number of cells

were observed in the frontolateral part of the neostriatum (NFL). Injections of CTb into this

area did indeed confirm a reciprocal connection with the ventrolateral part of the NCL (data not

shown). The large majority of fibers from this part of the anterior neostriatum, however,

terminated within the area corticoidea dorsolateralis (CDL) and the area temporo-parieto-

occcipitalis (TPO) overlying the NCL. Therefore we can not fully exclude that part of the

labeling within NFL was attributable to tracer spread into CDL.

In addition to labeling in sensory areas, retrograde transport revealed a number of labelled

neurons in the dorsal and ventral parts of the intermediate archistriatum. Cells in the ventral

archistriatum were the only telencephalic connection found to project bilaterally onto the NCL.

In contrast to Leutgeb et al., (1996) we never observed any labeling within the nucleus taeniae

(Tn). Furthermore, a few cells within the ventral striatum just dorsal to the fasciculus

prosencephali lateralis (FPL) were also found to project to the NCL.


HV

PA

NCL

NCm

Aidd

Ai

AvTn

Field L

APH

Hp

NCL

TPc

A 6.75

A 3.00

HV

Aidd

Ai

Av

APH

HpCDL

NCL

CDL

Nd

A 5.75

A 4.25AVT

rostral NCLcaudal NCL Nd

NCm

TnDLP

DIP

SRt

Ov

HpAPH

Figure 2: Schematic representation of retro- and anterograde labeling following injections of Cholera Toxin b (CTb) into variouslocations within NCL. Retrogradely labeled cells are represented by filled symbols (dots, triangles or squares for the differentinjection sites) and anterogradely labeled fibers and terminals are represented by short lines. Anterograde labeling is shown onlyfor selected regions. The left and middle column illustrate the topographic ordering of NCL afferents and efferents afterinjections into the caudalmost NCL (left column) or the far rostral NCL (middle column), respectively. The right column showsan example of an injection into Nd. Labeling after Nd injections was very similar to that observed after tracer deposits intomedial NCL. In this and all following figures medial is to the left and dorsal is at the top.


Figure 2 (continued)

HA

NFT

NFL

HVvv

HD

A 12.50

A 14.00

A 10.50

A 8.25

HA

Bas

NFT

NFL

LPO

Va

HV

HD

HA

NILE

Ep

NIMm

NIMl

HV

Va

HD

LPOPAPP

E

Ep

HA

TPO

NIMlNIMm

HV

PAPP

FPL

AC

rostral NCLcaudal NCL Nd


In addition to labeling in sensory areas, retrograde transport revealed a number of labelled

neurons in the dorsal and ventral parts of the intermediate archistriatum. Cells in the ventral

archistriatum were the only telencephalic connection found to project bilaterally onto the NCL.

In contrast to Leutgeb et al., (1996) we never observed any labeling within the nucleus taeniae

(Tn). Furthermore, a few cells within the ventral striatum just dorsal to the fasciculus

prosencephali lateralis (FPL) were also found to project to the NCL.

In general, injections into NCL resulted in a similar pattern of retrograde labeling, but regional

differences exist. With respect to a mediolateral topography our results resemble those of

Leutgeb and coworkers (1996) and are thus not described in detail here. However, we found a

strong rostro-caudal topography in the organization of afferent and efferent connections of the

NCL. The nature of this topography was further elucidated in anterograde tracing experiments.

Figure 2 illustrates the results after two injections into the far caudal and rostral aspects of

NCL.

Subtelencephalic sources of afferent input to NCL were observed in the thalamus, midbrain and

tegmentum. In the lower brainstem the avian equivalent of the substantia nigra, the nucleus

tegmenti pedunculopontinus, pars compacta (TPc), the locus coeruleus (LoC) and the area

ventralis tegmentalis (AVT) were labeled. These cell groups have been shown to be

immunoreactive for TH and dopamine and contain the sources of dopaminergic afferents to the

telencephalon (Waldmann and Güntürkün, 1993; Metzger et al., 1996; present study). The

projection from the LoC and some cells in the formatio reticularis was bilaterally organized. In

the thalamus retrogradely labeled cells were mainly found in the nucleus subrotundus (SRt) and

the DLP. Yet, we also found a projection from the n. dorsointermedius posterior thalami (DIP)

and a band of cells that seemed to intersperse between DLP and DIP and also extended ventral

to the latter two. Therefore we termed these cells collectively the ventrointermediate area of the

posterior nuclei (VIP). Figure 3 shows examples of retrogradely labeled neurons from sensory

forebrain areas (A - F) and the sources of thalamic input to NCL following injections of CTb

into NCL.

Anterograde labeling of efferents of the NCL

The pattern of anterograde labeling resulting from injections into NCL was studied within the

basal ganglia and the archistriatum. Fibers showing terminal-like varicosities were found

throughout most parts of the archistriatum. For the description of archistriatal subdivisions we

adopted the terminology of Wynne and Güntürkün (1995).


Figure 3: Photomicrographs of retrogradely labeled cells resulting from injections of CTb into the NCL similar to those shown infigure 2. Afferents arise from secondary sensory „belt“ regions adjacent to primary thalamorecipient areas. See text for details.Retrogradely labeled cells in the telencephalon include the ectostriatal belt (A), field L 1 (B), HA and HD (C), the borderbetween HD and HVdv (D), the presumed parasensory area NIMm (E) and the NFT (F). Labeling in the thalamus was mostabundant in the DLP and SRt (G). H: Cells in the ventrointermediate area of the posterior nuclei (VIP) labeled after an injectioninto far lateral NCL. Scale bars are 100 µm (B, H) and 250 µm (A, C - G). See list for abbreviations.


Injections into the very caudal aspect of NCL predominantly labeled fibers in parts of the

central (Ai) and ventral parts (Av) of the intermediate archistriatum and in some cases also in

the posterior archistriatum (Ap). Injections into more rostral NCL resulted in terminal labeling

that was confined to the dorsal and central parts of the intermediate archistriatum. After

injections into the caudal NCL we observed terminal-like labeling in the medial lobus

parolfactorius (LPO) and the ventral striatum, i.e. the nucleus accumbens (AC) and the bed

nucleus of the stria terminalis (BNST), which was not attributable to tracer spread into the

piriform cortex (CPi) or the caudal archistriatum. Injections into rostral NCL on the other hand

resulted in dense terminal labeling in the lateral aspects of LPO and large parts of the

paleostriatum augmentatum (PA). Labeling in the PA was most abundant in its caudal extent.

Both termination patterns within the archistriatum and the basal ganglia thus suggest that the

anterior NCL may be connected to areas concerned with somatomotor functions, i.e. the dorsal

and lateral striatum (Veenman et al., 1995a,b), as well as the dorsal and central archistriatum

(Zeier and Karten, 1971), while the caudal aspect of NCL might project to limbic parts of the

striatum (Veenman et al., 1995b) and the presumed avian homologue of the amygdala, e.g. the

ventral and caudal archistriatum (Zeier and Karten, 1971, Dubbeldam et al., 1997). Figure 4

shows examples of anterograde labeling in the basal ganglia and the archistriatum following

injections of CTb into NCL.

Labeling after Nd injections

Labeling after CTb injections into Nd largely paralled the results after medial NCL injections.

A large number of retrogradely labeled cell bodies and fibers could be observed throughout the

NIM, which seemed to cluster in two distinct cell groups as also observed after injections into

the NCL. Strong labeling was also present in the caudomedial neostriatum and areas L1 and L3

of the field L complex. Furthermore, cells were found in the medial HV and along the border of

HV and HD, the caudal and rostral aspects of HA, the valleculla (Va), and the NFL. Finally,

somata and terminal-like labeling could be observed in the ventromedial archistriatum and

lateral parts of the NCL. Terminal-like labeling in the basal ganglia after Nd injections was

restricted to the medial and central parts of the caudal PA. In the thalamus labeling was most

prominent in the SRt and DLP. A number of cells was also found within the shell of the nucleus

ovoidalis (Ov) and the nucleus semilunaris parovoidalis. In the midbrain AVT, TPc and LoC

contained labeled cell bodies. The results following an injection into Nd are summarized

schematically in figure 2 (right column).


Figure 4: Photomicrographs of anterograde labeling of NCL efferents to the basal ganglia and archistriatum. A: Labeled fibers inthe presumed „limbic“ parts of the avian striatum (i.e., nucleus accumbens and medial LPO) after an injection of CTb into farcaudal NCL. B and C: Labeling in the presumed „somatomotor“ parts of the basal ganglia, the lateral LPO (B) and central PA(C). D and E: Labeled fibers and somata in somatomotor and limbic parts of the archistriatum. Injections into the most caudalNCL label numerous fibers in the central part of the intermediate archistriatum, but also extend into the ventral archistriatumwhich forms part of the avian amygdala (D). In contrast, injections into the rostral NCL lead to labeling in the dorsal and centralparts of the intermediate archistriatum which are considered somatomotoric in function. Cells from most parts of thearchistriatum project back to the NCL. The projection from Av is the only bilaterally organized connection of the NCL. Scalebars are 250 µm (A), 50 µm (B), 100 µm (C), and 500 µm (D, E).


Anterograde tracing experiments

Injections into the ectostriatal belt

After injections into the dorsal and lateral ectostriatal belt and the adjoining neostriatum

intermedium (NI) underneath the external pallium (Veenman et al., 1995b) a massive projection

of labeled fibers was observed to move caudally through the NI and to innervate the anterior

and lateral NCL. This projection seemed to be topographically organized and extended in the

case of caudal injections posteriorly to about A 5.00. In addition to terminals, a number of

retrogradely labeled cell bodies also were found, indicating a reciprocal connection between

the NCL and the ectostriatal belt region. As shown in figure 5 a massive reciprocal connection

of the Ep exists with the deeper hyperstriatal layers, especially with the lateral part of HV.

Cells within HV and the NI or NFL, respectively, were distributed up to far rostrally within the

hemisphere. Relatively few cells and fibers were found within the outer rind of the pallium, i.e.

the lateral NI (NIL), TPO and CDL, the larger portion of them remaining underneath the border

that separates the NI and anterior NCL from the overlying external pallium. Furthermore, in the

dorsal and ventral aspects of the rostral archistriatum small clusters of cells were labeled, as

well as a number of somata in the paleostriatum primitivum (PP). Within the ectostriatal core

labeled cells were found in a narrow area below the injection site. In the thalamus large

numbers of neurons were labeled in nucleus rotundus which probably are due to tracer spread

into the ectostriatal core.

Injections into field L

In accordance with previous reports (Bonke, et al., 1979; Wild et al., 1993) tracer deposits into

the auditory field L complex which were mainly centered on the secondary sensory field L1

lead to massive terminal-like labeling in Nd. Yet, injections of CTb revealed a diffuse

projection that also extended further laterally in caudal NCL (figure 6). In the medioventral

aspect of NCL also a small number of retrogradely labeled cells could be observed. Field L

injections also resulted in retrograde labeling of neurons throughout the HA, NIM and HV, as

well as in ventral and intermediate archistriatum. Diencephalic neurons were labeled within

shell and core regions of n. ovoidalis and the n. semilunaris parovoidalis.


LPO

PA

PP

NILNIM

HA

HD

HV

E

A 9.75

NCL

APH

PA

Tn Av

Hp

AiddAi

CDL

NCm

RtA 7.50

Hp

TPO

CDL

PA

PP

Av

Ai

Aidd

NCm

A 6.50

NCL

APH

PA

TnAv

Hp

Aidd

Ai

CDL

Rt

NCm

A 6.00

TnAv

Aidd

Ai

Aidv

SRt Rt

NCm

NCL

CDLHp

APHA 7.00

NCL

APH

Ap

TO

NCm

Nd

A 6.50

A 5.00

Rt

Ov

NCL

NCm

CDL

Hp

APH

Nd

SRt

Av

Aidd

AiAidv

Am

A 5.50

HV

HV

1 mm

A

Figure 5: Results of a 15 nl CTb injection into the ectostriatal belt region, which receives afferents from the ectostriatum, theprimary visual forebrain area of the tectofugal pathway. A: Schematic representation of labeling witin the caudal telencephalon.B: Photomicrograph showing the innervation of the anterior NCL and retrogradely labeled cell bodies at about A 7.00. Scale baris 200 µm.


Injections into the hyperstriatum accessorium

After injections into the Wulst that were centered on the medial aspect of the dorsal

hyperstriatum accessorium, fibers were found to travel through the lamina frontalis superior

(LFS) and further laterally along the border that separates the anterior NCL from the overlying

LPO and CDL. Fibers showing terminal-like varicosities remained in close apposition to the

ventricle as they moved further caudally through NCL. Neurons that project back to the injected

site generally followed the distribution of terminating fibers but extended somewhat ventrally to

these. Injections along the complete rostro-caudal extent of the Wulst revealed a topographic

pattern of termination within NCL with more caudal injections leading to stronger labeling in

the posterior NCL, while afferents from most rostral HA were confined to more anterior

aspects of the NCL. This termination pattern may thus reflect the previously described

functional segregation of the Wulst into a caudal visual and a rostral somatosensory area. Yet,

the projections from HA onto NCL were not restricted to separate termination fields but

showed considerable overlap (compare figures 7 and 8).

Injections into all parts of the HA evinced retrogradely labeled cells in the HD, frontolateral

neostriatum, field L1, external pallium (CDL and TPO), a caudomedial portion of the area

parahippocampalis (APH), and the ventral and central archistriatum. Injections into rostral HA

additionally labeled cells in the NIM. In several cases in which tracer spread occured into the

thalamorecipient granular layer of the intercalated nucleus of the HA (IHA), retrogradely

labeled cells were observed in the thalamic relay stations of the thalamofugal visual and

somatosensory system, respectively: deposits of tracer into the caudal Wulst evinced cells in

the dorsal aspect of the lateral geniculate nuclei (GLd; Karten et al., 1973; Güntürkün et al.,

1993). Injections into the rostral HA on the other hand predominatly labeled cells in the

somatosensory nucleus dorsalis intermedius ventralis anterior (DIVA; Wild, 1987b; Funke,

1989; Medina et al., 1997).

Injections into the neostriatum frontotrigeminale

After injections into the NFT dorsal to the nucleus basalis (Bas), fibers gathered in the tractus

fronto-archistriatalis and passed caudally until they terminated within the lateral part of the

anterior archistriatum and the rostral and ventral part of NCL. As can be seen in figure 9 the

terminal field of this projection occupied a large part of the anterior and lateral NCL and

extended caudally up to about A 5.00. Within this dense terminal field a comparatively small

number of neurons were found to project back to the injection site. In those cases that showed


A 6.50Tn

Av

Aidd

Ai

Aidv

SRtRt

NCm

NCL

CDLHp

APH

LH

Ov

Av

Aidd

Ai

Aidv

DLP AmDIP

Ov

RtSRt

NCL

NCm

CDLHp

APH

NdNCL

APH

PA

Tn Av

Hp

Aidd

Ai

CDL

Rt

HV

NCm

LH

L1

NCL

APH

Ap

TO

NCm

Nd

TO

NCL

APH

Nd

NCL

APH

TO

Nd

A 5.00 A 4.50 A 4.001 mm

L2

L3

A 6.00A 5.50

HV

A

Figure 6: Results of a 9.2 nl CTb injection centered onto the secondary sensory field L1 of the auditory field L complex withslight tracer spread into the HV. A: Schematic representation of labeling within Nd and the adjoining parts of NCL. B:Retrogradely labeled neurons in the nucleus ovoidalis, the main thalamic relay of auditory input to the telencephalon, resultingprobably from tracer spread into the primary field L2. C: Photomicrograph of labeled fibers in Nd and NCL. Scale bars are 500µm (B) and 250 µm (C).


no or minimal tracer spread into Bas, retrogradely labeled fibers and cells within Bas were

restricted to the area directly ventral to the injection site. Furthermore, cells were labeled in the

HV dorsal to the injection site, the anterior archistriatum, along the tractus fronto-archistriatalis,

and in medial HA.

Injections into the neostriatum intermedium, pars medialis

Injections into the intermediate neostriatum attempted to evaluate whether two separate

projections from this area onto NCL exist as indicated by the results from our retrograde tracing

experiments reported above. Tracer deposits were thus aimed at the terminal field of the DLP

just medially and caudally to the ectostriatum, termed NIMl here, and the medialmost aspect of

NI, the NIMm, respectively.

Injections into NIMm consistently produced dense terminal labeling and a considerable number

of retrogradely labeled cell bodies in the most caudal aspect of NCL. Fiber bundles left the

injection site laterally and travelled caudally within the LH. They reached the caudal

neostriatum at the rostral end of Nd where they made massive terminations. The remainder of

labeled fibers coursed laterally and caudally within the periventricular roof of NCL. On their

way towards the caudal pole of the hemisphere they gradually fanned out towards the midline

until they occupied a large proportion of the hemisphere (figure 10).

Small injections that were entirely confined to NIMl resulted in anterograde labeling that was

largely restricted to Nd and the dorsal roof of the medial NCL. Labeling in these cases was

virtually identical to that seen after field L injections (compare figure 6). However, labeling

within the NCL varied considerably with the amount of tracer used. Larger NIMl injections that

also extended into one of the adjacent areas HV, NIMm or Ep, led to a much higher number of

labeled fibers and somata within NCL, while at the same time the terminal field also expanded

further lateral and medial (figure 11). However, the distribution of these fibers did not

resemble the terminal fields resulting from injections into NIMm or Ep. In addition, all of these

cases showed a similar labeling in the thalamus that was distinct from the pattern observed after

NIMm injections (below).

Deposits of CTb into the medialmost NI yielded numerous neurons within the medial HV,

dorsal to the injection site, HA, HD, in Ai and Av. Retrogradely labeled cells from Av were

also found on the contralateral side. Massive anterograde labeling was seen in medial LPO.

Injections into lateral NIM resulted in a similar pattern of labeling which included the


somatosensory and visual Wulst, medial HV dorsal to the injection site, as well as Av and Ai.

In addition, cells were labeled in field L1 and L3, as well as in large aspects of the

caudomedial neostriatum. In these cases labeled fibers were seen in the PA ventral to the

injection site, but not within LPO. Labeling in the thalamus further strengthened the notion of

TnAv

AiddAi

Aidv

SRt

Hv

Rt

NCm

NCL

CDLHp

APH

DLL

Av

Aidd

Ai

Aidv

DLP Am

OvRt

SRt

NCL

NCm

CDLHp

APH

Nd

A 5.50A 6.00

NCL

APH

TO

NdNCL

APH

TO

Nd

1 mm A 4.50 A 4.00

NIL

LPO

PAPP

NI

NIM

HA

Hd

HVdv

HV

A 11.00

E

NCL

APH

PA

Tn Av

Hp

Aidd

Ai

CDL

Rt

Hv

NCm

DLL

L1

NCL

APH

Ap

TO

NCm

Nd

A 6.50

A 5.00

A

Figure 7: Labeling observed after a 40 nl injection of CTb into the caudal HA. A: Schematic representation of labeled fibers andcell bodies in the dorsal NCL. Injections of BDA revealed a similar amount of anterogradely labeled fibers but only very fewretrogradely labeled cells. B: Photomicrograph showing retrogradely labeled cells in the nuclei of the dorsolateral geniculatecomplex, as could be observed in some cases in which the injection site involved the thalamorecipient intercalated nucleus of theHA. Scale bar is 200 µm.


NCL

APH

TSM

L 1

PA

TnAv

Hp

Aidd

Ai

CDL

NCm

1 mm

TPO

L 1

CDL

PA

PP

Av

Ai

Aidd

AL

Hp

NCm

A 14.00

NCL

APH

TSM

L 1

PA

TnAv

Hp

Aidd

AiDIVA

DLL

CDL

SPCDLM

VIA Rt

HV

NCm

Av

Aidd

Ai

Aidv

TSM

AL

AmT

Rt

NCL

NCm

DIVA

TnAv

Aidd

Ai

Aidv

DLL

SPC

DLM

TSM

AL

VIASRt

Hv

Rt

NCm

NCL

A 5.50A 6.00

NCL

APH

ApTSM

TO

NCm

A 5.00

A 7.00A 6.50

A 7.50

NFL

HIS

HD

HA

HVHV

A

Figure 8: Results following a 13.8 nl injection of CTb into the rostral HA. A: Schematic representation of labeled fibers and cellbodies in the NCL. The projection from rostral HA is largely restricted to the anterior NCL. B: Photomicrograph showingretrogradely labeled cells in the somatosensory thalamic n. dorsalis intermedius ventralis anterior resulting from tracer spreadinto the intercalated nucleus of the HA. Scale bar is 200 µm.


two distinct subareas within NIM: injections into the NIMl confirmed the projection from DLP

(Kitt and Brauth, 1982; Gamlin and Cohen, 1986; Wild, 1994) and in addition labeled cells in

SRt, the shell of Ov and DIP. In contrast, after tracer deposits into the medialmost NI with slight

spread into the overlying HV, but not into PA or LPO, a large number of retrogradely cells

were found in the medial part of the dorsal thalamus, namely the n. dorsomedialis posterior

thalami (DMP) and DIP, and only very few within DLP.

Injections into the archistriatum

One of the most intriguing results of our retrograde labeling experiments was the extremely high

number of cells seen in the NCL that project to the archistriatum and the close match between

the distribution of these cells and the distribution of TH-immunoreactive fibers (figure 12).

After injections of CTb into the anterior two-thirds of the archistriatum that involved the central

and ventral parts of the intermediate archistriatum, cells along the whole rostro-caudal extent of

the NCL were labeled (figure 13). These cells clustered underneath the dorsal roof of the NCL

and gradually fanned out towards the medial neostriatum. This distribution thus follows closely

the original description of the NCL based on the distribution of dopaminergic and

catecholaminergic fibers (Waldmann and Güntürkün, 1993; Wynne and Güntürkün, 1995;

Metzger et al., 1996; present study) and it thus seems that all parts of the NCL, including Nd,

project upon the archistriatum.

According to the injection site within the archistriatum this projection showed an anterior to

posterior gradient with more rostral archistriatal injections labeling relatively more cells in

anterior NCL. Yet, in general, the highest absolute number of cells was seen in the caudalmost

aspect of NCL where labeled cells extend far medially and ventrally. There might exist a

further parcellation with respect to dorsal and ventral subdivisions of the archistriatum as

suggested from the anterograde tracing experiments, yet with our injections which always

involved various subdivisions of the intermediate archistriatum we could not confirm this. A

different pattern was seen after injections into the posterior archistriatum which is the source of

the tractus occipitomesencephalicus, pars hypothalami (HOM) and which together with parts of

the ventromedial archistriatum has been suggested to constitute the avian homologue of the

amygdala (Zeier and Karten, 1971, Dubbeldam et al., 1997; Davies et al., 1997). In these cases,

a small numer of retrogradely labeled cells could be observed in the periventricular rim of the

NCL and Nd (figure 14). Their number increased towards the caudal tip of the hemisphere

supporting the notion that the most caudal NCL projects to the limbic/amygdaloid part of the


1 mm

LPO

NFL

HA

HVvv

Bas

NFT

A 6.50

NCL

APH

PA

Tn Av

Hp

Aidd

Ai

CDL

Rt

HV

NCm

A 7.00Hp

NCL

APH

PA

Tn Av

Aidd

Ai

CDL

NCm

HV

A 7.50

TPO

CDL

PA

PP

Av

Ai

Aidd

NCm

A 6.00

TnAv

Aidd

Ai

Aidv

SRt

HV

Rt

NCmNCL

CDLHp

APH

A 5.50

Hp

Av

Aidd

AiAidv

DLP Am

RtSRt

NCL

NCm

CDL

APH

Nd

A 5.00

NCL

APH

Ap

TO

NCm

Nd

A 12.50

A

Figure 9: Results of a 13.8 nl CTb injection into the neostriatum frontotrigeminale overlying the n. basalis. B: Photomicrographshowing the dense anterograde labeling in the rostral and lateral aspects of NCL. C: Detail from B showing the border betweenNCL and the adjacent medial neostriatum. Scale bars are 500 µm (B) and 100µm (C).


A 9.25

LPO

PAPP

NIL

NI

NIM

HAHd HVdv

HV

LH

TnAv

Aidd

AiAidv

SRt

Hv

Rt

NCm

NCL

DMA

Hp

LH

Av

Aidd

Ai

Aidv

DLPAmDIP

Ov

RtSRt

NCL

NCm

CDL

DMP

Hp

APH

Nd

NCL

TO

APHNd

APH

TO

NCL

NdNCL

APH

Ap

TO

NCm

Nd

A 5.001 mm

A 6.00

A 5.50

A 4.00A 4.50

DIP

DLP

DMP

AC

E

A

Figure 10: Pattern of labeling observed after a 11.5 nl injection into the medialmost NIM. As shown in A fibers travel from theinjection site in a compact bundle caudally within the lamina hyperstriatica until they reach the medialmost NCL from wherethey gradually fan out to diffusely innervate the caudalmost part of NCL. B, C: Photomicrographs showing the abundantdistribution of anterogradely labeled fibers and retrogradely labeled cells in the caudalmost NCL, following injections ofbiotinylated dextran amine (B) and CTb (C) into the medialmost NIM. D: Retrogradely labeled cells in the posterior thalamus,specifically in DMP. In B and C dorsal is to the right. Scale bars are 100 µm (B, C) and 300 µm (D).


archistriatum. A very large number of cells, however, was labeled in the ventralmost

neostriatum surrounding the archistriatum dorsally and caudally.

After injections into the central and ventral archistriatum a large number of retrogradely labeled

cells was seen in the ventral hyperstriatum and the intermediate neostriatum. Other areas that

project to the central archistriatum include the caudomedial neostriatum, medial HA, rostral HD

and the vallecula, the field L complex as well as cells within NFT and NFL. Strong terminal-

like labeling was seen in LPO and medial PA as well as in HV. Descending fibers leave the

intermediate archistriatum through the tractus occipitomesencephalicus (OM), from which they

eventually split off to provide massive input to the deep layers of the optic tectum. Terminal-

like labeling was furthermore seen in the dorsal thalamus, predominantly in the posterior

thalamic nuclei DLP, DIP and SRt, all three of which also contained labeled cell bodies, as

well as in the nucleus spiriformis medialis (SPm) and the nucleus intercollicularis (ICo). At

midbrain level, terminal-like labeling and retrogradely labelled neurons could be observed in

TPc, AVT, and LoC. The OM further descended onto various nuclei in the brainstem (see Zeier

and Karten, 1971, Dubbeldam et al., 1997; Davies et al., 1997).

Injections into the posterior archistriatum labeled somata in the mediodorsal HV and along the

LFS. The Ap seemed to be reciprocally connected with cells in CDL and APH, as well as the n.

accumbens. Within the archistriatal complex, cells in the „limbic“ subcomponents, i.e. the

nucleus taeniae and Av, were labeled. The Ap was also found to project heavily onto the BNST

and tuberculum olfactorium. Descending fibers of the HOM terminated mainly within the lateral

hypothalamus (lHy). In the diencephalon cell bodies were observed in SRt and along the whole

track of the HOM. At the level of the midbrain cells could be found within TPc, AVT, and LoC.

Injections into the paleostriatum augmentatum

Injections into medial and central parts of the paleostriatum augmentatum between A 7.75 and A

9.00 labeled neurons along the complete rostrocaudal extent of NCL and Nd. As illustrated in

figure 15 these cells could be found throughout the whole depth of the NCL although their

number seemed to be highest within or just medially of the tractus archistriatalis dorsalis, at the

medial border of the TH-rich caudal neostriatum. A large number of retrogradely labeled cells

were seen in the dorsal part of the archistriatum. A smaller bilateral projection further-

Following page - Figure 11: Results of a comparatively large (23 nl) CTb injection into lateral NIM which also extended slightlyinto HV. See text for details. B: Photomicrograph of retrogradely labeled cells in the thalamus. C: Labeled fibers and cells in thedorsal NCL at about A 6.00. D: Detail showing the border of the densest innervation. Scale bars are 200 µm (B), 500 µm (C)and 50 µm (D).


DIP

Ov

Rt

NCm

Hp

APH

Nd

Av

Aidd

AiAidv

Am

SRt

DLP

A 5.50

NCL

TO

NCL

APH

Nd

A 4.00

TnAv

AiddAidv

RtSRt

DLP

Av

Aidd

AiAidv

Am

NCm

Hv

CDLHp

APH

A 6.00

NCL

APH

Nd

TO

DLP

NCL

A 4.50

TPO

Hp

PA

PP

FPL

HA

Hv

E

NCL

APH

PA

Tn

Hp

AiddAi

CDL

Rt

Hv

NCm

Av

A 6.50APH

Ap

TO

NCm

Nd

DLPDIP

A 5.00

NCL

A

Figure 11


more seemed to originate from portions of the anterior archistriatum. In addition, our injections

into PA labeled cells in HV, lateral HD, and the caudal pallium (e.g. NIL, TPO and CDL). A

few cells were also seen along the needletrack in NI, the caudomedial HV and the NCm. In the

thalamus cells were labeled within the lateral „somatic“ area of the dorsal thalamic zone,

namely DIP, VIP and to a lesser extent in DLP, SRt, n. suprarotundus (SpRt), and nuclei of the

ansa lenticularis. Sometimes marked neurons were also observed within the ventrointermediate

area of the thalamus (VIA) and n. rotundus (Rt) as well as n. triangularis (T). While labeling

within VIA was most likely due to tracer spread into the dorsal pallidum, i.e. PP (Medina and

Reiner, 1997), labeling in Rt and T derived from tracer spread into the ectostriatum (Benowitz

and Karten, 1976). At the level of the midbrain AVT and TPc also contained a number of

labeled cells.

Figure 12: Photomicrographs illustrating the close correspondence between NCL projection neurons to the archistriatum and thecatecholaminergic innervation of the caudal telencephalon. Cells in A were retrogradely labeld after a CTb injection into theintermediate archistriatum similar to that shown in figure 13. B shows an adjacent slice from the same animal processed for THimmunoreactivity. Scale bar is 1 mm.


Figure 13: Results from a large (81 nl) injection of CTb into the archistriatum which involved parts of the central and ventralintermediate archistriatum. A: Diagrammatic representation of retrogradely labeled projection cells within NCL. B:Photomicrograph showing cells in the far caudal NCL. C: Retrogradely labeled cell in the NCL following a BDA injection in theintermediate archistriatum colocalized with TH immunoreactive fibers. Scale bars are 1 mm (B) and 50 µm (C).

A 6.25

Ai

A 3.75A 4.25

A 4.75

A 5.25

A 7.25

A 6.00A 5.75

A 6.75

NCL

Nd

OMCDL

APH

Aidd

Av Aidv

Tn

AVT

DIP

DLP

HV

Am

SRt

PA

NCm

APH

OM

OMOM

NCL

NCL

NCL

Nd

OM

CDL

APH

Aidd

Av

AidvTnDIP

DLP

HV

Am

PA

NCm

APH

OM Av

Aidd

Aidd

NCL

DLP

Nd

Nd

NCm

SRt

DIP

AiAi

APHCDL

SGC

SACSGP

SGC

SACSGP

SGC

SACSGPPT

SPL

SCI

DMP

ICo

DMP

OM

NCL

APH

APH

APH

CDL CDL

NCmNCL

Aidd

AiTn

L1L1L1

NCm

Aidd

Ai

Av Av

PA

HV

HV

NCLNCL

CDL

Av

Ov

Tn

Am

NCm

HV

A


Injections into the medial lobus parolfactorius

Labeling within NCL after deposits of CTb into medial LPO between A 9.50 and A 11.50 was

very similar to that observed after PA injections. Again, retrogradely labeled neurons were

distributed diffusely along the medioventral and rostrocaudal extension of NCL, yet cells were

predominantly located along the medial border (figure 16). This was also true for two cases in

which the injection accidentally was centered onto lateral LPO and PA. Regarding a possible

rostrocaudal topography of NCL projections to the basal ganglia, we found a relative larger

number of retrogradely labeled neurons projecting to the LPO in caudal NCL while neurons

projecting to the PA tended to be located more rostrally and medially. However, as noted

APH

NCL

Nd

CDL

A 5.25

Hp

CPi

SPC AiAm Aidd

Aidv

NCL

APH

AVT

Hp CDL

A 4.25

HOM

NCL

Hp

SRt

SPC

Nd CDL

CPi

APH

A 5.75

AiddAidvAm

Ai

Av

NCL

APH

LHy

Ap

NdHp

A 4.75

CDL

Figure 14: Schematic representation of labeling in the caudal telencephalon after a deposit of 15 nl CTb into the posteriorarchistriatum. All injections into the archistriatum were placed using a lateral approach. Cells within NCL project only relativesparsely onto Ap. Their number increases towards the caudal pole of the hemisphere. A strong projection arises from cellsimmediately dorsal of the archistriatum. This area seems to intersperse between NCL and archistriatum, in contrast to the lattertwo it also labels considerably weaker for TH (compare also figure 1).


above, these projections largely overlapped. Labeling in other areas of the telencephalon

included the NIM and HV dorsal to the injection site, medial HA and HD, and the area

prehippocampalis (APrH). Many cells were also seen in TPO, CDL and up to far rostral levels

within NFL, but only few in NIL. In the archistriatum large parts of the Av and Ai were labeled.

The projection from the Av seemed to be bilaterally organized. We found that in the thalamus

CTb injections which were centered on medial LPO yielded numerous neurons within n.

dorsomedialis anterior thalami (DMA) and DMP, and to a lesser extent also in the n.

subhabenularis (SHL) and DIP. In the mesencephalon cells were found in AVT, TPc and LoC.

DISCUSSION

The main results of this study can be summarized as follows: The NCL has reciprocal

connections with the secondary sensory areas of all modalities, as well with at least two

parasensory areas in the NIM and the deep layers of HV. The afferents from these areas have

diffuse and largely overlapping termination fields within NCL (see figure 17). These

connections are restricted to the plexus of the densest innervation by catecholaminergic and

presumably dopaminergic fibers as demonstrated here by TH immunoreactivity. Furthermore,

all parts of the so defined NCL project onto the basal ganglia and the archistriatum. The

importance of these findings for sensorimotor integration in the avian telencephalon and their

possible relevance for the evolution of cortex-equivalent structures will be discussed

separately in detail below.

Sensory integration within NCL

All five secondary sensory areas were found to project to NCL. The organization of these

pathways conforms with a general pattern of sensory processing in the avian telencephalon

(summarized in Veenman et al., 1995b): the primary receptive fields of subtelencephalic input

Following pages:Figure 15: Results following an injection of 28 nl CTb into the „somatic“ central PA. A: Schematic representation of retrogradelylabeled cells in the caudal telencephalon. The most abundant labeling can be seen in the dorsal part of the archistriatum and therostral NCL. See text for details. B: Photomicrograph illustrating the location of projection neurons at the medial border of theNCL and within Aidd. C: Photomicrograph of labeling in the posterior dorsal thalamus. In addition to labeled neurons within DIP,cells could often be observed in the VIP, DLP and SRt. Scale bars are 500 µm (B) and 200 µm (C).

Figure 16: Projections to the limbic medial striatum. Results of a 22 nl injection of CTb into the medial LPO with no obvioustracer spread into AC or lateral LPO. Labeling in the caudal telencephalon was complementary to that observed after injectionsin the somatic parts of the basal ganglia. Most abundant labeling was seen in the ventral part of the intermediate archistriatumand the caudal NCL. B: Photomicrograph showing the location of NCL projection neurons (arrows) at the medial border of thetractus archistriatalis dorsalis as well as in Ai and Av in a slice double labeled for TH. The DA shows only weak staining forTH. C: Retrogradely labeled cells in the posterior neostriatum cluster at the medial extent of the NCL. D: Photomicrograph ofretrogradely labeled thalamic neurons within DMA.


TPO

NIM

Hp

PAPP

FPL

HA

HV

A 7.00

NCL

APH

PA

Tn Av

Hp

AiddAi

CDL

NCm

A 7.50

TPO

CDL

PA

PP

AvAi

Aidd

Hp

NCm

NCL

APH

PA

TnAv

Hp

Aidd

Ai

CDL

Rt

Hv

NCm

TnAv

Aidd

AiAidv

SRt

Hv

Rt

NCm

NCL

CDL

DMA

Hp

APH

Av

Aidd

AiDLPAmDIP

OvRt

SRt

NCL

NCm

CDL

DMP

Hp

APH

Nd

NCL

APH

TO

NdNCL

APH

TO

Nd

A 8.25

A 6.50 A 6.00A 5.50

A 4.50 A 4.00

DLPDIP

AVT

1 mm

NCL

APH

Ap

TO

NCm

Nd

A 5.00

A

Figure 15


NCL

APH

L 1

PA

Tn Av

Hp

AiddAi

CDL

NCm

FPL

LPO

PA

PP

NIL

NI

NIM

HA

Hd HVdv

HVL 1 TPO

CDL

PA

PP

Av

AiAidd

OM

Hp

NCm

NCL

APH

L 1

PA

Tn Av

Hp

Aidd

Ai

CDL

SPCDLM

VIA

HV

NCm

FPL

DMA

TnAv

Aidd

Ai

Aidv

HV

Rt

NCm

NCL

CDL

DMA

Hp

APH

SRt

SPC

DLPDIP

Ov

Rt

SPC

NCL

NCm

CDL

DMP

Hp

APH

Av

Aidv AiddAiAm

SRt

A 11.25 A 7.50

NCL

APH

OM

TO

Nd

NCL

TO

OM

APH

Nd

NCL

APH

Ap

TOOM

NCm

Nd

A 7.00

A 6.50 A 6.00A 5.50

A 4.00A 4.50A 5.001 mm

AVT

DIP

DMP

DLP

HVHV

A

Figure 16


relay the information to adjacent secondary sensory structures, which in turn project to areas of

the external pallium. In the case of the tectofugal visual pathway the ectostriatum is the primary

sensory structure (Kondo, 1933, Benowitz and Karten, 1976). From there intratelencephalic

projections lead to the ectostriatal belt (Karten and Hodos, 1970; Watanabe et al., 1985). The

present study demonstrates in accordance with Leutgeb et al. (1996) that the NCL is the tertiary

telencephalic component of the tectofugal complex. This pattern also holds for the other sensory

systems. Caudal IHA and the caudolateral component of HD are the primary sensory areas of

the thalamofugal visual pathway (Karten et al., 1973), from where projections lead to HA

(Shimizu et al., 1995), which by itself projects to NCL (Shimizu et al., 1995, Leutgeb et al.,

1996). More rostrally, somatosensory thalamic fibers terminate in HD/HIS and IHA (Delius

and Bennetto, 1972, Funke, 1989, Wild, 1997) which in turn project to rostral HA (Wild,

1987b). Field L2 is the primary telencephalic area of the auditory system (Wild et al., 1993),

which projects to L1 and L3, from where efferents lead to Nd/NCL (Bonke et al., 1979, Wild et

al., 1993, Leutgeb et al., 1996). N. basalis is the primary telencephalic area of the trigeminal

system (Schall et al., 1986), which projects to NFT (Wild et al., 1985; Wild and Farabaugh,

1996), from where efferents lead to NCL (Schall et al., 1986, Wild et al., 1985; Wild and

Farabaugh, 1996). Thus, the NCL is the tertiary telencephalic component of all sensory

modalities examined in the present study. Due to this multimodal input and the broad overlap of

terminations, the NCL is a true associative forebrain structure. Additionally, parts of the caudal

neostriatum may also have access to olfactory information: after injections into the posterior

archistriatum we found a large number of cells in the ventralmost neostriatum overlying the

archistriatum. This region has been suggested to be reciprocally connected with the avian

piriform cortex and olfactory bulb, which, as their mammalian counterparts process olfactory

information (Reiner and Karten, 1985; Haberly and Bower, 1989; Bingman et al., 1994). In

concert with the CPi and limbic archistriatum this area might thus be concerned with

viscerolimbic functions. However, from our combined immunocytochemical and pathway

tracing data it seems that this part of the NC is not a genuine subdivision of the NCL, although it

might be ventrally continuous with it. Olfactory information can reach the NCL nevertheless via

its connections with the HD. In addition to afferents from the visual dorsolateral thalamus

(Karten et al., 1973; Güntürkün and Karten, 1991; Güntürkün et al., 1993) and non-specific

input from the mediodorsal thalamus (Bagnoli and Burkhalter, 1983), the HD is also

reciprocally connected with the hippocampal formation and the CPi (Bingman et al., 1994;

Casini et al., 1986; Shimizu et al., 1995).


The pattern of afferents within NCL might provide clues to the functional architecture of the

pigeon´s nervous system. The only two pathways which show no or little overlap within NCL

are those from the auditory and the trigeminal systems. Indeed, conditioning paradigms have

shown that pigeons have severe constraints to associate auditory signals with food reward in an

appetitive learning paradigm in which pecks on a key serve as an operant (LoLordo and

Furrow, 1976, Delius and Emmerton, 1978). The same animals are rapidly able to associate in

a classical conditioning paradigm the very same auditory signals with a mild shock applied to

the body (Delius and Emmerton, 1978). Delius and Emmerton (1978) speculated that a

granivorous animal like a pigeon is simply not in need of associating acoustic cues with food

objects since grains are mostly silent. However, grains can be quite noisy during pecking due to

the direct auditory feedback generated by the impact of the beak on the substrate. Indeed, Delius

(1985) could show that pigeons are easily able to learn an auditory tone discrimination in an

appetitive paradigm if the acoustic signals are generated by the pecks. Yet, delaying the

acoustic feedback by more than 0.5 s abolishes learning. Since the n. basalis not only receives

trigeminal projections but also some auditory afferents from the lemniscal nuclei (Delius et al.,

1979, Schall et al., 1986, Arends and Zeigler, 1986; Wild and Farabaugh, 1996) it is likely that

the specialised auditory-trigeminal associations occuring during pecking are directly processed

within the basalis system. Indeed, lesions of the n. basalis eliminate the ability to discriminate

acoustic signals generated by the animals own pecks (Schall and Delius, 1991). Thus, it is

conceivable that the separation of auditory and trigeminal afferents within NCL is related to the

severe constraints of pigeons in associating acoustic and trigeminal stimuli. Both the tectofugal

and the thalamofugal systems project to NCL, but their common area of termination seems

limited. This fact might be related to differences in visual field representations between the

thalamo- and the tectofugal pathway. The retina of pigeons has two areas of enhanced vision

(Galifret, 1968, Binggeli and Paule, 1969). One is the central fovea which looks into the

monocular lateral field. The other is the so called “red field“ in the dorsotemporal retina with

which the animal views stimuli in the inferior frontal visual field (Güntürkün, 1998). The

representation of the red field within the GLd of the thalamofugal pathway is extremly sparse,

while the central fovea is heavily represented (Remy and Güntürkün, 1991). Thus, the

thalamofugal pathway mainly processes stimuli from the lateral rather than the frontal visual

field. Accordingly, thalamofugal lesions produce minor deficits when tested with frontal

stimuli, but severe deficits when tested with lateral stimuli (Güntürkün and Hahmann, 1998).

Contrarily, there is evidence that tectofugal rotundus lesions impair frontal acuity, while


Figure 17: Schematic representation of terminal fields within NCL showing the large overlap of afferents from secondarysensory areas. Note that the afferents from the presumed parasensory areas within NIM, which occupy large parts of thecaudalmost NCL are not included in this figure. See text for details.

1 mm

A 6.50

A 6.00

A 5.50A 5.00

A 7.50

visual afferents from Ectostriatal belt

trigeminal afferents from frontal neostriatum

somatosensory afferents from rostral HA visual afferents from caudal HA

auditory afferents from Field L

"border" of the NCL as defined by dopaminergic innervation

A 7.00


leaving lateral acuity intact (Güntürkün and Hahmann, 1998). This in turn may be due to the fact

that ventral tectal cells, which represent the lower frontal visual field, project heavily onto

rotundus, while the contribution of dorsal tectal cells to the tectofugal pathway is limited

(Hellmann and Güntürkün, 1997). Thus, thalamo- and tectofugal visual pathways in pigeons

seem to differentially represent lateral and frontal vision, respectively. Their largely

complementary representation within NCL might thus be related to their differential

representation of the visual field. Despite the small overlap of the tecto- and the thalamofugal

visual fields within NCL, these domains extensively overlap with the trigeminal and the

auditory projections, respectively. The large common territory of the tectofugal visual and the

trigeminal system might be related to the specialization of the tectofugal pathway to the lower

and frontal visual field (Hellmann and Güntürkün, 1997, Güntürkün and Hahmann, 1998) which

within the egocentric space overlaps with trigeminal inputs from the beak. Thus, a common

sensory focus of both systems could create a need for common coding which might be

accomplished by extensive areas of terminal overlap in an associative structure like NCL. The

large overlap between thalamofugal visual and auditory domains might be similarly interpreted.

As outlined above, the thalamofugal pathway in pigeons is oriented laterally (Remy and

Güntürkün, 1991, Güntürkün and Hahmann, 1998). Like most other birds, pigeons fixate distant

objects laterally with their central fovea (Blough, 1971, Martinoya et al., 1981, Bischof, 1988),

since their lateral monocular acuity is about twice as high as their frontal monocular one

(Hahmann and Güntürkün, 1993, Güntürkün and Hahmann, 1994). Thus, the thalamofugal visual

system is more related to distant visual objects and might therefore need common processing

with audition, which also is a distance sense.

Connections of the NCL with the NIM and ventral hyperstriatum

Our retrograde and anterograde tracing experiments have suggested the existence of two

different, albeit continuous, sources of afferents from NIM to NCL. While the NIMl projection

innervates the Nd region and the laterally adjacent parts of the NCL, afferents from NIMm

terminate abundantly throughout the caudalmost NCL. Both areas receive differential input from

the dorsal thalamus (Kitt and Brauth, 1982; Gamlin and Cohen, 1986; Wild, 1987a, 1994;

Metzger et al., 1996; present study) and differ in their cytoarchitectonic characteristics

(Rehkämper et al., 1985). The NIMl has been shown to receive visual and somatosensory

information from telencephalic areas (Funke, 1989; Wild, 1987b, 1994; Shimizu et. al., 1995;

present study). Since we found retrogradely labeled cells in L1 and L3 after injections into


NIMl, this structure probably also integrates auditory information. In addition, polysensory

input reaches the NIMl via DLP (Kitt and Brauth, 1982; Gamlin and Cohen, 1986;

Korzeniewska and Güntürkün, 1990; present study).

The NIMm, as defined here, resembles in its location and connectivity the mediorostral

neostriatum/hyperstriatum ventrale (MNH) of the chick which has been extensively studied in

the context of imprinting (e.g. Metzger et al., 1996; Bredenkötter and Braun, 1997; Gruss and

Braun, 1996; for review: Scheich, 1987). Injections into NCL which labeled cells in NIMm

also always produced cells in the overlying HV, suggesting a similar translaminar organization

of this area in the pigeon as is characteristic for the MNH. Data on the connections of either the

MNH or NIMm with telencephalic regions are very limited (Metzger et al., 1996). Our results

indicate that the NIMm, as the NIMl, might also receive input from the visual and

somatosensory areas of the Wulst. Thalamic input to NIMm originates mainly in the DMP and to

a much lesser extent in the DMA, a situation which is reversed for the MNH of chicks (Kitt and

Brauth, 1982; Wild, 1987a; Metzger, et al., 1996; present study). The sensory properties of

these afferents are not entirely clear, but the nuclei of the dorsomedial thalamus receive input

from the hypothalamus and tractus solitarius (Berk and Butler, 1981, Berk and Hawkin, 1985,

Arends et al., 1988; Wild et al., 1990) and in turn project to the viscerolimbic striatum and CPi

(Metzger et al., 1996; Bingman et al., 1994; Kitt and Brauth, 1982; Wild, 1987a; Veenman et

al., 1995a, 1997). Furthermore, the NIMm and to a lesser extent also the NCL of pigeons

display strong immunoreactivity for the calcium-binding protein parvalbumin (S. Kröner,

unpublished observations), as has previously been shown for the MNH of chicks (Braun et al.,

1991a). Parvalbumin is associated with fast spiking GABAergic neurons (DeFelipe et al.,

1989; Kawaguchi and Kubota, 1993), and in songbirds it is characteristic of nuclei displaying

high metabolic activity, notably all nuclei of the vocal motor system (Braun et al., 1985,

1991b). As outlined above, both areas of the NIM appear capable of multisensory processing.

This draws attention to our observation that their afferents terminate predominantly in those

aspects of the NCL which receive relatively little input from secondary sensory regions

(compare figure 17). Although it is possible that other telencephalic areas which send

projections to NCL (e.g. HA or NFT) are also parasensory in nature (Deng and Wang, 1992,

1993; Schall et al., 1986; Wild and Farabaugh, 1996) the nuclei within NIM clearly are a step

progressed in an assumed hierarchy of stimulus processing, as they, like the NCL, receive their

afferents from secondary sensory structures. It thus seems that parasensory inputs from NIM to

the NCL complement the pattern of multimodal integration.


It should be noted, that our injections into NCL also labeled cells in the HV which based on

their location resemble the intermediate and medial part of the hyperstriatum ventrale (IMHV),

another area which in the chick has been extensively studied in the context of imprinting and

memory formation (Bradley et al., 1985; Davies et al., 1988; review by Horn, 1998). Although

we did not target these cells in the anterograde tracing experiments, results from the study of

Shimizu and coworkers (1995) suggest that afferents from the HV also terminate abundantly

throughout the caudalmost neostriatum (Shimizu et al., 1995).

Projections to the archistriatum

We found a continous band of cells in the NCL which project in a topographically ordered

manner onto most parts of the intermediate archistriatum. In addition, the posterior archistriatum

receives a weak projection from cells in the caudal- and dorsalmost aspects of the NCL. The

connections with the intermediate archistriatum are reciprocally organized (Davies et al., 1997;

Leutgeb et al., 1996; Metzger et al., 1998; present study) and bilaterally projecting cells in the

Av provide a means of interhemispheric comparison (Wild and Farabaugh, 1996; Metzger et

al., 1998; present study). Functionally, the avian archistriatum has generally been divided into

two main subdivisions (Zeier and Karten, 1971; Davies et al., 1997; Dubbeldam et al., 1997).

These are a somatic sensorimotor part which in the pigeon comprises parts of the anterior and

intermediate archistriatum (Aa, Ai, Aidd, Aidv), and a viscerolimbic division that includes the

posterior and medial archistriatum, as well as the ventral parts of the intermediate archistriatum

(Ap, Am, Av). The sensorimotor archistriatum receives widespread afferent projections from

higher order sensory areas of the telencephalon (e.g., Ritchie, 1979; Wild et al., 1985, 1993;

Shimizu et al., 1995) and is involved in high-level motor control (Zeier, 1971; Knudsen et al.,

1995) and memory function (Knudsen and Knudsen, 1996). Furthermore, via the OM the

archistriatum projects to premotor areas of the brainstem (e.g. reticular formation and lateral

pontine nuclei) in all birds (Zeier and Karten, 1971; Wild et al., 1985, 1993; Davies et al.,

1997; Dubbeldam et al., 1997; present study) and to some specific groups of brainstem

motoneurons involved in respiration and vocalization in songbirds (Nottebohm et al., 1976,

Wild, 1993). The limbic portion of the archistriatum on the other hand is considered

homologous to the mammalian amygdala (Zeier and Karten, 1971; Davies et al., 1997;

Dubbeldam et al., 1997). It seems to be crucially involved in such viscerolimbic functions as

agonistic behaviour and homeostasis. (e.g. Cohen, 1975; Ramirez and Delius, 1979; Lowndes

and Davies, 1995). The archistriatum´s functional segregation is reflected in the


extratelencephalic projections via the OM or HOM, respectively (Zeier and Karten, 1971;

Davies et al., 1997; Dubbeldam et al., 1997), and in its connections with limbic or somatic

striatum (Veenman et al., 1995b; present study). While the main output from the NCL might be

directed to the sensorimotor part of the archistriatum (Leutgeb et al., 1996), especially the

caudalmost NCL also appears to be capable of modulating the amygdaloid division of the

archistriatum (present results). The very large number of neurons from NCL that project to the

archistriatum, and the close correspondence between the position of these output neurons and

the densest catecholaminergic innervation, underline the importance of this projection for an

understanding of the functions of the NCL. In addition, we believe that the projection from the

caudal neostriatum to the archistriatum might serve as a criterion for the delineation of NCL.

Given the fact that injections into the archistriatum labeled a continuos band of cells that

extended from the ventrolateral NCL up to Nd, we think it reasonable to consider Nd the

´auditory subcomponent` of NCL. This view is also supported by the facts that the Nd shares

afferents with the adjacent NCL from areas other than the field L complex, shows the same

dense dopaminergic innervation and shows a similar organization of efferent connections other

than those to the archistriatum.

Projections to the basal ganglia

The connectivity, neurotransmitter content, and cytoarchitecture of the avian basal ganglia are

highly similar to that in mammals (Karten and Dubbeldam, 1973; Reiner et al., 1984; Veenman

and Reiner, 1994; Veenman et al., 1995b). In mammals corticostriatal inputs from association,

sensorimotor and limbic cortices project in a segregated manner onto three distinct striatal

regions referred to as associative, sensorimotor and limbic striatal territories (Parent, 1990;

Parent and Hazrati, 1995). These corticostriatal inputs provide the basal ganglia with

exteroceptive sensory information as to the location of objects in space, interoceptive sensory

information on relative body position in space, and neural feedback and feedforward

information on ongoing and impending body movements (McGeorge and Faull, 1989; Alexander

et al., 1990; Romo et al., 1992; Schultz et al., 1992; Aldridge and Berridge, 1998). Based on its

palliostriatal connections a similar functional segregation has recently been suggested for the

avian striatum (Veenman et al., 1995b): ventral striatal structures such as the AC, the BNST,

and olfactory tubercle constitute the „limbic“ parts of the avian striatum. These „limbic“ parts

also include the medial LPO and lateralmost PA. The remaining dorsal striatal parts (lateral

LPO, medial PA) on the other hand are considered sensorimotor in nature. We show here that


all parts of the NCL project to the ventral and dorsal aspects of the striatum. Our combined

retro- and anterograde tracing experiments suggest a rostrocaudal topography of these

projections, with caudal parts of NCL projecting predominantly onto limbic medial LPO and

more rostral parts of NCL projecting to lateral LPO and mainly medial PA (figure 18). The

projection onto the limbic striatum had already been indicated by the study of Veenman and

coworkers (1995b); but as their injections also involved the CPi and Ap (see their figure 18

and discussion), which also project heavily onto the limbic parts of the avian basal ganglia

(Bingman et al., 1994; Veenman et al., 1995b; present study), until now the existence of this

connection from the NCL was not entirely clear. In addition to this direct pathway, the NCL is

also capable of influencing the ventral basal ganglia via its connection with the limbic

archistriatum. The ventral striatum of mammals receives input from limbic cortical regions such

as hippocampus, cingulate cortex, amygdala and piriform cortex (McGeorge and Faull, 1989;

Alexander et al., 1990), and similar projections have been shown for the avian homologues of

these structures (Bingman et al., 1994; Veenman et al., 1995b). In mammals, a further prominent

input to the ventral striatum originates in the frontal cortex (Berendse et al., 1992, Sesack et al.,

1989; Alexander et al., 1990), and particularly the prefronto-accumbal network plays a crucial

role in motivated behaviour (Apicella et al., 1991; Schultz et al., 1992; Floresco et al., 1997).

The demonstration of a projection from caudal NCL onto the AC thus further strengthens the

notion that the NCL might be functionally equivalent to the PFC (Veenman et al., 1995b; present

study). As in mammals, this projection might provide the sensory information required, e.g. for

the evaluation of appetitive and aversive stimuli. In addition to the projection upon limbic

striatum we found a prominent projection onto large parts of PA and lateral LPO which

comprise the somatomotoric divisions of the avian striatum (Veenman et al., 1995b), providing

the NCL with a further direct pathway for movement initiation. It is unclear why the projection

from NCL to the somatic PA was not described by Veenman and coworkers (1995b), but it

seems possible that this reflects differences in the sensitivity of the different tracers used, or,

more likely, that their injections into PA were not centered on those regions in which we found

the most abundant anterograde labeling, namely the caudal aspects of the central and lateral PA

(compare their figures 5 and 6 with figures 2 and 15 of the present study).

The palliostriatal projections form asymmetric and probably excitatory synapses with spiny

striatal neurons (Veenman and Reiner, 1996), a situation comparable to that in mammals (Dube,

et al., 1988). Based on these and similar findings large parts of the external pallium, including

the NCL and the archistriatum, have been compared with a subpopulation of corticostriatal


projection neurons from layer III and V in the rat (Zeier and Karten, 1971; Veenman et al.,

1995b; Akintunde and Buxton, 1992; Cowan and Wilson, 1994). In sum, the NCL is positioned

to amalgamate sensory inputs from all modalities and to relay them to the sensorimotor division

of the basal ganglia, providing the latter with information that has been further processed and

integrated with input from other modalities.

Comparison with vocal control pathways in songbirds:

In songbirds much work has focused on the pathways that control the acquisition and production

of learned song. Within oscines, the forebrain vocal system comprises two major pathways that

converge on the same premotor nucleus of the archistriatum. The two pathways diverge from a

nucleus in the caudolateral neostriatum, the „high vocal center“ (HVC, Nottebohm et al., 1976,

1982). The main descending vocal motor pathway leads from the HVC to the robust nucleus of

the archistriatum (RA), which in turn projects onto mesencephalic and medullary nuclei

involved in vocalization (Nottebohm et al., 1976, 1982; Bottjer et al., 1989; Vates et al., 1997).

rostral NCL

Ai

Av

caudal NCL

PA

PA

lateral LPO/ PA

medial LPO/ventral striatum

Figure 18: Diagrammatic illustration of the organization of presumed „limbic“ (grey) and „somatic“ (black) pathways from theNCL to the archistriatum and the basal ganglia. All parts of the NCL send projections to both limbic and somatomotor areas, butwith a complementary rostral to caudal topography (see text for details and compare figures 15 and 16). The relativecontribution of the caudal and rostral NCL to these pathways is indicated by the width of the arrows. In addition to directconnections indirect pathways exist, as limbic and somatomotor parts of the archistriatum also project to corresponding areaswithin the basal ganglia.


HVC also projects to RA by a second pathway that involves nuclei in the anterior forebrain.

This circuit sequentially connects HVC, area X of the LPO; the dorsolateral thalamus, the

magnocellular nucleus of the anterior neostriatum (MAN) and RA (Bottjer et al., 1989; Vates et

al., 1997). Auditory input is thought to reach both major pathways of the vocal system via HVC

and the adjacent „shelf“ region, which receive auditory information from the field L complex

and associated regions of the caudal forebrain (Kelley and Nottebohm, 1979; Fortune and

Margoliash, 1995; Vates et al., 1996). In nonsongbirds the projection from field L 1 and L3 to

the Nd/NCL is similar to this ascending auditory pathway in oscine songbirds (Wild et al.,

1993; present study), while the projection from NIMl to the Nd/NCL has been compared with

another polysensory projection upon HVC which arises from the nucleus interfacialis

(Nottebohm et al., 1982; Wild, 1994; Fortune and Margoliash, 1995). The present

demonstration of a prominent input from NIMm to Nd/NCL suggests a correspondence of this

pathway with yet another ascending projection onto HVC: in oscine songbirds MAN shares a

similar location in the intermediate neostriatum with the NIM. The MAN can also be divided

into a lateral (lMAN) and a medial (mMAN) part which receive differential input from the

medial portion of the dorsolateral thalamus (DLM) and the DMP, respectively (Bottjer et al.,

1989; Johnson, et al., 1995; Foster et al., 1997; Vates et al., 1997), thus resembling the thalamic

projections to NIMl and NIMm (Kitt and Brauth, 1982; Wild, 1987a; present study). The

mMAN then gives rise to a projection onto HVC and the adjacent „shelf“ region (Nottebohm et

al., 1982; Fortune and Margoliash, 1995; Vates et al., 1996, 1997; Foster et al., 1997). A

similar recursive pathway through the anterior forebrain also exists in budgerigars: the central

nucleus of the lateral neostriatum (NLc) integrates information from other forebrain nuclei and

relays it to the premotor neurons of the archistriatum (Brauth et al., 1994; Striedter, 1994;

Durand et al., 1997). It has previously been suggested that the HVC complex, as might be true

for the NLc of parrots, is concerned not only with translating auditory signals into vocal output

but with the integration of multimodal inputs, and that the song system arose from an elaboration

of pathways generally present in all birds (e.g. Wild, 1994; Margoliash et al., 1994). The fact

that the shelf region adjacent to HVC shares the afferent inputs of HVC (Fortune and

Margoliash, 1995; Vates et al., 1996; Foster et al., 1997) further strengthens the notion that

HVC represents a specialized nucleus for the control of learned song, while neural populations

in the adjacent neostriatum might be concerned with other aspects of multisensory integration

and motor control. However, judging from the presently available data important differences

seem to exist: the HVC of Passeriformes, or NLc of parrots, respectively, do not receive


similarly abundant sensory input as described for NCL (e.g. Hall et al., 1993; Brauth et al.,

1994; Striedter, 1994; Fortune and Margoliash, 1995; Leutgeb et al., 1996; present study).

Furthermore, in pigeons the NCL and the archistriatum have direct reciprocal connections with

all sensory and parasensory areas investigated here, and in addition both project onto the

striatum (Veenman et al., 1995b; present study).

Comparison with mammalian prefrontal cortex

The PFC of mammals is actually not a single functional region but, rather, a group of

anatomically and functionally heterogenous areas (e.g. Goldman-Rakic, 1987). It is commonly

thought to involve the dorsolateral PFC, as well as the orbitofrontal and anterior cingulate

cortices (Walker, 1940; Akert, 1964; discussed in Preuss, 1995; Condé et al., 1995). These

areas participate to different extents in the complex cognitive, limbic and premotor functions

that the PFC subserves, depending on their respective afferent and efferent connections (Sesack

et al., 1989; Pandya and Yeterian, 1990). Thus, in primates, deficits resulting from damage to

the lateral frontal region involve mainly attentional, temporal and integrative functions and are

different from changes relating to emotion following orbitofrontal, and motivational changes

following medial frontal damage (summarized in Fuster, 1989). Sensory input reaches the PFC

via a set of interconnected pathways (for reviews: Jones and Powell, 1970; Pandya and

Yeterian, 1990). The primary sensory area of each modality projects first to an adjacent area in

parietal, occipital, or temporal cortex. This is the beginning of a sequential order of cortical

areas that make up a pathway for that modality. Each area in the sequence projects not only to

the next in line but also to a discrete area of the frontal cortex, which in turn reciprocates by

sending fibers back to the projecting area. The fields that constitute the third link in each of the

three pathways - namely, parietal area 7 (somatic), temporal area 22 (auditory), and

inferotemporal area 21 (visual) - project to the prefrontal cortex and, in addition, to the cortex

in the dephts of the superior temporal sulcus, which constitutes another multimodal area.

Although the termination fields show considerable overlap, especially around the principal

sulcus of primates, a strong topography of these afferents exists, resulting in a distinct pattern of

connections for each of the subdivisions of PFC (primates: Jones and Powell, 1970; Pandya

and Yeterian, 1990; rodents: Condé et al., 1995; Reep et al., 1996).


Taken together, the organization of sensory input to the NCL strongly resembles the pattern

described above for mammalian PFC (Leutgeb et al., 1996; present study). In addition, it is

reciprocally connected with the avian homologue of the amygdala (present study), a connection

which is also characteristic for mammalian PFC (Condé et al., 1995; Reep et al., 1996; Sesack,

et al., 1989; Barbas and DeOlmos, 1990), and projects to most parts of the somatic and limbic

striatum, as well as the sensorimotor archistriatum (c.f. figure 19; Veenman et al., 1995b;

Leutgeb et al., 1996). Furthermore, the NCL and PFC share a dense dopaminergic input (Divac

and Mogensen, 1985, Waldmann and Güntürkün, 1993, Divac et al., 1985, 1994, Wynne and

Güntürkün, 1995), and a functional importance for processes which serve executive functions

(working memory: Mogensen and Divac, 1982, 1993, Gagliardo et al., 1996, 1997, Güntürkün,

1997; behavioral flexibility: Hartmann and Güntürkün, 1998; behavioral inhibition: Güntürkün,

1997). With regard to memory functions it has been proposed that the PFC performs a unitary

function, that of holding information gathered by the other association cortices „on-line“ in short

term memory (Goldman-Rakic, 1987; Kimberg and Farah, 1993). Judging from the anatomical

LPO

TeO

LHy

thalamus

brainstem

Acc/BNST

OM HOM

IHA

Ai

OMDLP/DIP

vis. HA

som. HA

BasNFT

PA

PP

L2

L1/L3NCLNIMIMHVEp

E

Ap

Figure 19: Diagrammatic summary of the circuitry in which the NCL participates as it emerges from this study. The NCL isreciprocally connected with the secondary sensory of all modalities as well as at least two parasensory areas. It projects to theavian equivalents of the amygdala (Ap, Av) and premotor cortex (Ai), as well as limbic and somatic parts of the avian striatumwhich taken together by means of their long descending projections exert control over limbic and motoric centers in thediencephalon, midbrain and brainstem.


data and the behavioural studies cited above the NCL seems well capable to occupy the

position of such a central excecutive in the avian brain.

However, at least one important difference hampers comparisons between NCL and PFC: As a

common hallmark for the delination of the PFC across mammalian species serve the afferents

from the mediodorsal (MD) nucleus of the thalamus (Rose and Woolsey, 1948; Akert, 1964;

Preuss, 1995). Thalamic afferents to the NCL arise mainly from DLP and surrounding

structures. DLP, however, is very likely not the avian version of mammalian MD. Based on its

afferents and electrophysiological properties, DLP was suggested to be equivalent to the

posterior complex of nuclei and especially its suprageniculate component (Korzeniewska,

1987, Korzeniewska and Güntürkün, 1990). Like the posterior nuclei, DLP is dominated by

afferents from the vestibular nuclei (mammals: Mickle and Ades, 1954; birds: Wild, 1988,

Korzeniewska and Güntürkün, 1990), dorsal column nuclei (mammals: Feldman and Kruger,

1980, Berkley et al., 1986; birds: Funke, 1989, Wild, 1989, Korzeniewska and Güntürkün,

1990), superior colliculus (mammals: Hicks et al., 1986, Katoh and Benedek, 1995; birds:

Gamlin and Cohen, 1986, Korzeniewska and Güntürkün, 1990), and reticular formation

(mammals: Hicks et al., 1986; birds: Korzeniewska and Güntürkün, 1990). Amygdalar or

hypothalamic afferents to mammalian posterior thalamic nuclei (Jones, 1985) or avian DLP

(Korzeniewska and Güntürkün, 1990) have not been reported. Contrary to this pattern,

mammalian MD is reported to receive, among others, afferents from the amygdaloid complex,

diagonal band of Broca, and different hypothalamic and preoptic structures (Siegel et al., 1977,

Sapawi and Divac, 1978, Velayos and Reinoso-Suarez, 1982, Irle et al., 1984, Russchen et al.,

1987, Cornwall and Phillipson, 1988, Groenewegen, 1988). Based on this evidence, any

general connectional similarity between MD and DLP could be denied. This view results from

a comparison between DLP and the whole MD. If, however, DLP is compared with only lateral

MD (rats: paralamellar MD portion, Cornwall and Phillipson, 1988, Groenewegen, 1988; cats:

intermediate and lateral MD component, Velayos and Reinoso-Suarez, 1982; monkeys:

parvocellular and multiform lateral MD portion, Russchen et al., 1987) the picture changes. The

lateral MD portion does not receive afferents from the amygdala but is reached by inputs from

vestibular nuclei, superior colliculus, and reticular formation. It thus resemble the pigeon's DLP

in many aspects. It is the lateral portion of MD in cats and monkeys which heavily projects onto

the dorsolateral component of prefrontal cortex (Markowitsch and Pritzel, 1979, Giguere and

Goldman-Rakic, 1988, Ray and Price, 1993), which is known to be of prime importance in

delay tasks. The paralamellar MD in rats is innervated by deep layers of superior colliculus


and projects to medial precentral cortex (Groenewegen, 1988), an area which has been

compared with the frontal eye field in the monkey (Reep, 1984), a caudal portion of prefrontal

cortex (Fuster, 1989; Pandya and Yeterian, 1990). But despite these similarities between lateral

MD of mammals and avian DLP, important differences also exist. For example, the pigeon's

DLP receives a prominent innervation by the cuneatus-gracilis complex and consequently the

majority of neurons in this structure respond to somatosensory stimuli either as uni-, bi- or

multimodal units (Korzeniewska and Güntürkün, 1990, Wild, 1994), which is clearly different

from the anatomical and physiological pattern found in lateral MD. Thus, at the present state of

knowledge, the initial suggestion of Korzeniewska and Güntürkün (1990) that the DLP is

equivalent to the mammalian posterior complex of nuclei and not to MD still seems to be the

most valid concept. Nevertheless the pigeon’s DLP might serve functions similar to MD, since

DLP-lesions were shown to disrupt working memory tasks (Güntürkün, 1997).

Veenman and coworkers (1997) have suggested that DLP might be homologous to the

intralaminar nuclei in mammals. In our view this seems not very likely since the mammalian

intralaminar nuclei receive extensive afferents from medial hypothalamus, preoptic area, lateral

septum, bed nucleus of the stria terminalis, ventral subiculum, perirhinal cortex, and amygdala

(Price, 1995). Additionally they are characterized by a dense cholinergic input (Tohyama and

Takatsuji, 1998), and their lesion effects are well characterized (Orem et al., 1973). A

comparable pattern does not exist for DLP (Korzeniewska and Güntürkün, 1990, Güntürkün,

1997, Veenman et al., 1997). The suggestion of Miceli and Repérant (1985) that the avian SRt,

which also projects onto NCL, is an intralaminar structure thus seems to be more likely, given

the hodological data provided by these authors.

Recently, Metzger et al. (1996, 1998) suggested that the MNH of chicks might constitute the

avian PFC. This hypothesis is based on the observation that MNH is characterized by a medium

density of dopaminergic afferents (Metzger et al., 1996), a high density of dopaminoceptive

neurons (Schnabel et al., 1997), and afferents from the DMA (Metzger et al., 1996). If avian

DMA should prove homologous to mammalian MD (Veenman et al., 1997), the hypothesis of

Metzger et al. (1996, 1998) would obviously gain weight. Yet, as discussed above, data on the

connections of DMA and MNH are limited. However, the DMA is known to project not only

onto MNH, but also to the accumbens (Veenman et al., 1997), medial LPO (Székely et al.,

1994), and piriform cortex (Bingman et al., 1994), projections which have not been shown for

mammalian MD (Groenewegen, 1988, Price, 1995). Given these contradictions in the DMA -

MD comparison and the complete lack of functional data on DMA, clearly more information is


needed to clarify the comparison between avian MNH and mammalian PFC. These discussions

on the organization of avian associative structures and mammalian PFC demonstrate that simple

equivalencies between avian and mammalian prefrontal systems will not hold, even if two

entities resemble each other with regard to their anatomical organization and function.

However, difficulties in establishing homologous areas are not specific to comparisons

between vertebrate classes but also apply to intraclass comparisons, as reflected in the recent

discussion about the criteria that delineate prefrontal cortex in rats (Condé et al., 1995; Preuss,

1995).

Chapter 2: Projections of the DLP to the forebrain of the pigeon 68

A polysensory pathway to the forebrain of the pigeon: the ascending

projections of the n. dorsolateralis posterior thalami (DLP)

Onur Güntürkün and Sven Kröner

ABSTRACT

Visual evoked potentials (VEPs) in the associative neostriatum caudolaterale (NCL) have shorter latencies than

those recorded in other visual forebrain areas. Therefore visual input into NCL probably stems from a

subtelencephalic relay. Tracing experiments revealed a projection of the n. dorsolateralis posterior thalami

(DLP) into those portions of NCL in which visual, auditory, and somatosensory afferents from

intratelencephalic parasensory areas terminate. Since VEPs in NCL are abolished after DLP-lesions, this

structure has to be the critical relay. However, DLP also projects to other associative forebrain areas and parts

of the basal ganglia. Previous experiments had furthermore revealed that DLP-neurons integrate visual,

auditory, and somatosensory inputs. Thus, the DLP-projection onto various associative forebrain areas

represents a true polysensory thalamo-telencephalic system.

European Journal of Morphology 37:124-128 (1999)

INTRODUCTION

This study aims to clarify two questions regarding the organization of the avian neostriatum

caudolaterale (NCL), an area thought to be equivalent to the prefrontal cortex (PFC). First, are

visual evoked potentials (VEPs) that can be recorded in and around NCL due to a projection of

the n. dorsolateralis posterior thalami (DLP) onto the caudal forebrain? Second, what are the

detailed patterns of projections established by ascending DLP-fibers?

The thalamo- and the tectofugal pathways which terminate in the wulst and the ectostriatum,

respectively, constitute the main ascending visual systems in birds. Latencies of VEPs are about

22 ms for the wulst and more than 50 ms for the ectostriatum (Parker and Delius, 1972). Since

Güntürkün (1984) could record VEPs with latencies of 15 ms in the caudal neostriatum, and

since these VEPs were unaffercted by wulst lesions, it was concluded that an independent ‘third

primary visual system’ had to exist which terminated in the caudal forebrain. However, the

thalamic relay of this pathway remained unknown.

Due to its dense dopaminergic innervation (Divac and Mogensen, 1985, Waldmann and

Güntürkün), and its multimodal organization (Metzger et al., 1998, Kröner and Güntürkün,


1999), the NCL was suggested to be comparable to the mammalian PFC. Indeed, behavioral

studies could demonstrate NCL-lesions to cause deficits in working memory and behavioral

flexibility tasks (Gagliardo et al., 1996, Hartmann and Güntürkün, 1998), similar to the

situation in PFC. The main thalamic afferents of NCL are the DLP and cells in the adjacent

ventrointermediate area of the posterior nuclei (VIP) (Kröner and Güntürkün, 1999). The DLP

has been shown to constitute a polysensory entity in which visual, somatosensory, and auditory

afferents converge on single neurons (Korzeniewska and Güntürkün, 1990) and which plays a

role in working memory tasks (Güntürkün, 1997). Since the DLP is known to project to NCL

(Waldmann and Güntürkün, 1993), the aim of the present study was to clarify if DLP is the

missing thalamic link of the ‘third primary visual system’ and whether all parts of NCL receive

input from DLP.

METHODS

Due to the multitude of approaches methods are described only briefly and we rather refer to studies, in which

these are described in more detail.

Anterograde tracing: A 2.5% solution of Phaseolus vulgaris leucoagglutinin (PHA-L) diluted in 0.1 M

phosphate buffer (PB, pH 7.2) was injected in two animals over 30 min into the DLP using pulsed positive DC

current of 4.5 µA. After 5 days animals were anesthetized and perfused first with 0.9% NaCl (40°C) followed

by 4% paraformaldehyde (PFA) in sodium acetate buffer (pH 6.5) and finally by 4% PFA in sodium borate

buffer (pH 9.5). Brains were postfixed in 4% PFA in acetate buffer (pH 6.5) for 1 hour and stored in 30%

sucrose in PB at 4°C. Then, 25 µm frozen frontal brain sections were cut and processed for ABC immuno-

cytochemistry. For details of the staining procedure see Gerfen and Sawchenko (1983).

Retrograde tracing: In 12 pigeons, small amounts of Fast Blue (FB) or rhodamine isothiocyanate (RITC), each

dissolved at 2% in 1 % DMSO in distilled water were injected stereotaxically into NCL or the terminal field of

DLP within the neostriatum intermedium medialis (NIM). An additional 16 animals received pressure

injections of cholera toxin b (CTb) into the same areas within NCL or NIM. After 2-3 days animals were

perfused with 0.9% NaCl followed by 4% PFA in PB (pH 7.2, 4°C). Brains were postfixed between 1 - 24 h

(4°C) in the fixative with 30% sucrose added, and cut frontally in frozen sections of 40 µm. Fluorescent tracers

were visualized using a fluorescence microscope. For details see Korzeniewska and Güntürkün (1990).

Immunohistochemistry for CTb followed the procedure described by Shimizu et al. (1995). In brief, slices

were incubated overnight (4°C) in a primary antibody against CTb (1:20.000, containing 0.3% Triton X),

followed by a goat biotinylated antibody (1:500 in 0.3% Triton PB; both Jackson) and finally the avidin-biotin

complex (1:100; Vector). Staining was achieved by the 3,3'-diaminobenzidine-(DAB) technique.

Electrophysiology: Eight adult pigeons were anesthetized with equithesin and four 0.12 mm thick insulated

stainless steel wires were implanted under stereotaxic guidance into the DLP-termination zone of the NCL of

each hemisphere. Electrodes tips were staggered in 1 mm steps. An uninsulated loop of wire placed under the


scalp served as reference. Free ends of electrodes terminated in a miniature connector glued onto the skull.

Recordings were performed in awake animals. A red light diode (665 nm, half-bandwidth: 30 nm, intensity: 0.8

mcd, stimulus duration: 5 ms, interstimulus interval: 2.3 s) was placed 5 mm from the eye contralateral to

recording. Potentials from electrodepairs were bandpass filtered between 7 and 120 Hz with a notch filter at 50

Hz, and were averaged over 32 trials. At the end of recordings small radiofrequency lesions were made at

electrode tips to histologically verify their locations. In three pigeons a lesion electrode was placed in a

second surgery into the DLP ipsilateral to the recording site and thermal coagulations (20 mA, 8 s) were made.

NCL-recordings were made a few days before and one week after lesions. For further details see Güntürkün

(1984).

RESULTS

Visual evoked potentials (VEPs) with first inflection latencies 16 to 22 ms were recorded from

NCL. The area examined reached from A 4.5 to A 6.75 and from L 2.00 to L 8.00. Structure and

latencies of VEPs were in close accordance with Güntürkün (1984). To test whether the initial

VEP components derived from the DLP-projection onto NCL, the DLP was thermally lesioned

(fig. 1a). Postoperative recordings revealed a complete absence of VEPs (fig. 1a).

PHA-L injections were centered on DLP but also slightly encroached upon surrounding VIP, n.

dorsointermedius posterior thalami (DIP) and n. dorsomedialis anterior thalami (DMA) (Fig.

1b). Labeled axons were observed to proceed rostrally to the fasciculus prosencephali lateralis

(FPL). On their ascending route, fibers made dense terminal endings on neurons of n.

dorsointermedius ventralis anterior (DIVA) and n. reticularis superior dorsalis (Rsd). Many

labeled axons could also be seen in the „somatic“ parts of the basal ganglia, namely the lateral

lobus parolfactorius (LPO) and the medial paleostriatum augmentatum (PA). Some fibers were

seen dorsal to the lamina hyperstriatica (LH) in the ventral hyperstriatum (HV), the largest

terminal field however extended underneath the LH just medially and dorsally of the

ectostriatum in the NIM. Some fibers coursed frontally and terminated on neurons of the lamina

frontalis superior (LFS). Others proceeded caudally to NCL, where fibers were seen to be

loosely distributed and making terminal endings (fig. 1c,f). These axons mainly concentrated in

the caudodorsal aspect of NCL (Waldmann and Güntürkün, 1993).

Following page - Figure 1: A: Diagrammatical summary of the electrophysiological experiments. The left part of the diagrammshows the location of the recording sites within NCL and the extent of the DLP lesion. The right part shows examples of VEPsbefore and after DLP lesion. B: Injection site of PHA-L in the dorsal thalamus. C: Anterogradely labeled fibers in the NCLfollowing the PHA-L injection shown in B. D: Terminal field of DLP afferents in NIM. E: Retrogradely labeled cells in DLPafter a CTb injection into the neostriatal area shown in D. F: Diagrammatic representation of the DLP terminal fields in NIMand NCL. Scale bars are 500 µm (B,D), 50 µm (C), and 250 µm (D).Abbreviations used in the figures: ectostriatum (E), n. dorsointermedius posterior thalami (DIP), n. dorsolateralis posteriorthalami (DLP), lobus parolfactorius (LPO), neostriatum caudolaterale (NCL), neostriatum intermedium medialis (NIM),paleostriatum augmentatum (PA), ventrointermediate area of the posterior nuclei (VIP).


20µV

25 ms stimulus artifact

postoperative

preoperative

E

NIM

LPOPA

HV

NCL

DLPVIP

DIP

A 5.00

NCL

A 6.50 A 10.50

1 mm

DLP

NCL

DLP

DLPE

HV

NIM

LPO

B C

D E

A

F


The majority of fibers terminated in an area reaching from A 4.75 to A 7.25 and L 4.5 to L 8.5

within NCL (Fig. 1d,f). Only very few fibers extended into far lateral and ventral parts of NCL.

Injections of retrograde tracers into NCL caudal to A 5.00 revealed few labeled cells in DLP.

Injections placed further rostrally labeled neurons in VIP and throughout the complete

rostrocaudal extent of DLP. Injections of retrograde tracers into the dense terminal field of the

DLP within NIM resulted in a very large number of cells within DLP (fig 1e). Injections into

LFS revealed a small number of labeled cells in dorsal DLP.

DISCUSSION

This study clearly reveals that the DLP projects to a number of forebrain somatic and

association areas. It could additionally be shown that the thalamic component of the ‘third

visual system’ proposed by Güntürkün (1984) is the DLP, projecting to large parts of NCL.

The prime argument to propose an independent third visual system to the avian forebrain had

been the latencies of VEPs which were shorter in caudal forebrain then in wulst and

ectostriatum (Güntürkün, 1984). DLP integrates besides auditory, and somatosensory afferents

also a massive visual input from the tectum. Since DLP projects onto NCL, the VEPs of the

caudal forebrain are likely due to ascending DLP-projections. This assumption is clearly

supported by the absence of VEPs in NCL after DLP-lesions. In addition to afferents from DLP,

NCL integrates further visual input from secondary visual forebrain structures like caudal

hyperstriatum accessorium (HA) and ectostriatal belt (Eb) (Metzger et al., 1998, Kröner and

Güntürkün, 1999). It is possible that these afferents contribute to the late VEP components.

However, DLP-lesions had completely diminished VEPs, including their late components. This

might be due to the cellular sensory properties of the structures involved. Korzeniewska and

Güntürkün (1990) demonstrated that DLP-neurons had broad receptive fields and responded to

simple stimulus properties. Given the possible complexity of reponse properties in secondary

visual association areas of the forebrain, the simple diode-flash induced VEPs are likely to

activate DLP-, but not HA- and Eb-neurons. The dorsal part of NCL which receives the densest

DLP innervation has also been shown to receive auditory input via field L, afferents from visual

and somatosensory parts of HA, as well as a projection from cells in the terminal field of the

DLP within NIM (Kröner and Güntürkün, 1999). It thus seems that the dorsal NCL is reached

by intratelencephalic input from those modalities that are also transported by DLP afferents,

and, in addition, those parts of NCL receive an indirect input from DLP via NIM. In sum, the

DLP supplies the NCL with both a fast and a slow copy of incoming sensory information from


all main modalities, which within NCL can be further compared with other input from the same

modalities that originates in higher sensory areas of the telencephalon.

Acknowledgements

Thanks to our old master Juan Delius for his continuing support, his never ending enthusiasm, and his bright ideas. Supported by

the Deutsche Forschungsgemeinschaft through its Sonderforschungsbereich 509 NEUROVISION.

Chapter 3: Cell types in the chick neostriatum caudolaterale 74

Characterization of cell types in the chick neostriatum caudolaterale:

Electrophysiological and morphological properties.

Sven Kröner, Kurt Gottmann, Hanns Hatt, and Onur Güntürkün

ABSTRACT

The neostriatum caudolaterale (NCL), in the chick also referred to as dorsocaudal neostriatal complex (dNc),

is a polymodal associative area in the forebrain of birds that is involved in sensorimotor integration and

memory processes. In chick it is part of an „imprinting network“, and might serve to integrate the various

sensory and emotional components of natural imprinting objects. We have used whole-cell patch-clamp

recordings in chick brain slices to characterize the principal cell types of the NCL. Electrophysiological

properties distinguished four classes of neuron. The morphological characteristics of these classes were

examined by intracellular injection of Lucifer Yellow. Type I neurons characteristically fired a brief burst of

action potentials. A number of these cells showed distinct low-threshold transient depolarizations.

Morphologically, type I neurons had lar somata and thick dendrites with many spines. Type II neurons were

characterized by their repetitive firing pattern from which conspicuous frequency-adaptation was observed.

Cells in this class could be subdivided by the presence or absence of a time-dependent inward rectification.

Type II neurons also had large somata and thick dendrites with many spines. Type III neurons showed a high-

frequency firing with little accommodation and a prominent time-dependent inward rectification. They had thin,

sparsely spiny dendrites and extensive local axonal arborizations. Type IV neurons had a longer action-potential

duration, a larger input resistance, and a longer membrane time constant than the other classes. Type IV neurons

had small somata and short dendrites with few spines. The axons of cells in all spiny cell classes (types I, II, IV)

showed similar projection patterns to targets outside the NCL. The classes of neuron described here may play

distinct roles in sensorimotor integration and learning.

Submitted

INTRODUCTION

The neostriatum caudolaterale (NCL), in the chick also referred to as dorsocaudal neostriatal

(dNc) complex (Metzger et al., 1998), is a multimodal association area in the forebrain of birds

(Leutgeb et al., 1996; Metzger et al., 1998; Kröner and Güntürkün, 1 999). In the chick it has

mainly been studied as part of a network responsible for early learning, specifically imprinting

(Metzger et al., 1996, 1998; Bock et al., 1997; Bock and Braun, 1999a,b; Braun et al., 1999),

whereas in the pigeon most work has focused on tasks which in mammals invoke „frontal“


executive functions, e.g. working memory (Mogensen and Divac, 1982, 1993, Gagliardo et al.,

1996, 1997, Güntürkün, 1997; Kalt et al., 1999), reversal learning (Hartmann and Güntürkün,

1998), response inhibition (Güntürkün, 1997, Aldavert-Vera et al., 1999), and spatial

orientation (Gagliardo and Divac, 1993). Anatomically, the NCL seems to be particularly well

suited for the association between external stimuli and the animals behaviour, as it integrates

information from all modalities and exerts influence over motor and limbic structures. In chicks

(Metzger et al., 1998) and pigeons (Leutgeb et al., 1996; Kröner and Güntürkün, 1999) the NCL

has been shown to be reciprocally connected with all major secondary sensory areas and at

least three further multimodal brain areas (i.e. the intermediate and medial part of the

hyperstriatum ventrale, IMHV, the mediorostral neostriatum/hyperstriatum ventrale, MNH, and

the archistriatum). Thalamic input reaches the NCL from several nuclei in the dorsal thalamus,

including the multimodal thalamic nuclei dorsolateralis posterior (DLP) and subrotundus (SRt)

(Korzeniewska and Güntürkün, 1990; Leutgeb et al., 1996; Metzger et al., 1998; Güntürkün and

Kröner, 1999; Kröner and Güntürkün, 1999). The NCL´s descending projections are directed

mainly to limbic and somatomotor regions within the basal ganglia and the archistriatum

(Kröner and Güntürkün, 1999). Thus, the NCL represents a nexus of sensorimotor integration in

the avian forebrain. In the chick, the NCL has therefore been postulated to be a polymodal

associative part of the „imprinting pathway“, in which the various sensory and emotional

components of natural imprinting objects are integrated during both, the learning process and

memory recall (Braun et al., 1999). While imprinting traditionally is conceptualized as a

relatively specialized form of learning in young animals that occurs during a sensitive phase

and which has permanent behavioural effects (Lorenz 1935), it shares certain features with

other forms of associative learning (Bolhuis, et al., 1990), such as stimulus generalization

(Bolhuis and Horn, 1992), stimulus comparison (Honey and Bateson, 1996), and the association

of stimuli in time (Bateson and Chantrey, 1972). If we are to explain the mechanisms that

underlie the sensorimotor integration and associative learning processes, we need to understand

in detail how sensory coordinates are mapped onto motor ones according to the learning history

of the animal and its current behavioural goals. Similar to the modules of mammalian cortical

structures (Douglas and Martin, 1992), the avian forebrain might be constructed from a

relatively small number of principal or canonical circuits, that are repeated in large quantities

to achieve parallel computing power. Furthermore, it has been shown in numerous preparations

that there is generally good correspondence between a neurons intrinsic firing pattern and its

morphology (e.g. McCormick et al., 1985; Chagnac-Amitai et al., 1990; Larkman and Mason,

1990; Kang and Kayano, 1994; Kasper et al., 1994a). Thus, to further our understanding of the

functions of the NCL on a cellular level, it is important to characterize its intrinsic neuronal


organization and the cell types that might be involved in the processing of different aspects of

information. In this report, we have used in vitro whole-cell patch-clamp recording in

combination with intracellular staining to characterize the principal cell types in the chick NCL

electrophysiologically and morphologically.


Preparation of slices and electrophysiological recording

Fertilized eggs were obtained from a commercial supplier (Sörries Trockels, Möhnesee, Germany) and chicks

(Gallus domesticus) were hatched and kept in small groups on a 12:12 hour dark/light cycle. Treatment of

animals conformed to NIH guidelines. A total of 29 chicks (2 - 11days post-hatch, median 6 d) were

decapitated and the brains were transferred to iced extracellular solution. Coronal slices (350 µm thick) of the

caudal telencephalon were cut on a vibratome. Slices were incubated in extracellular solution consisting of (in

mM): 119 NaCl, 2.5 KCl, 1 NaH2PO4, 26.2 NaHCO3 ,10 D-glucose, 4.5 CaCl2, and 1.3 MgCl2, saturated with

95% O2 /5% CO2, pH 7.3. Slices were allowed to recover for at least 1 hour, before being transferred to the

recording chamber. All recordings were made at room temperature (21 - 23°C) in a submerged slice chamber

perfused with extracellular solution.

Whole-cell patch-clamp recordings were obtained in the caudolateral subventricular region of the forebrain,

using the blind-patch technique (Blanton et al., 1989). In current-clamp experiments recording electrodes (6 –

8 MΩ resistance) were filled with an intracellular solution consisting of (in mM): 135 K-gluconate, 20 KCl, 2

MgCl2 , 10 N-(2-hydroxyethyl)piperazine-N'-(2-ethanesulfonic acid) (HEPES), 0.1 ethylene glycolbis (β-

aminoethyl ether)- N,N,N´,N´,-tertraacetic acid (EGTA), 4 Na2-ATP, and 0.5 Na2 -GTP, and adjusted to pH 7.3

with KOH. In some recordings 3 - 3.5 mg/ml of the dipotassium salt Lucifer Yellow (LY, Sigma, Deisenhofen,

Germany) were added to the intracellular solution and neurons were filled by diffusion during 30 - 90 minutes

of recording. In a further set of experiments K-gluconate and KCl in the intracellular solution were substituted

with 135 mM CsCl and 20 mM tetraethylamonium (TEA) to block potassium channels. Recordings were made

using a HEKA (Lambrecht, Pfalz, Germany) EPC-7 patch-clamp amplifier, filtered at 3 kHz, and sampled at 20

kHz. The voltage drop across the pipette (which was usually about 8 - 10 mV) could not be bridge-balanced.

Sampling was done with a Digidata 1200 interface using pClamp 6.0 software (Axon Instruments, Foster City,

CA), and data were stored on computer for off-line analysis with the CLAMPFIT module of pClamp. Input

resistance was determined from the voltage deflection induced by a hyperpolarizing current pulse in the linear

range of the current-voltage relation. Membrane time constants were calculated by fitting a single exponential

function to the voltage response to injections of hyperpolarizing currents in the linear range of the voltage

response. The percentage sag that occurred in some neurons in response to hyperpolarizing and depolarizing

current pulses was calculated as 100 x (Vmax -Vend)/Vmax, where Vmax was the peak voltage deflection and Vend

the voltage at the end of the current pulse. Resting membrane potentials were assessed in current-clamp mode

3 - 5 minutes after establishing the whole-cell configuration. Spike duration was measured at the half-maximal

amplitude from threshold, and the amplitude and time-to-peak value for the afterhyperpolarization (AHP) which

followed the action potential were measured from the equipotential-point on the repolarizing phase.


Data analysis. Data are presented as means ± standard error (SE). Statistical analyses were done using the

SPSS 8.0 software package. The differences among cell classes for the various parameters were compared by

analysis of variance (ANOVA) and the statistical significance of the differences was examined with a Scheffé

test for multiple comparisons. Subsequent comparisons between and within cell classes were done using

Student´s t test for unpaired samples.

Histological procedures

Following recording, slices were fixed in 4% paraformaldehyde and 0.2% glutaraldehyde in 0.12 M phosphate

buffer (4°C, pH 7.4) for about 15 hours. They were then transferred to a solution of 30% sucrose in phosphate

buffer containing 0.9% NaCl (PBS; pH 7.4) and 0.01% NaN3 as a preservative. Slices were resectioned at 70

µm on a freezing microtome and collected in PBS. For immunohistochemistry of LY, endogeneous

peroxidases were blocked by preincubating slices in 0.5% H2O2. Slices were then washed in PBS and floating

sections were incubated overnight at 4°C in biotynilated anti-LY from rabbit (Molecular Probes, Leiden, The

Netherlands; 1:200 working dilution) in PBS containing 0.3% Triton X-100 (Sigma). After washing, slices

were incubated for 2 hours in the avidin-biotin complex (ABC Elite, Vector Labs, Burlingame, CA; 1:100 in

PBS with 0.3% Triton X-100). Washes in PBS were followed by additional washes in 0.12 M acetate buffer

(pH 6). Staining was achieved by the 3,3'-diaminobenzidine (DAB) technique with heavy-metal amplification

(modified from Adams, 1981) by adding H8N2NiO8S2 (2.5 g/ 100 ml), NH4Cl and CoCl2 (both 40 mg/ 100

ml). After 20 minutes of preincubation the reaction was catalyzed using 0.3% H2O2. The reaction was stopped

by rinsing the tissue in acetate buffer and PBS. Slices were then mounted, dehydrated and coverslipped.

Quantative measures of cell morphology were done using a Leica DMR (Leica, Wetzlar, Germany) and the

analySIS software package (Soft-Imaging Software, Münster, Germany). Camera lucida reconstructions of

filled cells were drawn using a Leitz BioMed with a drawing tube at x 12.5 or x 50 magnification.

RESULTS

Stable whole-cell recordings were obtained from 90 cells in the NCL. Neurons were selected

only if they exhibited a resting membrane potential (RMP) more negative than -50 mV and an

overshooting action potential. In this voltage range, there was a general lack of spontaneous

firing in all NCL cells, however, in voltage clamp recordings spontaneous postsynaptic currents

could be observed in most cells, indicating synaptic activity (data not shown). Neurons were

classified into 4 distinct types according to differences in their intrinsic firing properties and in

the voltage responses to hyperpolarizing current pulses. For 36 of the neurons thus

characterized basic morphological features were analyzed after filling with LY. With a single

exception in which a second cell was weakly stained, we found no evidence of dye-coupling

between cells in the NCL. Dye-coupling results from neuronal signalling via gap junctions and

is a phenomenon commonly observed in adult (Connors et al., 1983; Benardo, 1997), but


particularly in developing mammalian (Connors et al., 1983; Kasper et al., 1994b) and avian

(Dutar et al., 1998) neurons, where it is believed to be involved in the functional specification

of brain areas (Peinado et al., 1993).

Type I neurons:

Type I cells displayed a characteristic firing pattern that consisted of a single action potential

or a brief burst of spikes. Type I cells represented 34 of the 90 cells characterized.

Firing properties: Cells in this class displayed the highest threshold for firing an action

potential, especially in relation to their low resting membrane potential (table 1). Type I cells

were marked with their tendency to fire a single action potential or 2 - 8 clustered spikes at

high frequency at the onset of depolarizing pulses (fig. 1A). The number of spikes in a burst and

the duration of the interspike-intervals strongly depended on the current amplitude injected (fig.

1A, C). No repetitive bursting was observed during long (1.1 sec.) depolarizing current pulses.

However, in the course of other experiments (data not shown) spontaneous occuring all-or-none

bursts were found when fast GABAergic transmission was abolished by adding picrotoxin (100

µM) to the extracellular solution (see below).

Type I cells varied with respect to their behaviour in the subthreshold voltage range: In

response to subthreshold depolarizing current pulses 13 type I cells showed prominent low-

threshold potentials that were not seen in the remaining neurons (c.f. fig. 1A1 and 1A2). This

subset of neurons also differed from the remaining type I cells in that their action potentials had

significantly larger amplitudes (60.9 ± 3.3 mV vs. 51.6 ± 2.9 mV) and shorter durations at half

maximal amplitude (1.72 ± 0.1 ms vs. 2.0 ± 0.08 ms), the latter being mostly due to shorter

rise-times of the action potentials (0.58 ± 0.05 ms vs. 0.74 ± 0.05 ms; Student´s t test for

unpaired samples; df = 32; all values for t > ± 2.1; p < 0.05). Further, 6 of the 13 cells that

showed these transient „hump“-like depolarizing responses also displayed a complex pattern of

afterpotentials that followed the first spikes. These consisted of an initial, fast

afterhyperpolarization (fAHP) that was followed by an intercalated depolarizing afterpotential

(DAP) and a late medium-duration AHP (mAHP; see fig. 1A1). The remainig type I cells

showed only a monophasic AHP (c.f. fig. 1A2).

Membrane rectification: Type I cells had low apparent input resistances and intermediate

membrane time constants (table 1). In response to hyperpolarizing current pulses that drove the

membrane potential more negative than about -80 mV, all type I neurons showed a pronounced

fast-activated inward rectification resulting in an upward bend in the current-voltage relation

(fig. 1D). Most type I cells also showed membrane outward rectification at the end of the


-75 mV

A1

-74 mV

A2

B

C

Spi

ke fr

eque

ncy

(Hz)

Time during stimulus (ms)

400 pA450 pA500 pA550 pA600 pA650 pA700 pA

0 200 400 600 800 1000 1200

10

20

30

40

50

60

70 D

Pot

entia

l (m

V)

Current (pA)

-120

-100

-140

-80

-60

-40

-400 -200-300 -100 100 200

0.5 nA 0.5 nA

Figure 1: Intrinsic firing properties and membrane rectification of type I neurons. A: Examples of voltage responses tohyperpolarizing and depolarizing current pulses (-400 - 700 pA in A1 and -300 - 800 pA in A2) in two neurons that showed rapidadaptation of action potential firing. Type I cells typically responded with „bursts“ of action potentials to current pulses ofincreasing intensity. Both cells show membrane inward rectification to relatively large hyperpolarizing currents and outwardrectification to depolarizing currents in the spike-threshold range. The cell in A1 displays a prominent subthreshold „hump“-likepotential and a depolarizing afterpotential following single spikes. The cell in A2 shows only monophasic afterhyperpolarizationsof comparatively long duration. Note that there is also apparently less membrane rectification in this cell than in the cell shown inA1. B: Camera lucida reconstruction of a type I neuron. The soma and dendrites are drawn in black, the axon collaterals areshown in grey. Scale bar is 50 µM. C: Relationship between the instantaneous spike frequency (as the inverse of inter-spike-interval) and time for different current amplitudes for the cell shown in A1. D: Current-voltage relationship of the cell shown inA. Deviations of the potential from the extrapolated ohmic responses close to resting potential reflect inward and outwardrectification, respectively.


current pulse, that was responsible for a downward bend in the I-V plot at depolarized

potentials (fig. 1A).

Morphology: Of the 34 cells designated

as type I cells 15 were successfully

recovered and reconstructed. Single or

burst firing neurons had large round or

fusiform somata (diameter 15 - 22 µm;

table 2) and multipolar arranged

dendrites. Dendrites were usually long

(dendritic field diameters 225 - 320 µm;

table 2) and displayed medium to high

spine densities (figs. 1B, 2, and 10). The

main axon branched near the soma and

gave rise to thin collaterals that ramified

mainly in a loose local plexus (figs. 1B,

2, and 10). In addition to these recurrent

collaterals which displayed numerous

varicose-like swellings, often 2 - 5

longer collaterals were observed that

projected outside the NCL. These axonal

arborizations showed less varicosities

and followed one of three general

directions: In most cases 1 - 3

collaterals moved ventrally within the

plane of the slice to terminate within the

underlying archistriatum. Other

collaterals travelled through the NCL

either dorsomedially, towards the

auditory subunit of the NCL, the

neostriatum dorsale (Nd), and possibly

other sensory forebrain areas, or ventromedially, in the direction of the the basal ganglia (c.f.

Metzger et al., 1998; Kröner and Güntürkün, 1999). Ventromedially directed arborizations,

however, could be followed only over relatively short distances before they left the sagittal

plane of the slice.

Figure 2: Morphological features of type I cells. A: Photomicrograph ofa Lucifer Yellow-filled neuron that responded to depolarizing currentpulses with rapidly adapting bursts of action potentials. B: Detail of atype I cell showing a dendrite with numerous spines. C: Dendritic treeand local axon arborizations of a type I neuron. The axon showsnumerous varicose-like swellings in the vicinity of the dendritic field. Thesoma of the cell was separated from the axon and dendrites duringresectioning of the slice. Scale bars are 50 µm (A, C), and 20 µm (B).


Type II neurons

Type II cells were characterized by their repetitive firing pattern. They represented 35 of the 90

neurons studied.

Firing properties: Type II had similarly low resting membrane potentials, spike amplitudes

and action potential durations, as well as AHP amplitudes and time-to-peak values as type I

neurons (table 1). In contrast to type I neurons, however, small somatically injected

depolarizing current pulses readily elicited relatively tonic firing in type II cells. Two typical

examples of type II cells are shown in figure 3A. Current pulses that depolarized type II cells

just above threshold usually evoked a single action potential that often had long latencies. With

larger depolarizing currents a pattern of phasic-tonic firing emerged, in which a short burst,

was followed by a longer train of action potentials. This train of action potentials showed

conspicuous frequency adaptation (fig. 3A lower traces, and C), but the repetitive firing pattern

could be sustained at relatively low frequencies (between 5 -15 Hz) over long (1.1 sec)

depolarizing current pulses. This firing behaviour was facilitated by a comparatively low

action potential threshold (table 1). A number of type II cells (n = 8) showed the same sequence

of afterpotentials following an action potential as described for the subset of type I neurons

above.

Membrane rectification: Among the four types identified, type II neurons had intermediate

input resistances and time constants (table 1). Type II cells differed with regard to the existence

of a hyperpolarization-activated time-dependent inward rectification: When hyperpolarizing

current pulses were applied to type II cells usually (n = 22) a fast-activated inward rectifying

conductance appeared at membrane potentials more negative than about -80mV (fig. 3A1). This

inward rectification resulted in a slight upward bend in the hyperpolarized portion of the

current-voltage plot when compared with the extrapolated linear portion (fig. 3D). In another

13 neurons, however, a small sag in the voltage response (2.1 - 6.7 % sag) could be seen at

potentials more negative than about -80 mV. This delayed inward rectifying conductance tended

to push the membrane potential back towards resting values and lead to a rebound overshoot at

the termination of hyperpolarizing current pulses (fig. 3A2, 13B). The amplitude of the sag

increased with more hyperpolarized membrane potentials. These cells in addition seemed to

possess the same fast-activated inward rectification seen in the remaining type II neurons

(above) as their peak voltage responses exhibited inward rectification over the same range of

membrane potentials as the time dependent inward rectification (fig. 3D). At depolarizing

voltage steps the membrane potential often depolarized more than expected from the ohmic

response around resting potential, indicating the contribution of an inward rectification in the

voltage range close to spike threshold (see below).


B

A1

-73 mV -68.5 mV

A2

sag

200 ms

20 mV

C

Spi

ke fr

eque

ncy

(Hz)


10

20

30

40

50

60

70

80

0 200 400 600 800 1000 1200

250 pA350 pA450 pA550 pA650 pA750 pA800 pA

D

-120

-100

-140

-80

-60

-40

Current (pA)

Pot

entia

l (m

V)

-400 -200-300 -100 100 200

voltage at peak

voltage at end

0.5 nA 0.5 nA

Figure 3: Intrinsic firing properties and membrane rectification of type II neurons. A: Examples of voltage responses tohyperpolarizing and depolarizing current pulses (-300 - 700 pA) in two neurons that showed a phasic-tonic firing of actionpotentials over long depolarizing current. The cell in A1 displays fast inward rectification to hyperpolarizing current pulses. Thecell in A2 shows the same fast inward rectification and additionally a sag in the voltage response to relatively largehyperpolarizing current pulses, resulting in a rebound overshoot at cessation of current pulses. See text for detail. B: Cameralucida reconstruction of a type II neuron. The soma and dendrites are drawn in black, the local axon collaterals are shown ingrey. Scale bar is 50 µM. C: Relationship between the instantaneous spike frequency and time for different current amplitudesfor the cell shown in A2. After a phasic response the spike frequency rapidly adapts to a steady-state tonic firing. D: Current-

voltage relationship of the cell shown in A2. , voltage response measured at the end of the current pulse; , voltage at thepeak.


During long current pulses the

membrane potential sagged back

towards rest (4.8 %- 19.4 % sag)

resulting in a rebound undershoot at the

termination of the current pulse (figs.

3A2, 12B).

Morphology: The gross morphology of

type II cells resembled that of type I

cells (figs. 3B, 4, 10). They had large

oblique somata (diameter 16 - 23µm;

table 2) and long, thick multipolar

dendrites. In 4 neurons that were

thoroughly analyzed, type II cells

displayed the highest number of spines,

but this value was not significantly

different from that obtained for type I

neurons (table 2). Dendrites of type II

neurons had the largest number of

branch points and covered large areas

(diameter of dendritic field 229 - 296

µm; table 2). The axons of type II

neurons often formed a radial axonal

plexus in the vicinity of the soma

giving rise to several local

ramifications as well as long

collaterals that projected to targets

beyond the borders of the NCL. As in

type I cells, most neurons sent 1 - 3

descending collaterals towards the

archistriatum, but also had additional arborizations that seemed to travel ventromedially and

dorsomedially (fig. 5), probably making numerous contacts with other cells within NCL on their

way.

Figure 4: Morphological features of type II cells. A, B:Photomicrographs of two Lucifer Yellow-filled cells that responded withphasic-tonic firing of action potentials in response to depolarizingcurrents. The arrow in A points to where the axon leaves the cell body. Itcourses ventrally and ramifies extensively at the bottom of thephotomicrograph. The cell in B is the same cell as shown in figure 3B. C:Detail of a type II cell showing a dendrite with numerous spines and partof an axon (top left). Scale bars are 50 µm (A, B), and 20 µm (C).


Type III neurons:

The key features of type III neurons were the ability to fire action potentials at a high frequency

and a prominent sag in the response to hyperpolarizing current pulses, as well as a short action

potential duration. Type III cells represented 8 of the 90 cells characterized.

Firing properties: In response to suprathreshold depolarizing current pulses all neurons

designated as type III responded with regular tonic firing that showed only relatively little

frequency adaptation. Thus, type III cells were able to initially fire action potentials at

frequencies of about 100 Hz. After several spikes the firing rate usually accomodated to some

degree, but sustained firing over long depolarizing current pulses continued at frequencies much

higher than found in any other cell class (fig. 6C). These cells also showed the most positive

resting membrane potentials, while generally having a very low threshold for the initiation of

action potentials (table 1). The duration of action potentials in type III cells was by far the

shortest among all classes, which was also reflected by short rise- and fall-times (table 1). The

action potential was followed by an AHP that was large in amplitude and had a comparatively

Figure 5: Example of the axonal arborizations of a type II cell within NCL. In addition to recurrent local arborizations mostneurons from all spiny cell classes possessed one or several axon collaterals that left the NCL ventrally towards thearchistriatum, ventromedially, in the direction of the basal ganglia, and/or dorsomedially, possibly to provide feedback to sensoryand associative forebrain areas. The insert shows the location of the cell in the dorsocaudal neostriatum. Scale bars are 500 µM.


Table 1: Electrophysiological characteristics of classes of NCL neurons (means ± standard error)

Type I

(n = 34)

Type II

(n = 35)

Type III

(n = 8)

Type IV

(n = 13)

Significant differences*

Resting potential (mV) -75.1 (± 0.9) -73.9 (± 0.9) -63.5 (± 1.4) -68.3 (± 1.8) Type III, IV > type I, II

Input resistance (MΩ) 210.5 (± 12.1) 241.6 (± 13.1) 458.6 (± 28.6) 568.5 (± 28.7) Type IV > type III > type I, II

Membrane time constant (ms) 59.5 (± 2.8) 67.6 (± 5.3) 46.1 (± 6.2) 91.6 (± 8.3) Type IV > type I, II, III

Spike threshold (mV) -25.0 (± 1.2) -30.3 (± 1.0) -33.6 (± 1.7) -27.5 (± 0.7) Type I > type II, III

Max. number of spikes

(1.1s depolarizing pulse)

3.4 (± 0.5) 10.6 (± 0.9) 46.3 (± 5.9) 7.6 (± 0.8 ) Type III > type II > type I, IV

Spike duration at half

amplitude (ms)

1.9 (± 0.07 ) 1.86 (± 0.1) 0.76 (± 0.1 ) 2.32 (± 0.1) Type IV > type I, II > type III

Rise time (10-90%) ms 0.67 (± 0.03) 0.72 (± 0.04) 0.43 (± 0.06) 1.15 (± 0.07) Type IV > type I, II > type III

Fall time (90-10%) ms 1.41 (± 0.1) 1.64 (± 0.05) 0.94 (± 0.1) 1.9 (± 0.08) Type IV > type I > type III;

type II > type III

Spike amplitude (mV) 54.7 (± 2.3) 51.9 (± 2.0) 49.4 (± 6.7) 45.2 (± 2.3) −

AHP amplitude (mV) -17.8 (± 0.7) -17.5 (± 0.7) -20.3 (± 1.7) -15.8 (± 1.1) −

AHP time to peak (ms) 3.4 (± 0.2) 4.1 (± 0.2) 5.6 (± 0.6)= 5.8 (± 0.5) Type III, IV > type I, II

* Significant differences indicated by the use of a Scheffé test for multiple comparisons after analysis of variance (P < 0.05).

long time-to-peak (table 1). Both the amplitude and the time-to-peak could be reduced to a

much larger degree than in the other cell types when increased depolarizing currents drove type

III neurons to high-frequency tonic firing.

Membrane rectification: Type III cells had comparatively high input resistances but relatively

short membrane time constants (table 1). In response to hyperpolarizing current pulses

(membrane potential more negative than -80 mV) a prominent sag in the voltage response (8 -

20.4% sag) occurred that was followed by a rebound depolarization at cessation of the current

pulse. The time dependent inward rectification and the rebound overshoot increased as the

hyperpolarizing current pulses increased. In response to larger negative current pulses the

rebound depolarization could initiate action potentials (fig. 6A). Type III neurons also appear

to possess an additional fast-activated inward rectifier, as the peak voltage responses exhibited

inward rectification over the same range of membrane potentials as the time dependent inward

rectification, (fig. 6D).

Morphology: Type III cells were characterized by small fusiform somata (diameter 10 - 14µm;

table 2) and thin, aspiny or sparsely spiny dendrites (figs. 6B, 7, 10). In fact, in 3 out of 6

neurons that were successfully reconstructed the dendrites were so thin that they could not be


BA

- 66.5 mV

200 ms

20 mV

C

Spi

ke fr

eque

ncy

(Hz)


10

20

30

40

50

60

70

0 200 400 600 800 1000 1200

40 pA80 pA

120 pA140 pA180 pA220 pA240 pA

80

90

100

110 D

Pot

entia

l (m

V)

Current (pA)

-120

-100

-140

-80

-60

-40

-400 -200-300 -100 100 200

voltage at peak

voltage at end

0.2 nA

Figure 6: Intrinsic firing properties and membrane rectification of a type III cell. A: Voltage responses to hyperpolarizing anddepolarizing current pulses (-120 - 240 pA). There is a large sag in response to strong hyperpolarizing current pulses, whichresults in a rebound overshoot that is large enough to trigger spikes. Small depolarizing currents elicit continuous firing atcomparatively high frequencies. B: Camera lucida reconstruction of a type III neuron. Scale bar is 50 µM. C: Relationshipbetween the instantaneous spike frequency and time for different current amplitudes for the cell shown in A. There is somefrequency adaptation after the first few spikes, but the cell maintains a high-frequency firing without further attenuation over along current pulse. D: Current-voltage relationship of the cell shown in A. , voltage response measured at the end of the

current pulse; , voltage at the peak.


clearly distinguished from the extensive local axonal ramifications. The remaining three cells

had relatively few long multipolar dendrites (diameter of dendritic field 227 - 267 µm). The

dendrites showed a low number of total branch points and especially few higher order (≥ 4th

branches) branchings (table 2). The axon appeared to have a large number of varicosities, and

arborized extensively to form a dense plexus of terminals in the vicinity of the soma (fig. 7).

Axons of type III neurons could never be followed beyond the borders of the NCL. The

diameter of axonal arborizations ranged between 545 - 736 µm and the mean brain area that

was covered by these arborizations was 320,400 µm2 (s.e. ± 35,503 µm2).

Type IV neurons:

The key features of type IV cells are a

high input resistance and a long

membrane time constant, as well as a

long action potential duration. Of the

90 cells characterized, 13 were found

to belong to this class.

Firing properties: Among the classes

identified here, type IV cells had an

intermediate RMP and spike-threshold.

Their bursting spike pattern resembled

that of type I cells (fig. 8), but type IV

neurons were able to initially fire a

larger number of action potentials

(table 1). However, in response to

large depolarizing currents the ability

of type IV neurons to fire repetitively

was markedly attenuated (fig. 8C). A

characteristic of cells in this class was

the long action potential duration,

which was also reflected in the

significantly longest rise- and fall-

times (table 1). Furthermore, the action

potentials of type IV cells showed a

prominent progressive spike

broadening during repetitive firing.

Figure 7: Morphological features of type III cells. A: Photomicrographof a Lucifer Yellow filled sparsely spineous type III cell. B: Detail ofthe cell shown in A. Through the middle of the photomicrograph runs along thin dendrite; parallel to this run axon collaterals above and below.C: Part of the local axonal arborizations of another type III cell. Scalebars are 50 µm (A, B), and 20 µm (C).


Type IV cells had comparatively small monophasic AHPs, but long times-to-peak (table 1).

Membrane rectification: The key feature required for type IV cells was a long membrane time

constant (> 60 ms). They also had the highest input resistance (> 350 MΩ) among all classes of

cells (table 1) and showed steep current-voltage relationships (fig. 8C). Only at membrane

potentials more negative than about - 95 mV in most neurons a weak, fast-activated inward

A

- 67 mV

200 ms

20 mV

B

C


0 200 400

10

20

30

40

50

Spi

ke fr

eque

ncy

(Hz)

150 pA210 pA240 pA270 pA300 pA330 pA360 pA

600 800 1000 1200

DP

oten

tial (

mV

)

Current (pA)

-120

-100

-140

-80

-60

-40

-400 -200-300 -100 100 200

0.2 nA

Figure 8: Intrinsic firing properties and membrane rectification of a type IV cell. A: Voltage responses to hyperpolarizing anddepolarizing current pulses (-180 - 360 pA). The cell shows large voltage responses to small hyperpolarizing current pulses withrelatively little inward rectification. Small depolarizing current pulses readily elicit short bursts of action potentials. B: Cameralucida reconstruction of a type IV neuron. The soma and dendrites are drawn in black, the axon collaterals are shown in grey.Scale bar is 50 µM. C: Relationship between the instantaneous spike frequency and time for different current amplitudes for thecell shown in A. D: Current-voltage relationship of the cell shown in A. The input resistance is high, and the voltage responseshows only little inward rectification to large current pulses.


inward rectification became evident (fig. 8D). All type IV cells showed a prominent membrane

outward rectification at depolarized potentials.

Morphology: Type IV cells were characterized by small, oblique somata (diameter 12 - 16,5

µm; table 2) and multipolar dendrites that possessed few thin spines. Usually, proximal

dendrites were short and thin, but sometimes 2 or 3 main dendrites were seen that had thick

stems but tapered considerably with distance from the soma (figs 8B, 9). The otherwise

spherical form of the dendritic field (diameter 180 - 225µm) was tilted by these dendrites as

they also extended further than the others. The dendritic tree of type IV neurons showed

relatively few branches and covered the smallest area among the four cell types (table 2).

Axonal arborizations of type IV cells were not as extensive as in type I or type II cells, but

projections followed the same pattern seen in the other classes of spiny neurons. One to three

main collaterals descended ventrally in the direction of the archistriatum, while other

collaterals travelled dorsomedially and/or ventromedially within the NCL.

Table 2: Morphological characteristics of classes of NCL neurons (means ± standard error)

Type I

(n = 15)

Type II

(n = 11)

Type III

(n = 6)

Type IV

(n = 5)

Significant differences*

Soma size (µm²) 217.8 (± 7) 252.6 (± 9) 121.8 (± 7) 142.6 (± 9) Type I, II > type III, IV

Total dendritic length (µm) 4832 (± 281) 5921 (± 354) 2228 (± 48)† 22212 (± 589) Type I, II > type III, IV

Area of dendritic

arborization (µm²)

56449 (± 4251) 59845 (± 4294) 44412 (± 538)† 28782 (± 3133) Type I, II > type IV

Number of branch points 51.7 (± 2.2) 61.9 (± 4.6) 24.6 (± 1.4)† 28.7 (± 0.3) Type I, II > type III, IV

Number of spines (n/10µM)initial segment

1.47 (± 0.59)‡ 2.91 (± 0.89)‡ − † 2.18 (± 0.78)‡ −

2nd branching 5.06 (± 0.22)‡ 6.60 (± 1.08)‡ 0.89 (± 0.32)† 2.87 (± 0.20)‡ Type II > type III, IV;Type I > type III

3rd branching 5.46 (± 0.35)‡ 6.89 (± 0.74)‡ 1.32 (± 0.31)† 2.83 (± 0.41)‡ Type I, II > type III, IV

4th branching 5.11 (± 0.87)‡ 6.59 (± 0.70)‡ 0.55 (± 0.02)† 2.96 (± 0.55)‡ Type II > type III, IV;Type I > type III

5th and 6th branching 5.71 (± 2.32)‡ 7.23 (± 1.03)‡ − † 2.82 (± 0.65)‡ n.a.

Direction of axon ventrally,dorsomedially,

andventromedially

ventrally,dorsomedially,

andventromedially

locally ventrally,dorsomedially,

andventromedially

* Significant differences indicated by the use of a Scheffé test for multiple comparisons after analysis of variance (P < 0.05). †

These values could be obtained in only three type III neurons (see text for details). ‡ Spine frequencies were counted in four

neurons of each class. n.a.: not enough cases for statistical comparison.


Mechanisms underlying membrane rectification and firing properties in type I and type II

neurons:

In a number of neurons the mechanisms

underlying membrane rectification and

possibly burst generation were studied.

Due to the relatively low incidence of

recordings that could be obtained from

type III and type IV neurons under blind-

patch conditions these analyses were

restricted to type I and type II cells.

In NCL neurons inwardly rectifying

properties in the hyperpolarized voltage

range seem to depend on K+

conductances. Blockade of K+ channels

by extracellular CsCl (5 mM) abolished

fast inward rectification to

hyperpolarizing current pulses in 5 type

II neurons which showed rapidly

adapting burst firing and 2 regular

spiking neurons classified as type II

cells. Similarly, in a third type II neuron

which showed sag in the voltage

response to hyperpolarizing current

pulses under control conditions, both the

fast and the time-dependent inward

rectifying conductances were blocked by

CsCl (data not shown).

Rectification in the depolarized voltage range was studied in one type I neuron and two type II

neurons. The firing characteristics of these cells were recorded before and after adding 1 µM

tetrodotoxin (TTX) to the perfusate (figure 11). As expected for a Na+ dependent action

potential, TTX completely eliminated the generation of spikes in all cells tested. Under control

conditions the type I cell showed transient depolarizing „hump“ potentials in the subthreshold

range which preceeded the generation of spikes (fig 11A, left column). These hump-potentials

Figure 9: Morphological features of type IV cells. A, B: Twoexamples of Lucifer Yellow filled cells that had high input resistance,small action-potentials of long duration and showed only little inwardrectification to hyperpolarizing current pulses. C: Detail Scale bars are50 µm (A, B), and 20 µm (C).


Figure 10: Comparison of the dendritic morphology and the appearance of local axonal arborizations in the 4 cell typesdescribed here. Somata and dendrites are drawn in green, axon collaterals in red. With regard to spine density and dendriticthickness the examples shown here for the type I and type II neurons represent extremes of what appeared to be a continuum.Scale bars are 50 µM.


persisted in the presence of TTX (fig 11A, right column), indicating that they might be mediated

by a Ca++ conductance. Like most type I neurons this cell also showed membrane outward

rectification, measured at the end of the current pulse, both under control conditions and in the

presence of TTX. In the type II neurons there was a small anomalous rectification in the

depolarized voltage range under control conditions. The membrane potential tended to „sag“

back towards rest and an undershoot of the potential occured after cessation of the current pulse

(fig. 11B, left column). The underlying current that mediates the inward rectification in the

depolarized voltage range is sensitive to TTX blockade. After bath application of TTX

depolarizing pulses unveiled activation of an outward rectifying conductance (fig 11B, right

column).

In the course of other experiments (data not shown) we had observed spontaneously occuring

all-or-none bursts in type I and type II neurons when picrotoxin (100 µM), a Cl- channel blocker

that eliminates the fast inhibitory effects of GABAergic interneurons, was added to the external

solution. In these spontaneous bursts usually 4 - 6 spikes rode on a depolarizing envelope that

was followed by an afterhyperpolarization of relatively long duration (tens to hundreds of ms)

Type I200 ms

20 mV

A control

- 69.5 mV

1µM TTX

Type II

- 70.5 mV

B control

200 ms

20 mV

1µM TTX

0.5 nA

0.5 nA

Figure 11: Membrane rectification in a type I neuron (A) and a type II neuron (B) before (left traces) and after blockadeof sodium channels by bath-application of tetrodotoxin (TTX, right traces). A: Under control conditions single spikes werefollowed by depolarizing afterpotentials in this type I neuron. Application of TTX abolishes the fast action potentials, butsmall transient depolarizing potentials persist at the beginning of strong depolarizing current pulses. This neuron also showedoutward rectification at depolarized potentials under both conditions. B: Under control conditions this type II neuron showedinward rectification and sag in the response to depolarized current pulses, and a voltage undershoot after the cessation ofcurrent pulses. Application of TTX abolishes the inward rectification and the undershoot.


(fig. 12). In addition, when depolarizing

current steps were applied from

hyperpolarized potentials (about -90 mV)

both type I and type II cells responded

with an increase in the number of spikes,

indicating the contribution of a voltage-

gated Ca++ conductance, that is normally

inactive at more positive holding

potentials (e.g. Llinás and Yarom, 1981).

However, changing the cells holding

potential with DC injections never

changed the cell´s firing pattern (c.f.

McCormick and Feeser, 1990; Livingston

and Mooney, 1997).

Taken together, these observations indicated that calcium conductances contribute to the firing

behaviour of cells in the NCL. To test this assumption an additional 7 neurons (8 days post-

hatch) were recorded under conditions suitable to evoke calcium spikes. Electrodes were filled

with CsCl (135 mM) and tetraethylamonium (TEA; 20 mM) to block most K+ conductances.

Immediately following establishment of the whole-cell configuration, a cells firing pattern was

assessed in voltage-clamp. Of the 7 cells

tested, 4 cells showed firing

characteristics of type I cells, while the

remaining 3 neurons maintained repetitive

spiking throughout the duration of the

depolarizing potential step. Following

theses measurements TTX (1 µM) was

bath applied. When K+ conductances were

blocked all neurons tested were able to

generate slow action potentials in the

presence of 1 µM TTX (fig. 13). Such

slow action potentials are most likely

mediated by Ca++ influx (c.f. McCormick

and Prince, 1987).

mAHP

DAP

mAHP

100 ms

20 mV

Figure 12: Spontaneous bursts of action potentials occurred in typeI and type II cells when GABAergic inhibition was diminished byadding picrotoxin (100 µM) to the extracellular solution. In thesebursts 4 - 6 fast action-potentials rode on a slowly developingdepolarizing envelope that was followed by a fastafterhyperpolarization (fAHP) and a medium-durationafterhyperpolarization (mAHP), often with an intercalateddepolarizing afterpotential (DAP), see insert.

Figure 13: Examples of Ca++ spikes that could be evoked in largespiny neurons (n = 7) when K+ conductances were blocked byadding tetraethylamonium (TEA; 20 mM) and CsCl (135 mM) to thepipette solution, and sodium conductances were blocked byextracellular tetrodotoxin (TTX; 1µM).


DISCUSSION

We have shown the existence of at least 4 celltypes within the NCL of chicks that differ in their

intrinsic electrophysiological and morphological properties. In general, there was a good

correspondence of physiological and morphological criteria, although the two main celltypes,

type I and II, could not be clearly differentiated by morphological criteria. Neurons from

animals between 2 and 11 days post hatch were studied. As chicks are relatively mature at

hatching and capable of independent action most behavioural studies that test the animals ability

to learn are undertaken within the first two weeks after hatch. As discussed below, several

electrophysiological characteristics that differentiated among neurons between (and within)

celltypes showed some degree of variation related to age, thus probably reflecting

developmental changes. In addition, our finding of 4 main celltypes does not exclude the

existence of further classes of neurons within NCL.

Cell types:

Type I: Type I neurons possessed a high action potential threshold, particularly in relation to

their low resting potential, and they showed prominent outward rectification in response to

depolarizing current pulses. These properties indicate that strong, temporally or spatially

integrated excitatory inputs are neccessary for type I neurons to fire. If sufficiently depolarized,

cells in this class preferentially responded with bursts of action potentials. Interestingly, in the

mammalian cortex it has also been observed that the threshold of bursting neurons is sometimes

higher than that for neurons generating single spikes (Mason and Larkman 1990).

We have used the occurrence of transient „hump“-like depolarizing responses to describe

subsets among the neurons which were grouped as type I cells, according to their burst-firing

behaviour. Neurons that displayed prominent subthreshold „hump“-potentials and those that did

not, differed in several basic parameters, i.e. input resistance, spike amplitude, spike duration

and rise time. In mammals these properties have been shown to be particularly prone to

developmental changes during early postnatal development (McCormick and Prince, 1987;

Kasper et al., 1994b). Thus it is conceivable that the differences among type I neurons reflect at

least in part developmental differences. Cells showing transient „hump“-like depolarizing

responses were indeed recorded in slices from animals with higher mean ages (7.1 days vs. 5.4

days), a difference that was statistically significant (df = 32; t = -2.7; p < 0.05; Student´s t test

for unpaired samples). Similarly, developmental differences might explain the occurrence of a

complex pattern of afterpotentials in some, but not all, type I and type II neurons. Single action

potentials or bursts were sometimes followed by three distinct afterpotentials: an initial fast,

and a late slow after-hyperpolarization, separated by an intercalated depolarizing


afterpotential. In mammalian cortical neurons the two phases of the AHP appear to be due to the

activation of Ca++-dependent K+ currents (Lancaster and Adams, 1986; Schwindt et al., 1988)

whereas the DAP results from the activation of a variety of Ca++ dependent currents (Higashi,

1993; Yang et al., 1996; Haj-Dahmane and Andrade, 1997). The expression of the DAP and the

emergence of a tripartionate AHP have been shown to depend on changes during postnatal

development (Kasper et al., 1994b). Similarly, in mammals burst firing does not develop

before the third postnatal week (Kasper et al., 1994b).

A common factor pertinent to the occurrence of either the „humps“ and the DAPs, as well as the

expression of full bursting behaviour in our preparations might be an increase in Ca++

conductances during development. Calcium currents are believed to underlie burst firing in

mammalian central neurons (Wong and Prince, 1981; Llinás and Jahnsen 1982; Friedman and

Gutnick, 1987; Schwindt and Crill, 1999). Notably, the generation of transient depolarizing

potentials by low-threshold Ca++ currents can result in bursts of fast Na+/K+ action potentials

(Llinás and Jahnsen 1982; McCormick and Feeser, 1990; Yang et al., 1996).

When the network activity in the slice was increased by diminishing the inhibitory control of

GABAergic interneurons, spontaneous burst could be seen in type I and II cells that were

clamped at or just below resting potential. In these bursts short-duration action potentials rode

on a slowly activated depolarizing envelope, resembling the plateau potentials generated by

low-threshold Ca++ currents in mammalian central neurons (Llinás and Jahnsen 1982;

McCormick and Feeser, 1990; Yang et al., 1996; Magee and Carruth, 1999). A contribution of a

low threshold Ca++ current to the firing behaviour of the large projection neurons was further

indicated by the fact that preconditioning hyperpolarization of the membrane potential led to an

increased number of spikes in response to a depolarizing current pulse. Also, in addition to

high-threshold Ca++ spikes, transient „hump“-like depolarizing potentials similar to those

occuring in some type I neurons under control conditions could be evoked in the presence of

TTX and TEA. Furthermore, about half of the type I neurons that showed „hump“-like potentials

(6 of 13) also showed prominent DAPs following single spikes. As briefly outlined above the

expression of DAPs depends either directly on low-threshold Ca++ currents (Zhang et al., 1993;

Yang et al., 1996) or the activation of other Ca++ dependent currents (Higashi, 1993; Haj-

Dahmane and Andrade, 1997). Functionally, depolarizing afterpotentials appear to be related to

a neurons ability to fire bursts of action potentials (Wong and Prince, 1981; Chagnac-Amitai et

al., 1990; Mason and Larkman, 1990; Yang et al., 1996; Magee and Carruth, 1999). And it has

been postulated that the Ca++ entry resulting from the DAP is responsible for long-term-

potentation (LTP) because manipulations that eliminate the DAP also block the induction of

LTP (Komatsu and Iwakiri, 1992). While in sum these findings strongly implicate the


contribution of Ca++ influx in the emergence of burst firing in type I and type II cells, it should

be noted that other components such as a sodium current (Schwindt et al., 1988; Franceschetti et

al., 1995; Magee and Carruth, 1999) or outward currents (Schwindt et al., 1988: Kang and

Kayano, 1994) might also be involved in the generation of bursts. Taken together, subthreshold

potentials appear to serve an important role in sculpturing a cells final output (Hamill et al.,

1991), and their occurrence in the large spiny neurons of type I and II might reflect the

integrative role these cells have in combining the diverse multimodal synaptic inputs that reach

the NCL from various sensory and associative areas (Metzger et al., 1998; Kröner and

Güntürkün, 1999). Furthermore, a neurons ability to fire bursts of action potentials will

probably enhance its integrative role within a network and aide the plasticity of synaptic

connections: Burst firing increases the probability to drive the postsynaptic cell beyond spike

threshold (Snider et al., 1998). The high frequency firing of action potentials during a burst is

thus thought to amplify a neural signal and to synchronize the activity in a population of

postsynaptic cells, both temporarily and spatially (Hablitz, 1986; Snider et al., 1998).

Type II: An initial tonic firing and a relatively low action potential threshold that characterize

type II neurons may indicate that the firing of these cells is readily elicited by weak excitatory

inputs. The occurrence of spontaneous burst in the absence of GABAergic inhibition and the

phasic-tonic firing pattern that often emerged with large depolarizing currents indicates that

type II neurons will respond strongly but transiently to a brief, new stimulus, yet produce a

sustained response to a maintained input. In contrast to the behaviour of neurons in the burst

firing mode, a pattern of single action potentials can give a more faithful representation of the

characteristics of depolarizing inputs. Functionally, the ability of type II neurons to generate a

tonic firing mode could also enable these cells to retain informations of their input for a short

time period. It is known that NCL-lesions in pigeons cause specific impairments in a variety of

delay tasks (Mogensen and Divac, 1982, 1993, Gagliardo et al., 1996, 1997, Güntürkün, 1997,

Güntürkün and Durstewitz, in press). In these experiments, the animal has to hold online a

specific information provided during a previous cue period to perform a correct response after

delay offset. In mammals, neurons in the deep layers of the prefrontal cortex (PFC) have been

identified as a cellular substrate for short term memory (see Goldman-Rakic, 1996 for review).

When the activity of these cells is recorded in awake rats or primates performing a delayed-

response task, a significant enhancement in their firing rate can be observed during the delay

(e.g. Fuster and Alexander, 1971; Funahashi et al., 1989; Sawaguchi and Yamane, 1999). Thus,

sustained delay activity of PFC neurons may provide the animal with the ability to hold an

internal representation of relevant aspects of the external world which is needed for planning or


organizing subsequent responses (Goldman-Rakic, 1996). Neurons with similar delay activities

to those recorded from rat or primate PFC have been observed in vivo in the pigeon’s NCL

during a delayed auditory-visual Go/No-Go task (Kalt et al., 1999). Like their mammalian

counterparts, these cells showed significant alterations of their firing mode during the delay

period. It is conceivable that the type II neurons of the present study may hold the information of

their synaptic inputs within a recurrent network and thus provide a part of the cellular substrate

for the delay activity within the NCL.

At least some type II neurons furthermore showed a small transient sag in the voltage response

to hyperpolarizing currents, which is likely to be mediated by a voltage- and time-dependent

hyperpolarization activated mixed cationic current called Ih (Pape, 1996; see also discussion of

type III below). It has been shown that a time-dependent-inward rectifier produces membrane

resonance (Hutcheon et al., 1996). When resonance is present, neurons fire action potentials

preferentially to rhythmic (oscillatory) inputs. The membrane resonance produced by Ih can be

amplified by a persistent sodium current (Hutcheon et al., 1996). In type II neurons we

observed inward rectification in the depolarized voltage range that may be indicative of a

persistent sodium current (Stafstrom et al., 1985, Yang et al., 1996). The existence of such a

current has already been demonstrated directly in neurons from the caudal forebrain of

songbirds which possess similar firing properties and morphological characteristics as the type

II neurons described here (Kubota and Saito, 1991; Kubota and Taniguchi, 1998). Thus type II

neurons might represent members of a common class of spiny projection neurons in the dorsal

forebrain of birds that can respond preferentially to input of a specific pattern.

Type III: The electrophysiological and morphological characteristics of the type III cells

resemble those of γ-aminobutyric acid-containing inhibitory interneurons found in the

mammalian telencephalon (Schwartzkroin and Mathers, 1978; McCormick et al., 1985;

Kawaguchi, 1995); i.e. little or no accomodation of spike frequency, and beaded aspiny

dendrites. Type III cells offer a wide-band responsiveness and sustained high-frequency firing

if neccessary. They are thus able to perform a very reliable input-output conversion, retaining

the temporal pattern of their synaptic inputs. These firing characteristics are facilitated by the

short rise and fall times of the action potential (table 1). In addition, the relatively pronounced

AHPs of type III cells might also contribute to the ability to fire actions potentials at high

frequencies in that they reduce the accumulation of depolarization-dependent Na+ channel

inactivation, which would otherwise reduce action potential amplitudes during long trains of

high frequency firing (Hamill et al., 1991). It will also reduce the influx of Ca++ into the cell

and thus will diminish the effects of Ca++-activated K+ channels which otherwise might lead to


prolonged hyperpolarization and spike frequency adaptation (Lancaster and Adams, 1986;

Storm, 1990).

A prominent feature in all type III cells was the existence of the strong inward-rectifying current

Ih that initiates slow depolarization if the membrane potential has become negative (e.g. by a

GABAB receptor-mediated IPSP or a postspike AHP). In central neurons of mammals Ih is a

mixed Na+/K+ current that has been implicated in the determination of the resting potential and

the generation of „pacemaker“ potentials. It also serves to decrease the propagation of

subthreshold voltage potentials in dendritic trees, thereby regulating the integration of synaptic

inputs (McCormick and Pape, 1990; Magee, 1998, 1999; see Pape, 1996 for review). In

mammalian cortex GABAergic interneurons are long recognized for such various functions as

maintaining the normal integrative properties of the cortex (Sillito, 1984), the control of spread

of activity (Chagnac-Amitai and Connors, 1989) and the synchronization of adjacent projection

neurons via inhibitory phasing (Cobb et al., 1995; Benardo, 1997), which might impose a

rhythm on the activity of the principal neurons, providing a „context“ to the „content“ of their

operation (Buzsáki and Chrobak, 1995). In the prefrontal cortex interneurons also appear to

posses own sensory or „mnemonic“ fields that enable them to fine tune the delay activity of

adjacent pyramidal neurons in a working memory task (Rao et al., 1999). With respect to the

extensive local axonal arborizations of each type III neuron these putative interneurons of the

NCL presumably also can control a large number of adjacent spiny „principal“ neurons (i.e.

types I and II), as is the case in mammalian cortex (e.g. Sik et al., 1995). In sum, given their

basic intrinsic electrophysiological and morphological properties type III neurons of the NCL

are equipped to play a similarly important role for signal integration as their mammalian

counterparts.

Type IV: Type IV neurons were characterized by high input-resistances, long time constants and

a long duration of action potentials. These electrophysiological properties are also

characteristic for developing cortical neurons in mammals (Kasper et al., 1994b; McCormick

and Prince, 1987) and possibly birds (Kubota and Taniguchi, 1998). McCormick and Prince

(1987) have argued that higher input resistances and longer membrane time constants of

immature neurons could increase the magnitude and duration of their responses to imposed

synaptic currents. Such cells should also be electrotonically compact, showing relatively little

cable attenuation of distal dendritic synaptic inputs, thus compensating for similarly poorly

developed synaptic inputs in newborn animals. This is in good agreement with the

morphological appearence of our type IV cells identified here which possesed small somata,

and short, sparsely spineous dendrites with relatively few branchings.


Similarly, the long duration of action potentials seen in type IV neurons may serve to facilitate

the output of developing cells: Activity-dependent spike broadening due to a reduction in

specific K+ currents has been correlated with increased efficacy of synaptic transmission onto

the postsynaptic target cell (Mercer et al., 1991; Byrne and Kandel, 1996 for review).

Similarly, the significantly longer action potentials of developing neurons might result in an

enhancement of synaptic transmission onto their postsynaptic cells (Bottjer et al., 1998). An

open question remains whether type IV cells constitute a distinct class of late maturing neurons

or simply represent developing cells of the other spiny celltypes (types I and II, respectively).

However, all physiological and morphological characteristics of type I and II neurons were

already present at the earliest ages recorded. Moreover, type IV cells indeed were recorded in

slices from animals with the highest median age (7 d post hatch). Bradley and coworkers

(1996) have proposed yet another mechanism that may account for developmental and

experience-dependent changes in input resistance that occur in neurons of the chicks IMHV.

They argue that existing large neurons with numerous spines might be morphologically

remodelled to produce cells with significantly higher resistance to render them more

susceptible to presynaptic inputs. It has been shown that in the NCL auditory imprinting reduces

the spine densities on the medial and distal segments of large multipolar neurons, which

correspond most likely to both types I and II of the present study (Bock and Braun, 1999a,b).

While the resulting decrease in membrane surface should lead to a decrease in capacitance and

an increase of apparent resistance, these changes are probably too subtle to account for the

large differences in morphology and electrophysiology observed between type IV neurons and

neurons of the other spiny cell classes.

Connectivity and sensorimotor integration

An important requirement for sensorimotor integration and associative learning is the ability of

a brain structure to combine sensory information from various sources and to affect the animals

behaviour either directly or via motoric and limbic structures downstream. Tracing studies in

vivo have revealed that the NCL receives converging input from the secondary sensory areas of

all modalities, as well as from other associative areas which themselves appear to integrate

information from more than one modality (Metzger et al., 1998; Kröner and Güntürkün, 1999).

Although some degree of regional specialization within NCL might exist, the pattern of afferents

strongly indicates that in large areas of the NCL single neurons might receive multimodal input.

In the pigeon single unit recordings in vivo have shown that neurons in the NCL can encode


aspects of a memory task across modalities (Kalt et al., 1999). The same neurons might be

involved in a cell assemblies that „hold“ information over a delay or initiate a motor response

via the archistriatum. In the chick Metzger and coworkers (1998) have shown that single

neurons within the NCL send axon collaterals to the archistriatum and another multimodal

forebrain area, the intermediate, medial part of the hyperstriatum ventrale (IMHV). Our findings

provide further insight into the anatomical arrangement that might underlie the integration of

various sensory features and the propagation of the NCL´s output to other forebrain areas.

Morphologically, single cells in all spiny cell classes of the NCL seem to be able to relay their

transformed input not only to neighbouring cells within NCL, but also to sensorimotor and

limbic areas downstream, namely the archistriatum and the basal ganglia, and possibly also

provide feedback to the sensory areas the original inputs come from. With the exception of the

small type IV cells neurons in the NCL had fairly large dendritic fields (see table 2). The extent

of the local axon collaterals for the spiny cell classes (types I, II, and IV) can be estimated at a

lower limit of about two or three times the area of a given cells dendritic field (c.f. figs. 5, 10).

Type III neurons had relatively symmetric axonal fields and their collaterals did not leave the

NCL. The diameter of the area covered by axonal arborizations of these cells ranged between

545 - 736 µm (mean area 320,400 µm2). A hypothetical basic „circuit“ consisting of a type III

interneuron and two postsynaptic spiny neurons at the rim of the axonal field would thus cover

approximately 650,000 µm2. Such a basic circuit could thus participate in intra-NCL

computations which surpass areas of representation of a single sensory modality. In addition,

most spiny cells were seen to send several long collaterals to targets outside the NCL. On their

way these collaterals often span the whole medio-lateral and/or dorso-ventral extent of the

NCL (c.f. fig. 5). Although the determination of the rostro-caudal extent of these intra-NCL

connections remains a critical topic which was constrained by the thickness of the slices used

here, it appears that these connections enable a stimulus comparison across most, if not all,

possible sensory compartments of the NCL (c.f. Metzger et al., 1998; Kröner and Güntürkün,

1999).

Based on similarities in the pattern of connections and the location in the dorsocaudal forebrain

it has been suggested that the Nd, the auditory subunit of the NCL might be related to the

sensorimotor „high-vocal center“ (HVC) of songbirds (Wild, 1994; Metzger et al., 1998;

Kröner and Güntürkün, 1999). Nucleus HVC is essential for song production and shows

selective responses to complex auditory stimuli. Two pathways arise from within HVC that

converge on the same premotor nucleus of the archistriatum. The first is a direct projection from

HVC to the robust nucleus of the archistriatum (RA), which serves as the primary motor

pathway for song production, and can also carry auditory information to RA. The other


projection of HVc begins a pathway through the anterior forebrain that is crucial for song

learning but, although active during singing, is not essential for adult song production. In this

pathway Area X of the lobus parolfactorius, a motor structure of the basal ganglia, is the first

target of the HVC (Nottebohm et al., 1982; Fortune and Margoliash, 1995). In the HVC of zebra

finches Area X and RA projecting neurons constitute two separate populations which resemble

type I and II neurons of the present study with regard to intrinsic electrophysiological

characteristics and morphology (Dutar et al., 1998; Kubota and Taniguchi, 1998). Neurons that

project to RA show inward and outward rectification in the subtreshold range (Dutar et al.,

1998; Kubota and Taniguchi, 1998) and possibly a strong accomodation in the firing pattern

(Dutar et al., 1998; but see Kubota and Taniguchi, 1998), comparable to that seen in at least

some type I neurons described here. Area X projecting neurons show a time-dependent inward

rectification and regular spiking pattern very similar to that described for type II cells here

(Dutar et al., 1998; Kubota and Taniguchi, 1998). However, our finding that almost all spiny

neurons of the NCL sent at least one axon collateral in the direction of the archistriatum and

sometimes also in the direction of the basal ganglia makes it unlikely that a similar distinction

between cell types exists as in the songbird forebrain. These differences might reflect the

specialization of the HVC relevant for singing and the need to exert control over specific

aspects of singing related motor behaviour. Yet in the songbird brain parts of the lateral

neostriatum and the „shelf“ region which adjoin the HVC share most afferents with the HVC

(Fortune and Margoliash, 1995; Vates et al., 1996; Foster et al., 1997; Iyengar et al., 1999) and

also show a similar pattern of projections (Fortune and Margoliash, 1995; Mello et al., 1998;

Iyengar et al., 1999), therefore these areas might resemble the NCL more closely.

In sum, our findings provide a framework for further study into sensorimotor integration and

associative learning in chicks and possibly other birds.

Acknowledgement: The authors would like to thank Markus Hausmann, Burkhard Hellmann, Thomas Kalt, Ralf

Mohrmann, Simon Rumpel and Markus Werner, who all gave valuable comments and technical support at

various stages of this study. This work was supported by grants of the DFG (GU 227/5 - 1 and SFB 509).

Chapter 4: Dopaminergic innervation of the avian telencephalon 102

The Dopaminergic Innervation of the Avian Telencephalon

Daniel Durstewitz, Sven Kröner and Onur Güntürkün

Table of contents

1. Introduction

2. Overview over the neuroanatomy of the avian telencephalon

3. Origin of dopaminergic fibers in the avian telencephalon: Mesencephalic A8-A10 cell groups

4. General features of the dopaminergic innervation of the avian telencephalon

4.1. 'Basket'- and 'en passant'-type of dopaminergic innervation

4.2. Tyrosine hydroxylase as a marker of dopaminergic fibers in the avian telencephalon

4.3. DARPP-32 as a marker of D1-receptors and dopaminoceptive neurons

4.4. Ultrastructural features of the dopaminergic innervation and of dopaminoceptive neurons

5. Distribution of dopaminergic fibers and dopaminoceptive elements in the avian telencephalon

5.1. Primary sensory areas

5.1.1. Dopaminergic fibers and projections

5.1.2. D1 receptors

5.1.3. D2 receptors

5.1.4. DARPP-32

5.1.5. Colocalizations

5.2. Secondary sensory and multimodal areas


5.2.2. D1 receptors

5.2.3. D2 receptors

5.2.4. DARPP-32


5.3. Septum, hippocampal complex, archistriatum


5.3.2. D1 receptors

5.3.3. D2 receptors

5.3.4. DARPP-32


5.4. Basal ganglia


5.4.2. D1 receptors

5.4.3. D2 receptors


5.4.4. DARPP-32


6. Comparative aspects of the dopaminergic system in mammalian and avian species

6.1. Basal ganglia

6.2. Cortex

6.3. Laminae-specific distribution of dopaminergic fibers

6.4. Ultrastructural features and postsynaptic targets of the dopaminergic innervation

7. Behavioral studies and functional implications

7.1. Dopaminergic modulation of (unconditioned) motor functions

7.2. Dopaminergic involvement in arousal and wakefulness

7.3. Dopaminergic modulation of appetitive and consummatory behavior

7.4. Dopaminergic modulation of learning and conditioned behavior

7.5. Dopaminergic modulation of working memory

7.6. The possible function of dopamine in sensory-motor processes

Abstract

The present review provides an overview of the distribution of dopaminergic fibers and dopaminoceptive

elements within the avian telencephalon, the possible interactions of dopamine with other biochemically

identified systems as revealed by immunocytochemistry, and the involvement of dopamine in behavioral

processes in birds. Primary sensory structures are largely devoid of dopaminergic fibers, dopamine receptors,

and the D1-related phosphoprotein DARPP-32, while all these dopaminergic markers gradually increase in

density from the secondary sensory to the multimodal association and the limbic and motor output areas.

Structures of the avian basal ganglia are most densely innervated but, in contrast to mammals, show a higher D2

than D1 receptor density. In most of the remaining telencephalon D1 receptors clearly outnumber D2

receptors. Dopaminergic fibers in the avian telencephalon often show a peculiar arrangement where fibers coil

around the somata and proximal dendrites of neurons like baskets, probably providing them with a massive

dopaminergic input. Basket-like innervation of DARPP-32-positive neurons seems to be most prominent in the

multimodal association areas. Taken together, these anatomical findings indicate a specific role of dopamine in

higher order learning and sensory-motor processes, while primary sensory processes are less affected. This

conclusion is supported by behavioral findings which show that in birds, as in mammals, dopamine is

specifically involved in sensory-motor integration, attention and arousal, learning, and working memory. Thus,

despite considerable differences in the anatomical organization of the avian and mammalian forebrain, the

organization of the dopaminergic system and its behavioral functions are very similar in birds and mammals.

Progress in Neurobiology 59:161-195 (1999)


1. Introduction

The neuromodulatory dopaminergic system of mammals and birds is critically involved in

numerous cognitive and behavioral functions, including appetitive and aversive behaviors

(Balthazart et al., 1997; Bertolucci-D'Angio et al., 1990b; Deviche, 1984; Salamone, 1992,

1994; Sokolowski et al., 1994), states of arousal and wakefulness (Ferrari and Giuliani, 1993;

Ongini, 1993), motor control (Goodman et al., 1983; Rieke, 1981, 1982; Salamone, 1992;

Sokolowski & Salamone, 1994; Waddington and Daly, 1993), learning and memory (Beninger,

1993; Gruss and Braun, 1997; McDougall et al., 1987; Schultz et al., 1995), as well as working

memory and attention (Güntürkün & Durstewitz, in press; Montaron et al., 1982; Roberts et al.,

1994; Sawaguchi and Goldman-Rakic, 1991, 1994; Schultz et al., 1993; Seamans et al., 1998;

Zahrt et al., 1997). Since structure and function are tightly coupled in neural systems,

information about the neuroanatomical organization of the dopaminergic system is required to

unravel the functional role of dopamine (DA) in these various cognitive and behavioral

processes. A large body of evidence indicates that the principal function of dopamine is very

similar in mammals and birds. Since members of these two classes of vertebrates have highly

developed cognitive skills (e.g., Epstein et al., 1984; von Fersen et al., 1990; Lanza et al.,

1982; Lubow, 1974) but radically different forebrain organizations, comparative studies may

uncover the invariant structural features which enable the dopaminergic system to exert its

specific functions.

The present article provides an overview over the structural organization of the dopaminergic

system in avian species, compares it to mammals, and discusses behavioral findings and

possible functional implications. However, this review will be restricted in two ways: First,

almost all studies that dealt specifically with markers of the dopaminergic system in the avian

brain (i.e., DA, DA receptors, DARPP-32) have been conducted within the last decade, and the

present review will focus mainly on these studies. Older literature on the general distribution of

markers of the monoaminergic or catecholaminergic systems are covered in the excellent

review by Reiner et al. (1994). Second, most of our knowledge on the dopaminergic

innervation of the bird brain derives from studies in pigeons and chicks. However, the few

available studies from other avian species make it likely that the results obtained in these avian

species generalize to other birds as well.


2. Overview over the neuroanatomy of the avian telencephalon

The avian telencephalon consists of nuclear structures, and mostly lacks the layered

organization that is characteristic of mammalian isocortex (Figure 1). This lack of lamination

led early investigators to assume that most of the avian telencephalon consists merely of a

hypertrophied striatum, and accordingly most forebrain areas were termed with the suffix '-

striatum', and this nomenclature remains still in use (Ariëns-Kappers et al., 1936; Edinger,

1903). On the basis of embryological (Källén, 1953, 1962; Kuhlenbeck, 1938; see also

Striedter et al., 1998) cytoarchitectonic (Rehkämper et al., 1984, 1985, Rehkämper and Zilles,

1991), hodological (Karten, 1969; Karten and Dubbeldam, 1973; Veenman et al., 1995) and

Figure 1. Morphology of the avian telencephalon (brain of a pigeon, Columba livia). (A-C) A series of hemispheres stained forcresyl-violet, indicating the location of several major structures. Rostro-caudal positions of the slices shown are approximately A4.25 (A), A 7.25 (B), and A 10.00 (C) according to the pigeon brain atlas of Karten and Hodos (1967). (D) Detail from C,showing part of the basal ganglia (PA, PP), the primary recipient area of the visual tectofugal pathway (ectostriatum), and apresumed multimodal area (NIM). (E-G) Prototypical examples of cells from the pigeon telencephalon that were labeled byintracellular filling (E) or Golgi impregnation (F, G). All cells display multipolar dendritic fields, a feature that seems to beassociated with the organization of the avian forebrain in nuclei rather than layers. Scale bars represent 1 mm (A, B, C), 50 µm(D), 25 µm (E, F, G). See list for abbreviations.


histochemical (Juorio and Vogt, 1967) criteria, however, it became clear that large parts of the

telencephalon could be regarded as pallial, and as such might constitute possible cortex

equivalents.

The flow of sensory information in the avian telencephalon follows a common pattern for all

modalities: The primary sensory areas which receive subtelencephalic input, relay their

information to adjacent secondary sensory structures, which in turn project in parallel to

presumed multimodal areas as well as to motor and limbic structures, including the basal

ganglia (Figure 2).

As in all amniotes, two main visual pathways convey visual information from the retina to the

telencephalon in birds. These are the thalamofugal and the tectofugal pathway. The tectofugal

pathway may correspond to the mammalian colliculo-thalamo-extrastriate pathway, whereas the

thalamofugal pathway is comparable to the mammalian geniculostriate pathway (review by

Güntürkün, 1991). The ectostriatum (E) is the primary sensory structure of the tectofugal

pathway (Benowitz and Karten, 1976; Kondo, 1933). From there intratelencephalic projections

lead to the surrounding ectostriatal belt (Karten and Hodos, 1970; Ritchie, 1979). The

thalamofugal visual pathway consists of a bilateral projection from a retinorecipient nucleus in

the dorsal thalamus onto a thin band of granule cells, the intercalated nucleus of the

hyperstriatum accessorium (IHA), as well as to the lateral part of the hyperstriatum dorsale

(HD), which together constitute the primary sensory areas within the so-called Wulst. The IHA

and HDl in turn project to secondary sensory structures within the Wulst, mainly to the

hyperstriatum accessorium (HA; Karten et al., 1973; Shimizu et al., 1995).

The same organization applies to a part of the somatosensory system which occupies the rostral

extent of the Wulst, and that seems to be comparable to the mammalian primary somatosensory

cortex (Karten, 1971; Medina et al., 1997): Somatosensory fibers from the dorsal thalamus

terminate in the HD and the IHA (Delius and Bennetto, 1972, Funke, 1989, Wild, 1997) which

in turn project to the rostral HA (Wild, 1987). Another primary somatosensory telencephalic

area is located within the nucleus basalis (Bas). The Bas mainly processes trigeminal

information (Schall et al., 1986), but also receives additional somatosensory and auditory input

(Arends and Zeigler, 1986; Wild and Farabaugh, 1996; Wild et al., 1997). It relays this

information to the overlying frontal neostriatum (NF; Wild and Farabaugh, 1996; Wild et al.,

1985).

The main telencephalic area reached by auditory input from the thalamus is the so-called Field

L complex in the caudomedial neostriatum (Karten, 1968). Field L2 is the primary sensory


structure of the auditory system, which projects to adjacent fields L1 and L3 (Bonke et al.,

1979; Karten, 1968; Wild et al., 1993).

The secondary sensory areas of all modalities are reciprocally connected with a region in the

caudolateral neostriatum, which in the pigeon is referred to as NCL (Kröner and Güntürkün,

1999; Leutgeb et al., 1996; Metzger et al., 1998). Other multimodal areas probably exist in the

intermediate aspect of the hyperstriatum ventrale (IMHV) and the intermediate neostriatum

(NIM; or mediorostral neostriatum/hyperstriatum ventrale, MNH, in chicks). Both these areas

have been studied extensively in chicks in the context of imprinting (Bradley et al., 1985;

Bredenkötter and Braun, 1997; Metzger et al., 1996; reviews: Horn, 1998; Scheich, 1987). In

addition, a system of "cortico"-striatal projection neurons which occupies the most dorsal and

lateral extent of the external pallium (PE; Brauth et al., 1978; Veenman et al., 1995) includes,

from rostral to caudal, the lateral frontal neostriatum (NFL), the lateral intermediate neostriatum

(NIL), the area temporo-parieto-occcipitalis (TPO) and the area corticoidea dorsolateralis

(CDL).

Figure 2. Schematic overview of the functional anatomy of the pigeon telencephalon as described in Section 2. Brain areas thatare still poorly understood with regards to their connections and functions were left blank. It should be noted, however, that thehyperstriatum intercalatus superior (HIS) probably receives secondary and, to a limited extent, possibly also primary visual input.The hyperstriatum ventrale (HV) is a highly heterogeneous region which probably encompasses multiple polysensorysubdivisions.


At the next level of information processing in the avian brain, sensory information is funneled in

parallel from both, the tertiary as well as the secondary sensory areas to the basal ganglia and,

in addition, to the ventrolaterally located archistriatum (e.g. Bradley et al., 1985; Csillag et al.,

1994; Kröner and Güntürkün, 1999; Leutgeb et al., 1996; Metzger et al., 1998; Phillips, 1966;

Shimizu et al., 1995; Wild et al., 1985, 1993). Functionally, the avian archistriatum has been

divided into two main subdivisions (Davies et al., 1997; Dubbeldam et al., 1997; Zeier and

Karten, 1971). These are a somatic sensorimotor part and a viscerolimbic division that is

considered to be equivalent to the mammalian amygdala (Davies et al., 1997; Dubbeldam et al.,

1997; Zeier and Karten, 1971).

The connectivity, neurotransmitter content, and cytoarchitecture of the avian basal ganglia are

highly similar to that in mammals (Anderson and Reiner, 1991b; Karten and Dubbeldam, 1973;

Reiner and Anderson, 1990; Reiner et al., 1983, 1984a; Veenman and Reiner, 1994; Veenman

et al., 1995). Based on the palliostriatal connections, a functional segregation of the avian

striatum into associative, sensorimotor and limbic territories has recently been suggested

(Veenman et al., 1995), similar to the situation in mammals (Parent, 1990). The paleostriatum

augmentatum (PA) and lobus parolfactorius (LPO) make up the dorsal (somatomotor) striatum.

Ventral striatal structures include the nucleus accumbens (Acc), the bed nucleus of the stria

terminalis (BNST), and the olfactory tubercle (TO) which constitute the ’visceral-limbic’ parts

of the avian striatum. The pallidal parts of the avian basal ganglia consist of the paleostriatum

primitivum (PP), which is homologous to the globus pallidus of mammals, and the ventral

pallidum (VP) which is comparable to the "limbic" ventral pallidum of mammals (Karten and

Dubbeldam, 1973; Medina and Reiner, 1997).

The description of the dopaminergic innervation of the avian telencephalon given in chapter 5

will follow the functional classification outlined above. It will turn out that the distribution of

DA and DA receptors is closely related to these functional subdivisions. However, certain

exceptions exist, as, for example, components of the song system of songbirds subserve sensory

as well as motor functions.

3. Origin of dopaminergic fibers in the avian telencephalon: Mesencephalic A8-A10 cell

groups

Comparable to the situation in mammals, the dopaminergic innervation of the avian

telencephalon arises mainly from three mesencephalic dopaminergic cell populations located in

the n. tegmenti pedunculopontinus pars lateralis and the area ventralis tegmentalis of Tsai


(AVT) (Figure 3; Kitt and Brauth, 1986; Metzger et al., 1996; Waldmann and Güntürkün, 1993).

The former nucleus is homologeous to the mammalian substantia nigra pars compacta (SNC;

Brauth et al., 1978; Karten and Dubbeldam, 1973; Kitt and Brauth, 1986; Reiner et al., 1983;

Rieke, 1981, 1982), where according to Reiner et al. (1994) the rostro-ventral part comprises

the A9 group and the caudo-dorsal part the A8 group. The medial part of the n. tegmenti

pedunculopontinus is considered to be equivalent to the mammalian substantia nigra, pars

reticulata (SNR; Kitt and Brauth, 1981; Reiner et al., 1983). The avian AVT is homologeous to

the ventral tegmental area of mammals (A10 group) (Kitt and Brauth, 1981, 1986; Reiner et al.,

1983, 1994; Waldmann and Güntürkün, 1993). In the AVT and SNC, intense perikaryal

immunoreactivity (ir) for tyrosine hydroxylase (TH) and DA (Figure 3A-C) but not for

dopamine-beta-hydroxylase (DBH), noradrenaline (NA), or phenylethanolamine-N-methyl

transferase (PNMT) can be observed (Bailhache and Balthazart, 1993; Moons et al., 1995;

Reiner et al., 1994; Waldmann and Güntürkün, 1993; Wynne and Güntürkün, 1995). TH is the

rate-limiting enzyme in the synthesis of all catecholamines and, with a few exceptions (Smeets

and Gonzales, 1990), must be present in dopaminergic neurons. The presence of DBH which

converts DA to noradrenaline, or of PNMT which converts noradrenaline to adrenaline, would

indicate non-dopaminergic catecholaminergic neurons and fibers. An A8-A10 group has been

recognized in many avian species, including zebra finches (Bottjer, 1993; Lewis et al., 1981),

quail (Bailhache and Balthazart, 1993; Bons and Olivier, 1986), warbling grass parakeet

(Takatsuki et al., 1981), chicks (Metzger et al., 1996; Moons et al., 1994), and pigeons (Kitt

and Brauth, 1981, 1986; Waldmann and Güntürkün, 1993; Wynne and Güntürkün, 1995). The

SNR, on the other hand, exhibits comparatively little perikaryal but high neuropil labeling for

DA (Waldmann and Güntürkün, 1993; Wynne and Güntürkün, 1995), as it is the case in

mammals where the SNR mainly contains GABAergic neurons (Fallon and Loughlin, 1995).

From the AVT and SNC, fibers ascend mainly ipsilaterally within two tracts of the mediolateral

part of the lateral forebrain bundle (Karten and Dubbeldam, 1973; Kitt and Brauth, 1986;

Metzger et al., 1996; Wynne and Güntürkün, 1995). In general, there is substantial overlap in

the telencephalic projection zones of the SNC and AVT (Kitt and Brauth, 1986; Metzger et al.,

1996; Waldmann and Güntürkün, 1993), although some degree of specialization and

topographical organization seems also to be present (see, e.g., Section 5.4.1).


Vice versa, the A8-A10 group receives pallidal inputs from the PP and striatal inputs from the

LPO, PA, and Acc (Karten and Dubbeldam, 1973; Kitt and Brauth, 1981; Reiner et al., 1994).

By far most of the striatonigral projection neurons co-contain substance P and dynorphin, while

the remainder contain enkephalin (Anderson and Reiner, 1990, 1991b; Reiner et al., 1983,

1984b; Reiner and Anderson, 1990). These two apparently disjunctive groups of neurons

project to different populations of dopaminergic and overlapping populations of non-

dopaminergic tegmental and nigral neurons (Reiner et al., 1994; Smeets, 1991), where

Figure 3. Dopaminergic neurons in the avian midbrain. (A-C and inset) Cells in the nucleus tegmenti pedunculopontinus and thearea ventralis tegmentalis of Tsai constitute the avian homologues of the substantia nigra, pars compacta (SNC) and reticulata(SNR), and the ventral tegmental area (AVT), respectively. (D) shows a similar view as seen in the right part of C labeled forDARPP-32. In contrast to dopamine-immunoreactivity as shown in C, only fibers are labeled with DARPP-32ir. Scale barsrepresent 1 mm (A, B) and 500 µm (C, D).


substance P-ir fibers make symmetric synaptic contacts predominantly on thin dendrites and

spines (Anderson et al., 1991), whereas enkephalinergic fibers contact dopaminergic somata as

well (Medina et al., 1995).

Both, D1 and D2 receptor densities seem to be quite low in the pigeon SNC compared to the

'striatal' parts of the avian basal ganglia (Dietl and Palacios, 1988; Richfield et al., 1987).

Thus, D2 receptors seem not to play such a prominent role as autoreceptors in avian

dopaminergic midbrain neurons (at least not within the local SNC/VTA circuit) as they do in

mammals (Brock et al., 1992; Bunney et al., 1987; Cooper et al., 1996; Yung et al., 1995).

Furthermore, DARPP-32 neuropil labeling (indicating the presence of D1 receptors, see

Section 4.3) but no soma labeling could be detected in the VTA, SNC, and SNR (Figure 3D).

This observation agrees well with the mammalian situation (Ouimet et al., 1984, 1992).

containing a small number of dopaminergic neurons around and below the glomeruli (Wynne

and Güntürkün, 1995), which probably correspond to the mammalian A16 group described by

Björklund and Lindvall (1984).

4. General features of the dopaminergic innervation of the avian telencephalon

Before getting into a detailed description of the distribution of dopaminergic fibers and

dopaminoceptive elements in the avian telencephalon, some general remarks on specific

features of the dopaminergic innervation of the avian brain, and on markers of the dopaminergic

system seem to be in place.

4.1. 'Basket'- and 'en passant'-type of dopaminergic innervation

The dopaminergic innervation in the avian telencephalon takes two different, and possibly

discrete, forms. The first form, which is found also in the mammalian cortex (Oades and

Halliday, 1987; Williams and Goldman-Rakic, 1993), is characterized by dopaminergic fibers

contacting the somata and dendrites of neurons 'en-passant' while passing through their target

region. These dopaminergic axons often travel in close vicinity along the somata and dendrites

of target-neurons while forming a large number of bouton-like axonal swellings. The second

one, called the 'basket'-type, is a very peculiar arrangement of dopaminergic fibers in the avian

telencephalon. In this case, single fibers densely coil around the somata and intial dendrites of

postsynaptic targets, enwrapping them in basket-like structures (Figure 4). In some

telencephalic regions, this type of innervation can be so dense that unlabeled postsynaptic

neurons and their initial dendrites can virtually be seen by labeling of catecholaminergic fibers


alone (Figure 4A, B). These fibers exhibit many varicosities in the vicinity of the soma and

proximal dendrites, which have been shown to contain large numbers of round clear vesicles

and also a few dense core vesicles (Metzger et al., 1996; Karle et al., 1992, 1994; see 4.4.).

Baskets seem to contact predominantly bigger neurons, whereas smaller neurons are more

likely innervated en-passant (Wynne and Güntürkün, 1995). However, basket-type structures

are not a speciality of the dopaminergic system, as they can also be demonstrated with

antibodies against NA and DBH (Moons et al., 1995).

4.2. Tyrosine hydroxylase as a marker of dopaminergic fibers in the avian telencephalon

Many studies have been carried out using THir as a marker of dopaminergic fibers in the avian

telencephalon. As noted above, TH is the rate-limiting enzyme in the synthesis of all

catecholamines and, with few exceptions (Smeets and Gonzales, 1990), must be present in

dopaminergic neurons. In general, the distribution of THir fibers closely follows that of DAir

fibers and might be a better indicator for the distribution of dopaminergic fibers than for the

much sparser noradrenergic innervation in the avian telencephalon, which predominantly

distributes along the VP, septum, HA, and the hippocampal region (Bailhache and Balthazart,

1993; Karle et al., 1996; Moons et al., 1994, 1995; Reiner et al., 1994; Wynne and Güntürkün,

1995). However, there are also some exceptions like the area parahippocampalis (APH) or the

n. taeniae (Tn) which are quite high in THir but low in DAir fibers (Wynne and Güntürkün,

1995). Hence, THir might be a good marker of dopaminergic fibers in most but not all

telencephalic areas.

4.3. DARPP-32 as a marker of D1-receptors and dopaminoceptive neurons

DARPP-32 is a Dopamine- and cAMP-Regulated PhosphoProtein of molecular weigth 32,000

that plays a role as a "third messenger" in the intracellular cascade induced by D1-receptor

stimulation (Hemmings et al., 1987a, 1987b, 1995). Via activation of the adenylyl cyclase, D1-

receptor action stimulates cAMP synthesis, which through protein kinase A leads to

phophorylation of DARPP-32. Phosphorylated DARPP-32 in turn inhibits protein phosphatase-

1, and thus affects the state of various target proteins. In rat striatal neurons for instance, DA

affects via phosphorylation of DARPP-32 the gating of Na+ and various high-voltaged-activated

Ca2+ channels (Hemmings et al., 1987a, 1995). DARPP-32 is itself dephosphorylated by

Ca2+/calmodulin-dependent protein phosphotase 2B (calcineurin), thus opening possibilities for

the interaction of glutamatergic transmission and Ca2+ influx with


DA-induced effects (Hemmings et al., 1987a, 1995; Nishi et al., 1997). In mammals the

distribution of DARPP-32 has been shown to be closely related to that of D1-receptors (Berger

et al., 1990; Hemmings and Greengard, 1986; Ouimet et al., 1984; Walaas and Greengard,

1984). In general, this is also the case in the pigeon (Durstewitz et al., 1998) and chick

(Schnabel et al., 1997) brain.

Figure 4. Modes of dopaminergic innervation in the avian telencephalon. (A, B) Examples of the ‘basket-type’ of dopaminergicinnervation. (C-G) Examples illustrating various densities and relative contributions of en-passant and basket-type innervations.Photomicrographs were taken from the NCL (A, C), PP (B), Aid (D), Ai (E), NIM (F), and Av (G). Scale bars represent 100µm (C, D, E), 50 µm (F, G) and 25 µm (A, B).


The use of antibodies against DARPP-32 as a marker of dopaminoceptive structures has the

advantage that postsynaptic neural targets of dopaminergic fibers are often labeled to a large

extent (Figure 5). In some cases even fine dendritic branches and spines are clearly visible at

the light-microscopic level (Figure 5D-F), revealing additional information about the

morphological properties of dopaminergic targets which are not apparent using D1 receptor-ir

(see Schnabel et al., 1997). The clear perikaryal labeling obtained with DARPP-32-ir also

enables morphometric approaches since cell counts can easily be performed.

Figure 5. DARPP-32 immuno-labeling in the pigeon forebrain. (A) Frontal section at level A 9.00, illustrating the differentintensities of DARPP-32-positive neuropil staining. (B, C) Most DARPP-32-positive cells displayed only labeling of the somaand initial dendritic segments. (D-F) In some neurons also fine dendritic processes and spines are visible. Photomicrographswere taken from the n. basalis (B), neostriatum frontolaterale (C), ventromedial hippocampus (D), border of ectostriatum andectostriatal belt (E), and medial septum (F). Scale bars represent 250 µm (A), 50 µm (B, D, E), 20 µm (C), 15 µm (F).Scalebars represent 250 µm (A), 50 µm (B, D, E), 20 µm (C), 15 µm (F). Taken from Durstewitz et al. (1998).


4.4. Ultrastructural features of the dopaminergic innervation and of dopaminoceptive

neurons

At the ultrastructural level, dopaminergic and THir fibers in the avian basal ganglia and in the

neostriatum form many varicosities which often exhibit multiple active zones and contain many

round clear small to medium-sized vesicles together with some large dense core vesicles

(Karle et al., 1992, 1994, 1996; Metzger et al., 1996), indicating the co-release of DA and

neuropeptides as it is the case in mammalian dopaminergic neurons (Cooper et al., 1996;

Seroogy et al., 1988a, 1988b). Some of these varicosities (about 50% in the avian basal ganglia

according to Karle et al., 1996) show small and flat synaptic specializations, most often of the

symmetrical type, but occasionally also asymmetrical ones (Karle et al., 1996; Metzger et al.,

1996). Dopaminergic synapses within the neostriatum, LPO, and PA are most often found on

thin dendritic shafts (about 50% in the basal ganglia), less frequently on dendritic spines, and

occasionally (< 20% in the basal ganglia) on thick dendritic shafts or perikarya (Karle et al.,

1996; Metzger et al., 1996). Thus, in the vicinity of the soma, unspecialized varicosities seem

to prevail. On dendritic spines, dopaminergic synapses are sometimes engaged in 'triadic

complexes' with other unlabeled symmetric or asymmetric synapses terminating on the same

spine, or where they make axo-axonic contacts on these unlabeled synapses (Metzger et al.,

1996). Thus, in the avian brain as in mammals (Calabresi et al., 1987; Hernández-López et al.,

1997; Law-Tho et al., 1994; Pralong and Jones, 1993; Yang and Seamans, 1996), DA might

control both, synaptic transmission via pre- and/or postsynaptic mechanisms as well as

postsynaptic somatic and dendritic membrane properties. Furthermore, both these aspects of

neural functioning may be modulated via D1 receptors, as D1 receptor-ir and DARPP-32ir

have been observed not only within the dendrites but occasionally also in axons and axon

terminals (Schnabel et al., 1997).

5. Distribution of dopaminergic fibers and dopaminoceptive elements in the avian

telencephalon

The distribution of dopaminergic fibers is closely related to functional subdivisions of the

avian telencephalon. Most prominently, structures of the avian basal ganglia receive by far the

densest dopaminergic input, whereas all primary sensory structures seem to be devoid of DA,

L-DOPA, DA receptors, and DARPP-32 (Dietl and Palacios, 1988; Durstewitz et al., 1998;

Juorio, 1983; Juorio and Vogt, 1967; Metzger et al., 1996; Moons et al., 1994; Schnabel and

Braun, 1996; Wynne and Güntürkün, 1995). All other telencephalic areas fall somewhere in


Figure 6. Schematic illustration of the distribution of dopaminergic fibers in the avian telencephalon.

between, however, with clear regional differences. In the following, we will describe the

distribution of dopaminergic fibers and dopaminoceptive elements according to functional

subdivisions, following the pathway of sensory information from the primary sensory to the

motor output structures as outlined in Section 2. Figures 6 to 9 give an overview over the

distribution of dopaminergic fibers, D1 receptors, D2 receptors, and DARPP-32 neuropil and

soma labeling throughout the avian telencephalon. In addition, findings on colocalizations of

markers of the dopaminergic system with other biochemically identified systems will be

discussed.

However, before going into details, it should be mentioned that although the general pattern may

be the same across avian species, some strain and species differences exist. Thus, Divac et al.

(1988) observed an about two-fold higher DA concentration in the posterior telencephalon of

mixed breed than of white Carneaux pigeons, while in the anterior telencephalon the DA

concentration was only slightly higher. Nevertheless, the relative DA concentration in six

telencephalic regions examined remained qualitatively the same in both strains. Similarly,

none - very low

high

very highmedium

low

Hp

SM

PAPP

CDL

HV

APH

SL

TPO

A 7.75

HA

LPOBas

NFL

NF

HVLFS

A 12.50

CDL

APH

NCL

AvAp

Hp

NC

A 5.25

A 9.25

HD

Hp

HA

ENIM

LPO PA

PP

HV

Ep

INP

NIL

VP

A 6.75

APH

HpCDL

NCL2

L1L3

AvAi

Aid

Tn

PA

NCL

HV

A 10.50

HA

NIL

Hp

Ep

HDHV

E

PALPO

NIM

NC

COS

AccAcc

LFSLFS


telencephalic DA concentrations were noted to vary by factors of up to five between chicks,

ducks, finches, pigeons, fowls, and quails (Juorio and Vogt, 1967; Juorio, 1983).

5.1. Primary sensory areas


Primary sensory areas of the pigeon and chick telencephalon display by far the lowest DAir and

L-DOPAir (Figure 10A; Metzger et al., 1996; Wynne and Güntürkün, 1995; Moons et al.,

1994), and no projections from the VTA or SNC to these areas have been demonstrated, except

for a restricted input to the HD (Kitt and Brauth, 1986). Whereas the E (visual tectofugal), Field

L2 (auditory), and the Bas (trigeminal) are devoid of dopaminergic fibers, some DAir fibers

are present in the rostral (somatosensory) and caudal (visual thalamofugal) IHA, and lateral

HD, the two input laminae of the Wulst. Interestingly, the specific sensory nuclei of the thalamus

(e.g., the n. rotundus which projects to the E) also contain little or no THir (Reiner et al., 1994).

Figure 7. Schematic illustration of the distribution of D1 receptors in the avian telencephalon.

none - very low

low

medium

high

very high

A 12.50

A 6.75

CDL

NCL2

HV

L3 NCL

Hp

PAAid

AvAi

Tn

APH

L1

A 10.50A 9.25

NC

AvAp

CDLHp

APH

NCL

A 5.25

ENIM

HDHA

Hp

PALPO

PP

Ep

HV

NIL

INP

A 7.75

SMTPO

Hp

APH CDL

HV

PAPPSL

NC

COS

LPO

LPOBas

NFLHV

NF

HA HIS

HV

HD

HA

Ep

ENIL

Hp

NIM

PALPO


Thus, primary sensory processes do not seem to be modulated, at least not to a significant

extent, by DA in the adult avian brain.

5.1.2. D1 receptors

Given the virtual absence of a dopaminergic innervation, it comes with no surprise that the D1

receptor density in the primary sensory structures is very low, with the E exhibiting the lowest

D1-specific binding in the whole telencephalon (Dietl and Palacios, 1988; Schnabel et al.,

1997; Schnabel and Braun, 1996; Stewart et al., 1996). However, as it is the case for DAir

labeling, the IHA and HD are higher in D1-specific binding, although still significantly lower

than the respective secondary projection zones in the HA and most of the neostriatum (Schnabel

et al., 1997).

5.1.3. D2 receptors

In general, D2 receptor densities as measured by [3H]spiperone and [3H]CV205-502 are low

and homogeneously distributed throughout the whole avian telencephalon, except for the basal

ganglia and the hippocampal complex as discussed below (Dietl and Palacios, 1988; Schnabel

and Braun, 1996; Stewart et al., 1996).

5.1.4. DARPP-32

The distribution of DARPP-32ir confirms the extremely low or absent dopaminergic input to

the primary sensory structures. The E and Field L2 are largely devoid of DARPP-32ir (Figure

10B; Durstewitz et al., 1998; Schnabel et al., 1997). The IHA and HD display low levels of

neuropil labeling and a low percentage of labeled neurons, while, however, scoring in both

respects again much lower than the respective secondary sensory zones in the HA and most of

the neostriatum (Durstewitz et al., 1998; Schnabel et al., 1997). Hence, primary sensory regions

in the avian forebrain could be sharply delineated and differentiated from bordering areas by

virtue of their low DARPP-32ir (Figure 5A; Figure 10B).

The only primary sensory forebrain area that might deviate from this general pattern, at least in

pigeons, is the trigeminal Bas. Although it is very low in terms of DARPP-32ir neuropil

labeling, the pigeon’s Bas contains a high proportion of labeled perikarya, many of them

exhibiting clear neuronal properties (Figure 5B; Durstewitz et al., 1998; but see Schnabel et al.,

1997, for chicks). This finding might be related to other peculiarities of the Bas and to the

specific demands imposed on the neural circuit controlling pecking behavior, in which the Bas


Figure 8. Schematic illustration of the distribution of D2 receptors in the avian telencephalon.

is critically involved (Schall, 1987; Wild et al., 1985). The Bas is the only avian forebrain

structure with direct sensory inputs bypassing the thalamus (Schall et al., 1986). Moreover,

within the trigeminal neural circuit controlling pecking, sensory signals possibly must be

integrated very fast and relayed rather directly to motor outputs. As the dopaminergic system

probably plays a central role in sensory-motor integration and learning (see Section 7), the

dopaminergic modulation might start earlier in this trigeminal somatosensory circuit than within

other forebrain systems. This notion is furthermore underpinned by the fact that the Bas receives

auditory and possibly also vestibular input, in addition to trigeminal information (Schall et al.,

1986; Schall and Delius, 1986; Wild et al., 1997), and might thus actually be regarded as a

multimodal structure. Furthermore, behavioral evidence supports the hypothesis of a

dopaminergic modulation within the pigeon Bas: Lindenblatt and Delius (1988) showed that

local apomorphine-injections into the Bas induce pecking bouts, while Wynne and Delius

(1996) made it likely that Bas lesions decrease pecking fits induced by peripheral

apomorphine-injections. However, the high number of DARPP-32ir neurons in the pigeon Bas

CDL

TPO

Hp HV

APH

SMSL

PAPP

A 7.75

A 10.50 A 12.50LPO

HA

LPOBas

NFL

NF

HV

high

very highmedium

low

very low

A 9.25

Av

Hp

NC

APH

CDL

NCL

Ap

A 5.25

HD

Hp

HA

ENIM

LPO PA

PP

HV

Ep

INP

NIL

VP

A 6.75

APH

Hp

CDL

NCL2

L3

AvAi

Aid

Tn

PA

NCL

HV

L1

HA

NIL

Hp

EpNIM

HD

HV

E

PALPO

NC


is clearly at odds with the lack of D1-receptors and dopaminergic fibers reported so far, as

well as with the lack of DARPP-32 in the chick Bas (Schnabel et al., 1997). In this context, it

should be pointed out that DARPP-32 is not exclusively involved in the D1-induced cascade

but is regulated by other neuromodulator/neurotransmitter pathways as well, e.g. by D2- or

NMDA-receptor stimulation via Ca2+ influx (Hemmings et al., 1995; Nishi et al., 1997). Thus,

DARPP-32 in the Bas might be linked to intracellular pathways other than the one coupled to

the D1 receptor, and differences between pigeons and chicks might exist in this respect.


Durstewitz et al. (1998) found that glutamate decarboxylase (GAD) ir neurons were never

colocalized with DARPP-32 or located in THir baskets throughout the whole pigeon

telencephalon, including the primary sensory areas.

5.2. Secondary sensory and multimodal areas


In general, the dopaminergic innervation of both, the secondary sensory and tertiary association

areas is clearly higher than that of the primary areas (Figures 10, 11). In addition, most of the

multimodal areas score higher than the secondary sensory areas due to an increasing

rostrocaudal and mediolateral gradient of DAir structures (Figure 6; Metzger et al., 1996;

Wynne and Güntürkün, 1995). Thus, in the secondary visual Ep and the trigeminal NF, DA

concentration is very low. The HA which includes the secondary projection fields of the

thalamofugal visual and the somatosensory system receives a moderate dopaminergic input

(Metzger et al., 1996; Moons et al., 1994; Wynne and Güntürkün, 1995).

Among the presumed multimodal forebrain structures, the rostrally located NFL is only

moderate and thus lowest in its DA content (Metzger et al., 1996; Moons et al., 1994; Wynne

and Güntürkün, 1995), while the medial portion of the lamina frontalis superior seems to be

most densely innervated by dopaminergic fibers (Figure 11A), with the fibers arising mainly

from the SNC (Bottjer, 1993; Kitt and Brauth, 1986; Wynne and Güntürkün, 1995). High DAir

and basket-like structures are also present in the MNH (Figure 11A; Metzger et al., 1996;

Moons et al., 1994; Wynne and Güntürkün, 1995), an area that has been implicated in filial

imprinting (Gruss and Braun, 1996, 1997). Metzger et al. (1996) also demonstrated a prominent

input to these areas from VTA by retrograde tracing, and to a lesser extent from SNC.


Considerable numbers of basket-like structures and dopaminergic fibers are also present in

multimodal areas comprising the pallium externum, especially in the TPO, which receives its

major input from SNC (Kitt and Brauth, 1986). Within the caudal neostriatum which is

generally very high in DA input, the NCL has received special attention because of its

presumed equivalency to the prefrontal cortex (PFC) of mammals. This assumption was first

derived from behavioral, radioenzymatical, and histofluorescence studies conducted by Divac,

Mogensen and Björklund (Divac et al., 1985; Divac and Mogensen; 1985; Mogensen and

Divac, 1982). These authors demonstrated that the NCL contains a high DA-to-NA-ratio and an

especially high number of catecholaminergic fibers of presumed dopaminergic origin, as it is

characteristic for the PFC of mammals (Berger et al., 1988, 1991; Björklund et al., 1978; Divac

et al., 1978; Glowinski et al., 1984; Van Eden et al., 1987). Using immunocytochemical

methods, Waldmann and Güntürkün (1993) and Wynne and Güntürkün (1995) confirmed that

DARPP-32neuropil labeling

medium - high

none - very lowlow

very high

Inserts: Percentage of DARPP-32+ neurons < 10 %20 - 40 %40 - 70 %> 70 %

Hp

SL

HV

TPO

CDL

PAPP

SM

A 9.25

A 7.75

A 10.50 A 12.50LPO

HA

LPO

NFLHV

Bas

NF

HAHD

Ep

HV

NILE

LPO

PPPA

Acc INP

VP

NIM

CDL

NC

AvAp

NCL

APH

Hp

A 5.25 A 6.75

NC NCL

APHCDL

L2

L1

HV

TnAv

Ai

AidPA

L3

Hp

HV

HD

HA

NILEp

NIM

LPO

E

PA

HA

NC

Figure 9. Schematic illustration of the distribution of the D1-receptor related phosphoprotein DARPP-32 in the aviantelencephalon.


the NCL could be differentiated from the surrounding caudal neostriatum by its denser

dopaminergic innervation and by the higher number of DAir fibers which constitute a basket,

while the number of baskets per se does not increase in the NCL (Figure 11B, C; see also

Divac et al., 1994). The dopaminergic input to the NCL arises from the dopaminergic midbrain

nuclei AVT and SNC (Waldmann and Güntürkün, 1993).

In chicks, a region in the dorsocaudal neostriatum shares the location, connections and dense

dopaminergic innervation with the pigeon´s NCL (Metzger et al., 1996, 1998). This tertiary

area has also been studied in the context of imprinting (Bock et al., 1997; Schnabel and Braun,

1996). Its dopaminergic innervation is also characterized by a higher occurrence of basket-like

structures and by a dopaminergic input from AVT and SNC (Metzger et al., 1996). However, it

appears that the area of densest DAir and THir is shifted somewhat ventromedially from the

ventricle compared to pigeons (Metzger et al., 1996; own unpublished observations).

5.2.2. D1 receptors

In general, the distribution of D1 receptors closely mimics that of dopaminergic fibers, although

D1 receptors seem to be more homogeneously distributed throughout the entire neostriatum and

hyperstriatum than dopaminergic fibers, which seem to be more locally restricted (Figure 7).

The HA, HD, HV, and neostriatum contain comparable, medium concentrations of D1

receptors, with the hyperstriatal areas scoring a bit higher than at least the rostral and

intermediate neostriatum (Ball et al., 1995; Dietl and Palacios, 1988; Schnabel and Braun,

1996; Schnabel et al., 1997; Stewart et al., 1996). In addition, two obvious tendencies in

distribution are apparent (Schnabel and Braun, 1996; Schnabel et al., 1997): First, the number

of D1 receptors increases from rostral to caudal in both, hyperstriatal and neostriatal areas.

Second, in the neostriatum, the number of D1 receptors increases from medial to lateral. Thus,

the NCL and the caudal HV are the telencephalic areas highest in D1-specific binding outside

the basal ganglia. This finding also implies that secondary sensory structures in the frontal

neostriatum, like the Ep and the NFL, but also the secondary auditory fields L1 and L3, are less

dense in D1 receptors than most tertiary telencephalic regions. Hence, in general, and possibly

with some species differences (Ball et al., 1995), there is an increase of D1-receptor density

proceeding from primary sensory to tertiary multimodal structures. This pattern might be related

to an increase in sensory-motor integration and complex learning processes that take place

primarily in higher-order structures.


Figure 10. Distribution of dopaminergic fibers and dopaminoceptive elements in primary and secondary sensory areas asexemplified for the auditory Field L complex. (A) The primary sensory area L2 is almost devoid of dopaminergic fibers and (B)of DARPP-32 immunoreactivity, while the surrounding secondary sensory “belt” regions L1 and L3 show considerably morelabeling for dopamine and DARPP-32. Scale bars represent 500 µm.

Figure 11. Dopaminergic innervation and DARPP-32 immunoreactivity in higher order sensory and multimodal structures. (A)Distribution of dopaminergic fibers in the rostromedial forebrain. The neostriatum intermedium (NIM) and hyperstriatumventrale (HV), and possibly also the lamina frontalis superior (LFS), constitute multimodal areas. Note that in the NIM and HVthe density of dopaminergic fibers increases slightly from lateral to medial. (B) Localization of the neostriatum caudolaterale(NCL), the presumed prefrontal cortex of birds, in the caudal telencephalon. (C) Detail of the dopaminergic innervation of theNCL and the posterior archistriatum. (D) DARPP-32 immunoreactivity in the caudal neostriatum is evenly distributed atmedium levels, but is low in the archistriatum. Scale bars represent 1 mm (B) and 50 µm (A, C, D).


5.2.3. D2 receptors

D2 receptor densities in the neostriatum and hyperstriatum are low and homogeneously

distributed (Figure 8; Dietl and Palacios, 1988; Schnabel and Braun, 1996; Stewart et al.,

1996). For example, Schnabel and Braun (1996) found the density of D2 receptors to be about

one third or less (< 25 fmol/mg) of that of D1 receptors in these areas. Hence, outside the basal

ganglia, dopaminoceptive neurons in the avian telencephalon seem to be modulated

predominantly via receptors of the D1 class, although it should be kept in mind that even low

densities of D2 receptors might play an important physiological role.

5.2.4. DARPP-32

DARPP-32ir is distributed with medium to high densities throughout the entire neostriatum, HA,

CDL and TPO (Figure 9; Durstewitz et al., 1998; Schnabel et al., 1997). Thus, DARPP-32ir

reaches much higher levels in the secondary and tertiary areas than in the primary sensory

structures (Figures 5A, 10B). In addition, Schnabel et al. (1997) reported a rostrocaudal and a

mediolateral trend in the DARPP-32 distribution in the chick neostriatum as it has been

described for D1 receptors. However, Durstewitz et al. (1998) were not able to discriminate

between secondary sensory and multimodal pigeon forebrain structures based on DARPP-32ir

neuropil labeling and cell countings, although a slight trend in neuropil labeling was also

observed by these authors. Thus, differences between secondary sensory and multimodal areas

with regards to their dopaminergic input may not be so apparent in DARPP-32ir as they are in

D1 receptor and dopaminergic fiber density.

A relatively high percentage of neurons (up to about 60 %) exhibits DARPP-32ir in the

secondary sensory and multimodal areas, about the same as in structures of the basal ganglia

and the ventral archistriatum, and significantly higher than in the HD (receiving primary visual

input) and the dorsal archistriatum (Figure 11D; Durstewitz et al., 1998). The NCL has already

been introduced as the possible avian equivalent of the mammalian PFC, and as such receives a

prominent dopaminergic input. However, this structure cannot be differentiated from the

surrounding neostriatal areas by DARPP-32 neuropil labeling or by its relative number of

DARPP-32-positive neurons (Figure 11D; Durstewitz et al., 1998). In contrast, the percentage

of DARPP-32-positive neurons seems even to decrease proceeding from rostral to caudal.

Similar discrepancies in the distribution of dopaminergic fibers, D1 receptors and DARPP-32

have also been oberserved in the mammalian neocortex (Berger et al., 1990; see Section 6.2).

One of the most obvious discrepancies between the distributions of DARPP-32ir and D1


receptors, however, is the significantly higher amount of DARPP-32 in the neostriatum than in

the HV, while just the opposite is the case for the density of D1 receptors (Durstewitz et al.,

1998; Schnabel et al., 1997). Another possible discrepancy is the considerable DARPP-32ir in

the CDL where Ball et al. (1995) reported very low concentrations of D1 receptors. These

discrepancies might be partly related to the fact that ligand-binding autoradiographic

demonstrations of D1 receptors have different properties than the immunocytochemical

demonstration of a receptor-related phosphoprotein. Thus, while receptorautoradiography might

provide objective means to demonstrate receptor densities, immunocytochemistry reveals more

about the morphology and possible colocalizations of the cells under study. However, it should

be stressed that DARPP-32 generally is a good marker for the distribution of D1 receptors in

avian species as it is in mammals (Berger et al., 1990; Hemmings and Greengard, 1986; Ouimet

et al., 1992).


Figure 12A presents examples from the caudal neostriatum of DARPP-32ir neurons which are

located in TH-baskets. A quantitative examination of the number of DARPP-32ir neurons

located in TH-baskets showed that this percentage differs considerably between various areas

(Durstewitz et al., 1998). The highest percentage (15-30 %) of DARPP-32/TH-basket

colocalizations was observed in the multimodal lateral and caudal aspects of the neostriatum,

including the NCL, i.e. exactly in those regions, which expressed the highest numbers of D1

receptors according to Schnabel et al. (1997). Less DARPP-32/TH-basket colocalizations

were found in the rostral and intermediate neostriatum (5-15 %), including most of the

secondary sensory areas. The lowest percentage of DARPP-32/TH-basket colocalizations

occurred in the hyperstriatal and hippocampal regions (< 5 %), although the HA is strongly

innervated by dopaminergic fibers and exhibits a high concentration of D1 receptors. That some

areas are dense in D1 receptors and dopaminergic input but display only few basket-like

structures is a further hint to a functional dissociation between an 'en passant'-mode of

dopaminergic innervation and the basket-like DA input. According to the functional

considerations outlined in Section 6.3, one might speculate that the higher level lateral and

caudal neostriatal areas have a special role in synchronizing and coordinating network activity.

In the male quail telencephalon, especially within the neostriatum, colocalizations of THir

baskets and varicosities and aromatase-ir cells were observed (Balthazart et al., 1998). The

enzyme aromatase converts testosterone into 17β-estradiol which is involved in the activation


of male sexual behavior. These colocalizations may thus represent an anatomical basis for the

regulation of male sexual behavior by DA (see Section 7.3.) DA probably stimulates aromatase

activity and is, vice versa, possibly influenced itself by aromatase (Balthazart and Absil, 1997;

Pasqualini et al., 1995).

From a functional point of view, it is of great importance to identify the postsynaptic targets of

the dopaminergic innervation. Baskets provide a clear and convenient morphological marker to

do this. GABAergic inhibitory neurons constitute one of the prominent potential candidates

which might be dopaminergically modulated via baskets. Veenman and Reiner (1994) estimate

the fraction of GABAergic neurons to be about 10-12 %, homogenously distributed throughout

most of the telencephalon. Antibodies directed against GAD, the enzyme which converts

glutamate to GABA, labels approximately the same population of cells, and thus represents a

useful indicator for GABAergic neurons (Veenman and Reiner, 1994). Like in other

telencephalic structures, throughout the neostriatum and hyperstriatum no GADir neurons were

observed that were located in THir baskets (Figure 12B) or were colocalized with DARPP-32

(Figure 12C; Durstewitz et al., 1998). Thus, GABAergic neurons in the pigeon telencephalon

might not express D1 receptors and might not receive a strong dopaminergic input. As in

addition D2 receptors are very low in these areas, GABAergic neurons possibly do not receive

a significant dopaminergic input at all. Hence, an interesting functional conclusion might be that

mainly or solely excitatory neurons are the targets of the dopaminergic innervation in the pigeon

telencephalon. However, it has to be emphasized that D2 receptors despite their low densities

might nevertheless modulate the physiological activity of dopaminoceptice neurons in a

significant manner. Thus, it still has to be investigated whether GABAergic neurons express D2

receptors.

5.3. Septum, hippocampal complex, archistriatum


Prominent motor and limbic structures in the avian telencephalon include the septum, the

hippocampal complex, and the archistriatum. The lateral septum exhibits a high number of

DAir, L-DOPAir, and THir baskets and fibers (Bailhache and Balthazart, 1993; Balthazart et

al., 1998; Bottjer, 1993; Moons et al., 1994; Wynne and Güntürkün, 1995), whereas the medial

septum displays a high density of THir and NAir fibers (Moons et al., 1995) but only moderate

to low amounts of DAir and L-DOPAir (Figure 13A). The lateral septum, and especially the

ventrolaterally located region, receives most of its dopaminergic input from the VTA, while the


medial septum is more densely innervated by axons from the SNC (Kitt and Brauth, 1986). In

the hippocampus (Hp) and APH only a low number of DAir and L-DOPAir structures was

detected (Metzger et al., 1996; Moons et al., 1994; Wynne and Güntürkün, 1995), which mainly

stem from the SNC (Kitt and Brauth, 1986).

Fig. 12. Fluorescence double-labeling against (A) DARPP-32 and TH, (B) GAD and TH, and (C) DARPP-32 and GAD. Thephotomicrographs of each pair were taken from the same site in caudal neostriatum of the pigeon telencephalon. Two of threeDARPP-32ir neurons (A1) are surrounded by THir baskets (A2). GADir neurons (B1) do not receive input from THir baskets(B2). GADir neurons (C2) do not show DARPP-32 immunoreactivity (C1), although GADir and DARPP-32ir neurons mayoccur in close vicinity to each other. Arrows mark corresponding positions in each pair. In (C), open and filled arrows indicatepositions of DARPP-32ir and some GADir cells, respectively. Scale bars represent 30 µm (A, B), 40 µm (C). Taken fromDurstewitz et al. (1998).


Hence, the dense THir in these areas may be mainly due to a noradrenergic innervation, which

has also been demonstrated by DBHir and NAir (Bailhache and Balthazart, 1993; Moons et al.,

1995; Reiner et al., 1994).

The archistriatum displays a very heterogenous innervation by dopaminergic fibers (Figures

11C, 14A). A particular dense dopaminergic input, reaching levels as high as in the basal

ganglia, was described for the most dorsal archistriatum intermedium (Aid), which constitutes

part of the sensorimotor archistriatum, as well as for the posterior 'limbic-visceral'

archistriatum (Ap) (Bailhache and Balthazart, 1993; Balthazart et al., 1998; Metzger et al.,

1996; Wynne and Güntürkün, 1995). This input stems mainly from SNC, and to a lesser extent

from VTA (Kitt and Brauth, 1986). However, whereas Metzger et al. (1996) reported the rest

of the archistriatum in the chick to be low in dopaminergic structures, Wynne and Güntürkün

(1995) demonstrated a relatively high number of dopaminergic structures throughout the whole

pigeon archistriatum, with a decreasing gradient of dopaminergic fiber density from dorsal to

ventral. The ventromedially located Tn is devoid of DAir, but is relatively high in THir, NAir

and DBHir (Bailhache and Balthazart, 1993; Moons et al., 1995). Interestingly, despite the very

low DAir fiber density in the ventral archistriatum, there is nevertheless a high number of DAir

baskets present in this structure, possibly surrounding the large projection cells which give rise

to the descending tractus occipitomesencephalicus (Wynne and Güntürkün, 1995). Projections

to the intermediate archistriatum from VTA and SNC have also been demonstrated by Kitt and

Brauth (1986).

5.3.2. D1 receptors

In general, the medial and lateral septum seem to be very low in D1 receptor binding, at least in

the quail (Ball et al., 1995). However, in accordance with its relatively dense input from SNC

(Kitt and Brauth, 1986), the n. commissuralis of Baylé et al. (1974) within the medial septum is

relatively high in D1 density (Ball et al., 1995), and this seems also to be the case in the

domestic chick as judged from Figure 1 of Stewart et al. (1996). Both, the Hp and the APH are

very low in D1-specific binding (Schnabel and Braun, 1996; Schnabel et al., 1997; Stewart et

al., 1996). The pattern of [3H]SCH23390 labeling in the archistriatum in general follows that

described for dopaminergic fibers above. In particular, the dorsal archistriatum intermedium

contains a very high number of D1 receptors, while the density strongly drops in the ventral

archistriatum (Dietl and Palacios, 1988; Schnabel and Braun, 1996; Schnabel et al., 1997;

Stewart et al., 1996).


5.3.3. D2 receptors

The dorsal Hp and medial APH are the only

telencephalic regions besides the basal ganglia

(see below) which display considerable

amounts of D2 receptors (Schnabel and Braun,

1996; Stewart et al., 1996). As it is the case in

the basal ganglia, D2 receptors even outnumber

D1 receptors in the avian hippocampal areas.

5.3.4. DARPP-32

The septum is devoid of DARPP-32ir except

for the n. commissuralis of Baylé et al. (1974)

and a circumscribed region in the lateral

septum (Figure 13B), both of which contain

DARPP-32ir somata as well as fibers

(Durstewitz et al., 1998).

DARPP-32ir neuropil labeling in the Hp is heterogenous, with the medial part being completely

blank, while the ventromedial part and the lateral border show considerable neuropil labeling

and a high number of DARPP-32ir somata (Durstewitz et al., 1998). These regions probably

correspond to the V-shaped region and area 7 of Krebs et al. (1991) and Erichsen et al. (1991),

respectively. The APH displays moderate to high levels of DARPP-32ir neurons and neuropil

labeling.

With the exception of the Tn, which is virtually devoid of DARPP-32, the pigeon archistriatum

is homogenously stained by DARPP-32ir with densities comparable to the neostriatum and the

APH (Figure 14B; Durstewitz et al., 1998). Only the ventral but not the dorsal archistriatum in

addition contains a high percentage of DARPP-32ir neurons, while just the opposite is the case

for the distribution of DA fibers and receptors (Metzger et al., 1996; Wynne and Güntürkün,

1995; see Section 5.3.1). However, the finding of many large DARPP-32ir neurons in the

ventral archistriatum is consistent with the relatively high number of dopaminergic baskets in

this region (Wynne and Güntürkün, 1995), and may thus further establish a functional

dissociation between a ‘basket‘- and an ‘en passant’-mode of dopaminergic innervation. In

addition, a small, dense population of DARPP-32ir cells was observed at the dorsomedial

Figure 13. (A) Dopamine- and (B) DARPP-32immunoreactivity in the septum. (C) Dopamine- and (D)DARPP-32 immunoreactivityin the nucleus accumbens. Scalebars represent 50 µm (A, B, C) and 25 µm (D).


border of the intermediate archistriatum (Figure 14B; Durstewitz et al., 1998), lying within the

dorsal archistriatum intermedium dorsale which is also densely innervated by dopaminergic

fibers and is dense in D1-receptor content (see Sections 5.3.1 and 5.3.2). As judged from figure

3 in Schnabel et al. (1997), the situation seems to be comparable in chicks, with quite high

numbers of DARPP-32ir somata and fibers in the dorsal intermediate and most ventral

archistriatum, but low staining in between.


In the archistriatum, and especially in its ventral component, a large proportion of DARPP-32ir

neurons are located within THir baskets (15-40 %; Durstewitz et al., 1998). Since axons which

constitute baskets very likely exert a massive effect on their postsynaptic targets, it is

conceivable that many D1-positive archistriatal cells are to an important extent under

modulatory control of the dopaminergic system. Since in the quails archistriatum intermedium,

THir fibers have also been found to enwrap aromatase-ir neurons, it is possible that D1-

positive archistriatal cells are jointly modulated by DA and steroids and thus play an important

role in the motor control of sexual behavior (Balthazart et al., 1998). In contrast, in the

hippocampal complex, the percentage of DARPP-32ir neurons located in THir baskets is very

low (< 5%), further supporting the notion that most if not all THir baskets in this region may be

of non-dopaminergic and thus probably noradrenergic origin (Moons et al., 1995; Reiner et al.,

1994; Wynne and Güntürkün, 1995). In the archistriatal and hippocampal regions GADir

neurons were never observed to be located within TH baskets or to be colocalized with

DARPP-32 (Durstewitz et al., 1998). Thus, in these regions like in most other telencephalic

areas the avian dopaminergic system might mainly modulate non-GABAergic and thus

excitatory forebrain neurons. The situation may be different in the hippocampal region,

Figure 14. (A) Dopaminergic fibers and (B) DARPP-32 immunoreactivity in the archistriatum. DARPP-32 immunoreactivitywithin the archistriatum is largely restricted to the Av and a small rim of the Aid (arrows). Scale bars represent 500 µm.


however, where a higher density of D2 receptors has been observed (see Section 5.3.3). In the

mammalian cortex, mainly GABAergic interneurons seem to be modulated via D2 receptors,

while the activity of pyramidal cells seems to be mainly modulated via D1 receptors (Godbout

et al., 1991; Law-Tho et al., 1994; Pirot et al., 1992; Rétaux et al., 1991; Yang and Seamans,

1996). If this would hold true also in the dorsal HP and medial APH, where D2 receptors are

even more numerous than D1 receptors, then one might speculate that in the hippocampal areas

mainly the network of inhibitory interneurons is modulated by DA, in contrast to the situation in

other telencephalic areas.

5.4. Basal ganglia


The avian basal ganglia can be subdivided into four major functional regions: 1) the LPO and

the PA which together are comparable to the dorsal, somatomotor striatum, 2) the Acc, TO, and

BNST, which form the ventral, visceral-limbic striatum, 3) the PP, and 4) the VP, which

constitute the dorsal and ventral pallidum, respectively (Karten and Dubbeldam, 1973; Reiner

et al., 1984a; Veenman et al., 1995). Whereas the dorsal striatum contains small, densely

packed cells, the PP mainly contains sparsely scattered large projection neurons.

Comparable to the mammalian situation, anterograde and retrograde tracing studies

demonstrated that structures of the avian basal ganglia are the ones receiving the densest input

from the dopaminergic midbrain nuclei SNC and VTA, with some degree of topographical

organization (Bons and Oliver, 1986; Brauth et al., 1978; Karten and Dubbeldam, 1973; Kitt

and Brauth, 1986, Metzger et al., 1996). Thus, the LPO and the rostromedial PA receive input

from VTA and the central core of SNC (but see Metzger et al., 1996, in chicks), the medial and

caudal PA receive inputs mainly from the medial SNC, and the lateral PA from the lateral SNC

(Bons and Oliver, 1986; Brauth et al., 1978; Kitt and Brauth, 1986). Structures of the ventral

striatum also receive projections from the SNC, and to a somewhat lesser degree from VTA

(Balthazart and Absil, 1997; Kitt and Brauth, 1986). In contrast, notable mesencephalic

projections to the PP and n. intrapeduncularis (INP), i.e. pallidal parts of the avian basal

ganglia, were not observed (Brauth et al., 1978; Kitt and Brauth, 1986; Medina and Reiner,

1997).

The results from tracing studies have been confirmed by immunocytochemical investigations in

several avian species using antibodies directed against DA, L-DOPA, or TH (Figure 15;


Bailhache and Balthazart, 1993; Balthazart et al., 1998; Bottjer, 1993; Karle et al., 1996;

Metzger et al., 1996; Moons et al., 1994; Wynne and Güntürkün, 1995). The 'striatal' parts of

the avian basal ganglia are characterized by an extremely dense meshwork of dopaminergic

fibers (Figure 15A, C), which exhibit many bouton-like axonal swellings. Actually, this

meshwork is so dense that basket- and non-basket-types of dopaminergic innervation can hardly

be differentiated in these regions, except for the BNST and the Acc, where the innervation is

less dense (Karle et al., 1996; Wynne and Güntürkün, 1995). In the Acc, a high number of DAir

(but not THir) basket-like structures are observed in addition to DAir and THir fibers (Figure

13C). In contrast, DA- and TH-positive labeling in the 'pallidal' parts of the avian basal ganglia

is comparatively weak (Figure 15A, C; Bailhache and Balthazart, 1993; Balthazart et al., 1998;

Bottjer, 1993; Karle et al., 1996; Metzger et al., 1996; Moons et al., 1994; Wynne and

Güntürkün, 1995). Moons et al. (1994) were not able to detect any DAir or L-DOPAir in the PP

and INP in the chick brain at all. However, Wynne and Güntürkün (1995) observed that many

unlabeled large neurons in the pigeon PP are enwrapped by basket-like DA-positive structures

(c.f. Figure 4B). The light labeling of the PP may thus in part be due to the sparse distribution of

the large projection neurons in this area.

Considerable THir occurs in the VP, which must be largely due to labeling of noradrenergic

fibers (Bailhache and Balthazart, 1993; Karle et al., 1996; Moons et al., 1995; Wynne and

Güntürkün, 1995). In contrast, by far the most TH-positive elements in the PA and LPO seem to

be of dopaminergic origin, as only little DBHir and NAir were observed in these regions

(Karle et al., 1996; Moons et al., 1995). Furthermore, in the PA and LPO, DAir and THir

terminals make up for about 15-20% of all synaptic terminals in these areas (Karle et al.,

1996). A substantial decrease of DA concentration within the basal ganglia from rostral to

caudal has been noted by Juorio and Vogt (1967).

5.4.2. D1 receptors

Autoradiographic D1-receptor binding studies, all using [3H]SCH23390 as ligand, in chicks

(Schnabel et al., 1997; Stewart et al., 1996), pigeons (Dietl and Palacios, 1988; Richfield et

al., 1987), Japanese quails (Ball et al., 1995), and European starlings (Casto and Ball, 1994)

consistently demonstrated that D1 receptors are most abundant in structures of the avian basal

ganglia (Figure 7). However, the absolute binding of [3H]SCH23390 varied by several orders

of magnitude between studies, ranging in the LPO from 3.8 fmol/mg protein (Stewart et al.,

1996) to 2800 fmol/mg (Ball et al., 1995). This might partly be due to species and strain


differences (see Divac et al., 1988; Juorio and Vogt, 1967), but as large variations occur even

within a given species (e.g., compare studies in chicks of Schnabel et al., 1997, and Stewart et

al., 1996), differences in technical procedures are probably the main factor. Nevertheless, the

consistent result of all of these studies was an about 5- to 10-fold lower concentration of D1

receptors in the pallidal parts (PP and VP) than in the striatal parts (LPO, PA, BNST, Acc, TO)

of the avian basal ganglia.

With regards to subdivisons within the avian striatum, some authors reported higher D1

concentrations in the LPO than in the PA (Ball et al., 1995; Schnabel et al., 1997), while others

could not observe notable differences (Dietl and Palacios, 1988; Richfield et al., 1987; Stewart

et al., 1996). In addition, Schnabel et al. (1997) reported a decreasing rostro-caudal gradient of

D1 receptors in the LPO, in accordance with the rostro-caudally decreasing DA concentration

noted above (Juorio and Vogt, 1967; Juorio, 1983). In a developmental study by Schnabel and

Braun (1996) in chicks, D1 receptors showed a gradual but non-significant increase during the

first post-hatch week. Thus, the D1 receptor system seems to be largely developed in chicks at

the time of hatching, consistent with their precocial nature.

Figure 15. (A, C) Dopamine-immunoreactive fibers and (B, D) DARPP-32 immunoreactivity in the basal ganglia. Scale barrepresents 1 mm.


5.4.3. D2 receptors

The density of D2 receptors (Figure 8) in striatal parts of the avian basal ganglia (LPO, PA,

BNST, TO, Acc) seems to be even higher (about 1.5-fold) than that of D1 receptors (Dietl and

Palacios, 1988; Richfield et al., 1987; Stewart et al., 1996), although this was not consistently

found (Schnabel and Braun, 1996; Stewart et al., 1996). The lack of a higher D2 receptor

density in the Schnabel and Braun (1996) study is possibly due to a significant developmental

increase in the striatal parts of the basal ganglia during the first post-hatch week as reported by

these authors. Stewart et al. (1996), on the other hand, found that the relation of D1:D2 receptor

densities in the basal ganglia depended on the learning history of the animals, since after one-

trial avoidance learning the density of D1- but not D2-receptors was highly increased in the

LPO. The density of D2 receptors in the LPO is approximately the same as (Dietl and Palacios,

1988; Richfield et al., 1987; Schnabel and Braun, 1996) or slightly lower than in the PA

(Stewart et al., 1996). In contrast, D2 densities are 5- to 10-fold lower in the PP and INP than

in the LPO/PA (Dietl and Palacios, 1988; Richfield et al., 1987; Stewart et al., 1996).

5.4.4. DARPP-32

In pigeons (Durstewitz et al., 1998) and chicks (Schnabel et al., 1997), the striatal parts of the

basal ganglia exhibit very dense DARPP-32 labeling (Figure 15B, D), consistent with the

abundance of D1 receptors in these structures. In addition, and in contrast to the

autoradiographic D1-receptor binding studies, the pallidal PP (but not the INP) is as dense in

its DARPP-32 labeling as the striatal components (Figure 15B, D). However, DARPP-32-

positive labeling in the PP is almost exclusively constrained to fiber labeling, whereas a high

number of DARPP-32ir somata shows up in the neighbouring PA and LPO. It is conceivable

that the dense neuropil labeling in PP stems from descending D1-positive axons, as it is the

case in the globus pallidus of rats. In these animals, D1- and DARPP-32- but not D2-positive

fibers which probably belong to the striatonigral and striatoentopeduncular tracts have been

described surrounding unlabeled somata and dendrites in the globus pallidus (Ouimet and

Greengard, 1990; Yung et al., 1995). Finally, the avian Acc is relatively low in DARPP-32

content (Figure 13D; Durstewitz et al., 1998; Schnabel et al., 1997), in contrast to the high

densities of dopaminergic fibers (Wynne and Güntürkün, 1995) and D2 receptors (Dietl and

Palacios, 1988), but consistent with the low density of D1 receptors in the quail Acc according

to Ball et al. (1995).



Anderson and Reiner described two major populations of striatonigral projection neurons, one

that co-contains substance P and dynorphin and makes up for about 85-95 % of all projection

neurons, and the other that contains enkephalin and makes up for only 1-4 % of all projection

neurons (Anderson and Reiner, 1990, 1991b; Reiner et al., 1984b; Reiner and Anderson, 1990).

Projections from both of these neuronal populations in the medial striatum terminate on DAir

and THir, but also - to an about equal degree - on unlabeled cells in the SNC (Anderson et al.,

1991; Medina et al., 1995). Vice versa, both populations of striatonigral projection neurons

receive input from THir fibers as demonstrated by EM, targeting perikarya, dendritic shafts,

and spines (Karle et al., 1992, 1994). Furthermore, a subpopulation of substance P-ir

projection neurons (39%) also contained somatostatin, whereas neurons containing only

somatostatin or neuropeptide Y were not observed to project to the SNC (Anderson and Reiner,

1990). These findings demonstrate that in the avian like in the mammalian basal ganglia (Song

& Harlan, 1994; Takagi, 1986) subpopulations of striatal output neurons which interact with the

dopaminergic midbrain neurons could be differentiated by means of the neuropeptides co-

released by these neurons.

Reiner and Anderson (1990; Reiner et al., 1994) suppose that both, substance P containing and

enkephalin containing neurons are also colocalized with GABA. However, until now there is

no convincing evidence confirming this hypothesis. According to immunocytochemical studies,

the percentage of GADir and GABAir neurons in the avian basal ganglia seems to be low (< 15

%), even after pre-treatment with colchicine, and no significant regional variations could be

observed throughout the entire telencephalon (Durstewitz et al., 1998; Veenman and Reiner,

1994). If this pattern should hold true in the light of other methods such as in situ hybridization,

this would stand in striking contrast to what is known from the mammalian basal ganglia, where

the vast majority of neurons are of the medium spiny type which utilizes GABA (Chesselet et

al., 1987; Kita and Kitai, 1988; Surmeier et al., 1988). This apparent difference between the

mammalian and avian basal ganglia might also explain why no colocalizations of GAD and

DARPP-32 could be detected in the PA by Durstewitz et al. (1998). In contrast, in the

mammalian striatum, the very high percentages of GABAergic and DARPP-32ir neurons per se

implicate a high degree of overlap (Anderson and Reiner, 1991a; Ouimet et al., 1984, 1992).

However, it has to be pointed out that the extremely dense meshwork of THir and DAir fibers

and the very high concentration of DARPP-32 in the avian basal ganglia make the study of

possible colocalizations at the light-microscopic level extremely difficult if not impossible in


this structure. For two other reasons the apparent lack of DARPP-32/GAD colocalizations in

the paleostriatum does not exclude that GABAergic neurons in these regions are

dopaminoceptive: First, Veenman and Reiner (1994) reported a subpopulation of small

GABAergic neurons not detected by GAD immunochemistry. Second, as discussed above, D2

receptors are even more abundant than D1 receptors in the avian basal ganglia, so that the

possibility remains that GABAergic cells in these regions are equipped with receptors of this

class. Hence, the issue whether GABAergic neurons in the avian basal ganglia receive a

significant dopaminergic input awaits further investigation.

In the VP and the BNST of the male quail brain, aromatase-ir cells were found to be located in

THir baskets, and THir varicosities were found in close vicinity of almost all of these cells

(Balthazart et al., 1998). In fact, due to the particular high concentration of both of these

markers in the VP and BNST, these structures might be major loci of interaction between

catecholaminergic systems and forebrain steroid hormone activity.

5.5. Dopaminergic innervation of song nuclei in song birds

The brain of song birds contains several sexually dimorphic nuclei involved in the song system

(Nottebohm et al., 1976, 1982) which are adapted to the specific sensorimotor requirements of

song perception and production. In the context of the dopaminergic system, these specific nuclei

deserve a special discussion, because, interestingly, they receive a much denser dopaminergic

input as measured by THir and catecholamine histofluorescence than the surrounding tissue

(Bottjer, 1993; Lewis et al., 1981; Sakaguchi and Saito, 1989; Soha et al., 1996). Among these

nuclei are a nucleus in the dorsolateral LPO, the so-called Area X, the robust n. of the

archistriatum, the lateral magnocellular n. of the anterior neostriatum, the n. interfacialis, and

the high vocal center located in the caudal neostriatum. In addition to the higher THir exhibited

by these nuclei compared to the surrounding areas, a higher density of D1-specific binding has

been demonstrated in area X of male starlings compared to the surrounding LPO (Casto and

Ball, 1994). Furthermore, a dense input to area X from VTA and SNC has been shown by

retrograde tracing (Lewis et al., 1981). In addition, the sexual dimorphism of the song nuclei is

also reflected in their dopaminergic innervation as these nuclei in male birds which sing

receive a much denser dopaminergic input than in the non-singing females. In contrast, sex

differences with regards to DA receptor densities have not been observed in non-singing birds

like quails (Ball et al., 1995).

What makes these nuclei even more interesting in terms of their dopaminergic innervation is that


the density of the dopaminergic input to these nuclei as well as DA levels and turnover rates are

clearly correlated with the phases of song learning. Thus, in young zebra finches (< 30 d), THir

in the song nuclei is even less dense than in the surrounding tissue (Sakaguchi and Saito, 1989;

Soha et al., 1996), while a strong developmental increase occurs with the onset of the

sensorimotor phase of song learning around day 35. In accordance with these findings, Harding

et al. (1998) observed highly significant increases in DA levels and turnover rates in the song

nuclei within the LPO, the n. robustus of the archistriatum, the lateral magnocellular n. of the

anterior neostriatum, and the n. interfacialis in the neostriatum in zebra finches between

postnatal days 35-55, which strongly declined again at day 90. A similar peak around day 35 at

least in DA turnover was observed in the auditory Field L. Thus, the dopaminergic system

probably has a prominent role in song learning and production that might be related to the

requirement that sensory and motor aspects have to be integrated over time during song learning

and production (see 7.6).

6. Comparative aspects of the dopaminergic system in mammalian and avian species

From the description above, readers familiar with the dopaminergic system of mammals may

have recognized that the avian dopaminergic system shares many important organizational

features with its mammalian counterpart. Like in mammals (Berger and Gaspar, 1994;

Björklund and Lindvall, 1984; Fallon and Loughlin, 1995; Swanson, 1982), the dopaminergic

innervation of the avian telencephalon arises mainly from a small population of dopaminergic

midbrain nuclei in the VTA and SNC, where only a small percentage of fibers crosses into the

contralateral hemisphere (Metzger et al., 1996; Kitt and Brauth, 1986). The following

description focuses on some very prominent and interesting similarities and differences in the

dopaminergic innervation of the telencephalon of birds, rodents, and primates.

6.1. Basal ganglia

Like the avian 'striatum', the caudate-putamen in rats and primates receives the densest

dopaminergic fiber input and displays the highest concentrations of D1- as well as D2-

receptors (Ariano, 1997; Björklund and Lindvall, 1984; Brock et al., 1992; Camps et al., 1990;

Joyce et al., 1993; Ouimet et al., 1984; Yung et al., 1995). In addition, in mammals and birds

the caudate-putamen, or dorsal striatum, is higher in its dopaminergic innervation, DARPP-32

content, and DA receptor density (at least in D2-specific binding; Richfield et al., 1987) than


the n. accumbens/ ventral striatum (Ball et al., 1995; Brock et al., 1992; Durstewitz et al., 1998;

Hemmings and Greengard, 1986; Joyce et al., 1993; Karle et al., 1996; Yung et al., 1995;

Wynne and Güntürkün, 1995). In contrast, 'pallidal' structures in both amniote classes are much

weaker innervated. Furthermore, despite the lower D1 receptor densities in pallidal compared

to striatal structures, a high amount of DARPP-32 is present in the mammalian globus pallidus

(Hemmings and Greengard, 1986; Ouimet et al., 1984, 1992) and its avian equivalent PP (see

5.4.4). However, as discussed for the PP in Section 5.4.4, the high DARPP-32ir in the globus

pallidus of rats is almost exclusively due to neuropil labeling, in contrast to the high number of

DARPP-32ir neurons in the striatum (compare Figure 5B in Durstewitz et al., 1998, to Figure 8

in Ouimet et al., 1984), and may stem from intense staining of descending striatonigral and

striatopallidal fibers. Finally, a rostro-caudal gradient in DA receptor densities and DA

concentration as observed in birds (see Section 5.4) has also been noted in mammals (Bockaert

et al., 1976; Boyson et al., 1986; Richfield et al., 1987; Tassin et al., 1976)

However, some differences between birds and mammals are also apparent. First, direct

comparisons between pigeons, rats, and cats revealed that the concentration of D1 receptors in

the avian 'striatum' is about 5- to 10-fold lower, and that in the avian 'pallidum' is about 20-fold

lower than in the mammalian striatum (Dietl and Palacios, 1988, Richfield et al., 1987). In

contrast, the density of D2 receptors in the pigeons 'striatum' is only about half of that in the

mammalian striatum, while, however, this difference is larger for the pallidum (Dietl and

Palacios, 1988, Richfield et al., 1987). Second, the first point implies that the ratio of D1:D2

receptors is different in the avian and the mammalian basal ganglia. Whereas in mammals D1

receptors are much more prevalent than D2 receptors (Joyce et al., 1993), the opposite seems to

be the case in birds (see 5.4.3). As D1- and D2-receptors are linked to different G-proteins and

influence adenylate cyclase activity in opposite ways (Hemmings et al., 1987b; Robinson and

Caron, 1997), the different D1:D2 receptor ratios may imply important differences in basal

ganglia function between birds and mammals.

The possible lack of DARPP-32/GAD colocalizations in the pigeon PA has already been

discussed (Section 5.4.5), and has been related to other particularities of the avian basal

ganglia, namely to the fact that the number of GAD- and GABA-positive neurons in the avian

basal ganglia as assessed by immunocytochemical techniques (Durstewitz et al., 1998;

Veenman & Reiner, 1994) seems to be much lower than in the mammalian basal ganglia (Kita

and Kitai, 1988; Surmeier et al., 1988).


6.2. Cortex

Like in birds, in rodents and primates the dopaminergic input to cortical areas is clearly weaker

than that to the striatum, and the densities of dopaminergic receptors and of DARPP-32 sharply

decline outside the basal ganglia (Ariano, 1997; Berger et al., 1988, 1990, 1991; Berger and

Gaspar, 1994; Björklund and Lindvall, 1984; Brock et al., 1992; Hemmings and Greengard,

1986; Joyce et al., 1993; Ouimet et al., 1984, 1992). In the frontal cortices of rats and primates,

the density of D1 receptors has been estimated to be about five- to tenfold higher than that of D2

receptors (Joyce et al., 1993; Lidow et al., 1991), whereas this ratio seems to be in the range of

3:1 throughout most of the bird telencephalon (Dietl and Palacios, 1988; Schnabel and Braun,

1996; Stewart et al., 1996). In contrast, in the avian Hp the density of D2 receptors outnumbers

that of D1 receptors (Stewart et al., 1996). In the mammalian Hp, the D1:D2 ratio seems to be

at least lower than in the neocortex although in general D1 receptors may still be more

numerous (Dewar and Reader, 1989; Joyce et al., 1993). However, D1 and D2 receptors

distribute differentially across hippocampal layers (Köhler et al., 1991).

With respect to the finctional organization of the dopaminergic innervation of the avian and

mammalian 'neocortex', strong equivalencies can be observed. Like in birds, the primary visual,

auditory, and somatosensory cortices in mammals are only weakly innervated by dopaminergic

fibers, at least much weaker than the respective secondary areas (Berger et al., 1988, 1991;

Berger and Gaspar, 1994; Joyce et al., 1993). These areas also express lower levels of

DARPP-32 (Berger et al., 1990). In addition, layer IV in the granular cortices in rats and

primates, which is the major target zone of specific thalamic input, lacks a significant

dopaminergic input (Berger et al., 1988, 1991; Berger and Gaspar, 1994; Joyce et al., 1993;

Phillipson et al., 1987).

With respect to the 'higher order' neocortical areas, considerable differences between rodents

and primates have been pointed out by Berger and coworkers (Berger et al.,1991; Berger and

Gaspar, 1994). In general, the dopaminergic innervation of the primate neocortex is more

widespread than that of the rat neocortex, and may thus compare better to the dopaminergic

innervation of the avian neostriatal and hyperstriatal areas. In both, rodents and primates, the

prefrontal areas and the anterior cingulate cortex are densely innervated (Berger et al., 1991;

Berger and Gaspar, 1994; Joyce et al., 1993). Therefore, the dense dopaminergic innervation of

the NCL in pigeons was taken as an indication that this multimodal area may be comparable to

the mammalian PFC (Divac et al., 1985; Divac and Mogenson, 1985; Waldmann and

Güntürkün, 1994; Wynne and Güntürkün, 1995). However, the PFC of mammals is not a


homogeneous structure. In primates, only the orbitofrontal and ventrolateral portions receive a

dense dopaminergic input, while the dorsolateral part is lower in this respect (Berger et al.,

1988; Lewis et al., 1992; Williams and Goldman-Rakic, 1993), and is also very low in

DARPP-32 content in adult animals (Berger et al., 1990). In rats, however, the medial

prefrontal areas, in particular the pre- and infralimbic region, which are assumed to be

equivalent to the primate dorsolateral PFC, are the neocortical areas highest in DA fiber

density (Berger et al., 1991; Berger and Gaspar, 1994; Joyce et al., 1993). In fact, this was one

of the reasons for Preuss (1995) to question the equivalency between the rat pre- and

infralimbic cortex and the primate dorsolateral PFC. Similar subdivisions of the avian

'prefrontal cortex' were not identified yet. Given the dense dopaminergic input of the NCL, one

would compare this area to the prelimbic/infralimbic area of rats, or to the orbitofrontal or

ventrolateral PFC of primates. On the other hand, the percentage of DARPP-32ir neurons in the

NCL was lower than in other, more rostrally located neostriatal structures, reminiscent of the

relatively low DARPP-32ir in the adult primate dorsolateral PFC (Berger et al., 1990). The

NCL would also compare better to the primate dorsolateral PFC with respect to its functional

characteristics, as both areas seem to be involved especially in spatial working memory

(Funahashi and Kubota, 1994; Gagliardo and Divac, 1993; Gagliardo et al., 1996; Goldman-

Rakic, 1988; Güntürkün, 1997; Mogenson and Divac, 1982, 1993; Wilson et al., 1993; but see

Petrides, 1995).

Furthermore, in primates, in contrast to rats, the premotor and motor cortices are very densely

innervated by the dopaminergic system (Berger et al., 1991; Berger and Gaspar, 1994; Joyce et

al., 1993). Insofar, pigeons and chicks compare better to monkeys than to rats (see also Reiner

et al., 1994), as major parts of the avian 'motor cortex', in particular the dorsal archistriatum

intermedium, also receive a rich dopaminergic input. The comparison holds also for the

distribution of DARPP-32ir neurons: Despite its exceptionally high dopaminergic input, the

primate motor cortex displays only a sparse distribution of DARPP-32ir neurons (Berger et al.,

1990), similar to the very low number of DARPP-32ir neurons throughout most of the

intermediate archistriatum (Durstewitz et al., 1998). In addition, Berger et al. (1990) pointed

out a mismatch between the distribution of D1 receptors and DARPP-32ir neurons in the

primate motor cortex similar to that described for the dorsal archistriatum (Sections 5.3.2 and

5.3.4).

Finally, moderate to high numbers of dopaminergic fibers are present in the secondary sensory

and association areas of the primate neocortex (Berger et al., 1988, 1991; Berger and Gaspar,


1994). In this respect, the dopaminergic innervation of the avian telencephalon might also be

more similar to that of primates than to that of rats.

6.3. Laminae-specific distribution of dopaminergic fibers

An additional feature of the dopaminergic innervation of the mammalian neocortex is its

laminae-specific distribution. In the rat medial and orbital (lateral) prefrontal areas, the deep

layers V-VI are the major targets of dopaminergic fibers (Berger et al., 1991; Björklund and

Lindvall, 1984; Joyce et al., 1993). In the primate granular cortices, the superficial layers I-III

are generally even more densely innervated than the deep layers V-VI, whereas in the motor and

anterior cingulate (agranular) cortices all layers receive a dense dopaminergic input (Berger et

al., 1988, 1991; Berger and Gaspar, 1994; Goldman-Rakic et al., 1992; Lewis et al., 1992). As

outlined in Section 2, a cortical lamination is absent in the avian brain. Nevertheless, based on

hodological, histochemical, and functional criteria, some authors have compared subdivisions

of the avian telencephalon to specific laminae of the mammalian neocortex. It was argued that

the primary sensory areas (E, IHA,, Field L2, Bas) are equivalent to the thalamic input layer IV

of the respective neocortical areas (extrastriate and striate visual, auditory, somatosensory),

and the secondary sensory belts surrounding these primary areas and the HD to layers II-III

(Karten and Shimizu, 1989; Reiner and Karten, 1983; Veenman et al., 1995). Furthermore, the

sensorimotor archistriatum, the HA, structures of the PE, and possibly the NCL, which either

give rise to the major descending pathways or directly project onto the sensorimotor avian

striatum, have been identified with the neocortical output layers V-VI. This conception would

again render the dopaminergic innervation of the avian brain more similar to that of primates

than to that of rats, as both, the 'deeper layers' and the 'superficial layers' of the avain

'neocortex' would be high in DA and DARPP-32.

From a functional perspective, another interesting parallel between birds and mammals might

lie in the fact that DA-baskets seem to contact predominantely bigger neurons in the avian brain

(Ø > 15 µm), whereas smaller neurons are more likely innervated en-passant (Wynne and

Güntürkün, 1995). Likewise, in the mammalian neocortex, the bigger pyramidal neurons reside

in the deeper layers and are thus the ones probably most heavily innervated by DA fibers within

their deep proximal and, in primates in addition, upper layer distal apical dendrites. Moreover,

the bigger pyramidal cells are the ones which most likely exhibit repetitive oscillatory bursting-

characteristics in the neocortex (Mason and Larkman, 1990; Yang et al., 1996a), an

electrophysiological feature that - theoretically - could be exclusively due to their bigger


somata and dendrites (Mainen and Sejnowski, 1996). Thus, dopaminergic baskets may

innervate neurons in the avian telencephalon with specific functional characteristics as

dopaminergic fibers possibly do in the mammalian neocortex.

6.4. Ultrastructural features and postsynaptic targets of the dopaminergic innervation

With respect to the postsynaptic targets of the dopaminergic input to the mammalian neocortex,

pyramidal cells seem to be the predominant dopaminoceptive population, and most DARPP-

32ir neurons are of the pyramidal type (Berger et al., 1990; Goldman-Rakic et al., 1989; Smiley

and Goldman-Rakic, 1993). Thus, the finding that GADir neurons in the avian telencephalon do

not express DARPP-32 and are never located in THir baskets (Durstewitz et al., 1998), on a

first glance, fits well into the mammalian schema. However, dopaminergic fibers in the PFC

have also been observed to terminate on smooth, presumably GABAergic, stellate cells (Smiley

and Goldman-Rakic, 1993), and the activity of GABAergic neurons in the mammalian PFC has

been shown to be modulated by DA in vivo and in vitro (Penit-Soria et al., 1987; Pirot et al.,

1992; Rétaux et al., 1991; Yang et al., 1997). The dopaminergic effect on GABAergic activity

can be antagonized by D2- but not by D1-receptor blockers (Godbout et al., 1991; Pirot et al.,

1992; Rétaux et al., 1991, but see Yang et al., 1997), although D2 receptors are present in much

lower densities in the mammalian neocortex than D1 receptors (see above). Hence, GABAergic

neurons may be affected mainly via D2 receptors, and this might also be the case in the avian

brain. Thus, despite their low densities, D2 receptors might play an important functional role

also in the avian telencephalon, and further investigation of this subject is certainly very

important.

Dopaminergic fibers in the avian telencephalon also have most of their ultrastructural features

in common with their mammalian counterparts. In both animal classes, dopaminergic synapses

are relatively small, are mainly of the symmetric type, contact predominantly dendritic arbors

and spines, and sometimes converge with other unlabeled synapses on the same spine (i.e., form

triadic complexes) (Goldman-Rakic et al., 1989; Karle et al., 1996; Metzger et al., 1996;

Séguéla et al., 1988; Smiley and Goldman-Rakic, 1993; Yung et al., 1995; see 4.4). In addition

to a specialized synaptic release-mode, much of the neuromodulator seems to be released via

unspecialized axonal varicosities in both amniote classes (Cooper et al., 1996).

In conclusion, the general pattern of the distribution of dopaminergic fibers and receptors in the

avian telencephalon, as well as the laminar and biochemical characteristics of the

dopaminergic target neurons, are quite similar to that of mammals. Some specific differences


seem also to exist, but may partly be due to the fact that our knowledge about DA receptor

subtypes in the avian brain is still incomplete. Finally, considerable differences in the cortical

organization of the dopaminergic system have also been observed within the class of mammals.

7. Behavioral studies and functional implications

From the neuroanatomical features of the dopaminergic innervation in the avian brain, some

functional clues could be derived. For example, the fact that primary sensory areas are devoid

of DA, while higher sensory, associative, and motor areas that have a direct link to structures of

the avian basal ganglia receive a dense dopaminergic input and are high in DA receptors makes

it likely that DA plays a special role in sensory-motor integration and associative learning.

In fact, the dopaminergic system in mammals is implicated especially in motor, associative, and

higher order functions like aversive and appetitive learning, and working memory (Beninger,

1993; Salamone, 1992, 1994; Sawaguchi and Goldman-Rakic, 1991, 1994; Schultz et al., 1993,

1995; Seamans et al., 1998; Sokolowski et al., 1994; Zahrt et al., 1997). Although much less is

known about the involvement of DA in behavioral and cognitive functions in birds, from the

studies in avian species that do exist so far a similar functional involvement of the

dopaminergic system as in mammals is apparent. The next sections will deal with these studies.

7.1. Dopaminergic modulation of (unconditioned) motor functions

Dopamine in the avian brain has been shown to be critically involved in motor functions. Rieke

(1980, 1981) observed that unilateral kainic acid lesions of the paleostriatum, or its source of

dopaminergic input, the SNC, in pigeons induced persistent turning in one direction, postural

problems, arhythmic movements, and head or whole body tremors. These behavioral

dysfunctions were reproduced by unilateral injections of GABA agonists into the SNC (Rieke,

1982). Furthermore, the turning behavior elicited by unilateral GABA-agonistic injections into

the SNC or 6-hydroxydopamine (6-OHDA) destruction of dopaminergic neurons in this area

could be enhanced by apomorphine but was suppressed by haloperidol administration (Rieke,

1982; Yanai et al., 1995).

Apomorphine has also been shown to facilitate stereotypic pecking bouts in a dose-dependent

manner, alone or induced by feeding stress, and the avian basal ganglia is a likely site for these

dopaminergic effects (Cheng and Long, 1974; Deviche, 1985; Goodman et al., 1983; Nisticò

and Stephenson, 1979; Rieke, 1982; Zarrindast and Amin, 1992). In contrast, DA antagonists


like haloperidol, chlorpromazine, or clozapine inhibit behavioral stereotypies (Cheng and

Long, 1974; Goodman et al., 1983; Kostal and Savory, 1994), and could cause sedation and

tonic immobility that could be reversed by paleostriatal but not by neostriatal lesions (Sanberg

and Mark, 1983). The effects of apomorphine and amphetamine on pecking behavior can

furthermore be antagonized by specific D1- (SCH23390) as well as by high doses of specific

D2-receptor blockers (sulpiride), with a combination of both being most effective, while D1-

or D2-specific agonists enhance apomorphine induced pecking (Dehpour et al., 1998;

Zarrindast and Amin, 1992; Zarrindast et al., 1992; Zarrindast and Namdari, 1992). Low doses

of sulpiride, however, enhanced pecking, presumably by blocking D2 autoreceptors, and low

doses of apomorphine, on the other hand, might reduce pecking also via an autoreceptor

mechanism (Deviche, 1985). It is important to point out that DA antagonists like haloperidol do

not prevent all aspects of motor behavior although they reduce pecking bouts and turning, and

may lead to sedation. Hence, even after haloperidol administration, Goodman et al. (1983)

observed pigeons engaging in some normal behaviors like feeding, and Rieke (1982) was still

able to elicit orienting and escape responses, including flying, by stressful stimuli.

Stereotypic motor behavior and hyperactivity on the one hand as induced by supranormal DA

receptor stimulation or DA activity, and sedation on the other hand as induced by, e.g.

neuroleptic drugs, are also typically observed in mammals including humans (Bo et al., 1988;

Clark and White, 1987; Gessa et al., 1985; Le Moal and Simon, 1991; Ridley, 1994;

Sokolowski and Salamone, 1994; Waddington and Daly, 1993; Wickens, 1990). However, like

in birds, behavioral reactions to intense or significant stimuli may be left intact after DA

depletion or subnormal DA receptor stimulation in the basal ganglia (Salamone, 1992). Thus,

Salamone (1992) concluded that DA is not that much involved in motor processes per se but

more in sensory-motor integration that requires novel motor sequences or in processes that

extend significantly in time (see also Taylor and Saint-Cyr, 1995). DA might also be especially

important for motor behavior that is driven by internal states without being directly controlled

by external stimuli.

Lindenblatt and Delius (1988) demonstrated that apomorphine-induced pecking bouts could

also be elicited by local injections into the Bas, and could be reversed by 6-OHDA injections

into this structure. However, since dopaminergic fibers and receptors seem to be lacking in the

Bas, the behavioral effects of DA-agonistic drugs might also be attributed to ligand diffusion

into the adjacent PA and PP. Consistent with this interpretation, apomorphine injections into the

Bas also produced contralateral turning (Lindenblatt and Delius, 1988), a typical symptome of


altered DA activity in the avian basal ganglia (Rieke, 1981, 1982). Wynne and Delius (1996)

reported that apomorphine-induced pecking could be reduced by lesions of the Bas, thus

demonstrating that the Bas is involved in apomorphine-induced stereotypies. However, this

result does not necessarily imply that the Bas is in fact a site of apomorphine action.

Apomorphine might increase pecking rates via actions in the basal ganglia, while Bas lesions

could reduce pecking rates independently. On the other hand, the finding of Durstewitz et al.

(1998) that the Bas had high number of DARPP-32ir neural cell bodies may hint to some

anatomical basis for the apomorphine induced pecking bouts. Hence, the neuroanatomical site

of the apomorphine-induced pecking bouts remains unclear.

7.2. Dopaminergic involvement in arousal and wakefulness

Ferrari and Giuliani (1993) found that DA in the chick brain might also participate in the

regulation of states of arousal and wakefulness. Selective D2 agonists induced hypomotility and

sleep-like states. However, for some D2 agonists (like lisuride) or the mixed D1/D2 agonist

apomorphine, this effect reversed at higher doses, leading to behavioral excitation and

increased spontaneous pecking instead. Similar dose-dependent patterns were also observed in

rodents, where the sedatory effects of the drugs were attributed to predominant stimulation of

presynaptic D2 autoreceptors at lower doses (Ferrari and Giuliani, 1993; Ongini, 1993). In

accordance with this interpretation, microinfusion of apomorphine into the rat VTA, but not into

the Acc, caudate n., or PFC, induces behavioral and EEG signs of sleep, which could be

antagonized by sulpiride (Bagetta et al., 1988a, 1988b). In general, in mammals dopaminergic

drugs increase alertness, wakefulness and exploratory acitvity, with accompanying indications

in the EEG, while lesions of the dopaminergic system, DA antagonists, or subnormal DA

activity after preferential D2 autoreceptor stimulation have the opposite effect (Bagetta et al.,

1988a, 1988b; Gessa et al., 1985; Montaron et al., 1982; Ongini, 1993). Dopaminergic effects

on arousal and wakefulness on the one hand, and on behavioral excitation and stereotypies

(Section 7.1) on the other, may be mediated by the same neural mechanism (Section 7.6).

7.3. Dopaminergic modulation of appetitive and consummatory behavior

DA in the mammalian brain is well known to regulate aspects of food and water intake

(Bertolucci-D’Angio et al., 1990a, 1990b; Cooper and Al-Naser, 1993; Wilson et al., 1995),

and of sexual behavior (Pfaus and Phillips, 1989). Likewise, apomorphine and DA

administration in pigeons reduces food consumption (Deviche, 1984; Ravazio and Paschoalini,


1992), possibly because apomorphine and DA could substitute for the rewarding value of food.

Alternatively, the low doses of apomorphine used by Deviche (1984) might have reduced food

intake by very much the same mechanism that has been proposed for apomorphine-induced

sedation (see Section 7.2), i.e. primarily via D2 autoreceptors. In fact, Deviche (1984) noted

that the doses of apomorphine used in his study were too low to elicit pecking bouts, or even

inhibited pecking.

Apomorphine also reduces appetitive and consummatory aspects of sexual behavior in the male

quail (Absil et al., 1994; Castagna et al., 1997). These apomorphine-induced effects are

probably due to D2-receptor stimulation as D2 agonists reduce both appetitive and

consummatory sexual behavior, while D1 agonists do just the reverse, i.e. increase both aspects

of sexual behavior (Balthazart et al., 1997). Again, these actions of the dopaminergic system

may be related to a rewarding function of DA which has been proposed by several authors

(Schultz et al., 1993, 1995; see Beninger, 1993; see Salamone, 1992, 1994; Wickens, 1990;

Wickens and Kötter, 1995). According to studies in rats, part of the dopaminergic effects on

food consumption and sexual behavior might be regulated by dopaminoceptive hypothalamic

nuclei (Cooper and Al-Naser, 1993). However, additional structures are probably also

involved since increases of DA metabolites during food consumption or upon presentation of

appetitive stimuli have also been observed in the Acc, medial PFC, and in the amygdala in rats

(Bertolucci-D’Angio et al., 1990b; Wilson et al., 1995).

The ‘reward theory’ of DA function partly derived from studies on intracranial self-stimulation

and self-administration of dopaminergic drugs (Fibiger, 1978; Le Moal and Simon, 1991; Wise,

1978; Yokel & Wise, 1975). Self-administration of cocaine was also observed in pigeons,

where it could be antagonized by haloperidol (Winsauer and Thompson, 1991), pointing to a

similarly ‘rewarding’ action of DA in the avian as in the mammalian brain.

7.4. Dopaminergic modulation of learning and conditioned behavior

Dopamine in the avian telencephalon is involved in various forms of learning. Stewart et al.

(1996) found that after a one-trial taste aversion avoidance learning, D1- but not D2-specific

binding increased highly significantly in the LPO. A role for DA in parts of the striatum, namely

in the Acc, in passive avoidance training (and, more generally, in aversive situations) is also

well established in mammals (Beninger, 1993; Bertolucci-D’Angio et al., 1990a, 1990b;

Salamone, 1994). In an in vivo microdialysis study, Gruss and Braun (1997) demonstrated

decreased levels of the DA metabolite homovanillic acid in the MNH after auditory imprinting


in young chicks, while significant increased levels were observed after exposing the chicks to

handling stress. Stressful situations induce increases of DA metabolites also in rats in the Acc

and medial PFC (Abercrombie et al., 1989; Bertolucci-D’Angio et al., 1990b). The studies by

Soha et al. (1996) and Harding et al. (1998) reported in Section 5.5 suggest that DA is

additionally involved in the sensorimotor stage of song learning in zebra finches as the

postnatal development of the innervation of various forebrain song nuclei by THir fibers as

well as DA levels and turnover in these nuclei are correlated with this period. A direct

involvement of DA in learning mechanisms is made likely by a study by Lindenblatt and Delius

(1987) who showed that apomorphine (and thus probably DA) might serve as a positive

reinforcer when coupled with a conditioned stimulus. Furthermore, DA antagonists like

haloperidol reduce the rate of conditioned responding without affecting discrimination

performance (Tombaugh, 1981), while conditioned responding of young chicks for heat

reinforcement was shown to be enhanced especially after combined D1- and D2-agonist

injection (Dose and Zolman, 1994). Finally, following apomorphine administration, chicks

were found to be unable to suppress formerly rewarded but now punished responses, while this

can be reversed by haloperidol pretreatment (McDougall et al., 1987).

Although these studies show that DA is somehow involved in various learning processes, the

nature of the underlying mechanism is still far from clear. Thus, often it may not be easy to

decide whether the DA-induced effects can be attributed to a specific role of DA in learning, to

its motivational effect, to its role in motor behavior, or to its impact on attention and arousal

(see Salamone, 1992). This is especially true when DA agonists are administered systemically,

thus acting on many structures. For example, McDougall et al. (1987) hinted to the fact that

apomorphine in their experiments generally enhanced the response rate but that this effect may

have been obscured due to a ceiling effect as long as the response was associated with reward

only. Only when punishment was introduced, the behaviorally exciting effect of apomorphine

might have become apparent. Thus, although it is interesting to note that increased pecking rates

prevail even in the presence of punishment, the main effect of apomorphine may not be to inhibit

response suppression learning but to induce behavioral excitation.

7.5. Dopaminergic modulation of working memory

A failure to inhibit inadequate responses as demonstrated by McDougall et al. (1987) is also

often observed after lesions of the rat and primate PFC, or lesions of its dopaminergic afferents

(Dias et al., 1997; Fuster, 1989; Sokolowski and Salamone, 1994). A failure to suppress


irrelevant response options may also play a role in the inability of pigeons with NCL lesions -

depending on the extent to which the NCL was destroyed - to perform a reversal learning

(Hartmann and Güntürkün, 1998) or a go/no-go task (Güntürkün, 1997). In addition, lesions of

the NCL or its main source of thalamic afferents, the n. dorsolateralis posterior thalami, lead to

diminished delayed alternation performance (Gagliardo et al., 1996; Güntürkün, 1997;

Mogensen and Divac, 1982, 1993). Tasks which involve a delay are thought to be indicative of

working memory functions, as in these tasks an animal has to actively hold information in

memory for the guidance of responding in forthcoming situations (Fuster, 1989). Reduced

delayed alternation performance is also a characteristic deficit observed after PFC lesions in

mammals (Fuster, 1989). With regards to the transient nature of the spatial working memory

deficits after NCL lesions, the NCL compares better to the pre-/infralimbic region of rats than

to the dorsolateral PFC of primates, where lesions lead to more serious and longer-lasting

working memory deficits (Preuss, 1995).

DA and D1 receptors within the PFC are well known to be involved in working memory

functions in primates and rats (Brozoski et al., 1979; Müller et al., 1998; Sawaguchi and

Goldman-Rakic, 1991, 1994; Seamans et al., 1998; Simon et al., 1980; Zahrt et al., 1997). Also

in pigeons, blockade of dopaminergic transmission or receptors by systemic administration of

various neuroleptic drugs like chlorpromazine and clozapine has been shown to decrease

delayed matching to sample performance in a dose-dependent fashion (Picker and Massie,

1988; Watson and Blampied, 1989). Interestingly, both these drugs increased accuracy when

administered at low doses, possibly via predominant actions on D2 autoreceptors at these doses

(Picker and Massie, 1988).

Only recently our laboratory has gathered evidence that D1 receptors specifically in the pigeon

NCL might be involved in response inhibition and working memory. Thus, Diekamp et al.

(1998) presented evidence that local D1 receptor blockade by SCH23390 in the NCL impairs

go/no-go performance. After local SCH23390 application, pigeons tended to respond to all

stimuli irrespective of their reward value. Güntürkün and Durstewitz (in press) trained pigeons

in a labyrinth task which had a spatial working memory and a non-spatial reference memory

component. At the beginning of a daily session pigeons were randomly placed at one of two

starting positions in a labyrinth with 24 chambers, in which 12 arms contained red cups that

were never baited, while the remaining 12 chambers contained white cups that were baited only

at the onset of a session. Thus, as a reference memory component of the task, pigeons had to

learn to discriminate red from white cups as only the latter ones contained food pellets.


Moreover, they had to learn never to return to an arm visited previously as each white cup only

contained food once at the onset of a session. This was the working memory component of the

task as pigeons mentally had to keep track of all arms already visited, or - complementary - had

to keep in mind all arms still worth visiting. After local SCH23390 injections into the NCL,

pigeons had no problems with their reference memory, but were significantly impaired

compared to pigeons which received saline control injections in the working memory part. In

addition to true working memory errors, which were defined as re-entrances into arms after the

animal had already visited one or more other arms, there was also a significant increase in

'perseveration' errors. Perseveration in this task was defined as an immediate re-entrance into

an arm, most often due to the fact that a pigeon stayed in an arm and tried to access the same

food cup a couple of times in close temporal succession. Perseveration is also a well known

phenomenon in prefrontal mammals and patients with prefrontal disorders or lesions (Dias et

al., 1997; Iversen and Mishkin, 1970; Fuster, 1989; Milner and Petrides, 1984).

7.6. The possible function of dopamine in sensory-motor processes

Summarizing Sections 7.1-7.5, the dopaminergic system in birds - consistent with the

anatomical data - plays a role in motor functions, in arousal, in learning, and probably in

working memory, while DA levels do not change with visual experience (Davies et al., 1983)

and DA receptor antagonists, in general, do not seem to affect sensory discrimination per se,

although they might interfere with sensory-motor integration and instrumental responding

(Güntürkün and Durstewitz, in press; Mogensen and Divac, 1993; Tombaugh, 1981). Thus, in

the avian telencephalon the dopaminergic system is involved in very much the same motor and

cognitive functions as it is in mammals, supporting the notion of a tight link between structure

and function in the nervous system. That is, the similar structural/anatomical organization of the

dopaminergic system in birds and mammals seems to give rise to a similar functional

organization, despite the facts that the evolution of birds and mammals diverged more than 200

million years ago and that the organization of avian telencephalic areas is quite different from

that of mammals.

As already noted in Section 7.4, some of the behavioral functions of DA might be difficult to

discern. According to Salamone (1992), the apparent dopaminergic involvement in motor

functions and in motivation might actually refer to common neural processes, related to sensory-

motor integration evolving in time, and to instrumental or 'intentional' behavior. To unravel the

possible function of DA in various motor and cognitive processes, and to explicate the detailed


neural mechanisms by which DA achieves these functions, knowledge about DA-induced

manipulations of neural and synaptic biophysical properties is essential. However, to our

knowledge, until now nothing is known about dopaminergic modulation of electrophysiological

and biophysical properties of neurons in avian forebrain regions, neither in vivo nor in vitro.

In the PFC of rats, DA enhances a persistent Na+ current (Gorelova and Yang, 1997; Shi et al.,

1997; Yang and Seamans, 1996), reduces a slowly inactivating K+ and a dendritic high-voltage-

activated Ca2+ current (Seamans et al., 1997; Yang and Seamans, 1996; Yang et al., 1996b),

reduces synaptic currents through AMPA and NMDA channels (Law-Tho et al., 1994; Pralong

and Jones, 1993), and increases spontaneous activity of GABAergic interneurons (Godbout et

al., 1991; Penit-Soria et al., 1987; Pirot et al., 1992; Rétaux et al., 1991; Yang et al., 1997). It

could be shown by computational modeling that DA by inducing these biophysical changes

could act to stabilize active neural representations, and to protect them against interfering

stimulation and noise (Durstewitz et al., 1999). That is, DA ensures that delay-activity of PFC

neurons in working memory tasks, which is believed to represent the active holding of goal-

related information (Funahashi and Kubota, 1994; Fuster, 1989; Goldman-Rakic, 1995), is

sustained until the goal has been achieved. The finding that mainly or exclusively excitatory

neurons are equipped with D1 receptors (see Section 5) fits well into this functional schema,

as, at least in mammals, the D1 receptor-mediated effects on pyramidal cells are the ones

contributing most to the stabilizing effect of DA (Durstewitz et al., 1999).

The proposed function of DA may not just be important in working memory situations, but in

fact in any sensory-motor process, where a representation of the current goal-state of the

movement has to be kept upright and to be compared with incoming sensory information.

Indeed, sustained delay-activity has also been observed in the primate premotor and motor

cortices (Di Pellegrino and Wise, 1991), the striatum (Apicella et al., 1992), and the posterior

parietal cortex (Constantinidis and Steinmetz, 1996), which all receive a dense dopaminergic

input in primates. In addition, as pointed out by Salamone (1992), temporally extended sensory-

motor processes and sequences are especially susceptible to interference with the

dopaminergic system while brief responses are less easily disrupted. Sustained delay-activity

may be furthermore important in various, especially operant, learning processes where a

temporal gap between related stimuli has to be bridged in order to detect the relation. Thus,

various forms of learning, selective attention, working memory, and sensory-motor integration

may all depend on the stability (and thus maintenance) of neural representations, which might be

critically regulated by DA. Hence, some of the seemingly very different effects that DA excerts


in different anatomical structures may in fact be related to the same basic neural mechanism.

However, active maintenance of representations may be less important in primary sensory

processes, and this could explain why primary sensory areas are not or only weakly innervated

by DA in birds as well as in mammals.

The here proposed function of DA might also explain the peculiar finding that many DARPP-

32ir neurons show up in the Bas, although it is a primary sensory center. The Bas is the only

region of the avian telencephalon that receives direct sensory inputs without thalamic relay, and

it is involved in a forebrain circuit for the control of pecking where sensory and motor signals

have to be integrated within extremely short time spans. Thus, the Bas might be directly

involved in sensory-motor processes where sustained delay-activity plays an important role.

In conclusion, the structural and functional organization of the telencephalic dopaminergic

system in birds and mammals suggests that DA might have a common function across animal

classes and possibly within different telencephalic regions. DA seems to be particularly

involved in sensory-motor and associative processes, where it might act to stabilize and sustain

neural activity related to behavioral or motor goal states, or to linkage of temporal

discontingent stimuli.

Acknowledgements

Supported by grants of the Alfried Krupp-Stiftung (D.D. and O.G.) and by the Deutsche

Forschungsgemeinschaft through its Sonderforschungsbereich NEUROVISION (S.K. and O.G.).

Chapter 5: Distribution of DARPP-32 in pigeon telencephalon 152

The Dopaminergic Innervation of the Pigeon Telencephalon: Distribution of

DARPP-32 and Co-occurrence with Glutamate Decarboxylase and

Tyrosine Hydroxylase

Daniel Durstewitz, Sven Kröner, Hugh Hemmings JR, and Onur Güntürkün

Abstract

Dopaminergic axons arising from midbrain nuclei innervate the mammalian and avian telencephalon with

heterogeneous regional and laminar distributions. In primate, rodent, and avian species, the neuromodulator

dopamine is low or almost absent in most primary sensory areas and is most abundant in the striatal parts of the

basal ganglia. Furthermore, dopaminergic fibers are present in most limbic and associative structures.

Herein, the distribution of DARPP-32, a phosphoprotein related to the dopamine D1-receptor, in the pigeon

telencephalon was investigated by immunocytochemical techniques. Furthermore, co-occurrence of DARPP-

32-positive perikarya with tyrosine hydroxylase-positive pericellular axonal ‘baskets’ or glutamate

decarboxylase-positive neurons, as well as co-occurrence of tyrosine hydroxylase and glutamate decarboxylase

were examined. Specificity of the anti-DARPP-32 monoclonal antibody in pigeon brain was determined by

immunoblotting. The distribution of DARPP-32 shared important features with the distribution of D1-

receptors and dopaminergic fibers in the pigeon telencephalon as described previously. In particular, DARPP-

32 was highly abundant in the avian basal ganglia, where a high percentage of neurons were labeled in the

‘striatal’ parts (paleostriatum augmentatum, lobus parolfactorius), while only neuropil staining was observed in

the ‘pallidal’ portions (paleostriatum primitivum). In contrast, DARPP-32 was almost absent or present in

comparatively lower concentrations in most primary sensory areas. Secondary sensory and tertiary areas of the

neostriatum contained numbers of labeled neurons comparable to that of the basal ganglia and intermediate

levels of neuropil staining. Approximately up to one third of DARPP-32-positive neurons received a basket-

type innervation from tyrosine hydroxylase-positive fibers in the lateral and caudal neostriatum, but only about

half as many did in the medial and frontal neostriatum, and even less so in the hyperstriatum. No case of

colocalization of glutamate decarboxylase and DARPP-32 and no co-occurrence of glutamate decarboxylase-

positive neurons and tyrosine hydroxylase-basket-like structures could be detected out of more than 2000

glutamate decarboxylase-positive neurons examined, although the high DARPP-32 and high tyrosine

hydroxylase staining density hampered this analysis in the basal ganglia.

In conclusion, the pigeon dopaminergic system seems to be organized similar to that of mammals. Apparently,

in the telencephalon, dopamine has its primary function in higher level sensory, associative and motor

processes, since primary areas showed only weak or no anatomical cues of dopaminergic modulation.

Dopamine might exert its effects primarily by modulating the physiological properties of non-GABAergic and

therefore presumably excitatory units.

Neuroscience 83:763-779 (1998)


INTRODUCTION

The mesotelencephalic dopaminergic system of mammals (Berger et al., 1991; Björklund and

Lindvall, 1984; Joyce et al., 1993) and birds (Kitt and Brauth, 1986; Moons et al., 1994;

Waldmann and Güntürkün, 1993; Wynne and Güntürkün, 1995) consists of a relatively small

population of dopaminergic midbrain neurons in the ventral tegmentum and substantia nigra

(cell groups A8, A9 and A10) which project diffusely to various limbic, motor and association

areas. Primary sensory structures receive much weaker innervation (Berger et al., 1991; Berger

et al., 1988; Joyce et al., 1993; Lidow et al., 1986; Moons et al., 1994; Wynne and Güntürkün,

1995). In mammals, the most prominent telencephalic target areas of the dopaminergic

innervation are parts of the basal ganglia (caudate-putamen, n. accumbens), the prefrontal

cortex, and the olfactory, entorhinal and anterior cingulate cortex. There are considerable

differences in the distributions of dopaminergic fibers in the cortex of rodents and primates

(Berger et al., 1991). For example, in primates the cortical motor areas (supplementary motor,

premotor and motor cortex) are also densely innervated whereas in rats they are not. In addition

to these area-specific differences in dopaminergic intensity, differential patterns of laminar

distribution have also been observed (Berger et al., 1991).

The dopaminergic system plays an important role in instrumental learning and working memory

(Apicella et al., 1992; Baunez et al., 1995; Beninger, 1993; Brozoski et al., 1979; Salamone,

1992; Sawaguchi and Goldman-Rakic, 1994; Schultz et al., 1995; Shu et al., 1988; Sokolowski

and Salamone, 1994; Sokolowski et al., 1994). For example, blockade of D1-receptors in the

primate prefrontal cortex produces severe deficits in a spatial delayed response task

(Sawaguchi and Goldman-Rakic, 1994). There is evidence that dopamine is more involved in

the acquisition of instrumental responses than in their performance (Beninger, 1993).

Dopaminergic midbrain neurons are active in new situations where reward or reward

predicting stimuli occur (Ljungberg et al., 1992; Schultz and Romo, 1993; Schultz et al., 1993).

However, after acquisition of the operant, stimulus-triggered activity disappears in most

dopaminergic units. In accordance with these data, Montaron et al. (1982) showed that lesions

of the ventral tegmentum in the cat disrupt periods of explorative behavior and focal attention.

Also, the 35-45 Hz EEG rhythms normally associated with these states disappear in ventral

tegmentum-lesioned animals. Computational models of dopaminergic function suggest that

dopamine may enable fast learning in new or unexpected and biologically significant situations

by suppressing interference with previously learned associations and aiding ‘relearning’ of

connections. (Durstewitz and Güntürkün, 1996).


Although extensive data have been accumulated on the anatomy and functional relevance of the

dopaminergic system, detailed comparative information concerning the neural architecture

within which dopaminergic fibers interact with other chemically identified systems is lacking.

To gain further insight into the structural basis of dopaminergic function, we extended previous

studies on the dopaminergic innervation of pigeon telencephalon, which demonstrated

substantial similarities to mammals (Divac et al., 1985; Divac et al., 1994; Juorio and Vogt,

1967; Kitt and Brauth, 1986;Moons et al., 1994; Waldmann and Güntürkün, 1993; Wynne and

Güntürkün, 1995). Here, the distribution of DARPP-32 throughout the pigeon telencephalon

was analyzed and estimates of the percentage of DARPP-32-positive (DARPP-32+) neurons in

selected telencephalic regions are provided. Furthermore, co-occurrence of DARPP-32,

tyrosine hydroxylase (TH) and glutamate decarboxylase (GAD) was investigated. DARPP-32

is a dopamine- and cyclic AMP-regulated phosphoprotein (Mr = 32,000), which is closely

associated with cells expressing the dopamine D1-receptor (Berger et al., 1990; Hemmings et

al., 1984; Quimet et al., 1992; Qimet et al., 1984;Walaas and Greengard, 1984). The D1

receptor is the principal dopamine receptor subtype in mammalian cortex (Joyce et al., 1993)

and bird telencephalon outside the basal ganglia (Dietl and Palacios, 1988). It has been

implicated in working memory and associative learning (Beninger, 1993; Sawaguchi and

Goldman-Rakic, 1994). TH, the rate-limiting enzyme in the synthesis of catecholamines, is

present in catecholaminergic neurons and fibers. A special feature of the dopaminergic

innervation of pigeon telencephalon is the presence of ‘basket’-like structures (Veenman and

Reiner, 1994; Wynne and Güntürkün, 1995) in which dopaminergic fibers densely coil around

single somata and thereby provide a massive input to them (Metzger et al., 1996). Baskets make

possible the observation of postsynaptic targets of dopaminergic innervation at the light

microscopic level. We therefore investigated co-occurrence of DARPP-32+ neurons and TH

baskets. More interestingly from a functional perspective, we determined whether GABAergic

neurons, as assessed by GAD-immunoreactivity, receive this type of dense catecholaminergic

input. In addition, colocalization of GAD and DARPP-32 was analyzed, which would indicate

the expression of D1-receptors in GABAergic neurons.

EXPERIMENTAL PROCEDURES

Animals and tissue preparation. A total of 14 adult pigeons (Columba livia) from local stock were used.

Animals were injected with 1,000 IU heparin 15 min prior to perfusion and deeply anaesthetized with 0.3 - 0.4

ml Equithesin (Güntürkün et al., 1993) per 100 g body weight. Pigeons were then perfused intracardially with

400 ml 0.9 % saline (40°C) followed by 1,000 ml of a fixative consisting of 4 % paraformaldehyde in 0.12 M


phosphate buffer (PB; 4°C, pH 7,4). After perfusion, brains were dissected and stored for one hour in the same

fixative to which 30 % (w/v) sucrose was added, and then transferred to 30 % (w/v) sucrose in PB for 12 h at

4°C. Brains were cut in frontal slices of 35-40 µm and collected in PB containing 0.05 % (w/v) NaN3 as a

preservative. Representative sections were then processed either for a peroxidase-antiperoxidase (avidin-

biotin-conjugate, ABC, or Vector VIP) or the indirect immunofluorescence technique. Treatment of animals

conformed to the specifications of the German law for the prevention of cruelty to animals.

Immunocytochemistry: Avidin-biotin-peroxidase procedure. For the ABC technique, slices were treated

according to the following procedure: Free floating sections were incubated overnight at 4°C in anti-DARPP-

32 (C24-6a) from mouse (Hemmings and Greengard, 1986) (working dilution 1:10,000) or anti-TH from

mouse (Boehringer; working dilution 1: 200) in phosphate buffer containing 0.9 % (w/v) NaCl (PBS; pH 7,4)

and 0.3 % (v/v) Triton X-100 (Sigma). The following steps were carried out at room temperature, separated by

three washes in PBS of 10 min each. Slices were preincubated in 10 % (w/v) bovine albumin fraction V (Serva)

or sheep (Sigma) serum in PB. After washing, slices were incubated in the secondary antibody directed against

mouse from sheep (Boehringer) diluted 1:200 in PBS containing 0.3 % Triton X-100 for 1 hour. After three

washes, slices were put for 1 h in Vectastain ABC (Vector) in PBS and 0.3 % Triton X-100. Three washes in

PBS were followed by two additional washes in 0.12 M acetate buffer (pH 6). Staining was achieved by the

3,3'-diaminobenzidine-(DAB) technique with heavy-metal amplification (Adams, 1981; Shu et al., 1988) by

adding H8N2NiO8S2 (2.5 g/ 100 ml), NH4Cl and CoCl2 (both 40 mg/ 100 ml) or the VIP-(Vector) technique,

yielding a black or violet signal, respectively. For labeling with DAB, 400 mg/ 100 ml β-D-glucose were added

to the solution. After 15 min of preincubation the reaction was catalyzed with 100-200 U/ mg glucose-oxidase

(Sigma, type VII). In some cases, instead of β-D-glucose and glucose-oxidase, a solution of 0.3% H2O2 was

used to catalyze the reaction. Finally, slices were washed three more times for 5 min in 0.12 M acetate buffer

and two times in PBS. For labeling with VIP, 110 µl of reagent 1 of the VIP Substrate Kit were added to 50 ml

PBS and mixed well. Then 70 µl of reagent 2 were added and mixed well. Finally 70 µl of reagent 3 were added

and mixed well. The reaction was started by a 0.3% H2O2 solution. Washing of slices in PBS stopped the

reaction. Slices were then mounted and coverslipped.

Immunocytochemistry: Indirect immunofluorescence procedure. For indirect immunofluorescence, the

following basic procedure was used: Slices were incubated overnight at 4°C in the first primary antibody in

PBS. For DARPP-32 and TH, but not for GAD, 0.3 % (v/v) Triton X-100 was added; Triton X-100 yields

poorer staining with GAD-antibodies. As described above, slices were washed three times (per 10 min) in PBS,

preincubated in 10 % (w/v) bovine serum albumin, and then washed three more times. Slices were then

incubated for 1 h with fluorescence dye-coupled secondary antibody and the reaction was stopped by three

washes in PBS. This procedure was repeated for the second primary antibody. The primary antibodies were

diluted as follows (the respective dilutions of the secondary antibodies are given in brackets): anti-DARPP-32

from mouse 1:1,000 (1:200), anti-TH from mouse (Boehringer) 1:100 (1:50), anti-GAD from rabbit


(Chemicon) 1:500 (1:100). Fluoresceine FITC, Texas Red TRSC and Rhodamine LRSC or TRITC labeled

donkey antibodies directed against mouse or rabbit IgG (all Jackson) were used as secondary antibodies.

In some cases, the following procedure was used to amplify the DARPP-32, TH, or GAD immunosignal:

Instead of the respective fluorescence dye-coupled secondary antibody, a biotinylated secondary antibody

directed against mouse IgG from sheep (Boehringer) was used. Slices were then incubated in avidin-coupled

fluoresceine or Texas Red (avidin DCS; Vector) for 1h, followed by incubation in a biotinylated tertiary

antibody directed against avidin (anti-avidin D; Vector) for 1 h, and again fluoresceine or Texas Red avidin DCS

for 1h. All these steps were carried out in PBS with Triton X-100 omitted, and were separated by three washes

in PBS of 10 min each.

For immunofluorescence the following filters were used: Olympus BH - IB block with additional short-pass

emitter filter G520 (FITC). Olympus BH-G with additional long-pass exciter filter EO 530 (TRITC/LRSC).

Chroma exciter HQ 577/10 with dichromic beamsplitter Q 585 LP and emitter HQ 645/75 (TRSC).

In general, fluorescence methods yielded poorer staining with fewer cells labeled compared to the avidin-

biotin-peroxidase techniques.

Quantitative analysis. To estimate the percentage of DARPP-32+ neurons in different brain regions,

consecutive sections of 35 µm were stained alternately for DARPP-32 as described above or Nissl-stained

with cresyl violet. 20 anatomical structures, which are listed in the results section, were selected due to

functional considerations. For each of these structures, a center region was defined as illustrated in Fig. 3,

divided into thirds, and from each third one sample area was drawn randomly. These sample areas had a size of

108 × 87 µm2

, which was the maximum size processed by our image analysis system (analySIS, Münster) using

a 100 × 2.5 × 1.6 magnification. All DARPP-32 labeled neurons in this area were counted. By use of

landmarks, the corresponding area at the same location was selected in the succeeding Nissl-stained section,

and here also all neurons were counted. DAB sections were shrunk by a factor of approximately 1.43 compared

to Nissl preparations and cell countings were adjusted accordingly. The percentage of DARPP-32+ neurons in

a region was estimated by dividing the corrected mean DARPP-32+ count by the respective mean of Nissl-

stained neurons.

For calculating percentages of DARPP-32+ neurons located in TH-baskets, we used sections with TH+ fibers

labeled with DAB, and DARPP-32+ neurons labeled with VIP, with anti-TH used first. As both antibodies are

derived from mouse, cross-reactions might be expected. However, since TH labels only fibers in the

telencephalon, with the sole exception of a number of cell bodies in the olfactory bulb, (Wynne and Güntürkün,

1995) staining patterns of DARPP-32 and TH could easily be separated. Proceeding in steps of 500 µm from

rostral A 13.0 to caudal A 4.5, a frame of size 95 × 80 µm2

was moved across the whole latero-medial and

dorso-caudal extent of the telencephalon with a constant step size of 800 µm in vertical and horizontal

direction. All labeled neurons within the area enclosed by the frame were checked for being enwrapped by a

TH-basket. The same basic procedure was applied to the investigation of GAD/TH and GAD/DARPP-32

colocalizations. However, in these analysis, indirect immunofluorescence was used, since double labeling

against GAD and TH with ABC produced a high background such that the GAD signal could not be well

discriminated.


Control experiments. For GAD, a blocking experiment was performed. Slices were incubated in the primary

antibody together with the GAD antigen (10-3 M, Worthington) and then processed according to the ABC-

technique. This resulted in an unspecific light brown staining. The specificity of the antibody directed against

DARPP-32 was assessed by immunoblotting (see below). As controls for the fluorescence double-labeling

studies, in different trials either one of the two primary or one of the secondary antibodies was omitted. In all

of these cases, no specific staining or no signal at all could be detected for the respective target protein.

Immunoblotting. Samples of pigeon brain cerebellum and neostriatum were rapidly dissected and frozen.

Frozen brain samples were thawed and homogenized in 1% (w/v) sodium dodecyl sulfate (SDS) and 10 µg/ml

leupeptin (Chemicon). Aliquots (100 µg of total protein determined by the bicinchoninic acid method (Smith et

al., 1985) with bovine serum albumin as standard) were subjected to SDS/polyacrylamide gel electrophoresis

(SDS/PAGE) on gels cast from 13% (w/v) acrylamide and transferred to 0.2 µm nitrocellulose membrane

(Schlicher and Scheull) as described (Hemmings and Adamo, 1997). Immunoblotting with anti-bovine DARPP-

32 antibody C24-6a (Hemmings and Greengard, 1986) at a 1:500 dilution was performed as described

(Hemmings and Adamo, 1997) using prestained molecular weight markers (Amersham).

RESULTS

Immunoblotting

The specificity of the antibody to DARPP-32 was demonstrated by the selective labeling of a

single protein of Mr = 32,000 in pigeon neostriatum (Fig. 1). No immunoreactive protein was

detected in the cerebellum by immunoblotting or by immunocytochemistry.

Distribution of DARPP-32

The antibody directed against DARPP-32 stained a number of neurons to a large extent so that

fine dendritic processes with spines were clearly visible (Figs 2D-F, 4D). More often,

however, only somata with the first segments of stem dendrites were labeled (Fig. 2B-C).

Neurons of various sizes and shapes were stained, with soma diameters ranging from about 5 -

25 µm. Neurons also differed considerably in their degree of staining intensity.

Telencephalic regions differed with respect to the ratio of soma to neuropil staining. Whereas,

for example, in the paleostriatum primitivum the neuropil was densely stained and neurons

seemed to be selectively spared (Fig. 5B), in the n. basalis there was almost no neuropil

staining while a large number of somata were clearly DARPP-32+ (Fig. 4A). To acknowledge

the different contributions of neuropil and soma labeling, telencephalic subdivisions were

classified according to two different schemes with four categories each. The results are shown

in Fig. 3.


The first categorization was based on neuropil staining intensity. Structures were defined as "1"

when being virtually devoid of any labeling, as "2" for a light neuropil staining, as "3" for

considerably stronger staining, and as "4" for extremely dense staining (Fig. 2A). The second

categorization was based on the percentage of DARPP-32+ neurons. Percentage estimates were

obtained as described in Experimental Procedures for a subset of 20 areas, selected for

functional reasons.

Figure 1. Characterization of the specificity of anti-DARPP-32 monoclonal antibody C24-6a by immunoblotting. Samples ofpigeon cerebellum (left 2 lanes) and neostriatum (right 2 lanes) were dissected from the brain, homogenized in SDS, subjectedto SDS/PAGE (13% acrylamide), and electrophoretically transferred to a nitrocellulose membrane as described inExperimental Procedures. The nitrocellulose membrane was labeled with a 1:500 dilution of monoclonal antibody C24-6a,followed by 1:500 rabbit anti-mouse IgG and 125I-labeled protein A, and analyzed by autoradiography. Pigeon DARPP-32 (Mr= 32,600) is indicated by the arrow. Samples contained 100 µg of total brain protein.


Figure 2. DARPP-32 immuno-labeling in the pigeon forebrain. (A) Frontal section at level A 9.00 according to the pigeon brainatlas of Karten and Hodos (1967), illustrating the four categories of DARPP-32+ neuropil staining (see text). Paleostriatumaugmentatum (PA) and paleostriatum primitivum (PP), displaying the densest neuropil labeling, were classified as ‘4’,ectostriatum (E), showing almost no labeling, as ‘1’, neostriatum (N) as ‘3’, and hyperstriatum ventrale (HV) as ‘2’. (B, C) MostDARPP-32+ cells displayed only labeling of the soma and initial dendritic segments. (D-F) In some neurons, however, also finedendritic processes and spines were visible. Photomicrographs were taken from the n. basalis (B), neostriatum frontolaterale (C),ventromedial hippocampus (D), border of ectostriatum and ectostriatal belt (E), and medial septum (F). Scale bars represent 250µm (A), 50 µm (B, D, E), 20 µm (C), 15 µm (F).


Primary sensory forebrain areas included in our analysis were the ectostriatum of the visual

tectofugal, (Benowitz and Karten, 1976) the hyperstriatum dorsale pars lateralis of the visual

thalamofugal, (Karten et al., 1973) field L2 of the auditory, (Bonke et al., 1979) and n. basalis

of the trigeminal system (Schall et al., 1986). The intercalated nucleus of the hyperstriatum

accessorium, a further termination area of the visual thalamofugal pathway, (Güntürkün and

Remy, 1990) was not included in the analysis since percentage estimates of labeled neurons

could not be obtained reliably in this extremely thin structure. The secondary sensory forebrain

areas included in our analysis were the ectostriatal belt, which receives afferents from the

ectostriatal core, (Ritchie, 1979) the hyperstriatum accessorium as the projection zone of the

lateral hyperstriatum dorsalis, (Shimizu et al., 1995) field L1, which receives afferents from

L2, (Wild et al., 1993) and the neostriatum frontale, pars trigeminale, which is innervated by

the primary trigeminal n. basalis (Wild et al., 1985). Multimodal (tertiary) areas were the

neostriatum caudolaterale as defined by Waldmann and Güntürkün (Veenman and Reiner, 1994)

with its afferents from ectostriatal belt, (Ritchie, 1979) hyperstriatum accessorium, (Shimizu et

al., 1995) L1, (Rehkämper and Zilles, 1991) and neostriatum frontale trigeminale (Wild et al.,

1985). Three further multimodal structures, which comprise the pallium externum of Veenman

et al., (1995) were included in our analysis: the neostriatum frontolaterale as defined by

Shimizu et al (1995). with its afferents from hyperstriatum accessorium/hyperstriatum dorsale

(Shimizu et al., 1995) and ectostriatal belt; (Ritchie, 1979) the temporo-parieto-occipital area,

and the area corticoidea dorsolateralis. The latter two are known to receive visual afferents

from hyperstriatum dorsale (Shimizu et al., 1995) and ectostriatal belt, (Veenman et al., 1995)

as well as trigeminal afferents from neostriatum frontale trigeminale (Dubbeldam and Visser,

1987). The archistriatum intermedium with its descending projections to various brainstem

structures was considered as a motor output area (Zeier and Karten, 1971). Limbic structures

included in our analysis were the archistriatum ventrale and archistriatum posterior (Veenman

et al., 1995; Zeier and Karten, 1971). The lobus parolfactorius, the paleostriatum augmentatum,

and the paleostriatum primitivum were analyzed as parts of the avian basal ganglia (Karten and

Dubbeldam, 1973; Reiner et al., 1984; Veenman et al., 1995). The description of DARPP-32

distribution follows this functional classification. Figure 3 depicts the topographical

distribution of neuropil staining density and the DARPP-32+ neuron percentage classification.


Primary sensory areas. Primary sensory regions showed the lowest DARPP-32

immunoreactivity of all telencephalic areas (Figs 3, 4). The ectostriatum and field L2 appeared

almost completely blank, containing only few broadly scattered cells (Fig. 2A; Figs

4B, D). The lateral hyperstriatum dorsale had more neuropil staining and labeled neurons, but

was considerably lower in both respects compared to most secondary sensory and tertiary

areas, particularly its secondary projection zone, the hyperstriatum accessorium (Fig. 4C). The

LPOINP

PAPP

E

Ep

HVHD

HA

A 6.75

A 7.75

Hp

SLSM

HV

TPO

CDL

PA

CDL

NCNCL

AvAp

APH/Hp

A 9.25

A 10.50

A 12.50

A 5.25

Acc

PP

1

2

3

4

DARPP-32neuropil labeling

medium - high

none - very lowlow

very high

Percentage of DARPP-32+ neurons< 10 %20 - 40 %40 - 70 %> 70 %

E

PALPO

LPO

HA

NFL

BasNFT

LPO

Ep

HA

HVHD

N

Aid AivTn

PA

NCL

APH/Hp

HV

Av

CDL

L2L1

Figure 3. Distribution of DARPP-32 in pigeon telencephalon (levels according to the atlas of Karten and Hodos43). The left sideof each section shows the relative density of DARPP-32+ neuropil staining; on the right side shading codes for the percentageof DARPP-32+ neurons. Telencephalic regions, for which no percentage estimates were obtained (see text), appear blank.Frames indicate the areas from which samples for counting were actually drawn as described in Experimental Procedures.


primary sensory area of the trigeminal system, the n. basalis, was an exception. Although it

showed very light neuropil staining, it actually contained the highest percentage of DARPP-32+

neurons of all selected areas (Figs 2B, 4A).

Secondary sensory areas. The secondary sensory areas of the visual-tectofugal (ectostriatal

belt), visual-thalamofugal (hyperstriatum accessorium) and auditory (L1) system ranked clearly

higher in both neuropil labeling and the relative number of DARPP-32+ neurons when

compared to the respective primary sensory areas (Fig. 3). As mentioned above, the trigeminal

system was an exception with the neostriatum frontale trigeminale having a much denser

neuropil staining than the n. basalis, but containing a lower percentage of labeled neurons than

the latter one (Fig. 4A). In general, secondary structures displayed approximately equal relative

numbers of DARPP-32+ neurons. Only the caudally located field L1 scored lower than the

more rostral structures.

Figure 4. DARPP-32 immunoreactivity in primary and adjacent secondary sensory areas of the pigeon telencephalon. (A) Thetrigeminal n. basalis (Bas) displays a high number of DARPP-32+ neurons, although neuropil staining is very low in this area.(B) Primary auditory Field L2 shows almost no labeling in contrast to the adjacent secondary areas. (C) Hyperstriatum dorsale(HD), from which the lateral portion receives thalamic input and is part of the primary thalamofugal visual system, shows clearlyless soma and neuropil labeling than its secondary projection zone, the hyperstriatum accessorium (HA). (D) Single DARPP-32-positive cell in the otherwise unstained ectostriatum (E), surrounded by the ectostriatal belt and the neostriatum. Scale barsrepresent 500 µm (A-C), 200 µm (D).


Tertiary areas. The tertiary structures neostriatum frontolaterale, area temporo-parieto-

occipitalis, corticoidea dorsolateralis, neostriatum caudolaterale, and the caudal neostriatum

were stained homogeneously at medium to high densities. Based on the intensity of neuropil

labeling as well as on the percentage of DARPP-32+ neurons, these areas could not be

differentiated from the secondary sensory structures (Fig. 3). Among the tertiary regions, the

more rostral areas of the neostriatum (neostriatum frontolaterale, area temporo-parieto-

occipitalis, corticoidea dorsolateralis) showed more DARPP-32+ neurons than the caudal

components (neostriatum caudale, neostriatum caudolaterale).

Motor and limbic structures. According to neuropil staining intensity, structures of the avian

basal ganglia (paleostriatum augmentatum, paleostriatum primitivum, lobus parolfactorius)

were most heavily labeled and could not be distinguished from each other (Figs 2A, 3, 5A).

Lobus parolfactorius and paleostriatum augmentatum also contained a high percentage of

DARPP-32+ neurons, whereas in the paleostriatum primitivum only a very small number of

cells was labeled (Fig. 5B). On the basis of cell counting alone, no difference between lobus

Figure 5. DARPP-32 immunoreactivity of pigeon motor and limbic structures. (A) Avian basal ganglia at level A 8.50. (B)Border between the paleostriatum augmentatum (PA), in which both neuropil and cell bodies were extensively labeled, and thepaleostriatum primitivum (PP), which showed only neuropil labeling. (C) Archistriatum at level A 7.50. While the ventralarchistriatum (Av) contains many DARPP-32+ neurons, the dorsal archistriatum (Ad) displays only neuropil labeling, with theexception of a dorsomedial rim where also high perikaryal labeling could be observed (arrowhead). Unstained fibers of thetractus occipito-mesencephalicus travel through the archistriatum ventrale. (D) Septum, showing two DARPP-32+ cellpopulations in the lateral (SL) and ventromedial (SM) part. Scale bars represent 1000 µm (A), 250 µm (B), 500 µm (C, D).


parolfactorius/paleostriatum augmentatum and secondary or tertiary areas could be observed

(Fig. 3). The n. accumbens and bed nucleus of the stria terminalis, according to the definition of

Veenman et al., (1995) also showed a considerable number of labeled cells.

Neuropil labeling in the archistriatum was homogeneous, comparable to the neostriatal regions

and considerably lower than in the basal ganglia (Figs 3, 5C). The ventral portions of the

‘motor’ as well as of the ‘limbic-visceral’ archistriatum demonstrated a high percentage of

DARPP-32+ neurons (Fig. 3). In contrast, in the dorsal parts of the archistriatum mainly

neuropil was labeled, with exception of a small dorsomedial rim of labeled perikarya (Fig.

5C).

Other limbic structures in the telencephalon of the pigeon include the hippocampal system and

the septum. Neuropil labeling in most regions of the hippocampus was comparable to that of the

archistriatum and higher sensory and multimodal areas of the neostriatum. Also, a high number

of DARPP-32+ neurons was observed in the ventromedial area and at the lateral border, which

probably correspond to the V-shaped region and area 7 of Krebs et al. (1991) and Erichsen et

al. (1991) (Fig. 2D). The medial border, however, which is part of the ventromedial and

inferior dorsomedial region as defined by these authors, was found to be completely blank. In

the septum, DARPP-32 immunoreactivity strongly paralleled the distribution of TH+ fibers.

Only two restricted portions of the lateral and ventromedial septum, the latter probably

corresponding to the n. commissuralis of Baylé et al., (1974) contained numbers of DARPP32+

cells, while the whole septum demonstrated very low levels of DARPP-32+ neuropil staining

(Fig. 5D).

Co-occurrence of DARPP-32 and tyrosine hydroxylase baskets

‘Baskets’, which consist of fibers densely coiling around single perikarya and initial dendritic

segments, represent a striking feature of catecholaminergic innervation in avian telencephalon

(Fig. 6A, B). As illustrated in Fig. 6C-E and Fig. 7A, a number of DARPP-32+ neurons receive

input from TH+ baskets. To obtain estimates of the percentage of DARPP-32+ neurons located

in TH-baskets, a total of 3692 neurons were examined. Analysis of several regions were

pooled and classified into one of three categories. Because of the extremely dense meshwork of

TH+ fibers and the very high DARPP-32 abundance in regions of the avian basal ganglia

(paleostriatum augmentatum, paleostriatum primitivum, lobus parolfactorius), reliable countings

were impossible here and these areas were not included in the analysis.


The highest percentage of co-occurrence was seen in the multimodal lateral and caudal

neostriatum (area temporo-parieto-occipitalis / corticoidea dorsolateralis / neostriatum caudale

/ neostriatum caudolaterale: 15 - 30 %). This value was lower in the intermediate medial and

frontal neostriatum (including ectostriatal belt / neostriatum frontolaterale / neostriatum frontale

trigeminale: 5 - 15 %). The lowest percentage of co-occurrences was observed in the

hyperstriatal and hippocampal regions (hyperstriatum accessorium / area parahippocampalis /

hippocampus: < 5 %), although these areas were rich in DARPP-32+ neurons. Also, many TH-

baskets containing no DARPP-32-labeled neurons were observed throughout the telencephalon.

Forebrain regions differed with respect to the number of fibers constituting single baskets.

Baskets in the caudal neostriatum consisted of a large number of axonal loops surrounding

Figure 6. TH+ fibers and DARPP-32+ neurons in the pigeon caudal neostriatum. (A, B) TH+ baskets consist of axons denselycoiling around single perikarya and initial dendritic segments, so that neurons virtually become visible by the TH+ signal alone. (C-E) VIP-stained DARPP-32+ neurons enwrapped by DAB-stained TH+ fibers. Scale bars represent 30 µm (A), 20 µm (B, D, E),50 µm (C).


perikarya and initial dendrites, while TH+ fibers in the frontal neostriatum coiled only a few

times around single somata (Waldmann and Güntürkün, 1993).

Co-occurrence of glutamate decarboxylase and tyrosine hydroxylase baskets

To examine whether GAD+ neurons receive TH-basket-type input, 861 GAD+ neurons and 636

TH-baskets throughout the telencephalon were inspected. As demonstrated in Fig. 7B, no co-

occurrence was found. However, in the lobus parolfactorius, paleostriatum augmentatum and

paleostriatum primitivum reliable counting could not be obtained due to the dense net of

catecholaminergic fibers in these regions.

Co-localization of glutamate decarboxylase and DARPP-32

To investigate the possible colocalization of DARPP-32 and GAD antigen, a total of 939

DARPP-32+ and 1149 GAD+ neurons throughout the telencephalon were examined. No

colocalization was observed. Fig. 7C shows examples of neurons in the neostriatum, labeled

for GAD and DARPP-32, respectively. However, in the basal ganglia the possibility of

colocalization cannot be excluded with certainty, since the high DARPP-32 staining density in

these regions hampered the analysis.

DISCUSSION

The present study describes the distribution of DARPP-32 in the avian telencephalon and

provides a classification of telencephalic regions in terms of their neuropil staining intensity

and DARPP-32+ neuron numbers. Additionally, colocalization of GAD-and DARPP-32-

labeled cells and their co-occurrence with TH-labeled axonal baskets were investigated.

Distribution of DARPP-32

The strongest DARPP-32 immunoreactivity was observed in parts of the avian basal ganglia,

namely the paleostriatum augmentatum, lobus parolfactorius and paleostriatum primitivum,

which are regarded as equivalent to the mammalian caudate-putamen and globus pallidus,

respectively (Veenman et al., 1995). In the paleostriatum augmentatum/lobus parolfactorius

both fibers and somata were densely labeled. In these areas, TH+ and DA+ fibers have been

shown to contact predominantely shafts and necks of dendritic spines, (Karle et al., 1996)

suggesting possible sites for dopaminergic receptors. The strong staining in the paleostriatum

primitivum on the other hand resulted almost exclusively from neuropil labeling (see Fig. 5B).


This clear difference in somatic labeling between the paleostriatum augmentatum/lobus

parolfactorius and paleostriatum primitivum corresponds to results obtained with other markers

of the dopaminergic system in mammals and birds: the abundance of dopaminergic

Figure 7. Fluorescence double-labeling against (A) DARPP-32 and TH, (B) GAD and TH, and (C) DARPP-32 and GAD. Thephotomicrographs of each pair were taken from the same site in caudal neostriatum of the pigeon telencephalon. Two of threeDARPP-32+ neurons (A1) are surrounded by TH+ baskets (A2). GAD+ neurons (B1) do not receive input from TH+ baskets(B2). GAD+ neurons (C2) do not show DARPP-32 immunoreactivity (C1), although GAD+ and DARPP-32+ neurons mayoccur in close vicinity to each other. Arrows mark corresponding positions in each pair. In (C), open and filled arrows indicatepositions of DARPP-32+ and some GAD+ cells, respectively. Scale bars represent 30 µm (A, B), 40 µm (C).


fibers as well as of D1 and D2 receptors is highest in the caudate-putamen and the

paleostriatum augmentatum/lobus parolfactorius of the mammalian and avian basal ganglia, but

is relatively low in the globus pallidus and the paleostriatum primitivum, respectively (Ball et

al., 1995; Berger et al., 1991; Brock et al., 1992; Joyce et al., 1993; Moons et al., 1994;

Richfield et al., 1987; Schnabel and Braun, 1996; Stewart et al., 1996; Wynne and Güntürkün,

1995; Yung et al., 1995). The dense neuropil labeling in the paleostriatum primitivum may stem

from descending D1-positive axons, as in the globus pallidus of rats where DARPP-32- and

D1- but not D2 receptor-positive axons have been described surrounding unlabeled perikarya

and dendrites (Quimet and Greengard, 1990; Yung et al., 1995). These axons probably belong

to the descending striatonigral and striatoentopeduncular tract (Yung et al., 1995).

Besides the basal ganglia, several other structures of the avian motor and limbic system showed

medium-to-high DARPP-32 immunoreactivity, namely the archistriatum and the area

parahippocampalis/hippocampus. According to Zeier and Karten, (1971; 1973) the anterior two

thirds of the archistriatum (archistriatum pars dorsalis and archistriatum intermedium)

correspond to the mammalian motor cortex, whereas the archistriatum posterior and pars

ventralis have been compared to the mammalian amygdala. In the present study, neurons in the

ventral parts of the motor as well as of the limbic archistriatum were strongly DARPP-32

labeled while in the dorsal parts only neuropil was labeled. This finding stands in contrast to

the pattern described for dopaminergic fibers in the archistriatum by Wynne and Güntürkün,

(1995) who found DA baskets, presumably innervating somata, especially in those archistriatal

portions where no or few DARPP-32 cells were labeled. D2 receptors cannot account for the

observed mismatch, since these are known to have extremely low densities outside the basal

ganglia (Dietl and Palacios, 1988; Richfield et al., 1987; Stewart et al., 1996). However,

differences regarding the dopaminergic innervation and the distribution of dopamine receptors

also exist in motor output areas of various mammalian species. In primates, the motor cortex

receives an exceptionally strong dopaminergic innervation (Berger et al., 1991; Björklund and

Lindvall, 1984; Wiiliams and Glodman-Rakic, 1993) but displays only a sparse distribution of

DARPP-32 labeled cells and contains only medium densities of D1 receptors (Berger et al.,

1990; Cortés et al., 1989; Lidow et al., 1989; Lidow et al., 1991; Quimet et al., 1992; Richfield

et al., 1989).

DARPP-32 soma and neuropil labeling revealed a general pattern of low or no staining in

primary and much stronger labeling in secondary and multimodal association areas (see Fig. 3).

This is in good agreement with studies on the distribution of dopaminergic fibers and D1


receptors in birds (Dietl and Palacios, 1988; Moons et al., 1994; Wynne and Güntürkün, 1995).

It is also in accordance with data from rodents and primates, where the density of D1-receptors

and dopaminergic fibers is lowest in primary sensory but higher in secondary and association

cortices (Berger et al., 1991; Berger et al., 1988; Joyce et al., 1993; Lewis et al., 1986).

The n. basalis is a notable exception from the rule outlined above. Although showing weak

neuropil labeling, it contains a high percentage of DARPP-32+ neurons (see Fig. 4A). This

observation contrasts with results for other primary sensory regions and could be linked to

other peculiarities of the pigeon's trigeminal system. The n. basalis receives, among other

afferents, direct input from the n. sensorius principalis nervi trigemini and projects through

other forebrain areas to rhombencephalic structures related to pecking (Schall, 1987; Schall et

al., 1986; Wallenberg, 1903; Wild et al., 1985). It is thus the only avian forebrain structure with

sensory afferents bypassing the thalamus. The projection via the n. basalis, neostriatum frontale

trigeminale and archistriatum constitutes a trigeminal circuit for the control of pecking(Schall,

1987; Schall, 1989; Zeigler, 1976). Within this circuit, sensory signals and motor outputs

probably have to be integrated within extremely short time spans. Therefore, dopaminergic

modulation (see general discussion) might start earlier in this system than in other sensorimotor

circuits.

The presence of DARPP-32+ cells within the n. basalis contrasts with a number of previous

anatomical studies. According to Dietl and Palacios (1988) the levels of D1 receptors in the n.

basalis are negligibly low. Kusunoki (1969) could also not reveal monoaminoxidase activity in

the n. basalis, and Wynne and Güntürkün (1995) were unable to detect dopaminergic fibers in

this structure. In contrast to these anatomical results, behavioral data make it likely that

dopaminergic modulation exists within this nucleus. Apomorphine-induced pecking fits are

accompanied by an increase of glucose uptake in the n. basalis (Delius, 1985) and intracranial

apomorphine injections into this structure induce stereotypical pecking bouts (Lindenblatt and

Delius, 1988). Therefore, n. basalis was thought to be a central element of apomorphine

triggered pecking stereotypies and to possess dopamine receptors (Delius, 1985). This last

assumption was strengthened by the finding of Wynne and Delius (1996) that n. basalis-lesions

significantly reduce pecking fits after apomorphine injection. The present finding of a prominent

group of DARPP-32+ cells in the n. basalis does not fit easily with the chemoanatomic results

gathered up to now but may provide an anatomical clue to behavioral data.

According to biochemical, hodological and behavioral results, the neostriatum caudolaterale

has been suggested as an avian equivalent of the prefrontal cortex (Divac and Mogensen, 1985;


Divac et al., 1985; Divac et al., 1994; Güntürkün, 1997; Mogensen and Divac, 1982; Mogensen

and Divac, 1993; Reiner, 1986; Waldmann and Güntürkün, 1993; Wynne and Güntürkün, 1995).

The dopaminergic innervation of the neostriatum caudolaterale is among the strongest in the

pigeon telencephalon, (Divac et al., 1994; Waldmann and Güntürkün, 1993; Wynne and

Güntürkün, 1995) just as it is in the mammalian prefrontal cortex (Berger et al., 1988; Joyce et

al., 1993; Lewis et al., 1986; Williams and Goldman-Rakic, 1993). The percentage of DARPP-

32+ neurons in the neostriatum caudolaterale, as revealed in the present study, is lower than in

the secondary sensory areas ectostriatal belt or neostriatum frontale trigeminale, while this

pattern is reversed for the innervation by dopaminergic fibers in these regions (Wynne and

Güntürkün, 1995). Again, this finding parallels data from mammals: The dorsolateral prefrontal

cortex of adult monkeys also displays a low number of DARPP-32+ neurons despite its dense

dopaminergic innervation (Berger et al., 1990).

DARPP-32 as a marker for D1-receptors

In rodents and primates the distribution of DARPP-32 has been found to parallel the

distribution of cells possessing D1 receptors (Berger et al., 1990; Hemmings and Greengard,

1986; Quimet et al., 1992; Quimet et al., 1984; Walaas and Greengard, 1984). Receptors of the

D1-family are positively coupled to adenylyl cyclase activity and can thus stimulate cAMP

synthesis (Kebabian and Calne, 1979). This in turn leads to phosphorylation of DARPP-32 via

cAMP-dependent protein kinase, resulting in inhibition of phosphatase-1 (Hemmings et al.,

1984; Hemmings et al., 1987). DARPP-32 acting as an intracellular “third messenger” is thus

involved in mediating the D1 receptor dependent effects of dopamine.

Key features of the distribution of dopaminergic fibers and D1-receptors in avian, such as

sparing of primary sensory areas or dense innervation of structures of the basal ganglia, (Ball et

al., 1995; Dietl and Palacios, 1988; Moons et al., 1994; Richfield et al., 1987; Stewart et al.,

1996; Waldmann and Güntürkün, 1993; Wynne and Güntürkün, 1995) could also be observed in

the distribution of DARPP-32. There are some exceptions, however Richfield et al. (1987) and

Stewart et al. (1996) demonstrated higher D1-specific binding in hyperstriatum ventrale than in

neostriatum, while this pattern is reversed with DARPP-32. Mismatches in the distributions of

DARPP-32 and D1 receptors have also been noted in mammals, e.g. in primate motor cortex

(Berger et al., 1990). In general, however, these distributions overlap to a high degree, and

DARPP-32 thus seems to represent a useful marker for dopaminoceptive cells expressing D1

receptors within the mammalian as well as within the avian telencephalon.


It has been noted that glia may also contain DARPP-32 (Berger et al., 1990; Quimet et al.,

1984). Thus, the cell counting analysis presented here may contain overestimations. However,

most DARPP-32+ cells observed were larger than glia cells (Ø > 6 µm) or expressed clear

neuronal properties like dendritic elements (see Fig. 2B-F). Also, in Nissl-counterstained

preparations, very few DARPP-32+ cells were observed resembling glia cells. The estimation

error due to glia staining therefore was probably very low in our preparations.

Co-localization of glutamate decarboxylase with DARPP-32 and co-occurrence with

tyrosine hydroxylase

GAD is relatively homogeneously distributed throughout the pigeon telencephalon (Domenici et

al., 1988; Veenman and Reiner, 1994). Veenman and Reiner (1994) report that the distributions

of GAD+ and GABA+ cells overlap almost completely, and they estimate the fraction of

GABAergic neurons to be about 10-12 % in the neo- and hyperstriatum, showing only slight

regional differences. According to the present results, GAD+ neurons do not express DARPP-

32 and do not receive input from TH-baskets. Therefore, GABAergic cells, at least in the avian

‘neocortex’, do not seem to be the target of the dense type of dopaminergic innervation and

probably do not express D1 receptors. Yet the possibility remains that these cells express

receptors of the D2 receptor family. In mammalian neocortex, there is indeed

electrophysiological evidence that GABAergic neurons may be modulated by dopamine via D2

receptors (Pirot et al., 1992; Rétaux et al., 1991). However, receptor autoradiography studies

in birds have shown that, with exception of the basal ganglia, the abundance of D2 receptors in

the telencephalon is very low (Dietl and Palacios, 1988; Richfield et al., 1987; Schnabel and

Braun, 1996; Stewart et al., 1996). Thus, GABAergic cells in the pigeon telencephalon might

not, or at least not to an important extent, be innervated by dopamine. Therefore presumably

excitatory neurons are the main target of dopaminergic afferents. This interpretation would also

be consistent with data from mammalian frontal cortex, where dopaminergic axons mainly, but

not exclusively, (Smiley and Goldman-Rakic, 1993) synapse on presumably glutamatergic

pyramidal cells, (Berger et al., 1991; Goldman-Rakic et al., 1989; Smiley and Goldman-Rakic,

1993) which also constitute the major cell population expressing DARPP-32 (Berger et al.,

1990).

In the pigeon basal ganglia also no colocalization of GAD and DARPP-32 was detected. In

contrast, GABAergic neurons in mammalian caudate-putamen express D1-receptors and very

likely also DARPP-32 (Joyce et al., 1993; Quimet et al., 1992;Yung et al., 1995). This


discrepancy may be explained in terms of another difference between avian and mammalian

basal ganglia. The avian basal ganglia contain only 5-10 % GAD+ cells (Veenman and Reiner,

1994) as opposed to the vast majority of medium spiny neurons in mammalian basal ganglia,

which are supposed to utilitize GABA (Groves et al., 1995). Yet, for several reasons we

cannot exclude the possibility that some GAD/DARPP-32 colocalizations went unnoticed in the

present study. First, as noted in the results section, colocalizations were difficult to analyze in

the avian basal ganglia due to the high DARPP-32 neuropil staining. Second, it is conceivable

that some GABAergic cells were indeed colocalized with DARPP-32 but were not labeled by

GAD immunohistochemistry, since Veenman and Reiner (1994) report a subpopulation of small

cells which show GABA but not GAD immunoreactivity. Third, the methods used in the present

study may have been not sensitive enough to detect all GAD+ cells. However, although

Veenman and Reiner (1994) note that pre-treatment of pigeons with colchicine enhances GAD

perikaryal labeling, the percentage of GAD+ cells in the paleostriatum as estimated by these

authors (5-10 %) seems to be in the same range as in the present study.

In conclusion, the present results make it likely that dopamine in the pigeon's forebrain exert its

effects primarily by changing physiological parameters of non-GABAergic and thus presumably

excitatory units, although this conclusion has to be treated with caution with respect to the avian

basal ganglia.

Co-occurrence of DARPP-32 and tyrosine hydoxylase

In the present study, a large number of DARPP-32+ neurons localized in TH-baskets have been

observed, although the majority of labeled perikarya did not receive this type of dense

catecholaminergic innervation. The highest percentage of DARPP-32/TH co-occurences was

found in the lateral and caudal parts of the neostriatum and thus in parts of the neostriatal

association areas. In contrast to primary and secondary sensory areas, these association areas

are also known to have a direct link to the avian basal ganglia (Veenman et al., 1995). The

finding that there are many ‘empty’ TH-baskets could be due to noradrenergic fibers, and TH-

baskets of non-dopaminergic origin have indeed been demonstrated (Wynne and Güntürkün,

1995). Alternatively, neurons expressing other than D1 dopamine receptor subtypes could also

be the target of the basket-type input.


General discussion

In birds, dopaminergic innervation of the telencephalon with clear regional differences arises

from the midbrain nuclei tegmenti pedunculopontinus pars lateralis and area ventralis

tegmentalis, comparable to the projections from the mammalian substantia nigra pars compacta

and area ventralis tegmentalis, respectively (Kitt and Brauth, 1986; Waldmann and Güntürkün,

1993). Our study shows that dopaminergic innervation of the pigeon telencephalon is

comparable to that of mammals in three important respects. First, DARPP-32, and therefore

presumably the D1 receptor, is most abundant in parts of the avian basal ganglia and is present

in medium to high concentrations in most other telencephalic secondary sensory, associative,

limbic and motor regions. Second, the primary auditory and visual areas show no or only weak

DARPP-32 labeling. Third, GAD+ neurons are not colocalized with DARPP-32 and do not

receive TH-basket input. Thus, dopamine presumably mainly innervates excitatory forebrain

units in pigeons, a situation at least partly comparable to mammalian neocortex (Berger et al.,

1991; Gaspar et al., 1995; Goldman-Rakic et al., 1989; Smiley and Goldman-Rakic, 1993).

What are the possible functional implications of the present findings and the fact that a similar

organization of the dopaminergic innervation exists in mammals? The sparing of most primary

sensory regions by markers of dopaminergic innervation and the medium to high levels in most

other telencephalic structures imply that dopamine has its most prominent role in higher level

sensory and motor processes. Based on the findings by Schultz and coworkers, (Apicella et al.,

1992; Ljungberg et al., 1992; Schultz and Romo, 1990; Schultz et al., 1993; Schultz et al., 1995)

who showed that dopaminergic midbrain neurons are activated in new and unexpected

situations, and computational studies, (Durstewitz and Güntürkün, 1996) one may infer that

dopamine allows for rapid strengthening of couplings between higher sensory representations,

motor outputs, and the reward value of the behavior. By stabilizing such new configurations, the

dopaminergic system may enable the organism to develop predictions of behavioral

consequences (Montague et al., 1996; Schultz et al., 1995). This type of fast and highly flexible

learning may not be that important in primary sensory regions for two reasons. First, most

structural modifications in these areas occur during pre- and early postnatal development. In

contrast, in adult primary sensory structures there may be need for only slight incremental

modifications of, e.g., receptive field properties. In line with this assumption, considerable

amounts of DARPP-32 are present in the visual cortex of infant monkeys but not in that of adults

(Berger et al., 1990). Second, only higher elaborated sensory representations may participate in

those associations reinforced by the dopaminergic system because only these representations


may stand in a unique relationship to relevant aspects of the situation. In contrast, there may be

too many interpretational possibilities for activity patterns in primary sensory areas. Therefore,

these patterns possibly cannot become directly coupled to certain motor functions but must first

be resolved by higher processes.

In cortex, dopamine has been shown to inhibit spontaneous and evoked activity in vivo,

(Bernardi et al., 1982; Ferron et al., 1984; Godbout et al., 1991; Sesack and Bunney, 1989;

Verma and Moghaddam, 1996) to reduce excitatory synaptic transmission, (Law-Tho et al.,

1994; Pralong and Jones, 1993) and to shift the probability of long term modifications in favor

of long term depression as opposed to long term potentiation (Chen et al., 1995; Law-Tho et al.,

1995). Computational modeling suggests that dopamine, by these physiological mechanisms,

suppresses the recall of previously learned associations during learning and thereby enables the

fast formation of new couplings (Durstewitz and Güntürkün, 1996). This hypothesis implies that

dopaminergic fibers mainly innervate excitatory neurons and that these cells are the biggest

subpopulation expressing dopaminergic receptors. This prediction is consistent with the present

findings that GABAergic units do not express DARPP-32 and are not targets of a strong

dopaminergic input.

A remaining question concerns why only a certain fraction of DARPP-32+ neurons receives the

dense basket-type catecholaminergic input. It is conceivable that these neurons serve special

functions in the innervated neural networks. One hypothesis derives from the proposed role of

dopamine in fast associative learning (Durstewitz and Güntürkün, 1996). Fast learning

presumably requires the respective neurons or cell populations to be synchronously active for

some time to allow for long term synaptic modifications (Singer, 1993). The dopaminergic

system may be involved in keeping new patterns active by inducing oscillatory processes

(Montaron et al., 1982). In the avian telencephalon, the strong basket-type dopaminergic input

may terminate on those excitatory cells which can acquire pacemaker-like characteristics in

learning situations. Thereby, dopamine in the avian brain may excert some control over the long

term formation of associations between certain situations and profitable behaviors. Of course,

these notions are highly speculative and clearly more evidence is needed. However, our

assumptions may guide further research on the possible functions of dopamine in learning in

different species.

Chapter 6: Actions of dopamine on neurons in the chick NCL 175

Effects of dopamine on the excitability of neurons in the chick NCL

INTRODUCTION

In the previous chapters the importance of the NCL for sensorimotor integration and „frontal“

excecutive functions has been outlined repeatedly. Most of the evidence that supports the

involvement of the NCL in working memory (Mogensen and Divac, 1982, 1993, Gagliardo et

al., 1996, 1997, Güntürkün, 1997; Kalt et al., 1999), reversal learning (Hartmann and

Güntürkün, 1998), response inhibition (Güntürkün, 1997, Aldavert-Vera et al., 1999), and

spatial orientation (Gagliardo and Divac, 1993) stems from lesion studies. More recent studies

have begun to unravel the cellular mechanisms that underlie these functions of the NCL.

Electrophysiological single-unit recordings in awake pigeons which performed a delayed

Go/NoGo task, have elucidated the contribution of neurons in the NCL to response inhibition

and working memory (Kalt et al., 1999). In this study single neurons showed an increase in their

firing frequency relative to baseline conditions during the delay period of the task. Neuronal

activity changes also signalled the onset of the stimulus or the delivery of the reward. Some

neurons furthermore seemed to code the behavioral significance of the stimulus, such that the

neuronal activity, either an increase or a decrease in spike frequency, was statistically different

between Go and NoGo trials (Kalt et al., 1999).

In the prefrontal cortex (PFC) of mammals, task-related electrical activity is modulated by

dopamine (DA), mainly via the D1 receptor (Sawaguchi et al., 1988, 1990a,b; Williams and

Goldman-Rakic, 1995). Dopaminergic midbrain neurons are activated at the onset of working

memory tasks (Schultz et al., 1993) and DA levels in the PFC increase during delay-task

performance (Watanabe et al., 1997). Blockade of the dopaminergic input to the PFC or of

dopaminergic D1 receptors in the PFC disrupt delay-task performance (Brozoski et al., 1979;

Simon et al., 1980; Sawaguchi and Goldman-Rakic, 1994; Seamans et al., 1998). Similarly, in a

behavioural study in the pigeon, blockade of D1 receptors in the NCL by a competitive

antagonist (SCH-23390), strongly impaired performance in a working memory task (Güntürkün

and Durstewitz, in press).

One concept regarding the functions DA might play during working memory, is to increase the

signal-to-noise ratio and to stabilize neural representations (Sawaguchi et al., 1990a,b;

Durstewitz et al., 1999, 2000). This stabilizing effect on the network could enable sustained


task-related activity, even in the presence of interfering input (Durstewitz et al., 1999), which in

mammals appears to be a unique feature of the PFC (Miller et al., 1996).

In this chapter I will present preliminary data on the effects of DA on the firing behaviour of

chick NCL neurons recorded in-vitro, using both whole-cell and perforated patch-clamp

techniques. The electrophysiological findings are complemented with morphological data from

double-labeling experiments, in which neurons that were intracellularly filled during whole-

cell recordings were immunohistochemically stained and colocalized with labeling of tyrosine-

hydroxylase, as a marker of the dopaminergic innervation. It can be assumed that the basket-like

innervation provides a means by which the dopaminergic system could selectively modulate a

specific class of neurons, and/or that the strong somatic innervation renders these neurons

particularly susceptible for the effects of DA. The results show that DA can increase the

excitability of NCL neurons and profoundly alter their firing pattern, which in turn might sustain

neuronal delay activity during delay-task performance. However, the responses of neurons,

which were anatomically shown to receive basket-like dopaminergic innervation, to DA

application did not differ from those that received only „en-passant“ innervation, thus

suggesting that the different modes of dopaminergic innervation exert no qualitatively different

influence over their target neurons.


Preparation of slices and electrophysiological recording: For preparation of brain slices a total of 32 chicks

(5 - 12 days post-hatch, median 7 days) were decapitated and the brains were rapidly transferred to iced

extracellular solution. Coronal slices (350 µm thick) of the caudal telencephalon were cut on a vibratome.

Slices were incubated in extracellular solution consisting of (in mM): 119 NaCl, 2.5 KCl, 1 NaH2PO4, 26.2

NaHCO3 ,10 D-glucose, 4.5 CaCl2, and 1.3 MgCl2, saturated with 95% O2 /5% CO2, pH 7.3. Slices were

allowed to recover for at least 1 hour, before being transferred to the recording chamber. All recordings were

made at room temperature (21 - 23°C) in a submerged slice chamber perfused with extracellular solution.

Whole-cell and perforated-patch recordings were made from neurons in the caudolateral subventricular region

of the chick forebrain, using the blind-patch technique (Blanton et al., 1989). Electrodes (6 – 8 MΩ resistance)

were filled with an intracellular solution consisting of (in mM): 135 K-gluconate, 20 KCl, 2 MgCl2 , 10 N-(2-

hydroxyethyl)piperazine-N'-(2-ethanesulfonic acid) (HEPES), 0.1 ethylene glycolbis (β-aminoethyl ether)-

N,N,N´,N´,-tertraacetic acid (EGTA), 4 Na2-ATP, and 0.5 Na2 -GTP, and adjusted to pH 7.3 with KOH. In some

recordings 3 - 3.5 mg/ml of the dipotassium salt Lucifer Yellow (LY, Sigma, Deisenhofen, Germany) were

added to the intracellular solution and neurons were filled by diffusion. Perforated-patch recordings were made

by adding nystatin to the pipette solution (final concentration was 500 µg/ml, 0.3 % DMSO). The membrane of

a cell-attached patch is permeabilized by nystatin in the patch pipette, thus providing electrical continuity


between the pipette and the cytoplasm of the cell without the loss or alteration of cytoplasmic compounds

necessary for the maintenance of the effects under study.

Recordings were made using a HEKA (Lambrecht, Pfalz, Germany) EPC-7 patch-clamp amplifier, filtered at 3

kHz, and sampled at 10 kHz. The voltage drop across the pipette could not be bridge-balanced. Sampling was

done with a Digidata 1200 interface using pClamp 6.0 software (Axon Instruments, Foster City, CA), and data

were stored on computer for off-line analysis with the CLAMPFIT module of pClamp.

Experimentally produced changes in membrane potential (Vm) during current-clamp recordings were

compensated with a bias current to maintain Vm at baseline levels.

The apparent input resistance (Ri) was monitored continuously by applying brief (50 ms) hyperpolarizing

current pulses before each depolarizing current step. The amplitude of these hyperpolarizing current pulses (10

- 50 pA) was adjusted for each cell to be in the linear range of the current–voltage relationship, as estimated

from the cells response to a series of depolarizing and hyperpolarizing current injections.

Experimental paradigm: To study cell excitability, action potentials were evoked by 1.1 sec-long depolarizing

current injections with various amplitudes (100 - 800 pA). The amplitude of the current used for statistical

comparisons of drug effects was adjusted to be in the mid-range between the current that depolarized the

neuron just above spike-threshold and the close-to-maximal stimulation. In the majority of cells the time

course of drug action and wash-out of the effect was measured continuously by applying 5 current steps (at 0.2

Hz) every 30 or 45 seconds. After baseline measurements of action-potential number and Ri, dopamine (1 - 50

µM) was bath applied for approximately 1 minute. Two gravity-fed perfusion lines were used to change

solutions. Both control and drug ACSF were continuously oxygenated throughout experiments.

The dopamine D1-receptor agonist SKF-38393, a D1 antagonist (SCH-23390), and the D2-receptor agonist

(quinpirole) were prepared as stock solutions and stored frozen until diluted to the final concentration. In a

number of experiments DA stock solutions were prepared in a similar way and ascorbate (3 mM) was added to

the solution to reduce oxidation. In all other experiments DA solutions were prepared at the desired final

concentration immediately before application. Dopaminergic agonists and antagonists were obtained from

Research Biochemicals (Natick, USA). Only one neuron experiencing drug treatment was used from each brain

slice. In several instances different agonists were applied successively to one cell to test its differential

response to various pharmaka. Usually, either a D1 or D2 receptor agonist (SKF-38393, or quinpirol) were

followed by a single application of DA. In these cases only the response to the agonist that was first applied

was used for statistical analysis. In all cases, only cells with a resting potential of at least -60 mV and stable

baseline responses were given drug treatments.

Analysis and statistics. For each cell, the number of action potentials evoked by depolarizing current steps was

quantified by averaging the number of spikes produced in 5 consecutive traces before, at the end of, and

approximately 5 minutes after drug application. The input resistance (Ri) served as a measure for the stability

of a given recording. Thus, cells that showed a non-reversible change of input resistance larger than 15 % were

excluded from analysis to control for possible unspecific changes in recording conditions. Furthermore, for

cells that showed a change in the number of evoked spikes during a given drug treatment, only those that

exhibited a reversible effect (a washout of at least 50% of the effect) were used for analysis. The effects of DA


or its agonists on the number of evoked action potentials were expressed as the percentage change relative to

baseline conditions to account for the variation in the absolute number of spikes which could be evoked in

neurons of different cell types (c.f. chapter 3).

Comparisons of drug effects across concentrations relative to control recordings were made using univariate

analysises of variance (ANOVA). Subsequent tests for statistical significance of effects were done using

Student´s t test for unpaired samples. For multiple comparisons the alpha-niveau was Bonferroni-adjusted.

„Within-group“ comparisons used a Student’s t test for paired samples where noted.

Histological procedures

Following recording, slices were fixed in 4% paraformaldehyde and 0.2% glutaraldehyde in 0.12 M phosphate

buffer (4°C, pH 7.4) for about 15 hours. They were then transferred to a solution of 30% sucrose in phosphate

buffer containing 0.9% NaCl (PBS; pH 7.4) and 0.01% NaN3 as a preservative. Slices were resectioned at 70

µm on a freezing microtome and collected in PBS. For double-labeling immunohistochemistry of LY and

tyrosine-hydroxylase (TH), endogeneous peroxidases were first blocked by preincubating slices in 0.5% H2O2

for 30 minutes. Slices were then washed in PBS and floating sections were incubated overnight at 4°C in a

monoclonar anti-TH antibody from mouse (Boehringer; working dilution 1: 200) in PBS containing 0.3 %

Triton-X (Sigma, Deisenhofen). The following steps were carried out at room temperature, separated by three

washes in PBS of 10 min each. After washing, slices were incubated for 2 hours in the secondary antibody

directed against mouse from horse (Vector Labs, Burlingame, CA), diluted at 1:200 in PBS containing 0.3 %

Triton-X. After washing, slices were incubated for 2 hours in the avidin-biotin complex (ABC Elite, Vector

Labs; 1:100 in PBS with 0.3% Triton X-100). Washes in PBS were followed by additional washes in 0.12 M

acetate buffer (pH 6). Staining was achieved by the 3,3'-diaminobenzidine (DAB) technique with heavy-metal

amplification (modified from Adams, 1981) by adding H8N2NiO8S2 (2.5 g/ 100 ml), NH4Cl and CoCl2 (both

40 mg/ 100 ml), which yields a black signal. After 20 minutes of preincubation the reaction was catalyzed using

0.3% H2O2. The reaction was stopped by rinsing the tissue in acetate buffer and PBS. The same staining

procedure was then repeated for labeling of LY. Therefore, slices were incubated overnight at 4°C in a

monoclonar biotinylated anti-LY antibody from rabbit (Molecular Probes, Leiden, The Netherlands; 1:200

working dilution) in PBS containing 0.3% Triton X-100 (Sigma). After washing, slices were incubated for 2

hours in the avidin-biotin complex (1:100 in Triton-PBS). Washes in PBS were followed by additional washes

in acetate buffer. Labeling of LY was again achieved by incubating slices in DAB, but without addification of

heavy-metal compounds, thus resulting in a brown staining product. The reaction was stopped by rinsing the

tissue in acetate buffer and PBS. Slices were then mounted, dehydrated and coverslipped. Camera lucida

reconstructions of filled cells were drawn using an Olympus BH-2 with a drawing tube at x 125 magnification.

RESULTS

A total of 84 neurons that met the criteria specified in Materials and Methods were analyzed.

All neurons could be assigned unequivocally to one of the 4 cell classes described in chapter 3


by their responses to depolarizing and hyperpolarizing current pulses. Of these 84 neurons, 37

were characterized as type I neurons by their burst firing pattern of action potentials, and 33

showed a regular or phasic-tonic firing, characteristic of type II neurons. Five cells displayed

the high-frequency firing typical of type III interneurons, and the remaining 9 cells were

classified as type IV neurons by their burst firing of action-potentials and the absence of

prominent inward rectification.

10 µM DA washbaseline

200 ms

20 mVA


200 ms

20 mVB

200 ms

20 mV


C

200 ms

20 mV


D

Type I

Type II

Type III

Type IV

-30pA

180pA

-30pA

180pA

-100pA

550pA

-100pA

600pA

Figure 1: The effects of dopamine application in the 4 major celltypes of the chick NCL. Dopamine at 10 µM could largelyincrease the number of spikes evoked by a constant current injection in the three spiny cell types (types I, II and IV, shown inA, B, and D, respectively), but not in type III interneurons (C). The enhancement of neuronal excitability by dopamine appearedto be independent from changes in the passive properties of the cells.


Effect of dopamine application on neuron excitability

The effect of DA on the excitability of NCL neurons was studied by measuring changes in spike

number in response to 1.1-sec-long depolarizing current injections. The effect of DA did not

vary with the recording technique used, therefore, for comparison of dopaminergic effects

across concentrations data obtained during whole-cell and perforated-patch recordings were

pooled. In control recordings (n = 9), neurons that were recorded under the same conditions

and over comparably long periods as experimentally-treated cells (at least 20 minutes), showed

only a slight decrease of 4.5 ± 4% in the number of evoked spikes. This change from baseline

was neither reversible nor significant (df = 8; t = 1.05; p > 0.05; Student’s t test for paired

samples).

Application of DA reversibly increased the number of spikes in 32 of 43 neurons tested (74%).

However, the effect of DA differed with the cell-type studied. Dopamine increased the number

of evoked action potentials in neurons of all spiny cell-classes (i.e. types I, II, and IV), but not

in type III interneurons (figure 1). Thus, application of dopamine increased the number of spikes

in 13 of 16 type I neurons, in 14 of 17 type II neurons, and in 4 of 5 type IV neurons. In contrast,

in 4 of 5 type III neurons tested, application of DA resulted in no change in the number of

evoked spikes, or in a small decrease, similar to that observed in controls. In one neuron there

was a small increase of 11% (9 vs. 8 spikes). To account for these differences in responsivity,

type III neurons were excluded from further analysis of drug effects.

(n=7)(n=5)

**(n=17)

*(n=5)

(n=4)

(n=9)

control 1 µM 5 µM 10 µM 20 µM 50 µM

concentration of dopamine

% c

hang

e in

num

ber o

f spi

kes

140

120

100

80

60

40

20

0

-20

Figure 2: The effect of dopamine at various concentrations. In contrast to control recordings, in which a small decrease in thenumber of evoked spikes occured during prolonged recording, application of dopamine consistently increased the firing of spinyNCL neurons. The change in the number of action potentials was significantly different (p < 0.01) from changes in controls at 5µM and 10 µM dopamine. At higher concentrations the effect of dopamine did not saturate, but rather diminished, resulting in aninverted U-shape of the dose-response curve.


The increase in spike number induced by DA across the range of concentrations studied (i.e. 1,

5, 10, 20, and 50 µM; N = 38) was significantly different from changes observed in controls (df

= 46; F = 3.7; p < 0.01; ANOVA). When grouped according to the concentration of DA applied,

significant increases in the spike number compared to control recordings occured only at

concentrations between 5 and 10 µM DA, indicating a dose dependence of the effect of DA

(figure 2). In 5 µM DA the mean number of spikes increased by 74.5 ± 9.7 % (df = 12; t = -8.9;

p < 0.01, α-adjusted; Student´s t test for unpaired samples) and in 10 µM DA by 112.5 ± 22.6

% (df = 24; t = -5.1; p < 0.01, α-adjusted; Student´s t test for unpaired samples). The effect of

DA did not saturate, but rather decreased with higher concentrations, resulting in almost no

change in the number of spikes at 50 µM DA (n = 4; c.f. figure 2). This finding is in agreement

with an inverted U-curve that has been used to describe the range of optimal dopaminergic

activation (Williams and Goldman-Rakic, 1995; Murphy et al., 1996; Zahrt et al., 1997). The

effect of DA usually developed within 3 minutes and peaked about 5 - 10 minutes after the end

of bath-application (figure 3). The effect was long lasting, sometimes persisting up to about an

hour, depending on the concentration of DA that was applied.

0 2 40

2

4

6

8

10

12

14

16

18

20

22

24

6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

10 µM DA 10 µM DA

Time (min)

Num

ber o

f spi

kes

Figure 3: Example for the time course of dopamine-induced enhancement of cell excitability in a type IV neuron. The dopamineeffect usually developed within a few minutes after dopamine application in the bath. In most neurons the strongest effect onthe number of evoked action potentials was seen approximately 5 - 10 minutes after cessation of dopamine application, and itcould last for more than an hour. The short application-time of dopamine (∼1 minute) served to minimize rapid D1 receptordesenzitation.


To examine possible mechanisms by which DA might regulate the excitability of NCL neurons,

the near-threshold firing behaviour of 20 neurons that had received DA applications at doses

shown to effectively increase the number of spikes (i.e. 5 and 10 µM) was compared with

changes in controls. The effects of DA on the latency and the threshold of the first spike were

studied, and in addition the change in the duration of the first inter-spike-interval was analyzed.

The results show that application of DA led to a negligible reduction in the latency of the first

spike (mean changes -6.5 ± 3.8 % for DA, and +4.3 ± 6.1 % for controls; df = 27; t = 1.50; p >

0.05; Student´s t test for unpaired samples). In current-clamp recordings the threshold for

generation of an action-potential varied considerably over the duration of the experiment (a

change of +1.69 ± 1.4 mV in controls); application of 5 or 10 µM DA, however, resulted in a

small shift of the spike-threshold towards hyperpolarized potentials (-1.14 ± 0.72 mV), and this

change showed a statistical trend when compared to controls (df = 27; t = 1.98; p = 0.058;

Student´s t test for unpaired samples). Dopamine had its strongest impact on the duration of the

first inter-spike-interval, which usually occured at more depolarized potentials than the first

spikes. Application of DA significantly shortened the first inter-spike-interval when compared

to changes in control recordings (mean changes -23.5 ± 4.4 % for DA, and +32.6 ± 14.4 % for

controls; df = 27; t = 4.89; p < 0.01; Student´s t test for unpaired samples).

Application of DA had little effect on the passive membrane properties of NCL neurons. There

was only a negligible overall reduction in input resistance (-2.66 ± 1.4 %) after DA application

compared to controls (df = 46; F = 0.36; p > 0.05; ANOVA).

In addition to its effects on cortical pyramidal neurons in mammals, DA has been shown to

modulate the cell excitability of GABAergic interneurons (Penit-Soria et al., 1987; Retaux et

al., 1991; Pirot et al., 1992; Zhou and Hablitz, 1999). Because spontaneous synaptic input can

strongly modulate the passive properties (i.e. input resistance) of neurons (Paré et al., 1998), I

tested whether DA might increase the excitability of NCL neurons by decreasing spontaneous

inhibitory synaptic transmission. Therefore, in 6 principal neurons the GABAA receptor blocker

picrotoxin (100 µM) was added to the extracellular solution before the application of 10 µM

DA. In the presence of picrotoxin, DA still produced a significant increase in the number of

spikes of 82.6 ± 26.5 % (df = 13; t = -3.24; p < 0.05; Student´s t test for unpaired samples).

This result indicates that the effect of DA on evoked spikes is not mediated indirectly by a

decrease in spontaneous inhibitory synaptic transmission.


Pharmacology of dopaminergic enhancement of excitability

To determine which DA receptor subtype might mediate the observed changes in spike number,

the effects of selective agonists of either the D1 or D2 family of DA receptors were studied. In

the same line of evidence the effect of DA was challenged by a specific antagonist to

dopaminergic D1 receptors (figure 4).

Application of the D1 receptor agonist SKF-38393 mimicked the effects of DA in that it

significantly increased the number of spikes, both at 10 and 25 µM (df = 23; F = 5.75; p < 0.01;

ANOVA; figure 4A, D). The mean increase in the relative number of action potentials was 79.3

± 19.1 % compared to control recordings. At 10 µM (n = 7), SKF-38393 led to an increase in

the number of spikes of 98.9 ± 33.9 % (df = 14; t = -3.02; p < 0.05, α-adjusted; Student´s t test

for unpaired samples), while at 25 µM (n = 8) the increase was 67.4 ± 24.3 % (df = 15; t = -

3.04; p < 0.05, α-adjusted; Student´s t test for unpaired samples).

Conversely, when 10 µM DA was applied approximately 5 min after bath application of the D1

receptor antagonist SCH-23390 (45 µM), the change in the number of action potentials was not

significantly different from control recordings. In the presence of SCH-23390, DA (n = 5) had

only a small (mean increase 16.0 ± 21.3%), non-significant effect on the number of evoked

spikes (df = 12; t = -0.94; p > 0.05). Bath application of SCH-23390 for 5 - 10 minutes alone,

following 5 minutes of baseline measurements, had no effect on the number of evoked spikes (df

= 4; t = 0.87; p > 0.05; Student´s t test for paired samples; figure 4B, D).

The D2 receptor agonist quinpirole (n = 6) had no effect on the excitability of NCL neurons.

Bath application of 25 or 50 µM quinpirol resulted only in a small, non-significant decrease in

the number of action-potentials by 1.53 ± 1.5 % (df = 13; t = -0.57; p > 0.05; Student´s t test for

unpaired samples), similar to that observed in controls (figure 4C, D). In 3 of the 6 neurons 10

µM DA was applied subsequently to the establishment of the quinpirole effect (at least 20

minutes after application of quinpirole). In all 3 cells DA application led to an increase in the

number of spikes (mean 74.8 ± 41%), showing that these cells were not generally unresponsive

to dopaminergic transmission (figure 4 C). This result supports the assumption drawn from the

anatomical data reviewed in chapter 4, that D2 receptor mediated mechanisms have only

relatively little, if any, influence on neurons in the caudal forebrain of birds. Taken together,

these data suggest that DA modulates cell excitability in the NCL via activation of D1 type

receptors by a mechanism that appears to be independent of input resistance.


(n=9)

10 µM 25 µM

SKF 38393

% c

hang

e in

num

ber o

f spi

kes

140

120

100

80

60

40

20

0

-20 quinpiroleSCH 23390(+ 10 µM DA)

*(n=8)

*(n=7)

(n=6)

(n=5)

control

baseline

20 mV

200 ms

25 µM SKF 38393

wash

baseline

20 mV

200 ms

45 µM SCH 23390+ 10 µM DA

wash

20 mV

200 ms

baseline

50 µM quinpirole

wash

1 µM DA

A B C

D

400pA 600pA 550pA

-100pA-100pA -50pA

Figure 4: Pharmacology of the enhancement of cell excitability. Activation of the D1, but not the D2 receptor, mimicked theeffects of DA in that it led to a significant increase in evoked action potentials. A: Example of a regular spiking neuron in whichthe selctive D1 agonist SKF-38393 was bath applied. B: Dopamine D1 receptor blockade by application of the competitive D1receptor antagonist SCH-23390 before and during application of dopamine results in no change in the number of spikes in thisneuron, and only in a small non-significant increase across all neurons tested (c.f. D). C: Application of the selective D2receptor agonist quinpirole has no effect on cell excitability in this type I neuron, and in all neurons tested (c.f. D). Application ofDA in the same neuron, however, could largely increase the number of spikes in a burst, demonstrating that this neuron was notgenerally irresponsive to dopaminergic stimulation. D: Overview of the effects of different DA receptor agonists andantagonists on cell excitability compared to control recordings.


Cell type specifity

One aim of the present experiments was to determine whether the response to DA in cells that

receive the basket-type dopaminergic innervation differ from those that receive only „en-

passant“ innervation of their dendrites. Type I and type II neurons are the most plausible

candidates to receive basket-type innervation, as baskets preferentially innervate neurons with

large soma sizes (c.f. chapter 4) and the anatomical data presented in chapter 5 also had shown

that GABAergic interneurons do not receive dopaminergic basket input.

In addition, spikes are unitary events and I had expressed the effects of DA as the percentage

change occuring between baseline and drug condition. Thus, although the spike number was

averaged across 5 consecutive sweeps, this procedure might have led to a significantly larger

statistical increase in type I neurons, as they have a low basal firing rate and might therefore

„profit“ more from small variations in the number of spikes. However, a comparison of the

effect of 5 - 10 µM DA on type I and type II neurons, respectively, showed no significant effect

of cell type. Dopamine increased the mean number of spikes in type I cells (n = 10) by 120 ±

20.7 %, and in type II neurons (n = 11) by 73.5 ± 23.1 % (df = 19; t = 1.77; p > 0.05; Student´s

t test for unpaired samples). This result shows that DA appears to modulate the large principal

cell types in the NCL in the same manner.

Morphology

It was attempted to identify the class of neurons that receives basket-type dopaminergic input by

colocalizing Lucifer Yellow-intracellularly filled neurons with tyrosine-hydroxylase containing

fibers by means of double-label immunohistochemistry.

In 15 slices in which neurons were filled with Lucifer Yellow, TH-labeling was sufficiently

good enough to examine the catecholaminergic innervation of the soma in detail. Of these 15

neurons 12 showed only few TH immunopositive fibers in the vicinity of their soma or initial

dendrites, making no, or seemingly only „random“ contact. The remaining 3 neurons received

catecholaminergic input from TH-labeled fibers showing numerous varicose-like swellings in

close apposition to the soma and the initial segments of dendrites, resembling the basket-type

innervation (figure 5). All of these 3 neurons were characterized as type I neurons by their burst

firing pattern of action potentials. In each of these neurons application of DA (n = 2) or SKF-

38393 (n = 1) resulted in a marked increase in the number of spikes. This finding shows that

burst-firing neurons in the NCL appear to be a main target of the dense basket-like somatic

innervation, however, due to the small sample size, it would


Figure 5: Anatomy of the dopaminergic innervation of electrophysiologically characterized neurons that were shown to beresponsive to dopamine application. Neurons were intracellularly filled with Lucifer Yellow (LY) during recording, and double-labeled immunohistochemically for LY and tyrosine-hydroxylase (TH) as a marker of dopaminergic fibers. A: Camera lucidareconstruction of the basket-type innervation of a type I neuron by TH-positive fibers (red). For reason of clarity only theproximal innervation by dopaminergic fibers is shown. The axon of the cell is shown in light green. B, C: Photomicrographs ofthe cell shown in A. D: Another example of a neuron that received numerous TH-positive varicosities in close apposition to thesoma and initial dendrites (arrows). E: Photomicrograph of a cell that did not receive basket-like innervation. Although thenumber of labeled fibers is generally lower than in C (due to differences in staining quality), several basket-like structures can stillbe seen (arrows; c.f. also F). Furthermore, there is a large number of varicose-like swellings in close apposition to the shafts andspines of the cells dendrites. F: Detail of the TH-labeling shown in E. The arrow points to a basket-like structure that coilsarround the soma of an unlabeled neuron. Scale bars are 25 µM in A, B, D, F, and 50 µM in C, E.


be premature to conclude that they constitute the only cell class receiving basket-innervation.

Moreover, the observed changes in the spike-pattern following DA application in neurons

innervated by basket-forming dopaminergic fibers did not differ qualitatively nor quantitatively

from cells that received diffuse „en-passant“ innervation to their dendrites and somata. Thus, it

appears that the different modes of dopaminergic innervation exert no qualitatively different

influence over their target neurons.

DISCUSSION:

The data presented here, suggest that DA modulates cell excitability in the NCL via activation

of D1-type receptors. The enhancement of membrane excitability appears to be independent of

changes in input resistance, and an indirect effect on GABAergic transmission. The latter was

inferred from direct recordings of putative GABAergic fast-spiking interneurons, and by the

absence of an effect of the Cl- channel blocker picrotoxin on the DA-induced increase in

projection neurons.

The effect of DA appeared to be dose-dependent, being maximally in a narrow range at

concentrations of 5 - 10 µM DA in the sample of neurons recorded here. This finding is in

agreement with numerous studies which have demonstrated that an optimal degree of activation

of the dopaminergic system is necessary for the active retention of information, or the planning

of motor sequences in working memory (Sawaguchi and Goldman-Rakic, 1991; Williams and

Goldman-Rakic, 1995; Murphy et al., 1996; Sawaguchi, 1997; Zahrt et al., 1997; Arnsten and

Goldman-Rakic, 1998).

Similarly to the present results in birds, in mammals an enhancement of cell excitability in

vitro, or an increase in task-related activity in vivo, have been shown to depend on activation

of D1-like receptors (Sawaguchi et al., 1988, 1990,b; Sawaguchi and Goldman-Rakic, 1991,

1994; Williams and Goldman-Rakic, 1995; Yang and Seamans, 1996; Shi et al., 1997;

Sawaguchi, 1997; Kita et al., 1999). It should be noted, however, that several other studies

have observed an inhibitory action of DA on pyramidal cells (Geijo-Barrientos and Pastore,

1995; Gulledge and Jaffe, 1998; Zhou and Hablitz, 1999; Geijo-Barrientos, 2000), and a

possible involvement of D2 receptors in this depressant effect has been reported (Gulledge and

Jaffe, 1998).

One possible explanation for these conflicting results is a differential, and possibly also time-

dependent activation of D1- and D2-type receptors by DA, which might result in exactly


opposite effects (Zheng et al., 1999). Thus, activation of D2-like receptors may require higher

concentrations of DA than those needed to activate D1-like receptors, as DA has a higher

affinity for D1-like receptors than for D2-like receptors when measured at the high affinity state

of the receptors (Seeman and Van-Thol, 1993). Furthermore, in the neocortex there are much

more D1 -like receptors than D2-like receptors (e.g. Lidow et al., 1991; Joyce et al., 1993), a

situation also seen in the dorsal forebrain of birds (Schnabel and Braun, 1996; see chapter 4).

In addition, the depressant effects of DA may be time-dependent and/or related to receptor

desensitization. Thus, studies in PFC neurons that utilized long (5 min) DA or D2-like agonist

applications, noted a decrease of input resistance and a concurrent reduction in the number of

evoked action-potentials (Geijo-Barrientos and Pastore, 1995; Gulledge and Jaffe, 1998; Zhou

and Hablitz, 1999). In contrast, the short duration application time (about 1 min.) used in the

experiments presented here, served to minimize rapid D1 receptor desensitization (Jarvie et al.

1993; Ng et al., 1995). Thus, for a given depolarizing input, time-dependence (desensitization)

or concentration-dependence of the DA effect might in principle account for the different results

obtained in various studies. In this view, the D2-like receptor may induce an „early“ transient

suppression of neuronal excitability, which in-vitro becomes only evident after desensitization

of the D1 receptor (Geijo-Barrientos and Pastore, 1995; Gulledge and Jaffe, 1998), whereas

D1/5 receptor activation may induce a „late“, prolonged (tens of minutes) enhancement of

neuronal excitability in response to depolarizing inputs (Yang and Seamans, 1996; Shi et al.,

1997; present study). However, the cellular basis for a possible D2 receptor mediated effect in

the cortex is not known. More importantly, in the present experiments no depressant action of

DA, or an involvement of D2 receptors in the mediation of the dopaminergic effects was found.

The ionic mechanisms by which DA exerts its effects in the frontal cortex following stimulation

of the D1 receptor are much better studied. By reviewing some of these findings I hope to

reveal some parallels and differences regarding the possible effects of DA in mammalian

cortical neurons and cells of the NCL. However, voltage-clamp studies into the effect of DA on

isolated currents in NCL neurons are necessary to directly study the ionic mechanisms by which

DA appears to modulate cell excitability.

In the cortex DA, via its action at the D1 receptor, appears to shift the threshold of the persistent

Na+ current (INaP) towards more hyperpolarized potentials (Yang and Seamans 1996; Gorelova

and Yang, 1997; but see Geijo-Barrientos and Pastore, 1995; Geijo-Barrientos, 2000). This

current is responsible for inward rectification in the depolarized voltage range and serves

important functions in setting the firing threshold for repetitive spike firing (Schwindt, 1992)

and signal amplification (Stuart and Sakmann, 1995). The existence of such a current in neurons


of the bird´s caudal forebrain was shown in the zebrafinch HVC (Kubota and Saito, 1991), and

indicated by a TTX-sensitive inward rectification in NCL neurons (chapter 3). The tendency of

DA to lower the spike threshold and to prolong repetitive firing in NCL neurons, might reflect

an action of DA on the persistent Na+ current.

In contrast, the observed reduction in the duration of the first inter-spike-interval in the present

study might result from an action of DA on K+ currents, e.g. the Ca++ activated K+ current

termed IAHP, which is involved in spike-frequency adaptation (Madison and Nicoll, 1982;

Storm, 1990), or the slowly inactivating K+ current (IKS) that is responsible for membrane

outward rectification in the depolarized voltage range (Schwindt et al. 1988; Kitai and

Surmeier, 1993; Yang and Seamans, 1996). The latter will functionally counteract sustained

membrane depolarization and suppress repetitive spike firing. A strong outwardly rectifying

conductance was also seen in the spiny projection neurons of the NCL (see chapter 3). Thus in

the NCL, DA might shorten inter-spike-intervals and reduce accomodation by inactivating an

outwardly rectifying K+ conductance, as has been shown in the PFC and striatum (Yang and

Seamans, 1996; Nisenbaum et al., 1998).

The concurrent actions of DA on both, INaP and IKS, have been proposed to reduce spike treshold

and to contribute to the increased firing rate observed in vivo (Sawaguchi et al., 1988; 1990a,b)

and in vitro (Cepeda et al., 1992; Yang and Seamans, 1996; Shi et al., 1997; Yang et al., 1999).

In the present study, the facilitatory effects of DA could be seen most clearly at membrane

potentials considerably above threshold, i.e. in response to large depolarizing current pulses.

This might indicate a dopaminergic modulation of ionic currents that are maximally active at

depolarized potentials, such as high-voltage activated (HVA) Ca++ channels of the L, P, and N

type (Sayer et al., 1990; Seamans et al., 1997; but see Durstewitz et al., 2000). High-threshold

Ca++ channels are activated by suprathreshold synaptic inputs that elicit spike firing in the soma

and which then backpropagates into the dendritic arbor (Stuart and Sakmann, 1994; Spruston et

al., 1995). Dendritic high-voltage-activated Ca++ currents are likely to participate in multiple

functions that are important for burst firing mechanisms, sustaining repetitive firing, and

enduring forms of changes in synaptic plasticity (Yang et al., 1999). The existence of Ca++

currents that are capable of initiating slow Ca++ spikes has also been shown in neurons of the

NCL (c.f. chapter 3), but the voltage-dependence of these currents was not studied.

In neurons of the PFC and the striatum DA has been shown to activate D1/5 receptors on

pyramidal dendrites to directly attenuate the amplitude of high-voltage-activated Ca++ spikes, or

to augment the duration of a dendritic HVA Ca++ plateau, respectively (Surmeier et al., 1995;

Yang and Seamans, 1996; Hernández-López et al., 1997; Yang et al., 1998). At moderate


concentrations, DA application results in a weak suppression of N-type Ca++ channels, that

cluster in the distal dendrites (Westenbroek et al., 1992). This action of DA may serve to

„sharpen” incoming depolarizing synaptic signals (from layer I–II inputs) arriving at the distal

dendrite, such that only sufficiently strong depolarizing inputs are propagated to the soma. Large

suppression of N-type Ca++ channels, however, may „uncouple” the distal dendrites, preventing

distal signals from reaching the soma (Yang and Seamans, 1996; Yang et al., 1999). This latter

action may help to explain in part why dopamine has often been attributed as having an

“inhibitory” or “suppressive” effect on spontaneous firing or activities evoked by specific

inputs in the PFC1. At the same time DA augments the duration of a Ca++ plateau, which is

mediated by L-type Ca++ channels that cluster in the soma and proximal dendrites of pyramidal

neurons (Westenbroek et al., 1990; Yang et al., 1998). Functionally, activation of L-type Ca++

channels may enhance the amplitude and duration of subthreshold excitatory synaptic inputs

(Seamans et al., 1997), and the dopaminergic modulation of L-type Ca++ currents has been

shown to increase the excitability of striatal neurons to NMDA application and depolarizing

current injections (Hernández-López et al. 1997; Cepeda et al. 1998). Through its action on L-

type Ca++ channels, DA will thus complement the increased cell excitability caused by the

enhancement of INaP, eventually leading cortical neurons to sustained firing (Yang et al., 1999).

Both, L-type Ca++ channels (Westenbroek et al., 1990; Seamans et al., 1997), and the persistent

Na+ conductance (Westenbroek et al., 1989; Stuart and Sakmann, 1995) are found

predominantly in the soma and basal dendrites of cortical neurons, thus enabling a differential

regulation of distal vs. proximal inputs by DA, as originally proposed by Yang and Seamans

(1996). While distal signals may be inhibited through the actions of DA on N- and P-type Ca++

channels, proximal inputs may be facilitated by dopamine´s enhancement of INaP and L-type

currents.

In the mammalian PFC, the network-effects of DA also appear to involve a modulation of

GABAergic interneurons. Dopamine directly depolarizes interneurons in the PFC and enhances

spontaneous and evoked IPSPs recorded in pyramidal neurons (Penit-Soria et al., 1987;

Godbout et al., 1991; Rétaux et al., 1991; Pirot et al., 1992; Zhou and Hablitz, 1999; Durstewitz

et al., 2000; but see Law-Tho et al., 1994). Accordingly, the suppressive action of DA on

spontaneous firing of PFC neurons in vivo is blocked by previous application of a GABA

antagonist (Pirot et al., 1992), suggesting that DA acts through interneurons in the PFC to reduce

spontaneous activity of pyramidal neurons. In the NCL, however, a significant contribution of

1 This effect of DA might also in part explain the downward slope of the dose-response curve at highconcentrations of DA (inverted U-shape).


GABAergic interneurons is highly unlikely, as in the present study DA did not modulate the

firing-pattern of presumed GABAergic fast-spiking interneurons, and its actions on spiny

principal neurons were unaffected by depressing spontaneous inhibitory transmission.

It is important to note that in vivo the effect of DA within the PFC is determined by the activity

level of PFC neurons and the specific inputs that are driving this activity. Although DA inhibits

spontaneous activity within the PFC of the anesthetized rat (Ferron et al., 1984; Mantz et al.,

1988; Pirot et al., 1992), it enhances task-related single-unit activity more than background

activity in the behaving primate (Sawaguchi et al., 1988, 1990a). DA (but not VTA stimulation)

also enhances responses evoked by hippocampal stimulation in the anesthetized rat and use-

dependent changes in synaptic efficacy in the hippocampal-PFC pathway that may be associated

with learning (Doyère et al., 1993; Jay et al., 1995). Similarly, in vitro the effects of DA are

voltage-dependent, favouring excitatory actions (e.g. via INaP, HVA-Ca++, or NMDA currents),

when the membrane already is in a depolarized state (Cepeda et al., 1998; Durstewitz et al.,

2000). Thus, DA may enhance task- or learning-related activity relative to spontaneous or

background activity within the PFC, and very likely also in the NCL. In the context of a delay-

task, the DA-induced increase in action-potential firing might augment the reliability of

transmission and thus provide the stability needed to hold online goal-related information in a

recurrent network.

The cellular basis by which DA in the NCL could achieve such a selective enhancement of

task-related inputs or reverbarating activity in a local network, remains unclear. One working

model assumed that basket- and non-basket-type innervation are dichotomic modes of

dopaminergic innervation which control different postsynaptic targets, or which strongly differ

in the degree of modulation of cell excitability. From the present results, burst firing neurons of

type I appear to be a main target of the dopaminergic basket-type innervation. Cells in this class

display a high threshold for action-potential generation and may thus require strong spatially or

temporarily summated inputs to fire. On the other hand, their burst of action-potentials results in

a faithfull transduction onto their target cells (Snider et al., 1998). An increase in the phasic

response of type I neurons will effectively drive more postsynaptic cells and „bind“ them in a

network. Enhanced firing of bursting neurons by DA could therefore establish a „pacemaker“-

activity in a network.

However, neurons from the other cell types in the NCL were modulated in the same way by

DA, and there was no obvious difference in the size of the response between cells receiving

basket and non-basket innervation. In addition, due to the small sample size of neurons that


could morphologically be shown to receive basket-type innervation, it should not be excluded

that members of other cell-types also receive basket-type dopaminergic innervation. In any

way, the observed enhancement of firing in regular-spiking neurons of the type II, will also

strongly support sustained firing in a reverberating network, and thus functionally might enable

active retention of information in the NCL.


General discussion

The experiments presented in this thesis provide an attempt to delineate an associative region in

the forebrain of birds, which has been shown to be involved in executive functions, such as

working memory and response control. The input- and output relations of the neostriatum

caudolaterale (NCL) with telencephalic and thalamic structures were studied with pathway

tracing techniques (chapters 1 and 2), and the direct sensory projection from the dorsal

thalamus was tested in-vivo (chapter 2). The principal celltypes of the NCL were described

physiologically and anatomically in-vitro (chapter 3), and the results give hints about their

possible roles in a proposed canonical circuit. Furthermore, the relationship of the

dopaminergic innervation with immunohistochemically or electrophysiologically identified

elements of this circuit was analyzed at the light-microscopic level (chapters 5 and 6). Finally,

the effect of dopamine or its receptor-specific agonists on the excitability of NCL neurons were

studied in-vitro (chapter 6).

Elements of a proposed circuit involved in working memory and executive functions

Fuster (1989) has proposed that the temporal organization of behaviour is at the heart of

prefrontal cortex functions. The cognitive operations that underly this „memory for action“ are

essentially targeted by the logic of delay tasks: „If earlier that, then now this“. This logic

requires a dynamic process of internal transferring of information across time and of cross-

temporal matching, the contents of which must be constantly changing as behaviour evolves,

especially in situations in which the outcome of the behaviour is yet not fully predictable, e.g.

when learning a new task (Fuster, 1989).

The PFC seems essential for three cognitive functions necessary to mediate contingencies

across time, (i) a temporally retrospective function of short-term memory (working memory),

(ii) a temporally prospective function of anticipatory set or preparation (planning and

foresight), and (iii) a protective function of interference control (inhibition of false response

tendencies).

Behavioural studies have shown that the NCL serves at least two of these „prefrontal“

functions. Lesions of the NCL show behavioural impairments in a variety of working memory

tasks (Mogensen and Divac, 1982, 1993, Gagliardo et al., 1996, 1997, Güntürkün, 1997).


Similarly, the role of the NCL in interference control has been tested with Go/NoGo tasks and

reversal-learning (Hartmann and Güntürkün, 1998; Güntürkün, 1997, Kalt et al., 1999).

The prospective coding functions of the NCL have not been studied yet, but Kalt and coworkers

(1999) observed neurons that increased their firing towards the end of the delay, which, in

accordance to the findings in primates, has been interpreted as „reward expectancy“ or a

preparatory motor function of these cells. Therefore, it is highly likely that the anticipation of

events and the preparation for them, also takes part in the NCL.

As briefly outlined in chapter 1, the temporal organization of behaviour requires to bridge

temporally separate elements of action and sensorium. The anatomical data presented in

chapters 1 - 3 strongly suggest that the NCL may serve such a synthetic function in the

perception-action cycle. Similar to the PFC of mammals, the NCL receives converging afferents

from various telencephalic and thalamic areas that relay input from the major sensory

modalities of somesthesis, vision, and audition (chapters 1 and 2). At least the

intratelencephalic projections are reciprocal which might allow the NCL to control its own

input. The pattern of subtelencephalic projections is not known and need to be studied with

more sensitive anterograde tracers. In addition, the NCL is reciprocally connected with the

avian homologue of the amygdala (chapter 1), a connection which is also characteristic for the

PFC (Condé et al., 1995; Reep et al., 1996; Sesack, et al., 1989; Barbas and DeOlmos, 1990),

and projects to most parts of the somatic and limbic striatum, as well as the sensorimotor

archistriatum (chapters 1 and 3). Via the projections through the basal ganglia and the

archistriatum, the NCL does not only exert indirect influence over motor structures in the

brainstem (Wild, 1993; chapter 1), but may also control its own dopaminergic innervation from

the substantia nigra and AVT.

In the context of working memory tasks it is generally assumed that items from long-term

memory relevant to the task at hand are retrieved from stores outside the PFC. Functional

imaging studies in humans, and electrophysiological recordings in animals performing delay

tasks indeed show time- and domain-dependent activation within distributed networks related

to the PFC (Cohen et al., 1997, Belger et al., 1998; Chafee and Goldman-Rakic, 1998).

It is reasonable to believe that such a gross anatomical arrangement of different sensory-

specific networks also exists between the NCL and other associational or sencondary sensory

forebrain areas in the avian brain. This is inferred from the sheer necessity to save memory

„storage“, and the fact that both, at the areal level (chapter 1), and at the single cell level


(chapter 3) a large number of cells were seen to project back to sensory areas in the anterior

forebrain. The reciprocal connection of the NCL with virtually all of its sources of afferents

suggest the existence of multiple distributed networks within the avian forebrain. One such

functional network probably inolves the connection between the NCL and the parasensory area

in the neostriatum intermedium (NIM, or MNH in chicks). This area sends a massive projection

to large parts of the NCL and in turn receives considerable feedback from the NCL (chapter 1).

Both forebrain areas receive a distinct innervation from the polymodal DLP (chapter 2) and the

termination field of the DLP within NCL completely overlaps with the termination area of the

NIM (c.f. figs. 10, 11 in chapter 1, and fig 1f in chapter 2).

The exact homologue of the DLP remains unclear (c.f. detailed discussion in chapter 1),

functionally, however, the DLP-NCL projection has been implicated in the proper execution of

working memory tasks (Güntürkün, 1997), similar to the MD-PFC projection in mammals

(Sakurai and Sugimoto, 1985; Fuster, 1989 for review). It is therefore not unreasonable to

speculate that the DLP-NIM projection also conveys information pertinent to these tasks. The

NIM receives afferents from areas in the Wulst and archistriatum, and its projection onto NCL

could complement the parallel input the NCL receives from these areas. Via the projections of

the NCL and NIM to the basal ganglia and the archistriatum, this loop through the forebrain

could also influence activity in the dorsal thalamus (including DIP and DLP) and lower

mesencephalic areas (e.g. the optic tectum, or AVT; c.f. fig. 13 in chapter 1).

In the PFC of mammals, working memory is manifest in the activity of delay-related neurons

that are activated after offset of the cue and continue firing throughout the duration of the delay

(e.g. Fuster, 1973; Sawaguchi and Yamane, 1999). Most cellular models of working memory

assume that the activity of these neurons reflects reverberating activity in a cell assembly that

codes for the given memorandum (Amit, 1995; Goldman-Rakic, 1996; Durstewitz et al., 1999;

2000). In the PFC, reciprocal horizontal projections between neurons in layers III, and V-VI,

respectively, connect discrete „columns“ within the same cortical region, which are assumed to

form functional modules (e.g. Goldman-Rakic, 1996; Melchitzky et al., 1998; González-Burgos

et al., 2000). The local interaction between these cells provides a neural basis for positive

feedback circuits in the PFC that may play a role in sustaining neural activity (Amit 1995;

Goldman-Rakic, 1996; Durstewitz et al., 1999, 2000).

In the pigeon, Kalt and coworkers (1999) have recently shown that neurons in the NCL can

similarly enhance their firing throughout the delay period of a delayed response task to „hold“

information after the offset of the sensory cue. It can be assumed that at least part of the


sustained activity results from reverbarating activity within the NCL. In accordance with this

view, simultaneous recordings from two NCL neurons during a working memory task show a

significant correlation of task-related activity (Thomas Kalt, personal communication).

The electrophysiological analysis of cell types in-vitro, allows the description of certain

distinct features (e.g. firing pattern, or membrane time constant) that determine their integrative

properties and the impact on other cells (chapter 3). However, the conclusions that can be

drawn from this analysis about a cell´s role within a functional circuit are rather limited. This is

due to the fact that in-vivo the cell´s function might be determined much more by the properties

of the network, than by the properties of the cell itself.

In the NCL, spiny principal neurons of the type II, and smooth type III interneurons showed

sustained firing over a wide range of frequencies and throughout long depolarizing current

pulses. Thus, in principle, the firing patterns of these cells could represent a neural substrate for

active memory retention on the single cell level. Furthermore, Durstewitz and coworkers

(2000) could show, that in a network of the PFC stimulus-specific recurrent activity can be

maintained at frequencies below 20 Hz, which is in the range of the steady-state firing

frequency of type II neurons (c.f. chapters 3 and 6).

However, in contrast, the firing frequency of single NCL neurons in-vivo is not only highly

irregular (as might also be the case in cortical neurons), but furthermore it appears that a large

portion of neurons which show delay-related activity, respond with bursts of action potentials

(Thomas Kalt, personal communication). Thus, reverberating activity in an ensemble of burst-

firing neurons (i.e. cell types I and IV1) also seems to make a major contribution to delay

activity. As bursting neurons showed strong accomodation of firing and were not able to

generate repetitive bursts in response to continued depolarization, phases of refractory

inhibition, most likely provided by feedback from the type III interneurons, are also essentially

to produce the pattern observed in-vivo.

Functionally, burst firing increases the probability to drive the postsynaptic cell beyond spike

threshold (Snider et al., 1998). The high frequency of action potentials during a burst will thus

amplify the neural signal and synchronize the activity in a population of postsynaptic cells

(Hablitz, 1986; Snider et al., 1998). Therefore, burst firing pattern of type I cells might provide

for reliable transmission within an active cell assembly, which is needed for the stability of the

mnemonic representation.

1 As discussed in detail in chapter 3, neurons of type IV are likely to represent immature stages of other spinycell types, and therefore they are not considered here further.


Modulation by the dopaminergic system

As discussed in detail in chapters 4 and 5, the organization of the mesotelencephalic DA

projections and the regional distribution of dopaminoceptive elements is highly similar in

mammals and birds, despite approximately 250 million years of divergent evolution.

As also shown in chapter 4, the structural similarities extend to functional similarities with

regard to the control of sensorimotor and associative processes, that are controlled by the

dopaminergic system.

In mammals, the integrity of the active short-term memory trace appears to be critically

dependant on an optimal level of DA receptor stimulation. Both, prefrontal DA depletion

(Brozoski et al., 1979; Simon et al., 1980) and excessive stimulation of DA receptors (Murphy

et al., 1996; Sawaguchi, 1997; Zahrt et al., 1997; Arnsten and Goldman-Rakic, 1998), disrupt

performance on delay tasks. Similarly, the administration of high doses of D1 antagonists into

the PFC also impair performance on delay tasks and decrease delay-period activity of PFC

neurons (Sawaguchi et al., 1990b; Sawaguchi and Goldman-Rakic, 1994; Williams and

Goldman-Rakic, 1995; Zahrt et al., 1997), however, application of low doses of D1 antagonists

or DA into the PFC increases delay-correlated activity of PFC neurons relative to background

activity (Sawaguchi et al., 1988; 1990a; Williams and Goldman-Rakic, 1995; Müller et al.,

1998). Thus, too much or too little DA may be detrimental to cognition.

Thus, one major finding of the present studies is that DA also profoundly alters the excitability

of NCL neurons, and that the observed increase in the number of spikes was strongly dose

dependent. Moreover, this dose dependency was best described with an inverted U-curve,

demonstrating also that under physiological conditions in birds, as is the case in mammals, an

optimal level of DA might be required for the changes in neuronal excitability to take place

(chapter 6).

Together with the data by Kalt and coworkers (1999) and the voltage-dependence of the DA

effect (discussed in chapter 6), one can speculate that in-vivo the mesencephalic DA signal to

the NCL enhances the excitability of active cell assemblies, which might result in recurrent

activity that is strong enough to carry the memory trace over the delay.

So far, there exists only one study that has examined the contribution of the DA system on

working memory in birds (Güntürkün and Durstewitz, in press). In support for the hypothesis

outlined above, the results show a severe impairment in working memory functions, following

the blockade of D1 receptors by injections of high doses of SCH-23390 into the NCL.

Therefore, D1 receptor mediated enhancement of neuronal firing might be a requirement for the


active retention of memory contents both in mammals and birds. To test the validity of this

assumption, it could be attempted to pharmacologically alter the delay-related activity of NCL

neurons through the iontophoretic application of D1 receptor antagonists, or supranormal doses

of DA (c.f. Sawaguchi and Goldman-Rakic, 1991; Williams and Goldman-Rakic, 1995).

However, a general, non-specific increase in activity will not produce a stable representation

in short-term memory. As indicated above, the voltage- (i.e. activity-) dependance of the DA

effects (c.f. discussion in chapter 6) might serve to enhance only those cell assemblies that are

currently activated. Yet, the question remains, how the representation of memory items within

the recurrent activity might be protected from other interfering stimuli.

In mammals, the bilaminar pattern of prefrontal cortical DA innervation might be an inherent

aspect of the dopaminergic modulation of working memory. In the PFC, the innervation by

dopaminergic fibers is densest in the upper layers I-II and V-VI (c.f. figure 4 in the

introduction). The arrangement of long-range associational and local input, respectively, shows

a somewhat parallel pattern. Synaptic inputs from other association cortical regions mainly

arrive in layers I-II, where the apical tuft of the layer III, and V-VI pyramidal neurons extends

(Sesack et al., 1989; Berendse et al., 1992; Condé et al. 1995), and, as outlined above, local

synaptic connections that contact the soma and proximal dendrites arise from neurons from

within the same lamina.

As discussed in detail in chapter 6, DA might modulate distal signals in a manner different from

its effect on proximal inputs. In sum, DA acting on pyramidal neurons of the PFC (i) restricts the

effects of inputs to the apical dendrites of these neurons by attenuating the dendritic HVA Ca++-

mediated amplification of such inputs (Westenbroek et al., 1992; Surmeier et al., 1995; Yang

and Seamans, 1996); and (ii) increases cell excitability through its actions on conductances that

are located mainly in the soma and proximal dendrites (Yang and Seamans, 1996; Yang et al.,

1998). Thus, one effect of DA might be to selectively attenuate distal EPSPs relative to

proximally arriving EPSPs, resulting in a differential „selection“ of synaptic inputs (Yang et al.,

1996b; see also Law-Tho et al., 1994).

In the absence of a layered structure that enables the spatial separation of functionally distinct

inputs, another mechanisms is needed to „switch“ between inputs from long range associational

fibers and from intrinsic connections. In the avian brain, the dichotomic modes of DA

innervation might serve this function.


Figure 1: Summary of the anatomical data presented here, and the modulatory actions of dopamine on identified neuronalelements in the NCL. The input-output relations of the NCL have been described in chapter 1 - 3. Note that for reason of claritynot all possible afferent and efferent projections are depicted. However, neurons from all spiny cell types (i.e. types I, II, IV [notshown]), but not smooth interneurons (type III), project to the same targets outside NCL. In addition, it can be assumed, thatneurons of all cell types receive the same afferent input. The assignment of DARPP-32 immunoreactivity to certain cell types isinferred from a comparison of the morphology of DARPP-32 positive neurons described in chapter 5 with those identified inchapter 3, and this is complemented by the reaction of these cell types to the application of DA and the D1 receptor agonistSKF-38393. Part of the intrinsic connections of the NCL can be inferred from the pattern of local axonal collateralizations ofdifferent cell types (c.f. chapter 3). However, a detailed description of the intra-NCL connections is still lacking. The insertshows a hypothetical arrangement through which the dopaminergic system could achieve a spatial selection of functionallydistinct synaptic inputs. Local synaptic connections that arrive at the soma of neighbouring cells might be enhanced relative tosynaptic inputs from sensory areas arriving at the distal dendrites. In accordance with the data presented in chapter 6, a type Ineuron is shown to receive basket-type innervation. For the model proposed, however, it is not relevant whether a specific celltype receives basket-innervation, and there is evidence that projection neurons of the other celltypes also receive basket-innervation.

There are several working hypotheses regarding the role of the dopaminergic basket-type

innervation that have been tackled by the studies presented here. One assumption might have

been that the basket-innervation provides a means for the dopaminergic system to selectively

target a certain class of neurons, e.g. GABAergic interneurons, cells with distinct firing

patterns, or distinct axonal connections. Alternatively, it might have been thought that the strong

somatic basket innervation exerts a qualitatively or quantitatively different influence on the

feedback projectionsto secondary sensory areas

thalamic input(DLP, SRt, VIP)

archistriatum (limbic and motor) (limbic and motor)

basal ganglia

archistriatum (limbic and motor)

(limbic and motor)basal ganglia

basal gangliaand archistriatum

dopaminergic fibersfrom SN and VTA

telencephalic input fromsensory and associational areas

intra NCLconnections

type IIDARPP-32+DA and D1 agonistenhance firing

type IDARPP-32+DA and D1 agonistenhance firing

type IIIDARPP-32-no effect of DAon firing pattern


target cell, in the sense that it provides a higher concentration of DA, which in turn might result

in opposite or quantitatively different effects than comparatively lower concentrations of DA

(e.g. inverted U shape of effects).

Clearly, our understanding about the functional subtypes of NCL neurons is still too limited to

unequivocally rule out any of these hypothesis. However, from the data provided in chapter 6

no obvious qualitative or quantitative difference in the response to DA was evident for neurons

that received strong somatic basket input.

So what is the functional importance of the two modes of dopaminergic innervation then?

As indicated above, I would like to propose that baskets provide a means for spatial selection

of functionally distinct synaptic inputs. A differential effect of DA on distinct inputs might be

achieved if a specific class of connections, e.g. local connections between NCL neurons, make

synapses preferentially on the somata and initial dendrites of neighbouring cells. Such a

hypothetical arrangement is shown in the insert in figure 1. For this model it is not neccessary

that all intrinsic connections of NCL neurons selectively make contacts only with the somata of

other NCL neurons. Dopamine might still enhance these inputs relatively more if only a fraction

of intra-NCL connections synapses on the somata of neighbouring cells. For the differential

effects of DA to take place it is merely critical to assume that a larger number of proximal

inputs originates from intra-NCL projections relative to long-range sensory inputs. Also, the

large number of baskets found in the NCL is not an indicator for a larger number of intrinsic

connections. Other areas in the avian forebrain might possess a similarly large number of local

interconnections, but the functions of these circuits may not depend as much on dopaminergic

modulation as it appears to be the case for the NCL.

In the NCL, however, the proposed „input-selection“ effect of DA could account for

reverbarating activity between NCL neurons, and a concurrent protection of this memory trace

against interfering stimuli. This effect of DA could be brought about by several mechanisms.

The easiest model to assume is that the distribution of ionic conductances across the cell

membrane of NCL neurons varies in a manner similar to that found in mammalian neurons, so

that DA will attenuate distal („sensory“) inputs through dendritic HVA Ca++ channels, and will

enhance proximal inputs via its actions on INaP and L-type currents. Indeed it appears, that such

ionic „compartments“ exist not only in neocortical neurons that extend long apical dendrites.

For example, striatal (Surmeier et al, 1995) and cerebellar (Usowicz et al., 1992) neurons

show a similar distribution of N- and P-type currents in the dendritic segments, and a

predominance of L-type currents at the soma and proximal dendrites. Alternatively, DA could


have dose-dependent effects on these neurons, as DA released from baskets will probably

faster reach significant concentrations at a large number of receptors. However, such an effect

would have went undetected in the experiments presented in chapter 6, in which DA was bath

applied, and the recordings were made (most likely) only from the somata of NCL neurons.

Support for a differential dendritic vs. somatic action of DA may come from the observation by

Schnabel and coworkers (1997), who found that the majority of D1 receptor immunoreactivity

was localized to synaptic and extrasynaptic sites in spines and dendrites, but that in addition

some neurons also showed dense labeling of the somatic cytoplasm.

To test the ideas outlined above, anatomical and electrophysiological studies may be devised.

Clearly, an electronmicroscopic analysis of the synaptic types and the subcellular distribution

of intra-NCL connections is required. To this end, either small injections of an anterograde

tracer confined to the NCL might be utilized, or the axonal arborizations of single, intracellulary

labeled neurons might be studied. Functionally, these connections could also be analyzed by

paired recordings from neighbouring cells in NCL, and the effect of DA on the synaptic weight

may be compared relative to DA-induced changes in synaptic inputs from areas outside the

NCL. Furthermore, direct recordings from dendrites of NCL neurons and pharmacological

isolation of single ionic currents could adress the issue whether DA exerts its effects via a

differential modulation of specific conductances.

In sum, the data presented here describe the organisation of a polymodal associative structure in

the avian forebrain, and they provide a structural and physiological framework in which

sensorimotor integration and working memory may take place. Furthermore, based on a

comparison with the prefrontal cortex of mammals, it appears that functional similarities also

neccessitate some structural similarities. Thus, the localization of a cognitive function within a

neuronal circuit, and a further understanding of the circuit´s structural and physiological

requirements may ultimately help to identify the building blocks of mental computation in

various species.


Abbreviations used in the text and figures

6-OHDA 6-hydroxydopamine

AMPA α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid

AC or Acc nucleus accumbens

Ai archistriatum intermedium, pars centrale

Aid archistriatum intermedium, pars dorsale

Aidd archistriatum intermedium, parsdorsodorsale

Aidv archistriatum intermedium, parsdorsoventrale

AL ansa lenticularis

Am archistriatum mediale

Ap archistriatum posterior

APH area parahippocampalis

Av archistriatum intermedium, pars ventrale

AVT area ventralis tegmentalis; also see AVT

Bas nucleus basalis

BNST bed nucleus of the stria terminalis

Ca++ calcium

CDL corticoidea dorsolateralis

Cl- chloride

CPi cortex piriformis

DA dopamine

DARPP-32 dopamine- and cAMP-regulatedphosphoprotein, Mr = 32,000

DBH dopamine-beta-hydroxylase

DIP nucleus dorsointermedius posterior thalami

DIVA nucleus dorsalis intermedius ventralisanterior

DLL nucleus dorsolateralis anterior thalami,pars lateralis

DLM nucleus dorsolateralis anterior thalami,pars medialis

DLP nucleus dorsolateralis posterior

DMA nucleus dorsomedialis anterior thalami

DMP nucleus dorsomedialis posterior thalami

E ectostriatum

Ep ectostriatal belt

FPL fasciculus prosencephali lateralis

GABA γ-aminobutyric acid

GAD glutamate decarboxylase

GLd dorsolateral geniculate nuclei

HA hyperstriatum accessorium

HD hyperstriatum dorsale

HIS hyperstriatum intercalatus superior

HOM tractus occipitomesencephalicus, parshypothalami

Hp hippocampus

HV hyperstriatum ventrale

HVA high-voltage-activated (Ca++ currents)

HVdv hyperstriatum ventrale dorso-ventrale

HVvv hyperstriatum ventrale ventroventrale

ICo nucleus intercollicularis

IHA nucleus intercalatus of the hyperstriatumaccessorium

IMHV intermediate and medial part of thehyperstriatum ventrale

IKS slowly inactivating potassium current

INaP persistent sodium current

INP nucleus intrapeduncularis

ir immunoreactive

K+ potassium

L1 field L1

L2 field L2

L3 field L3

LFS lamina frontalis superior

LH lamina hyperstriatica

LHy nucleus lateralis hypothalami

LPO lobus parolfactorius

MNH mediorostral neostriatum/hyperstriatumventrale

n. nucleus


NA noradrenaline

Na++ sodium

NC neostriatum caudale

NCL neostriatum caudolaterale

NCm neostriatum caudomediale

Nd neostriatum dorsale

NF neostriatum frontale

NFL neostriatum frontolaterale

NFT neostriatum fronto-trigeminale

NI neostriatum intermedium

NIL neostriatum intermedium laterale

NIM neostriatum intermedium medialis

NIMl neostriatum intermedium medialis, parslaterale

NIMm neostriatum intermedium medialis, parsmediale

NMDA N-methyl-D-aspartate

OM tractus occipitomesencephalicus

Ov nucleus ovoidalis

PA paleostriatum augmentatum

PFC prefrontal cortex

PNMT phenylethanolamine-N-methyl transferase

PP paleostriatum primitivum

PT nucleus pretectalis

Rt nucleus rotundus

SAC stratum album centrale

SCI stratum cellulare internum

SGC stratum griseum centrale

SGP substantia grisea et fibrosaperiventricularis

SNC substantia nigra, pars compacta

SNR substantia nigra, pars reticulata

SPC nucleus tractus septomesencephalici

SPL nucleus spiriformis lateralis

SRt nucleus subrotundus

T nucleus triangularis

TH tyrosine hydroxylase

Tn nucleus taeniae

TO (chapter 1) optic tectum

TO (chapter 4) olfactory tubercle

TPc nucleus tegmenti pedunculopontinus, parscompacta

TPO area temporo-parieto-occcipitalis

TSM tractus septomesencephalicus

Va vallecula

VIA ventrointermediate area of the thalamus

VTA ventral tegmental area (A10); also seeAVT

VIP ventrointermediate area of the posteriornuclei

VP ventral pallidum

Literature 204

LITERATURE

Abercrombie, E.D., Keefe, K.A., DiFrischia, D.S., Zigmond, M.J. (1989) Differential effect ofstress on in vivo dopamine release in striatum, nucleus accumbens, and medial frontalcortex. J. Neurochem. 52:1655-1658.

Absil, P., Das, S., Balthazart, J. (1994) Effects of apomorphine on sexual behavior in malequail. Pharmacol. Biochem. Behav. 47:77-88.

Adams, J.C. (1981) Heavy metal intensification of DAB-based HRP reaction product. J.Histochem. Cytochem. 29:775.

Akert, K. (1964) Comparative anatomy of frontal cortex and thalamofrontal connections. In:J.M. Warren, K. Akert (eds) The Frontal Granular Cortex and Behavior. New York:McGraw-Hill, pp.372-396.

Akintunde, A. and Buxton, D.F. (1992) Origins and collateralization of corticospinal,corticopontine, corticorubral and corticostriatal tracts: A multiple retrograde fluorescenttracing study. Brain Res. 586:208-218.

Aldavert-Vera, L., Costa-Miserachs, D., Divac, I., Delius, J.D. (1999) Presumed 'prefrontalcortex' lesions in pigeons: Effects on visual discrimination performance. Behav. BrainRes. 102:165-70.

Aldridge, J.W. and Berridge, K.C. (1998) Coding of serial order by neostriatal neurons: a"natural action" approach to movement sequence. J. Neurosci. 18:2777-2787.

Alexander, G.E, Crutcher, M.D., DeLong, M.R. (1990) Basal ganglia-thalamocortical circuits:parallel substrates for motor, oculomotor, "prefrontal" and "limbic" functions. In: H.B.M.Uylings, C.G. van Eden, J.P.C. de Bruin, M.A. Corner, M.G.P. Feenstra (eds.), ThePrefrontal Cortex: Its Structure, Function, and Pathology. Prog. Brain Res. 85:119-146.

Amit, D.J. (1995) The Hebbian paradigm reintegrated: Local reverberations as internalrepresentation. Behav. Brain Sci. 18:617–657.

Anderson, K.D. and Reiner, A. (1990) Distribution and relative abundance of neurons in thepigeon forebrain containing somatostatin, neuropeptide Y, or both. J. Comp. Neurol.299:261-282.

Anderson, K.D. and Reiner, A. (1991a) Immunohistochemical localization of DARPP-32 instriatal projection neurons and striatal interneurons: implications for the localization ofD1-like dopamine receptors on different types of striatal neurons. Brain Res. 568:235-243.

Anderson, K.D. and Reiner, A. (1991b) Striatonigral projection neurons: a retrograde labelingstudy of the percentages that contain Substance P or enkephalin in pigeons. J. Comp.Neurol. 303:658-673.

Anderson, K.D., Karle, E.J., Reiner, A. (1991) Ultrastructural single- and double- labelimmunohistochemical studies of Substance P-containing terminals and dopaminergicneurons in the substantia. J. Comp. Neurol. 309:341-362.

Apicella, P., Ljungberg, T., Scarnati, E., Schultz, W. (1991) Responses to reward in monkeydorsal and ventral striatum. Exp. Brain Res. 85:491-500.

Apicella, P., Scarnati, E., Ljungberg, T., Schultz, W. (1992) Neuronal activity in monkeystriatum related to the expectation of predictable environmental events. J. Neurophysiol.68:945-960.

Arends, J.J.A. and Zeigler, H.P. (1986) Anatomical identification of auditory pathway fromnucleus of lateral lemniscular system to frontal telencephalon of pigeon. Brain Res.398:375-381.

Literature 205

Arends, J.J.A., Wild, J.M., Zeigler, H.P. (1988) Projections of the nucleus of the tractussolitarius in the pigeon (Columba livia). J. Comp. Neurol. 278:405-429.

Ariano, M.A. (1997) Distribution of dopamine receptors. In: The Dopamin Receptors, pp 77-103. Eds. K.A. Neve, R.L. Neve. Humana Press: Totowa.

Ariëns-Kappers, C.U., Huber, G.C., Crosby, E.C. (1936) The comparative anatomy of thenervous system of vertebrates, including man. Hafner Publishing company: New York.

Arnsten, A.F.and Goldman-Rakic, P.S. (1998) Noise stress impairs prefrontal corticalcognitive function in monkeys: Evidence for a hyperdopaminergic mechanism. Arch. Gen.Psychiatry. 55:362-368.

Asaad, W.F., Rainer, G., Miller, E.K. (1998) Neural activity in the primate prefrontal cortexduring associative learning. Neuron 21:1399-1407.

Atkinson, R.C. and Shiffrin, R.M. (1968) in The Psychology of Learning and Motivation:Advances in Research and Theory, ed. Spence, K.W. (Academic, New York) Vol. 2, pp89–195.

Baddeley, A.(1992) Working memory. Science 255:556-559.

Bagetta, G., DeSarro, G., Priolo, E., Nisticò, G. (1988a) Ventral tegmental area: Site throughwhich dopamine D2-receptor agonists evoke behavioural and electrocortical sleep inrats. Br. J. Pharmacol. 95:860-866.

Bagetta, G., Priolo, E., Nisticò, G. (1988b) Dopamine D-2 receptors in the ventral tegmentalarea mediate behavioural and electrocortical sleep of dopaminergic drugs in rats.Pharmacol. Res. Commun. 20:1091-2.

Bagnoli, P. and Burkhalter, A. (1983) Organization of the afferent projections to the Wulst inthe pigeon. J. Comp. Neurol. 214:103-113.

Bailhache, T. and Balthazart, J. (1993) The catecholaminergic system of the quail brain:immunocytochemical studies of dopamine beta-hydroxylase and tyrosine hydroxylase. J.Comp. Neurol. 329:230-256.

Ball, G.F., Casto, J.M., Balthazart, J. (1995) Autoradiographic localization of D1-likedopamine receptors in the forebrain of male and female japanese quail and theirrelationship with immunoreactive tyrosine hydroxylase. J. Chem. Neuroanat. 9:121-133.

Balthazart, J. and Absil, P. (1997) Identification of catecholaminergic inputs to and outputsfrom aromatase-containing brain areas of the japanese quail by tract tracing combinedwith tyrosine hydroxylase immunocytochemistry. J. Comp. Neurol. 382:401-428.

Balthazart, J., Castagna, C., Ball, G.F. (1997) Differential effects of D1 and D2 dopamine-receptor agonists and antagonists on appetitive and consummatory aspects of male sexualbehavior in japanese quail. Physiol. Behav. 62:571-580.

Balthazart, J., Foidart, A., Baillien, M., Harada, N., Ball, G.F. (1998) Anatomical relationshipsbetween aromatase and tyrosine hydroxylase in the quail brain: Double-labelimmunocytochemical studies. J. Comp. Neurol. 391:214-226.

Barbas, H. and DeOlmos, J. (1990) Projections from the amygdala to basoventral andmediodorsal prefrontal regions in the rhesus monkey. J. Comp. Neurol. 300:549-571.

Barbas, H., Henion, T.H., Dermon, C.R. (1991) Diverse thalamic projections to the prefrontalcortex in the rhesus monkey. J. Comp. Neurol. 313:65-94.

Bateson, P.P. and Chantrey, D.F. (1972). Retardation of discrimination learning in monkeys andchicks previously exposed to both stimuli. Nature 237:173-174.

Baunez, C., Nieoullon, A., Amalric, M. (1995) Dopamine and complex sensorimotorintegration: Further studies in a conditioned motor task in the rat. Neuroscience 65:375-384.

Literature 206

Baylé, J.D., Ramade, F., Oliver, J. (1974) Stereotaxic topography of brain of quail. J. Physiol.(Paris) 68:219-241.

Belger, A., Puce, A., Krystal, J.H., Gore, J.C., Goldman-Rakic, P.S., McCarthy, G. (1998)Dissociation of mnemonic and perceptual processes during spatial and nonspatialworking memory using fMRI. Hum. Brain Mapp. 6:14-32.

Benardo, LS. (1997). Recruitment of GABAergic inhibition and synchronization of inhibitoryinterneurons in rat neocortex. J. Neurophysiol. 77:3134-3144.

Benes, F.M., Vincent, S.L., Molloy, R. (1993) Dopamine-immu-noreactiveaxon varicositiesform nonrandom contactswith GABA-immunoreactive neurons of rat medial prefrontalcortex. Synapse 15:285–295.

Beninger, R.J. (1993) Role of D1 and D2 Receptors in Learning. In: D1:D2 Dopamine ReceptorInteractions, pp 115-157. Ed. J.L. Waddington. Academic Press: London.

Benowitz, L.I. and Karten, H.J. (1976) Organization of the tectofugal visual pathway in thepigeon: a retrograde transport study. J. Comp. Neurol. 167:503-520.

Benowitz, L.I. and Karten, H.J. (1976) The tractus infundibuli and other afferents to theparahippocampal region of the pigeon. Brain Res. 102:174-180.

Berendse, H.W., Galis-deGraaf, Y., Groenewegen, H.J. (1992) Topographical organization andrelationship with ventral striatal compartments of prefrontal corticostriatal projections inthe rat. J. Comp. Neurol. 316:314-347.

Berger, B. and Gaspar, P. (1994) Comparative anatomy of the catecholaminergic innervation ofrat and primate cerebral cortex. In: Phylogeny and development of catecholaminesystems in the CNS of vertebrates, pp 293-324. Eds. W.J.A.J. Smeets, A. Reiner, A.Cambridge University Press: Cambridge, UK.

Berger, B., Febvret, A., Greengard, P., Goldman-Rakic, P.S. (1990) DARPP-32, aphosphoprotein enriched in dopaminoceptive neurons bearing dopamine D1 receptors:Distribution in the cerebral cortex of the newborn and adult rhesus monkey. J. Comp.Neurol. 299:327-348.

Berger, B., Gaspar, P., Verney, C. (1991) Dopaminergic innervation of the cerebral cortex:Unexpected differences between rodents and primates. Trends Neurosci. 14:21-27.

Berger, B., Trottier, S., Verney, C., Gaspar, P., Alvarez, C. (1988) Regional and laminardistribution of the dopamine and serotonin innervation in the macaque cerebral cortex: Aradioautographic study. J. Comp. Neurol. 273:99-119.

Berk, M.L. and Butler, A.B. (1981) Efferent projections of the medial preoptic nucleus andmedial hypothalamus in the pigeon. J. Comp. Neurol. 203:379-399.

Berk, M.L. and Hawkin, R.F. (1985) Ascending projections of the mammillary region in thepigeon: Emphasis on telencephalic connections. J. Comp. Neurol. 239:330-340.

Berkley, K.J., Budell, R.J., Blomquist, A., Bull, M. (1986) Output systems of dorsal columnnuclei in cat. Brain Res. Bull. 11:199-225.

Bernardi, G., Cherubini, E., Marciani, M.G., Mercuri, N., Stanzione, P. (1982) Responses ofintracellularly recorded cortical neurons to the iontophoretic application of dopamine.Brain Res. 245:267-274.

Bertolucci-D'Angio, M., Serrano, A., Driscoll, P., Scatton, B. (1990b) Involvement ofmesocorticolimbic dopaminergic systems in emotional states. In: Progress in BrainResearch, pp 405-417. Eds. H.B.M. Uylings, C.G. van Eden, J.P.C. De Bruin, M.A.Corner, M.G.P. Feenstra. Elsevier: Oxford.

Bertolucci-D'Angio, M., Serrano, A., Scatton, B. (1990a) Mesocorticolimbic dopaminergicsystems and emotional states. J. Neurosci. Methods 34:135-142.

Literature 207

Binggeli, R. L. and Paule, W. J. (1969) The pigeon retina: Quantitative aspects of the opticnerve and ganglion cell layer. J. Comp. Neurol. 137:1-18.

Bingman, V.P., Casini, G., Nocjar, C., Jones, T.J. (1994) Connections of the piriform cortex inhoming pigeons (Columba livia) studies with Fast Blue and WGA-HRP. Brain Behav.Evol. 43:206-218.

Bischof, H.J. (1988) The visual field and visually guided behavior in the zebra finch(Taeniopygia guttata). J. Comp. Physiol. 163: 329-337.

Björklund, A. and Lindvall O. (1984) Dopamine-containing systems in the CNS. In Handbookof Chemical Neuroanatomy, Vol. 2: Classical Transmitters in the CNS, Part I (eds.Björklund A. and Hökfelt T.), pp 55-122. Elsevier Science Publishers, Amsterdam.

Björklund, A., Divac, I., Lindvall, O. (1978) Regional distribution of catecholoamines inmonkey cerebral cortex, evidence for a dopaminergic innervation of the primateprefrontal cortex. Neurosci. Lett. 7:115-119.

Blanton, M.G., Lo Turco, J.J., Kriegstein, A.R. (1989) Whole cell recording from neurons inslices of reptilian and mammalian cerebral cortex. J. Neurosci. Methods 30:203-2010.

Blough, P.M. (1971) The visual acuity of the pigeon for distant targets. J. Exp. Anal. Behav.15:57-67.

Bo, P., Ongini, E., Giorgetti, A., Savoldi, F. (1988) Synchronization of the EEG and sedationinduced by neuroleptics depend upon blockade of both D1 and D2 dopamine receptors.Neuropharmacol. 27:799-805.

Bock, J. and Braun, K. (1999a) Blockade of N-methyl-D-aspartate receptor activationsuppresses learning-induced synaptic elimination. Proc. Natl. Acad. Sci. 96:2485-2490.

Bock, J. and Braun, K. (1999b) Filial imprinting in domestic chicks is associated with spinepruning in the associative area, dorsocaudal neostriatum. Eur. J. Neurosci. 11:2566-2570.

Bock, J., Schnabel, R., Braun, K. (1997) Role of the dorso-caudal neostriatum in filialimprinting of the domestic chick: A pharmacological and autoradiographical approachfocused on the involvement of NMDA-receptors. Eur. J. Neurosci. 9:1262-1272.

Bockaert, J., Premont, J., Glowinski, J., Thierry, A.M., Tassin, J.P. (1976) Topographicaldistribution of dopaminergic innervation and of dopaminergic receptors in the ratstriatum. II. Distribution and characteristics of dopamine adenylate cyclase-interaction ofd-LSD with dopaminergic receptors. Brain Res. 107:301-15.

Bolhuis, J.J. and Horn, G. (1992) Generalization of learned preferences in filial imprinting.Anim. Behav. 44:185-187.

Bolhuis, J.J., deVos, G.J., Kruijt, J.P. (1990) Filial imprinting and associative learning. Q. J.Exp. Psychol. B. 42:289-312.

Bonke, B. A., Bonke, D., Scheich, H. (1979) Connectivity in the auditory forebrain nuclei in theGuinea fowl (Numida meleagris). Cell Tiss. Res. 200:101-121.

Bons, N. and Oliver, J. (1986) Origin of afferent connections to parolfactory lobe in quailshown by retrograde labelling with a fluorescent neurotracer. Exp. Brain Res. 63:125-134.

Bottjer, S.W. (1993) The distribution of tyrosine hydroxylase immunoreactivity in the brains ofmale and female zebra finches. J. Neurobiol. 24:51-69.

Bottjer, S.W., Brady, J.D., Walsh, J.P. (1998) Intrinsic and synaptic properties of neurons in thevocal-control nucleus IMAN from in vitro slice preparations of juvenile and adult zebrafinches. J. Neurobiol. 37:642-658.

Literature 208

Bottjer, S.W., Halsema, K.A., Brown, S.A., Miesner, E.A. (1989) Axonal connections of aforebrain nucleus involved with vocal learning in zebra finches. J. Comp. Neurol.279:312-326.

Boyson, S.J., McGonigle, P., Molinoff, P.B. (1986) Quantitative autoradiographic localizationof the D1 and D2 subtypes of dopamine receptors in rat brain. J. Neurosci. 6:3177-3188.

Bradley, P., Davies, D.C., Horn, G. (1985) Connections of the hyperstriatum ventrale of thedomestic chick (Gallus domesticus). J. Anat. 140:577-589.

Bradley, P.M., Burns, B.D., King, T.M., Webb, A.C. (1996) Morphological andelectrophysiological properties of neurons in an area of the chick brain involved inlearning. Brain Res. 727:125-132.

Bradshaw, C.M., Sheridan, R.D., Szabaldi, E. (1985) Excitatory neuronal responses todopamine in the cerebral cortex: Involvement of D2 but not D1 dopamine response. Br. J.Pharmacol. 86:483–490.

Braun, K., Scheich, H., Braun, S., Rogers, J.H., Heizmann, C.W. (1991a) Parvalbumin-,calretinin- and calbindin-D28k-immunoreactivity and GABA in a forebrain regioninvolved in auditory filial imprinting. Brain Res. 539:31-44.

Braun, K., Scheich, H., Heizmann, C.W., Hunziker, W. (1991b) Parvalbumin and calbindin-D28K immunoreactivity as developmental markers of auditory and vocal motor nuclei ofthe zebra finch. Neuroscience 40:853-869.

Braun, K., Scheich, H., Schachner, M., Heizmann, C.W. (1985) Distribution of parvalbumin,cytochrome oxidase activity and 14C-2-deoxyglucose uptake in the brain of the zebrafinch. I. Auditory and vocal motor system. Cell Tis. Res. 240:101-115.

Brauth, S.E., Ferguson, J.L., Kitt, C.A. (1978) Prosencephalic pathways related to thepaleostriatum of the pigeon (Columba livia). Brain Res. 147:205-221.

Brauth, S.E., Heaton, J.T., Durand, S.E., Liang, W., Hall, W.S. (1994) Functional anatomy offorebrain auditory pathways in the budgerigar (Melopsittacus undulatus). Brain Behav.Evol. 44:210-233.

Bredenkötter, M. and Braun, K. (1997) Changes of neuronal responsiveness in the mediorostralneostriatum/hyperstriatum after auditory filial imprinting in the domestic chick.Neuroscience 76:355-365.

Broca, P.P. (1861) In: Some Papers on the Cerebral Cortex, translated by von Bonin, G.(1960) (Thomas, Springfield, IL), pp 49–72.

Brock, J.W., Farooqui, S., Ross, K., Prasad, C. (1992) Localization of dopamine D2 receptorprotein in rat using polyclonal antibody. Brain Res. 578:244-250.

Brodmann, K. (1909) Vergleichende Lokalisationslehre der Grosshirnrinde in ihrenPrinzipien dargestellt auf Grund des Zellenbaues. Leipzig, Barth.

Brozoski, T.J., Brown, R.M., Rosvold, H.E., Goldman, P.S. (1979) Cognitive deficit caused byregional depletion of dopamine in prefrontal cortex of rhesus monkey. Science 205:929-932.

Brutkowski, S. (1964) Prefrontal cortex and drive inhibition. In: J.M. Warren, K. Akert (Hrsg.)The Frontal Granular Cortex and Behavior. McGraw-Hill New York pp 242-270.

Bunney, B.S., Sesack, S.R., Silva, N.L. (1987) Midbrain dopaminergic systems:neurophysiology and electrophysiological pharmacology. In: Psychopharmacology: TheThird Generation of Progress, pp 113-126. Ed H.Y. Meltzer. Raven Press: New York.

Buzsaki, G. and Chrobak, J.J. (1995) Temporal structure in spatially organized neuronalensembles: A role for interneuronal networks. Curr. Opin. Neurobiol. 5:504-510.

Literature 209

Byrne, J.H. and Kandel, E.R. (1996) Presynaptic facilitation revisited: State and timedependence. J. Neurosci. 16:425-35.

Calabresi, P., Mercuri, N., Stanzione, P., Stefani, A., Bernardi, G. (1987) Intracellular studieson the dopamine-induced firing inhibition of neostriatal neurons in vitro: Evidence for D1receptor involvement. Neuroscience 20:757-771.

Campbell, C.B.G. and Hodos, W. (1970) The concept of homology and the evolution of thenervous system. Brain Behav. Evol. 3:353-367.

Camps, M., Kelly, P.H., Palacios, J.M. (1990) Autoradiographic localization of dopamine D1and D2 receptors in the brain of several mammalian species. J. Neural. Transm. Gen.Sect. 80:105-127.

Casini, G., Bingman, V.P., Bagnoli, P. (1986) Connections of the pigeon dorsomedial forebrainstudied with WGA-HRP and 3H-proline. J. Comp. Neurol. 245:454-470.

Castagna, C., Ball, G.F., Balthazart, J. (1997) Effects of dopamine agonists on appetitive andconsummatory male sexual behavior in Japanese quail. Pharmacol. Biochem. Behav.58:403-414.

Casto, J.M. and Ball, G.F. (1994) Characterization and localization of D1 dopamine receptorsin the sexually dimorphic vocal control nucleus, area X, and the basal ganglia ofEuropean starlings. J. Neurobiol. 25:767-780.

Cepeda, C., Colwell, C.S., Itri, J.N., Chandler, S.H., Levine, M.S. (1998) Dopaminergicmodulation of NMDA-induced whole cell currents in neostriatal neurons in slices:Contribution of calcium conductances. J. Neurophysiol. 79:82–94.

Cepeda, C., Radisavljevic, Z., Peacock, W., Levine, M.S., Buchwald, N.A. (1992) Differentialmodulation by dopamine of responses evoked by excitatory amino acids in human cortex.Synapse 11:330–341.

Chagnac-Amitai, Y. and Connors, B.W. (1989) Horizontal spread of synchronized activity inneocortex and its control by GABA mediated inhibition. J. Neurophysiol. 61:747-758.

Chagnac-Amitai, Y., Luhmann, H.J., Prince, D.A. (1990) Burst generating and regular spikinglayer 5 pyramidal neurons of rat neocortex have different morphological features. J.Comp. Neurol. 296:598-613.

Chen, Z., Fujii, S., Ken-Ichi, I., Kato, H., Kaneko, K., Miyakawa, H. (1995) Activation ofdopamine D1 receptors enhances long-term depression of synaptic transmission inducedby low frequency stimulation in rat hippocampal CA1 neurons. Neurosci. Lett. 188:95-198.

Cheng, H.C. and Long, J.P. (1974) Dopaminergic nature of apomorphine-induced pecking inpigeons. Eur. J. Pharmacol. 26:313-320.

Chesselet, M.F., Weiss, L., Wuenschell, C., Tobin, A.J., Affolter, H.U. (1987) Comparativedistribution of mRNAs for glutamic acid decarboxylase, tyrosine hydroxylase, andtachykinins in the basal ganglia: an in situ hybridization study in the rodent brain. J.Comp. Neurol. 262:125-40.

Clark, D. and White, F.J. (1987) Review: D1 dopamine receptor-the search for a function: acritical evaluation of the D1/D2 dopamine receptor classification and its functionalimplications. Synapse 1:347-388.

Cobb, S.R., Buhl, E.H., Halasy, K., Paulsen, O., Somogyi, P. (1995) Synchronization ofneuronal activity in hippocampus by individual GABAergic interneurons. Nature 378:75-78.

Cohen, D.H. (1975) Involvement of the avian amygdalar homologue (archistriatum posteriorand mediale) in defensively conditioned heart rate change. J. Comp. Neurol. 160:13-35.

Literature 210

Cohen, J.D., Perlstein, W.M., Braver, T.S., Nystrom, L.E., Noll, D.C., Jonides, J., Smith, E.E.(1997) Temporal dynamics of brain activation during a working memory task. Nature386:604-608.

Condé, F., Maire-Lepoivre, E., Audinat, E., Crepél, F. (1995) Afferent connections of themedial frontal cortex of the rat. II. Cortical and subcortical afferents. J. Comp. Neurol.352:567-593.

Connors, B.W., Benardo, L.S., Prince, D.A. (1983) Coupling between neurons of thedeveloping rat neocortex. J. Neurosci. 3:773-782.

Constantinidis, C. and Steinmetz, M.A. (1996) Neuronal activity in posterior parietal area 7aduring the delay periods of a spatial memory task. J. Neurophysiol. 76:1352-1355.

Cooper, J.R., Bloom, F.E., Roth, R.H. (1996) The Biochemical Basis of Neuropharmacology,7th ed. Oxford UP: New York.

Cooper, S.J. and Al-Naser, H.A. (1993) D1:D2 dopamine receptor interactions in relation tofeeding reponses and food intake. In: D1:D2 Dopamine Receptor Interactions, pp 203-234. Ed J.L. Waddington. Academic Press: Toronto.

Cornwall, J. and Phillipson, O.T. (1988) Afferent projections to the dorsal thalamus of the ratas shown by retrograde lectin transport - I. the mediodorsal nucleus. Neuroscience24:1035-1049.

Cortés, R., Gueye, B., Pazos, A., Probst, A., Palacios, J.M. (1989) Dopamine receptors in thehuman brain: Autoradiographic distribution of D1 sites. Neuroscience 28:263-273.

Cowan, R.L. and Wilson, C.J. (1994) Spontaneous firing patterns and axonal projections ofsingle corticostriatal neurons in the rat medial agranular cortex. J. Neurophysiol. 71:17-32.

Csillag, A., Szekely, A.D., Davies, D.C. (1994) Termination pattern of medial hyperstriatumventrale efferents in the archistriatum of the domestic chick. J. Comp. Neurol. 348:394-402.

D´Esposito, M., Postle, B.R., Ballard, D., Lease, J. (1999) Maintenance versus manipulation ofinformation held in memory: an event-related fMRI study. Brain. Cogn. 41:66-86.

Dahlström, A. and Fuxe, K. (1964) Evidence for the existence of monoamine-containingneurons in the central nervous system. Acta Physiol. Scand. 62:1-55.

Davies, D.C., Csillag, A., Szekely, A.D., Kabai, P. (1997) Efferent connections of the domesticchick archistriatum: A phaseolus lectin anterograde tracing study. J. Comp. Neurol.389:679-693.

Davies, D.C., Horn, G., McCabe, B.J. (1983) Changes in telencephalic catecholamine levels inthe domestic chick. Effects of age and visual experience. Brain. Res. 312:251-255.

Davies, D.C., Taylor, D.A., Johnson, M.H. (1988) The effects of hyperstriatal lesions on one-trial passive-avoidance learning in the chick. J. Neurosci. 8:4662-4666.

DeFelipe, J., Hendry, S.H.C., Jones, E.G. (1989) Visualization of chandelier cell axons byparvalbumin immunoreactivity in monkey cerebral cortex. Proc. Nat. Acad. Sci. 86:2093-2097.

Dehpour, A.R., Samini, M., Aliebrahimi, F., Chavoushzadeh, M.A. (1998) The effect of acutelithium and rubidium pretreatment on apomorphine-induced pecking in pigeons.Pharmacol. Toxicol. 82:147-152.

Delius, J.D. (1985) The peck of the pigeon: Free for all. In C.F. Lowe, M. Richelle, D.E.Blackman, C.M. Bradshaw (eds): Behaviour Analysis and Contemporary Psychology.London: Lawrence Erlbaum.

Literature 211

Delius, J.D. and Bennetto, K. (1972) Cutaneous sensory projections to avian forebrain. Brain.Res. 37:205-221.

Delius, J.D. and Emmerton, J. (1978) Stimulus-dependent asymmetry in classical andinstrumental discrimination learning by pigeons. Psychol. Rec. 28:425-434.

Delius, J.D., Runge, T.E., Oeckinghaus, H. (1979) Short- latency auditory projection to frontaltelencephalon of pigeon. Exp.Neurol. 63:594-609.

Deng, C. and Wang, B. (1992) Overlap of somatic and visual response areas in the Wulst ofpigeon. Brain Res. 582:320-322.

Deng, C. and Wang, B. (1993) Convergence of somatic and visual afferent impulses in theWulst of pigeon. Exp. Brain Res. 96:287-290.

Deutch, A.Y., Lee, M.C., Gillham, M.H., Cameron, D., Goldstein, M., Iadorola, M.J. (1991)Stress selectively increases Fos protein in dopamine neurons innervating the prefrontalcortex. Cereb. Cortex 1:273-292.

Deviche, P. (1984) Administration of small doses of apomorphine attenuates feeding in non-deprived pigeons. Physiol. Behav. 33:581-585.

Deviche, P. (1985) Behavioral response to apomorphin and its interaction with opiates indomestic pigeons. Pharmacol. Biochem. Behav. 22:209-214.

Dewar, K.M. and Reader, T.A. (1989) Distribution of dopamine D1 and D2 receptors in rabbitcortical areas, hippocampus, and neostriatum in relation to dopamine contents. Synapse4:378-386.

Di Pellegrino, G. and Wise, S.P. (1991) A neurophysiological comparison of three distinctregions of the primate frontal lobe. Brain 114:951-978.

Dias, R., Robbins, T.W., Roberts, A.C. (1997) Dissociable forms of inhibitory control withinprefrontal cortex with an analog of the Wisconsin card sort test: restriction to novelsituations and independence from "on-line" processing. J. Neurosci. 17:9285-9297.

Diekamp, B., Ruhm, A., Kelc, M., Güntürkün, O. (1998) Effects of local application of a D1receptor antagonist into the "prefrontal cortex" of pigeons in a reversal learning task.Göttingen Neurobiology Report II:537.

Dietl, M.M. and Palacios, J.M. (1988) Neurotransmitter receptors in the avian brain. I.Dopamine receptors. Brain Res. 439:354-359.

Divac, I. and Mogensen, J. (1985) The prefrontal "cortex" in the pigeon: Catecholaminehistofluorescence. Neuroscience 15:677-682.

Divac, I., Björklund, A., Lindvall, O., Passingham, R.E. (1978) Converging projections fromthe mediodorsal thalamic nucleus and mesencephalic dopaminergic neurons to theneocortex in three species. J. Comp. Neurol. 180:59-72.

Divac, I., Mogensen, J., Björklund, A. (1985) The prefrontal "cortex" in the pigeon.Biochemical evidence. Brain Res. 332:365-368.

Divac, I., Mogensen, J., Björklund, A. (1988) Strain differences in catecholamine content ofpigeon brains. Brain Res. 444:371-373.

Divac, I., Thibault, J., Skageberg, G., Palacios, J.M., Dietl, M.M. (1994) Dopaminergicinnervation of the brain in pigeons. The presumed 'prefrontal cortex'. Acta Neurobiol.Exp. (Warsz) 54:227-234.

Domenici, L., Waldvogel, H.J., Matute, C., Streit, P. (1988) Distribution of GABA-likeimmunoreactivity in the pigeon brain. Neuroscience 25:931-950.

Literature 212

Dose, J.M. and Zolman, J.F. (1994) Separate and combined effects of dopamine D1 and D2receptor agonists on key pecking in the developing chick. Pharmacol. Biochem. Behav.49:73-78.

Douglas, R.J. and Martin, K.A.C. (1992) In search of the canonical microcircuits of neocortex.In: The Visual System from Genesis to Maturity (Lent R ed.) pp 213-232 Boston:Birkhäuser.

Doyère, V., Burette, F., Negro, C.R., Laroche, S. (1993) Long-term potentiation of hippocampalafferents and efferents to prefrontal cortex: Implications for associative learning.Neuropsychologia 31:1031-1053.

Dubbeldam, J.L. and Visser, A.M. (1987) The organization of the nucleus basalis-neostriatumcomplex of the mallard (Anas Platyhrynchos L.) and its connections with thearchistriatum and the paleostriatal complex. Neuroscience 21:487-517.

Dubbeldam, J.L., den Boer-Visser, A.M., Bout, R.G. (1997) Organization and efferentconnections of the archistriatum of the mallard (Anas platyrhynchos L.): An anterogradeand retrograde tracing study. J. Comp. Neurol. 388:632-657.

Dube, L., Smith, A.D., Bolam, J.P. (1988) Identification of synaptic terminals of thalamic orcortical origin in contact with distinct medium-size spiny neurons in the rat neostriatum. J.Comp. Neurol. 267:455-471.

Durand, S.E., Heaton, J.T., Amateau, S.K., Brauth, S.E. (1997) Vocal control pathways throughthe anterior forebrain of a parrot (Melopsittacus undulatus). J Comp. Neurol. 377:179-206.

Durstewitz, D. and Güntürkün, O. (1996) The possible function of dopamine in associativelearning: A computational model. In Artificial Neural Networks - ICANN 96, ConferenceProceedings, Lecture Notes in Computer Science (Eds. von der Malsburg C., von SeelenW., Vorbrüggen, J.C. and Sendhoff, B.), pp 667-672. Springer, Berlin, N.Y.

Durstewitz, D., Kelc, M., Güntürkün, O. (1999) A neurocomputational theory of thedopaminergic modulation of working memory functions. J. Neurosci. 19:2807-2822.

Durstewitz, D., Kröner, S., Hemmings, H.C., Güntürkün, O. (1998) The dopaminergicinnervation of the pigeon telencephalon: Distribution of DARPP-32 and co-occurrencewith glutamate decarboxylase and tyrosine hydroxylase. Neuroscience 83:763-779.

Durstewitz, D., Seamans, J.K., Sejnowski, T.J. (2000) Dopamine-mediated stabilization ofdelay-period activity in a network model of prefrontal cortex. J. Neurophysiol. 83: 1733-1750.

Dutar, P., Vu, H.M., Perkel, D.J. (1998) Multiple cell types distinguished by physiological,pharmacological, and anatomic properties in nucleus HVc of the adult zebra finch. J.Neurophysiol. 80:1828-1838.

Edinger, L. (1903) Untersuchungen über die vergleichende Anatomie des Gehirnes. 5. DasVorderhirn der Vögel. Abh. Senckenbergischen Naturforsch. Ges. 20:343-426.

Epstein, R., Kirshnit, C.E., Lanza, R.P., Rubin, L.C. (1984) 'Insight' in the pigeon: Antecedentsand determinants of an intelligent performance. Nature 308:61-62.

Erichsen, J.T., Bingman, V.P., Krebs, J.R. (1991) The distribution of neuropeptides in thedorsomedial telencephalon of the pigeon (Columba livia): A basis for regionalsubdivisions. J. Comp. Neurol. 314:478-492.

Fallon, J.H. and Loughlin, S.E. (1995) Substantia nigra. In: The Rat Nervous System, pp 215-237. Ed G. Paxinos. Academic Press: San Diego.

Feldman, S.G. and Kruger, L. (1980) Axonal transport study of ascending projection of mediallemniscal neurons in rat. J. Comp. Neurol. 192:427-454.

Literature 213

Ferrari, F. and Giuliani, D. (1993) Behavioural effects induced in rats and chicks by D2dopamine agonists. Physiol. Behav. 54:695-700.

Ferron, A., Thierry, A.M., Le Douarin, C., Glowinski, J. (1984) Inhibitory influence of themesocortical dopaminergic system on spontaneous activity or excitatory response inducedfrom the thalamic mediodorsal nucleus in the rat medial prefrontal cortex. Brain Res.302:257-265.

Fibiger, H.C. (1978) Drugs and reinforcement mechanisms: A critical review of thecatecholamine theory. Ann. Rev. Pharmacol. Toxicol. 18:37-56.

Fienberg, A.A., Hiroi, N., Mermelstein, P.G., Song, W., Snyder, G.L., Nishi, A., Cheramy, A.,O’Callaghan, J.P., Miller, D.B., Cole, D.G., Corbett, R., Haile, C.N., Cooper, D.C., Onn,S.P., Grace, A.A., Ouimet, C.C., White, F.J., Hyman, S.E., Surmeier, D.J., Girault, J.,Nestler, E.J., Greengard, P. (1998) DARPP-32: Regulator of the efficacy ofdopaminergic neurotransmission. Science 281:838-842.

Floresco, S.B., Seamans, J.K., Phillips, A.G. (1997) Selective roles for hippocampal,prefrontal cortical, and ventral striatal circuits in radial-arm maze tasks with or without adelay. J. Neurosci. 17:1880-1890.

Fortune, E.S. and Margoliash, D. (1995) Parallel pathways and convergence onto HVc andadjacent neostriatum of adult zebra finches (Taeniopygia guttata). J. Comp. Neurol.360:413-441.

Foster, E.F., Mehta, R.P., Bottjer, S.W. (1997) Axonal connections of the medial magnocellularnucleus of the anterior neostriatum in zebra finches. J. Comp. Neurol. 382:364-381.

Franceschetti, S., Guatteo, E., Panzica, F., Sancini, G., Wanke, E., Avanzini, G. (1995) Ionicmechanisms underlying burst firing in pyramidal neurons: Intracellular study in ratsensorimotor cortex. Brain. Res. 696:127-139.

Friedman, A. and Gutnick, M.J. (1987) Low-threshold calcium electrogenesis in neocorticalneurons. Neurosci. Lett. 81:117-122.

Funahashi, S. and Kubota, K. (1994) Working memory and prefrontal cortex. Neurosci. Res.21:1-11.

Funahashi, S., Bruce, C.J., Goldman-Rakic, P.S. (1989) Mnemoniccoding of visual space in themonkey’s dorsolateral prefrontal cortex. J. Neurophysiol. 61:331–349.

Funke, K. (1989) Somatosensory areas in the telencephalon of the pigeon II: Spinal pathwaysand afferent connections. Exp. Brain Res. 76:620-638.

Fuster, J.M. (1973) Unit activity in prefrontal cortex during delayed-response performance:Neuronal correlates of transient memory. J. Neurophysiol. 36:61-78.

Fuster, J.M. (1989) The Prefrontal Cortex: Anatomy, Physiology, and Neuropsychology ofthe Frontal Lobe. 2 ed.Raven Press: New York.

Fuster, J.M. and Alexander, G.E. (1971) Neuron activity related to short term memory. Science173:652-654.

Gagliardo, A. and Divac, I. (1993) Effects of ablation of the presumed equivalent of themammalian prefrontal cortex on pigeon homing. Behav. Neurosci. 107:280-288.

Gagliardo, A., Bonadonna, F., Divac, I. (1996) Behavioural effects of ablations of thepresumed 'prefrontal cortex' or the corticoid in pigeons. Behav. Brain Res. 78:155-162.

Gagliardo, A., Mazzotto, M., Divac, I. (1997) Memory of radial maze behavior in pigeons afterablations of the presumed equivalent of mammalian prefrontal cortex. Behav.Neurosci.111:955-962.

Galifret, Y. (1968) Les diverses aires functionelles de la retine de pigeon. Z. Zellforsch.Mikrosk. Anat. 86:535-545.

Literature 214

Gamlin, P.D.R. and Cohen, D.H. (1986) A second ascending visual pathway from the optictectum to the telencephalon in the pigeon (Columba livia). J. Comp. Neurol. 250:296-310.

Gaspar, P., Bloch, B., LeMoine, C. (1995) D1 and D2 receptor gene expression in the ratfrontal cortex: Cellular localization in different classes of efferent neurons. Eur. J.Neurosci. 7:1050-1063.

Geijo-Barrientos, E. (2000) Subthreshold inward membrane currents in guinea-pig frontalcortex neurons. Neuroscience 95:965-72.

Geijo-Barrientos, E. and Pastore, C. (1995) The effects of dopamine on the subthresholdelectrophysiological responses of rat prefrontal cortex neurons in vitro. Eur. J. Neurosci.7:358-366.

Gerfen, C. R. and Sawchenko, P. E. (1983) An anterograde neuroanatomical tracing method thatshows the detailed morphology of neurons, their axons and terminals:Immunohistochemical localization of an axonally transported plant lectin (Phaseolusvulgaris leucoagglutinin). Brain Res. 290:219-238.

Gessa, G.L., Porceddu, M.L., Collu, M., Mereu, G., Serra, M., Ongini, E., Biggio, G. (1985)Sedation and sleep induced by high doses of apomorphine after blockade of D-1receptors by SCH 23390. Eur. J. Pharmacol. 109:269-74.

Giguere, M. and Goldman-Rakic, P.S. (1988) Mediodorsal nucleus: areal, laminar andtangential distribution of afferents and efferents in the frontal lobe of rhesus monkeys. J.Comp. Neurol. 277:195-213.

Glowinski, J., Tassin, J.P., Thierry, A.M. (1984) The mesocortico-prefrontal dopaminergicneurons. Trends Neurosci. 7:415-418.

Godbout, R., Mantz, J., Pirot, S., Glowinski, J., Thierry, A.-M. (1991) Inhibitory influence ofthe mesocortical dopaminergic neurons on their target cells: electrophysiological andpharmacological characterization. J. Pharmacol. Exp. Therap. 258:728-738.

Goldman-Rakic, P.S. (1987) Circuitry of primate prefrontal cortex and the regulation ofbehavior by representational memory. In F. Plum and V.B. Mountcastle (Eds): Handbookof Physiology, Vol 5; Higher Cortical Function. Washington, DC: American PhysiologySociety, pp 373-417.

Goldman-Rakic, P.S. (1988) Topography of cognition: parallel distributed networks in primateassociation cortex. Annu. Rev. Neurosci. 11:137-156.

Goldman-Rakic, P.S. (1995) Toward a circuit model of working memory and the guidance ofvoluntary motor action. In: Model of Information Processing in the Basal Ganglia, pp131-148. Eds J.C. Houk, J.L. Davis, D.G. Beiser. MIT Press: Cambridge, Mass.

Goldman-Rakic, P.S. (1996) Regional and cellular fractionation of working memory. Proc.Nat. Acad. Sci. 93:13473–13480.

Goldman-Rakic, P.S. and Porrino, L.J. (1985) The primate mediodorsal (MD) nucleus and itsprojection to the frontal lobe. J. Comp. Neurol. 242:535-560.

Goldman-Rakic, P.S., Leranth, C., Williams, S.M., Mons, N., Geffard, M. (1989) Dopaminesynaptic complex with pyramidal neurons in primate cerebral cortex. Proc. Natn. Acad.Sci. USA 86:9015-9019.

Goldman-Rakic, P.S., Lidow, M.S., Smiley, J.F., Williams, M.S. (1992) The anatomy ofdopamine in monkey and human prefrontal cortex. J. Neural. Transm. Suppl. 36:163-77.

Goldman-Rakic, P.S., Selemon, L.D., Schwartz, M.L. (1984) Dual pathways connecting thedorsolateral prefrontal cortex with the hippocampal formation and parahippocampalcortex in the rhesus monkey. Neuroscience 12:719-743.

Literature 215

González-Burgos, G., Barrionuevo, G., Lewis, D.A. (2000)Horizontal synaptic connections inmonkey prefrontal cortex:an in vitro electrophysiological study. Cereb. Cortex 10:82-92.

Goodman, J., Zacny, J., Osman, A., Azzaro, A., Donovan, C. (1983) Dopaminergic nature offeeding-induced behavioral stereotypies in stressed pigeons. Pharmacol. Biochem.Behav. 18:153-158.

Gorelova, N. and Yang, C.R. (1997) Dopamine D1 receptor stimulation modulates a slowlyinactivating Na+ current in layer V-VI prefrontal cortical (PFC) neurons. Soc. Neurosci.Abstr. 23:1771.

Groenewegen, H.J. (1988) Organization of the afferent connections of the mediodorsal thalamicnucleus in the rat, related to the mediodorsal-prefrontal topography. Neuroscience24:379-431.

Groenewegen, H.J., Berendse, H.W., Wolters, J.G., Lohman, A.H. (1990) The anatomicalrelationship of the prefrontal cortex with the striatopallidal system, the thalamus and theamygdala: Evidence for a parallel organization. Prog. Brain. Res. 85:95-118.

Groves, P.M., Garcia-Munoz, M., Linder, J.C., Manley, M.S., Martone, M.E., Young, S.J.(1995) Elements of the intrinsic organization and information processing in theneostriatum. In Models of Information Processing in the Basal Ganglia (eds. Houk J.C.,Davis J.L. and Beiser D.G.), pp 51-96. MIT Press, Cambridge, Massachussets.

Gruss, M. and Braun, K. (1996) Stimulus-evoked increase of glutamate in the mediorostralneostriatum/hyperstriatum ventrale of domestic chick after auditory filial imprinting: an invivo microdialysis study. J. Neurochem. 66:1167-1173.

Gruss, M. and Braun, K. (1997) Distinct activation of monoaminergic pathways in chick brainin relation to auditory imprinting and stressful situations: A microdialysis study.Neuroscience 76:891-899.

Gulledge, A.T. and Jaffe, D.B. (1998) Dopamine decreases the excitability of layer Vpyramidal cells in the rat prefrontal cortex. J. Neurosci. 18:9139-9151.

Güntürkün, O. (1984) Evidence for a third primary visual area in the telencephalon of thepigeon. Brain Res. 294:247-254.

Güntürkün, O. (1991) The functional organization of the avian visual system. In: Neural andBehavioural Plasticity, pp 92-105. Ed R.J. Andrew. Oxford University Press: Oxford.

Güntürkün, O. (1997) Cognitive impairments after lesions of the neostriatum caudolaterale andits thalamic afferent in pigeons: Functional equivalencies to the mammalian prefrontalsystem? J. Brain Res. 1:133-144.

Güntürkün, O. (1998) Sensory Physiology: Vision. In G.C. Whittow (ed): Sturkie's AvianPhysiology. Orlando: Academic Press, pp 1-19.

Güntürkün, O. and Durstewitz, D. (in press) Multimodal areas of the avian forebrain -blueprints for cognition? In: Evolution of Brain and Cognition. Eds G. Roth, M.Wullimann. Scientific American Press: Heidelberg.

Güntürkün, O. and Hahmann, U. (1994) Cerebral asymmetries and visual acuity in pigeons.Behav. Brain Res. 60:171-175.

Güntürkün, O. and Hahmann, U. (1999) Functional subdivisions of the ascending visualpathways in the pigeon. Behav. Brain Res. 98:193-201.

Güntürkün, O. and Karten, H.J. (1991) An immunocytochemical analysis of the lateralgeniculate complex in the pigeon (Columba livia). J. Comp. Neurol. 314:721-749.

Güntürkün, O. and Kröner, S. (1999) A polysensory pathway to the forebrain of the pigeon: Theascending projections of the n. dorsolateralis posterior thalami (DLP). Eur. J. Morphol.37:124-128.

Literature 216

Güntürkün, O. and Remy, M. (1990) The topographical projection of the n. isthmi parsparvocellularis (Ipc) onto the tectum opticum in the pigeon. Neurosci. Lett. 111:18-22.

Güntürkün, O., Miceli, D., Watanabe, M. (1993) Anatomy of the avian thalamofugal pathway.In Vision, Brain and Behaviour in Birds (eds. Zeigler H.P. and Bischof H.-J.), pp 99-101. MIT, Cambridge.

Haberly, L.B. and Bower, J.M. (1989) Olfactory cortex: model circuit for study of associativememory? Trends Neurosci. 12:258-264.

Hablitz, J.J. (1986) Mechanisms regulating synchronization of hippocampal epileptiformactivity. In: Calcium electrogenesis and neuronal functioning (Heinemann U, Klee M,Neher E, Singer W. eds) pp 337-344, Berlin, Springer-Verlag.

Hahmann, U. and Güntürkün, O. (1993) The visual acuity for the lateral visual field in thepigeon (Columba livia). Vision Res. 33:1659-1664.

Haj-Dahmane, S. and Andrade, R. (1997) Calcium-activated cation nonselective currentcontributes to the fast afterdepolarization in rat prefrontal cortex neurons. J.Neurophysiol. 78:1983-1989.

Hall, W.S., Cohen, P.L., Brauth, S.E. (1993) Auditory projections to the anterior telencephalonin the budgerigar (Melopsittacus undulatus). Brain Behav. Evol. 41:97-116.

Hamill, O.P., Huguenard, J.R., Prince, D.A.. (1991) Patch-clamp studies of voltage-gatedcurrents in identified neurons of the rat cerebral cortex. Cereb. Cortex 1:48-61.

Harding, C.F., Barclay, S.R., Waterman, S.A. (1998) Changes in catecholamine levels andturnover rates in hypothalamic, vocal control, and auditory nuclei in male zebra finchesduring development. J. Neurobiol. 34:329-346.

Hartmann, B. and Güntürkün, O. (1998) Selective deficits in reversal learning after neostriatumcaudolaterale lesions in pigeons - possible behavioral equivalencies to the mammalianprefrontal system. Behav. Brain Res. 96:125-133.

Haxby, J.V., Petit, L., Ungerleider, L.G., Courtney, S.M. (2000) Distinguishing the functionalroles of multiple regions in distributed neural systems for visual working memory.Neuroimage 11:145-156.

Hellmann, B. and Güntürkün, O. (1997) Two ‘Visual Worlds’ in the pigeon: Segregated spacerepresentations in the ascending visual projections. In N. Elsner and H. Wässle (Eds):Göttingen Neurobiology Report 1997. Stuttgart: Thieme, p. 533.

Hemmings, H.C. and Adamo, A.I.B. (1997) Effect of halothane on conventional protein kinaseC translocation and down-regulation in rat cerebrocortical synaptosomes. Br. J. Anaesth.78:189-196.

Hemmings, H.C. and Greengard, P. (1986) DARPP-32, a dopamine- and adenosine 3':5'-monophosphate-regulated phosphoprotein: Regional, tissue, and phylogeneticdistribution. J. Neurosci. 6:1469-1481.

Hemmings, H.C. Jr., Nairn, A.C., Bibb, J.A., Greengard, P. (1995) Signal transduction in thestriatum: DARPP-32, a molecular integrator of multiple signaling pathways. In:Molecular and Cellular Mechanisms of Neostriatal Function, pp 283-297. Eds M.A.Ariano, D.J. Surmeier. R.G.Landes: Austin.

Hemmings, H.C. Jr., Walaas, S.I., Ouimet, C.C., Greengard, P. (1987a) Dopaminergicregulation of protein phosphorylation in the striatum: DARPP-32. Trends Neurosci.10:377-383.

Hemmings, H.C. Jr., Walaas, S.I., Ouimet, C.C., Greengard, P. (1987b) Dopamine receptors:Regulation of protein phosphorylation. In: Dopamine Receptors, pp 115-151. Eds I.Creese, C.M. Fraser. Alan R. Liss, Inc.: New York.

Literature 217

Hemmings, H.C., Greengard, P., Tung, H.Y.L., Cohen, P. (1984) DARPP-32, a dopamine-regulated neuronal phosphoprotein, is a potent inhibitor of protein phosphatase-1. Nature310:503-505.

Hernández-López, S., Bargas, J., Surmeier, D.J., Reyes, A., Galarraga, E. (1997) D1 receptoractivation enhances evoked discharge in neostriatal medium spiny neurons by modulatingan L-type Ca2+ conductance. J. Neurosci. 17:3334-3342.

Hicks, T.P, Stark, C.A, Fletcher, W.A. (1986) Origins of afferents to visual suprageniculatenucleus of the cat. J. Comp. Neurol. 246:544-554.

Higashi, H., Tanaka, E., Inokuchi, H., Nishi, S. (1993) Ionic mechanisms underlying thedepolarizing and hyperpolarizing afterpotentials of single spike in guinea-pig cingulatecortical neurons. Neuroscience 55:129-138.

Hof, P.R., Cox, K., Morrison, J.H. (1990) Quantitative analysis of a vulnerable subset ofpyramidal neurons in Alzheimer’s disease: I Superior frontal and inferior temporalcortex. J. Comp. Neurol .301:44-63.

Honey, R.C. and Bateson, P. (1996) Stimulus comparison and perceptual learning: furtherevidence and evaluation from an imprinting procedure. Q. J. Exp. Psychol. B. 49:259-269.

Horn, G. (1998) Visual imprinting and the neural mechanisms of recognition memory. TrendsNeurosci. 21:300-305.

Hutcheon, B., Miura, R.M., Puil, E. (1996) Subthreshold membrane resonance in neocorticalneurons. J. Neurophysiol. 76:683-697.

Irle, E., Markowitsch, H.J., Streicher, M. (1984) Cortical and subcortical, including sensory-related, afferents to the thalamic mediodorsal nucleus of the cat. J. Hirnforsch. 25:29-51.

Iversen, S.G. and Mishkin, M. (1970) Perseverative interference in monkeys followingselective lesions ofthe inferior prefrontal convexity. Exp. Brain Res. 11:376-386.

Iyengar, S., Viswanathan, S.S., Bottjer, S.W. (1999) Development of topography within songcontrol circuitry of zebra finches during the sensitive period for song learning. J.Neurosci. 19:6037-6057.

Jacobsen, C.F. (1935) Functions of the frontal association area in primates. Arch. Neurol.Psych. 33:558-569.

Jarvie, K.R., Tiber, M., Silvia, C., Gingrich, J.A., Caron, M.G. (1993) Molecular cloning,stable expression, and desensitization of the human dopamine D1B/D5 receptor. J.Receptor. Res. 13:573-590

Jay, T.M., Glowinski, J., Thierry, A.M. (1995) Inhibition of hippocampo-prefrontal cortexexcitatory responses by the mescocortical DA system. Neuroreport 6:1845-1848.

Johnson, F., Sablan, M.M., Bottjer, S.W. (1995) Topographic organization of a forebrainpathway involved with vocal learning in zebra finches. J. Comp. Neurol. 358:260-278.

Jones, E.G. (1985) The Thalamus. New York:Plenum Press.

Jones, E.G. and Powell, T.P.S. (1970) An anatomical study of converging sensory pathwayswithin the cerebral cortex of the monkey. Brain 93:793-820.

Joyce, J.N., Goldsmith, S., Murray, A. (1993) Neuroanatomical localization of D1 versus D2receptors: Similar organization in the basal ganglia of the rat, cat and human anddisparate organization in the cortex and limbic system. In D1:D2 Dopamine ReceptorInteractions (ed. J.L. Waddington), pp 23-49. Academic Press, London.

Literature 218

Juorio, A.V. (1983) Presence, distribution, storage and metabolism of dopamine and relatedmonoamines in the avian brain. In: Progress in Nonmammalian Brain Research. pp 63-80. Eds G. Nisticò and L. Bolis. CRC Press, Boca Raton..

Juorio, A.V. and Vogt, M. (1967) Monoamines and their metabolites in the avian brain. J.Physiol. (Lond) 189:489-518.

Källén, B. (1953) On the nuclear differentiation during ontogenesis in the avian forebrain andsome notes on the amniote strio-amygdaloid complex. Acta Anatomica 17:72-84.

Källén, B. (1962) Embryogenesis of brain nuclei in the chick telencephalon. Ergeb. Anat.Entwicklungsgesch. 36:62-82.

Kalt, T., Diekamp, B., Güntürkün, O. (1999) Single unit activity during a Go/NoGo task in the"prefrontal cortex" of pigeons. Brain. Res. 839:263-278.

Kang, Y. and Kayano, F. (1994) Electrophysiological and morphological characteristics oflayer VI pyramidal cells in the cat motor cortex. J. Neurophysiol. 72:578-591.

Karle, E.J., Anderson, K.D., Medina, L., Reiner, A. (1996) Light and electron microscopicimmunohistochemical study of dopaminergic terminals in the striatal portion of the pigeonbasal ganglia using antisera against tyrosine hydroxylase and dopamine. J. Comp. Neurol.369:109-124.

Karle, E.J., Anderson, K.D., Reiner, A. (1992) Ultrastructural double-labeling demonstratessynaptic contacts between dopaminergic terminals and substance P-containing striatalneurons in pigeons. Brain Res. 572:303-309.

Karle, E.J., Anderson, K.D., Reiner, A. (1994) Dopaminergic terminals form synaptic contactswith enkephalinergic striatal neurons in pigeons: an electron microscopic study. BrainRes. 646:149-156.

Karten, H.J. (1968) The ascending auditory pathway in the pigeon (Columba livia). II.Telencephalic projections of the nucleus ovoidalis thalami. Brain Res. 11:134-153.

Karten, H.J. (1969) The organization of the avian telencephalon and some speculations on thephylogeny of the amniote telencephalon. Ann. N.Y. Acad. Sci. 167:164-179.

Karten, H.J. (1971) Efferents projections of the Wulst of the owl. Anat. Rec. 169:353.

Karten, H.J. and Dubbeldam, J.L. (1973) The organization and projections of the paleostriatalcomplex in the pigeon (Columba livia). J. Comp. Neurol. 148:61-90.

Karten, H.J. and Hodos, W. (1967) Stereotaxic Atlas of the Brain of the Pigeon (Columbalivia). Baltimore:John Hopkins Press.

Karten, H.J. and Hodos, W. (1970) Telencephalic projections of the nucleus rotundus in thepigeon (Columba livia). J. Comp. Neurol. 140:35-52.

Karten, H.J. and Shimuzu, T. (1989) The origins of neocortex: Connections and lamination asdistinct events in evolution. J. Cog. Neurosci. 1:291-301.

Karten, H.J., Hodos, W., Nauta, W.H.J., Revzin, A.M. (1973) Neural connections of the “visualWulst” of the avian telencephalon. Experimental studies in the pigeon (Columba livia)and the owl (Speotyto cunicularia). J. Comp. Neurol. 150:253-277.

Kasper, E.M., Larkman, A.U., Lübke, J., Blakemore, C. (1994a) Pyramidal neurons in layer 5of the rat visual cortex. I. Correlation among cell morphology, intrinsicelectrophysiological properties, and axon targets. J. Comp. Neurol. 339:459-474.

Kasper, E.M., Larkman, A.U., Lübke, J., Blakemore, C. (1994b) Pyramidal neurons in layer 5of the rat visual cortex. II. Development of electrophysiological properties. J. Comp.Neurol. 339:475-494.

Literature 219

Katoh, Y.Y. and Benedek, G. (1995) Organization of the colliculo-suprageniculate pathway inthe cat: a wheat germ agglutinin-horseradish peroxidase study. J. Comp. Neurol.352:381-397.

Kawaguchi, Y. (1995) Physiological subgroups of nonpyramidal cells with specificmorphological characteristics in layer II/III of rat frontal cortex. J. Neurosci. 15:2638-2655.

Kawaguchi, Y. and Kubota, Y. (1993) Correlation of physiological subgroupings ofnonpyramidal cells with Parvalbumin- and Calbindin D28k-immunoreactive neurons inlayer V of rat frontal cortex. J. Neurophysiol. 70:387-396.

Kebabian, J.W. and Calne, D.B. (1979) Multiple receptors for dopamine. Nature 277:93-96.

Kelley, D.B. and Nottebohm, F. (1979) Projections of a telencephalic auditory nucleus-field Lin the canary. J. Comp. Neurol. 183:455-469.

Kesner, R.P., Hunt, M.E., Williams, J.M., Long, J.M. (1996) Prefrontal cortex and workingmemory for spatial response, spatial location, and visual object information in the rat.Cereb. Cortex 6:311-318.

Kimberg, D.Y. and Farah, M.J. (1993) A unified account of cognitive impairments followingfrontal lobe damage: The role of working memory in complex, organized behavior. J.Exp. Psychol. Gen. 122:411-428.

Kita, H. and Kitai, S.T. (1988) Glutamate decarboxylase immunoreactive neurons in ratneostriatum: their morphological types and populations. Brain Res. 447:346-52.

Kita, H., Oda, K., Murase, K. (1999) Effects of dopamine agonists and antagonists on opticalresponses evoked in rat frontal cortex slices after stimulation of the subcortical whitematter. Exp. Brain Res. 125:383-388.

Kitai, S.T. and Surmeier, D.J. (1993) Cholinergic and dopaminergic modulation of potassiumconductances in neostriatal neurons. Adv. Neurol. 60:40-52.

Kitt, C.A. and Brauth, S.E. (1981) Projections of the paleostriatum upon the midbraintegmentum in the pigeon. Neuroscience 6:1551-1566.

Kitt, C.A. and Brauth, S.E. (1982) A paleostriatal-thalamic-telencephalic path in pigeons.Neuroscience 7:2735-2751.

Kitt, C.A. and Brauth, S.E. (1986) Telencephalic projections from midbrain and isthmal cellgroups in the pigeon. II. The nigral complex. J. Comp. Neurol. 247:92-110.

Knudsen, E.I. and Knudsen, P.F. (1996) Disruption of auditory spatial working memory byinactivation of the forebrain archistriatum in barn owls. Nature 383:428-431.

Knudsen, E.I., Cohen, Y.E., Masino, T. (1995) Characterization of a forebrain gaze field in thearchistriatum of the barn owl: Microstmulation and anatomical connections. J. Neurosci.15:5139-5151.

Köhler, C., Ericson, H., Radesäter, A.C. (1991) Different laminar distributions of dopamine D1

and D2 receptors in the rat hippocampal region. Neurosci. Lett. 126:107-109.

Kolb, B. (1984) Functions of the frontal cortex of the rat: Acomparative review. Brain. Res.Rev. 8:65–98.

Kondo, T. (1933) Über die Verbindung des Sehhügels mit dem Streifenhügel beim Huhn.Okayama.J.Z. 45:797-817.

Korzeniewska, E. (1987) Multisensory convergence in the thalamus of the pigeon (Columbalivia). Neurosci. Lett. 80:55-60.

Korzeniewska, E. and Güntürkün, O. (1990) Sensory properties and afferents of the n.dorsolateralis posterior thalami (DLP) of the pigeon. J. Comp. Neurol. 292:457-479.

Literature 220

Kostal, L. and Savory, C.J. (1994) Influence of pharmacological manipulation of dopamine andopioid receptor subtypes on stereotyped behaviour of restricted-fed fowls. Pharmacol.Biochem. Behav. 48:241-52.

Krebs, J.R., Erichsen, J.T., Bingman, V.P. (1991) The distribution of neurotransmitters andneurotransmitter-related enzymes in the dorsomedial telencephalon of the pigeon(Columba livia). J. Comp. Neurol. 314:467-477.

Krimer, L.S., Jakab, R.L., Goldman-Rakic, P.S. (1997) Quantitative three-dimensional analysisof the catecholaminergic innervation of identified neurons in the macaque prefrontalcortex. J. Neurosci. 17:7450-7461.

Kritzer, M.F. and Goldman-Rakic, P.S. (1995) Intrinsic circuit organizationof the major layersand sublayers of the dorsolateral prefrontal cortex in the rhesus monkey. J. Comp.Neurol. 359:131-143

Kröner, S. and Güntürkün, O. (1999) Afferent and efferent connections of the caudolateralneostriatum in the pigeon (Columba livia): A retro- and anterograde pathway tracingstudy. J. Comp. Neurol. 407:228-260.

Kubota, M. and Saito, N. (1991) Sodium- and calcium-dependent conductances of neurones inthe zebra finch hyperstriatum ventrale pars caudale in vitro. J. Physio. 440:131-142.

Kubota, M. and Taniguchi, I. (1998) Characteristics of classes of neuron in the HVc of thezebra finch. J. Neurophysiol. 80: 914-923.

Kuhlenbeck, H. (1938) The ontogenetic development and phylogenetic significance of thecortex telencephali in the chick. J. Comp. Neurol. 69:273-301.

Kusunoki, T. (1969) The chemoarchitectonics of the avian brain. J. Hirnforsch. 11:477-497.

Lancaster, B. and Adams, P.R. (1986) Calcium-dependent current generating theafterhyperpolarization of hippocampal neurons. J. Neurophysiol. 55:1268-1282.

Lanza, R.P., Starr, J., Skinner, B.F. (1982) "Lying" in the pigeon. J. Exp. Anal. Behav. 38:201-203.

Larkman, A. and Mason, A. (1990) Correlations between morphology and electrophysiology ofpyramidal neurons in slices of rat visual cortex. I. Establishment of cell classes. J.Neurosci. 10:1407-1414.

Lashley, K.S. (1950) In search of the engram. Symp. Soc. Exp. Biol. 4:454–482.

Law-Tho, D., Desce, J.M., Crepel, F. (1995) Dopamine favours the emergence of long-termdepression versus long-term potentiation in slices of rat prefrontal cortex. Neurosci. Lett.188:125-128.

Law-Tho, D., Hirsch, J.C., Crepel, F. (1994) Dopamine modulation of synaptic transmission inrat prefrontal cortex: An in vitro electrophysiological study. Neurosci. Res. 21:151-160.

LeMoal, M. and Simon, H. (1991) Mesocorticolimbic dopaminergic network: Functional andregulatory roles. Physiol. Rev. 71:155-234.

LeMoine, C. and Gaspar, P. (1998) Subpopulations of cortical GABAergic interneurons differby their expression of D1 and D2 dopamine receptor subtypes. Molec. Brain. Res.58:231-236.

Leutgeb, S., Husband, S., Riters, L.V., Shimizu, T., Bingman, V.P. (1996) Telencephalicafferents to the caudolateral neostriatum of the pigeon. Brain Res. 730:173-181.

Levitt, J.B., Lewis, D.A., Yoshioka, T., Lund, J. (1993) Topography of pyramidal neuronintrinsic connections in macaque monkey prefrontal cortex (area 9 and 46). J. Comp.Neurol. 338:360–376.

Literature 221

Lewis, D.A., Campbell, M.J., Foote, S.L., Morrison, J.H. (1986) The monoamminergicinnervation of primate cortex. Human Neurobiol. 5:181-188.

Lewis, D.A., Hayes, T.L., Lund, J.S., Oeth, K.M. (1992) Dopamine and the neural circuitry ofprimate prefrontal cortex: Implications for schizophrenia research.Neuropsychopharmacol. 6:127-134.

Lewis, J.W., Ryan, S.M., Arnold, A.P., Butcher, L.L. (1981) Evidence for a catecholaminergicprojection to area X in the zebra finch. J. Comp. Neurol. 196:347-54.

Lidow, M.S., Goldman-Rakic, P.S., Gallager, D.W., Geschwind, D.H., Rakic, P. (1989).Distribution of major neurotransmitter receptors in the motor and somatosensory cortex ofthe rhesus monkey. Neuroscience 32:609-627.

Lidow, M.S., Wang, F., Cao, Y., Goldman-Rakic, P.S. (1998) Layer V neurons bear themajority of mRNAs encoding the five distinct dopamine receptor subtypes in the primateprefrontal cortex. Synapse 28:10–20.

Lidow, M.S.,Goldman-Rakic, P.S., Gallager, D.W., Rakic, P. (1991) Distribution ofdopaminergic receptors in the primate cerebral cortex: Quantitative autoradiographicanalysis using [³H]raclopride, [³H]spiperone and [³H]SCH23390. Neuroscience 40:657-671.

Lindenblatt, U. and Delius, J.D. (1987) Apomorphine induced pecking in pigeons classicallyconditioned to environmental cues. Psychopharmacol. 93:223-225.

Lindenblatt, U. and Delius, J.D. (1988) Nucleus basalis prosencephali, a substrate ofapomorphine-induced pecking in pigeons. Brain Res. 453:1-8.

Ljungberg, T., Apicella, P., Schultz, W. (1992) Responses of monkey dopamine neurons duringlearning of behavioral reactions. J. Neurophysiol. 67:145-163.

Llinás, R. and Jahnsen, H. (1982) Electrophysiology of mammalian thalamic neurones in vitro.Nature 297:406-408.

Llinás, R. and Yarom, Y. (1981) Properties and distribution of ionic conductances generatingelectroresponsiveness of mammalian inferior olivary neurones in vitro. J. Physiol.315:569-584.

LoLordo, V.M. and Furrow, D.R. (1976) Control by the auditory or the visual element of acompound stimulus: Effects of feedback. J. Exp. Anal. Behav. 25:251-256.

Lorenz, K. (1935) Der Kumpan in der Umwelt des Vogels. J. Ornithol. 83:137-213.

Lowndes, M. and Davies, D.C. (1995) The effect of archistriatal lesions on 'open field' andfear/avoidance behaviour in the domestic chick. Behav. Brain Res. 722:25-32.

Lubow, R.E. (1974) High-order concept formation in the pigeon. J. Exp. Anal. Behav. 21:475-483.

Madison, D.V. and Nicoll, R.A. (1982) Noradrenaline blocks accommodation of pyramidalcell discharge in the hippocampus. Nature 299:636-638.

Magee, J.C. (1998) Dendritic hyperpolarization-activated currents modify the integrativeproperties of hippocampal CA1 pyramidal neurons. J. Neurosci. 18:7613–7762.

Magee, J.C. (1999) Dendritic Ih normalizes temporal summation in hippocampal CA1 neurons.Nat. Neurosci. 2:508-514.

Magee, J.C. and Carruth, M. (1999) Dendritic voltage-gated ion channels regulate the actionpotential firing mode of hippocampal CA1 pyramidal neurons. J. Neurophysiol. 82:1895-1901.

Mainen, Z.F. and Sejnowski, T.J. (1996) Influence of dendritic structure on firing pattern inmodel neocortical neurons. Nature 382:363-366.

Literature 222

Mantz, J., Milla, C., Glowinski, J., Thierry, A.M. (1988) Differential effects of ascendingneurons containing dopamine and noradrenaline in the control of spontaneous activity andof evoked responses in the rat prefrontal cortex. Neuroscience 27:517-526.

Margoliash, D., Fortune, E.S., Sutter, M.L., Yu, A.C., Wren-Hardin, B.D., Dave, A. (1994)Distributed representation in the song system of oscines: Evolutionary implications andfunctional consequences. Brain Behav. Evol. 44:247-264.

Markowitsch, H.J. and Pritzel, M. (1979) The prefrontal cortex: Projection area of the thalamicmediodorsal nucleus? Physiol. Psychol. 7:1-6.

Martinoya, C., Rey, J., Bloch, S. (1981) Limits of the pigeon's binocular field and direction forbest binocular viewing. Vis. Res. 21:1197-1200.

Mason, A. and Larkman, A. (1990) Correlations between morphology and electrophysiology ofpyramidal neurons in slices of rat visual cortex. II. Electrophysiology. J. Neurosci.10:1415-1428.

McCormick, D.A. and Feeser, H.R. (1990) Functional implications of burst firing and singlespike activity in lateral geniculate relay neurons. Neuroscience 39:103-113

McCormick, D.A. and Pape, H.C. (1990) Properties of a hyperpolarization-activated cationcurrent and its role in rhythmic oscillation in thalamic relay neurones. J. Physiol.431:291-318.

McCormick, D.A. and Prince, D.A.. (1987) Post-natal development of electrophysiologicalproperties of rat cerebral cortical pyramidal neurones. J. Physiol. 393:743-762.

McCormick, D.A., Connors, B.W., Lighthall, J.W., Prince, D.A. (1985) Comparativeelectrophysiology of pyramidal and sparsely spiny stellate neurons of the neocortex. J.Neurophysiol. 54:782-806.

McDougall, S.A., Zolman, J.F., Mattingly, B.A. (1987) Effects of apomorphine and haloperidolon response suppression learning of young chicks. Behav. Neurosci. 101:254-263.

McGeorge, A.J. and Faull, R.L. (1989) The organization of the projection from the cerebralcortex to the striatum in the rat. Neuroscience 29:503-537.

Medina, L. and Reiner, A. (1997) The efferent projections of the dorsal and ventral pallidalparts of the pigeon basal ganglia, studied with biotinylated dextran amine. Neuroscience81:773-802.

Medina, L., Anderson, K.D., Karle, E.J., Reiner, A. (1995) An ultrastructural double-labelimmunohistochemical study of the enkephalinergic input to dopaminergic neurons of thesubstantia nigra in pigeons. J. Comp. Neurol. 357:408-432.

Medina, L., Veenman, C.L., Reiner, A. (1997) Evidence for a possible avian dorsal thalamicregion comparable to the mammalian ventral anterior, ventral lateral, and oralventroposterolateral nuclei. J. Comp. Neurol. 384:86-108.

Melchitzky, D.S., Sesack, S.R., Pucak, M.L., Lewis, D.A. (1998) Synaptic targets of pyramidalneurons providing intrinsic horizontal connections in monkey prefrontal cortex. J. Comp.Neurol. 390:211-24.

Mercer, A.R., Emptage, N.J., Carew, T.J. (1991) Pharmacological dissociation of modulatoryeffects of serotonin in Aplysia sensory neurons. Science 254:1811-1813.

Metzger, M., Jiang, S., Braun, K. (1998) Organization of the dorsocaudal neostriatal complex: aretrogarde and anterograde tracing study in the domestic chick with special emphasis onpathways relevant to imprinting. J. Comp. Neurol. 395:380-404.

Metzger, M., Jiang, S., Wang, J., Braun, K. (1996) Organization of the dopaminergicinnervation of forebrain areas relevant to learning: A combined

Literature 223

immunohistochemical/retrograde tracing study in the domestic chick. J. Comp. Neurol.376:1-27.

Miceli, D. and Repérant, J. (1985) Telencephalic afferent projections from the diencephalonand brainstem in the pigeon. A retrograde multiple-label fluorescent study. Exp. Biol.44:71-99.

Mickle, W.A. and Ades, H.W. (1954) Rostral projection pathway of the vestibular system. Am.J. Physiol. 176:243-246.

Miller, E.K., Erickson, C.A., Desimone, R. (1996) Neural mechanisms of visual workingmemory in prefrontal cortex of the macaque. J. Neurosci. 16:5154-5167.

Milner, B. and Petrides, M. (1984) Behavioural effects of frontal-lobe lesions in man. TrendsNeurosci. 7:403-407.

Mogensen, J. and Divac I. (1982) The prefrontal "cortex" in the pigeon: Behavioral evidence.Brain Behav. Evol. 21, 60-66.

Mogensen, J. and Divac, I. (1993) Behavioural effects of ablation of the pigeon-equivalent ofthe mammalian prefrontal cortex. Behav. Brain Res. 55:101-107.

Mogensen, J. and Holm, S. (1994) The prefrontal cortex and variants of sequential behaviourindications of functional differentiation between subdivisions of the rat's prefrontalcortex. Behav. Brain Res. 63:89-100.

Montague, P.R., Dayan, P. Sejnowski, T.J. (1996) A framework for mesencephalic dopaminesystems based on predictive hebbian learning. J. Neurosci. 16:1936-1947.

Montaron, M.-F., Bouyer, J.J., Rougeul, A., Buser, P. (1982) Ventral mesencephalic tegmentum(VMT) controls electrocortical beta rhythms and associated attentive behaviour in the cat.Behav. Brain Res. 6:129-145.

Moons, L., D'Hondt, E., Pijcke, K., Vandesande, F. (1995) Noradrenergic system in the chickenbrain: Immunocytochemical study with antibodies to noradrenaline and dopamine-beta-hydroxylase. J. Comp. Neurol. 360:331-348.

Moons, L., van Gils, J., Ghijsels, E., Vandesande, F. (1994) Immunocytochemical localizationof L-DOPA and dopamine in the brain of the chicken (Gallus domesticus). J. Comp.Neurol. 346:97-118.

Mrzljak, L., Bergson, C., Pappy, M., Huff, R., Levenson, R., Goldman-Rakic, P.S. (1996)Localization of dopamine D4 receptors in GABAergic neurons of the primate brain.Nature 381:245–248.

Müller, U., von Cramon, D.Y., Pollmann, S. (1998) D1- versus D2-receptor modulation ofvisuospatial working memory in humans. J. Neurosci. 18:2720-2728.

Muly, E.C., Szigeti, K., Goldman-Rakic, P.S. (1998) D1 receptor in interneurons of macaqueprefronzal cortex: Distribution and subcellar localizaytion. J. Neurosci. 18:10553-10565.

Murphy, B.L., Arnsten, A.F.T., Goldman-Rakic, P.S., Roth, R.H. (1996) Increased dopamineturnover in the prefrontal cortex impairs spatial working memory performance in rats andmonkeys. Proc. Nat. Acad. Sci. 93:1325-1329

Ng, G.Y.K., Trogadis, J., Stevens, J., Bouvier, M., O’Dowd, B.F., George, S.R. (1995)Agonist-induced desensitization of dopamine D1 recptor-stimulated adenylyl cyclaseactivity is temporally and biochemically separated from D1 receptor internalization.Proc. Nat. Acad. Sci. 92:10157-10161

Nisenbaum, E.S., Mermelstein, P.G., Wilson, C.J., Surmeier, D.J. (1998) Selective blockade ofa slowly inactivating potassium current in striatal neurons by (6) 6-chloro-APB-hydrobromide (SKF82958). Synapse 29:213-224.

Literature 224

Nishi, A., Snyder, G.L., Greengard, P. (1997) Bidirectional regulation of DARPP-32phosphorylation by dopamine. J. Neurosci. 17:8147-55.

Nisticò, G. and Stephenson, J.D. (1979) Dopaminergic mechanisms and stereotyped behaviourin birds. Pharmacol. Res. Commun. 11:555-570.

Nottebohm, F., Kelley, D.B., Paton, J.A. (1982) Connections of vocal control nuclei in thecanary telencephalon. J. Comp. Neurol. 207:344-357.

Nottebohm, F., Stokes, T.M., Leonard, C.M. (1976) Central control of song in the canary(serinus canarius). J. Comp. Neurol. 165:457-486.

Oades, R. and Halliday, G. (1987) Ventral tegmental (A10) system: Neurobiology. 1. Anatomyand connectivity. Brain Res. Rev. 12:117-165.

Ongini, E. (1993) Role of D1 versus D2 receptors in the modulation of states of arousal andsleep. In: D1:D2 Dopamine Receptor Interactions, pp 175-201. Ed J.L. Waddington.Academic Press: London.

Orem, J., Schlag-Rey, M., Schlag, J. (1973) Unilateral visual neglect and thalamic intralaminarlesions in the cat. Exp. Neurol. 40:784-797.

Ouimet, C.C. and Greengard, P. (1990) Distribution of DARPP-32 in the basal ganglia: Anelectron microscopic study. J. Neurocytol. 19:39-52.

Ouimet, C.C., LaMantia, A.S., Goldman-Rakic, P.S., Rakic, P., Greengard, P. (1992)Immunocytochemical localization of DARPP-32, a dopamine and cyclic-AMP-regulatedphosphoprotein, in the primate brain. J. Comp. Neurol. 323:209-218.

Ouimet, C.C., Miller, P.E., Hemmings, H.V., Walaas, S.I., Greengard, P. (1984) DARPP-32, adopamine- and adenosine 3':5'-monophosphate-regulated phosphoprotein enriched indopamine-innervated brain regions, III. Immunocytochemical localization. J. Neurosci.4:111-124.

Pandya, D.N. and Yeterian, E.H. (1990) Prefrontal cortex in relation to other cortical areas inrhesus monkey: Architecture and connections. In: H.B.M. Uylings, C.G. van Eden, J.P.C.de Bruin, M.A. Corner, M.G.P. Feenstra (Eds.), The Prefrontal Cortex: Its Structure,Function, and Pathology. Prog. Brain Res. 85:63-94.

Pape, H.C. (1996) Queer current and pacemaker: The hyperpolarization-activated cationcurrent in neurons. Annu. Rev. Physiol. 58:299-327.

Paré, D., Shink, E., Gaudreau, H., Destexhe, A., Lang.E.J. (1998) Impact of spontaneoussynaptic activity on the resting properties of cat neocortical pyramidal neurons in vivo.J. Neurophysiol. 79: 1450-1460.

Parent, A. (1990) Extrinsic connections of the basal ganglia. Trends Neurosci. 13:254-258.

Parent, A. and Hazrati, L.N. (1995) Functional anatomy of the basal ganglia. I. The cortico-basal ganglia-thalamo-cortical loop. Brain Res. Rev. 20:91-127.

Parker, D.M. and Delius, J. D. (1972) Visual evoked potentials in the forebrain of the pigeon.Exp. Brain Res. 14:198-209.

Pasqualini, C., Olivier, V., Guibert, B., Frain, O., Leviel, V. (1995) Acute stimulatory effect ofestradiol on striatal dopamine synthesis. J. Neurochem. 65:1651-1657.

Peinado, A., Yuste, R., Katz, L.C. (1993) Extensive dye coupling between rat neocorticalneurons during the period of circuit formation. Neuron 10:103-114.

Penit-Soria, J., Audinat, E., Crepel, F. (1987) Excitation of rat prefrontal cortical neurons bydopamine: An in vitro electrophysiological study. Brain Res. 425:263-274.

Literature 225

Petrides, M. (1995) Impairments on nonspatial self-ordered and externally ordered workingmemory tasks after lesions of the mid-dorsal part of the lateral frontal cortex in themonkey. J. Neurosci. 15:359-375.

Pfaus, J.G. and Phillips, A.G. (1989) Differential effects of dopamine receptor antagonists onthe sexual behavior of male rats. Psychopharmacol. 98:363-368.

Phillips, R.E. (1966) Evoked potential study of connections of avian archistriatum and caudalneostriatum. J. Comp. Neurol. 127:89-100.

Phillipson, O.T., Kilpatrick, I.C., Jones, M.W. (1987) Dopaminergic innervation of the primaryvisual cortex in the rat, and some correlations with human cortex. Brain Res. Bull.18:621-633.

Picker, M.J. and Massie, C.A. (1988) Differential effects of neuroleptic drugs on the delayedmatching-to-sample performance of pigeons. Pharmacol. Biochem. Behav. 31:953-957.

Pirot, S., Godbout, R., Mantz, J., Tassin, J.-P., Glowinski, J., Thierry, A.M. (1992) Inhibitoryeffects of ventral tegmental area stimulation on the activity of prefrontal cortical neurons:Evidence for the involvement of both dopaminergic and GABAergic components.Neuroscience 49:857-865.

Porrino, L.J. and Goldman-Rakic, P.S. (1982) Brain stem innervationof prefrontal and anteriorcingulate cortex in the rhesus monkey revealed by retrograde transport ofHRP. J. Comp.Neurol. 205:63-76.

Pralong, E. and Jones, R.S.G. (1993) Interactions of dopamine with glutamate- and GABA-mediated synaptic transmission in the rat enthorhinal cortex in vitro. Eur. J. Neurosci.5:760-767.

Preuss, T.M. (1995) Do rats have a prefrontal cortex? The Rose-Woolsey-Akert programreconsidered. J. Cog. Neurosci. 7:1-24.

Price, J.L. (1995) Thalamus. In: G. Paxinos (ed): The Rat Nervous System. 2nd ed. San Diego:Academic Press, pp 629-648.

Pucak, M.L., Levitt, J.B., Lund, J.S., Lewis, D.A. (1996) Patterns of intrinsic and associationalcircuitry in monkey prefrontal cortex. J. Comp. Neurol. 376:614–630.

Quintana, J., Yajeya, J., Fuster, J.M. (1988) Prefrontal representation of stimulus attributesduring delay tasks.1.Unit activity in cross-temporal integration of sensory and sensory-motor information. Brain Res. 474:211-221.

Rainer, G., Asaad, W.F., Miller, E.K. (1998a) Memory fields of neurons in the primateprefronzal cortex. Proc. Natl. Acad. Sci. 95:15008-15013.

Rainer, G., Asaad, W.F., Miller, E.K. (1998b) Selective representation of relevant informationby neurons inthe primate prefrontal cortex. Nature 393:577-579.

Ramirez, J.M. and Delius, J.D. (1979) Aggressive behavior of pigeons: Suppression byarchistriatal lesions. Aggress. Behav. 5:3-17.

Rao, S.C., Rainer, G., Miller, E.K. (1997) Integration of what and where in the primateprefrontal cortex. Science 276:821-824.

Rao, S.G., Williams, G.V., Goldman-Rakic, P.S. (1999) Isodirectional tuning of adjacentinterneurons and pyramidal cells during working memory: evidence for microcolumnarorganization in PFC. J. Neurophysiol. 81:1903-1916.

Ravazio, M.R. and Paschoalini, M.A. (1992) Modulation of food and water intake bycatecholamines injected into the lateral ventricle of the pigeon brain. Braz. J. Med. Biol.Res. 25:841-844.

Literature 226

Ray, J.P. and Price, J.I. (1993) The organization of projections from the mediodorsal nucleus ofthe thalamus to orbital and medial prefrontal cortex in macaque monkeys. J. Comp.Neurol. 337:1-31.

Redgrave, P., Prescott, T.J., Gurney, K. (1999) Is the short-latency dopamine response too shortto signal reward error? Trends Neurosci. 22:146-151.

Reep, R.L. (1984) Relationship between prefrontal and limbic cortex: A comparativeanatomical review. Brain Behav. Evol. 25:5-80.

Reep, R.L., Corwin, J.V., King, V. (1996) Neuronal connections of orbital cortex in rats:Topography of cortical and thalamic afferents. Exp. Brain Res. 111:215-232.

Rehkämper, G. and Zilles, K. (1991) Parallel evolution in mammalian and avian brains:Comparative cytoarchitectonic and cytochemical analysis. Cell Tiss. Res. 263:3-28.

Rehkämper, G., Zilles, K., Schleicher, A. (1984) A quantitative approach to cytoarchitectonicsIX. The areal pattern of the hyperstriatum ventrale in the domestic pigeon, (Columbalivia) f.d. Anat. Embryol. 169:319-327.

Rehkämper, G., Zilles, K., Schleicher, A. (1985) A quantitative approach to cytoarchitectonics,X. The areal pattern of the neostriatum in the domestic pigeon, (Columba livia) f.d., Acyto- and myeloarchitectonical study. Anat. Embryol. 171:345-355.

Reiner, A. (1986) Is prefrontal cortex found only in mammals? Trends Neurosci. 9:298-300.

Reiner, A. and Anderson, K.D. (1990) The pattern of neurotransmitter and neuropeptide co-occurence among striatal projection neurons: Conclusions based on recent findings. BrainRes. Rev. 15:251-265.

Reiner, A. and Karten, H.J. (1983) The laminar source of efferent projections from the avianwulst. Brain Res. 275:349-354.

Reiner, A. and Karten, H.J. (1985) Comparison of olfactory bulb projections in pigeons andturtles. Brain Behav. Evol. 27:11-27.

Reiner, A., Brauth, S.E., Karten, H.J. (1984a) Evolution of the amniote basal ganglia. TrendsNeurosci. 7:320-325.

Reiner, A., Davis, B.M., Karten, H.J., Brecha, N.C. (1984b) The distribution ofenkephalinelike immunoreactivity in the telencephalon of the adult and developingdomestic chicken. J. Comp. Neurol. 228:245-262.

Reiner, A., Karle, E.J., Anderson, K.D., Medina, L. (1994) Catecholaminergic perikarya andfibers in the avian nervous system. In: Phylogeny and development of catecholaminesystems in the CNS of vertebrates, pp 135-181. Eds W.J.A.J. Smeets, A. Reiner.Cambridge University Press: Cambridge, UK.

Reiner, A., Karten, H.J., Solina, A.R. (1983) Substance P: Localization within paleostriatal-tegmental pathways in the pigeon. Neuroscience 9:61-85.

Remy, M. and Güntürkün, O. (1991) Retinal afferents of the tectum opticum and the nucleusopticus principalis thalami in the pigeon. J. Comp. Neurol. 305:57-70.

Rétaux, S., Besson, M.J., Penit-Soria, J. (1991) Opposing effects of dopamine D2 receptorStimulation on the spontaneous and the electrically evoked release of [3H]GABA on ratprefrontal cortex slices. Neuroscience 42:61-71.

Richfield, E.K., Young, A.B., Penney, J.B. (1987) Comparative distribution of dopamine D-1and D-2 receptors in the basal ganglia of turtles, pigeons, rats, cats, and monkeys. J.Comp. Neurol. 262:446-463.

Richfield, E.K., Young, A.B., Penney, J.B. (1989) Comparative distributions of dopamine D-1and D-2 receptors in the cerebral cortex of rats, cats, and monkeys. J. Comp. Neurol.286:409-426.

Literature 227

Ridley, R.M. (1994) The psychology of perserverative and stereotyped behaviour. Prog.Neurobiol. 44:221-31.

Rieke, G.K. (1980) Kainic acid lesions of pigeon paleostriatum: A model for study ofmovement disorders. Physiol. Behav. 24:683-687.

Rieke, G.K. (1981) Movement disorders and lesions of pigeon brain stem analogues of basalganglia. Physiol. Behav. 26:379-384.

Rieke, G.K. (1982) The TPc, the avian substantia nigra: pharmacology and behavior. Physiol.Behav. 28:755-763.

Ritchie, T.L.C. (1979) Intratelencephalic visual connections and their relationship to thearchistriatum in the pigeon (Columba livia). Ph.D. Thesis, University of Virginia,Charlotsville.

Robbins, T.W. (1996) Dissociating executive functions of the prefrontal cortex. Phil. Trans. R.Soc. Lond. 351:1463-1471.

Roberts, A.C., De Salvia, M.A., Wilkinson, L.S., Collins, P., Muir, J.L., Everitt, B.J., Robbins,T.W. (1994) 6-hydroxydopamine lesions of the prefrontal cortex in monkeys enhanceperformance on an analog of the Wisconsin Card Sorting Test: Possible interactions withsubcortical dopamine. J. Neurosci. 14:2531-2544.

Robinson, S.W. and Caron, M.G. (1997) Interactions of dopamine receptors with G proteins.In: The Dopamin Receptors, pp 137-165. Eds K.A. Neve, R.L. Neve. Humana Press:Totowa.

Romanski, L.M., Tian, B., Fritz, J., Mishkin, M., Goldman-Rakic, P.S., Rauschecker, J.P.(1999) Dual streams of auditory afferents target multiple domains in the primateprefrontal cortex. Nature Neurosci. 2:1131-1136.

Romo, R., Scarnati, E., Schultz, W. (1992) Role of primate basal ganglia and frontal cortex inthe internal generation of movements. II. Movement-related activity in the anteriorstriatum. Exp. Brain Res. 91:385-395.

Rose, J.E. and Woolsey, C.N. (1948) The orbitofrontal cortex and its connections with themediodorsal nucleus in rabbit, sheep and cat. Res. Publ. Assoc. Res. Nerv. Ment. Dis.27:210-232.

Rosenkilde, C.E. (1979) Functional heterogeneity of the prefrontal cortex in monkey: A review.Behav. Neural. Biol. 25:301-345.

Russchen, F.T., Amaral, D.G., Price, J.L. (1987) The afferent input to the magnocellulardivision of the mediodorsal thalamic nucleus in the monkey (macaca fascicularis). J.Comp. Neurol. 256:175-210.

Sakaguchi, H. and Saito, N. (1989) The acetylcholine and catecholamine contents in songcontrol nuclei of zebra finch during song ontogeny. Brain Res. Dev. Brain Res. 47:313-317.

Sakurai, Y. and Sugimoto, S. (1985) Effects of lesions of prefrontal cortex and dorsomedialthalamus on delayed go/no-go alternation in the rat. Behav. Brain Res. 17:213-219.

Salamone, J.D. (1992) Complex motor and sensorimotor functions of striatal and accumbensdopamine: Involvement in instrumental behavior processes. Psychopharmacol. 107:160-174.

Salamone, J.D. (1994) The involvement of nucleus accumbens dopamine in appetitive andaversive motivation. Behav. Brain Res. 61:117-133.

Sanberg, P.R. and Mark, R.F. (1983) The effect of striatal lesions in the chick on haloperidol-potentiated tonic immobility. Neuropharmacol. 22:253-257.

Literature 228

Sapawi, R.R. and Divac, I. (1978) Connections of the mediodorsal nucleus of the thalamus inthe tree shrew. I. afferent connections. Neurosci. Lett. 7:183-189.

Sawaguchi, T. (1997) Attenuation of preparatory activity for reaching movements by a D1-dopamine antagonist in the monkey premotor cortex. J. Neurophysiol. 78:1769-1774.

Sawaguchi, T. and Goldman-Rakic, P.S. (1991) D1 dopamine receptors in prefrontal cortex:involvement in working memory. Science 251:947-950.

Sawaguchi, T. and Goldman-Rakic, P.S. (1994) The role of D1-dopamine receptor in workingmemory: Local injections of dopamine antagonists into the prefrontal cortex of rhesusmonkeys performing an oculomotor delayed-response task. J. Neurophysiol. 71:515-528.

Sawaguchi, T. and Yamane, I. (1999) Properties of delay-period neuronal activity in themonkey dorsolateral prefrontal cortex during a spatial delayed matching-to-sample task.J. Neurophysiol. 82:2070-2080.

Sawaguchi, T., Matsumura, M., Kubota, K. (1988) Dopamine enhances neuronal activity ofspatial short-term memory performance in primate prefrontal cortex. Neurosci. Res.5:465-473.

Sawaguchi, T., Matsumura, M., Kubota, K. (1990a) Catecholamine effects on neuronal activityrelated to a delayed response task in monkey prefrontal cortex. J. Neurophysiol.63:1385-1400.

Sawaguchi, T., Matsumura, M., Kubota, K. (1990b) Effects of dopamine antagonists onneuronal activity related to a delayed response task in monkey prefrontal cortex. J.Neurophysiol. 63:1401-1412.

Sayer, R.J., Schwindt, P.C., Crill, W.E. (1990) High- and low-threshold calcium currents inneurons acutely isolated from rat sensorimotor cortex. Neurosci. Lett. 120:175-178.

Schall, U. (1987) Vestibular, olfactory and vibratory responses of the nucleus basalisprosencephali neurons in pigeons. Neurosci. Res. 4:376-384.

Schall, U. (1989) The effects of lesions of the nucleus basalis prosencephali on feeding. In:Sensory Control of Pecking in the Pigeon (Columba livia). Pfaffenweiler CentaurusVerlagsgesellschaft, Pfaffenweiler, published thesis.

Schall, U. and Delius, J.D. (1986) Sensory inputs to the nucleus basalis prosencephali, afeeding-pecking centre in the pigeon. J. Comp. Physiol. 159:33-41.

Schall, U. and Delius, J.D. (1991) Grasping in the pigeon: Control through sound and vibrationfeedback mediated by the Nucleus basalis. Physiol. Behav. 50:983-988.

Schall, U., Güntürkün, O., Delius, J.D. (1986) Sensory projections to the nucleus basalisprosencephali of the pigeon. Cell Tiss. Res. 245:539-546.

Scheich, H. (1987) Neural correlates of auditory filial imprinting. J. Comp. Physiol. 161:605-619.

Schnabel, R. and Braun, K. (1996) Development of dopamine receptors in the forebrain of thedomestic chick in relation to auditory imprinting. An autoradiographic study. Brain Res.720:120-130.

Schnabel, R., Metzger, M., Jiang, S., Hemmings, H.C. Jr., Greengard, P., Braun, K. (1997)Localization of dopamine D1 receptors and dopaminoceptive neurons in the chickforebrain. J. Comp. Neurol. 388:146-168.

Schultz, W. (1998) The phasic reward signal of primate dopamine neurons. Adv. Pharmacol.42:686-690.

Schultz, W. and Romo, R. (1990) Dopamine neurons of the monkey midbrain: Contingencies ofresponses to stimuli eliciting immediate behavioral reactions. J. Neurophysiol. 63:607-624.

Literature 229

Schultz, W., Apicella, P., Ljungberg, T. (1993) Responses of monkey dopamine neurons toreward and conditioned stimuli during successive steps of learning a delayed responsetask. J. Neurosci. 13:900-913.

Schultz, W., Apicella, P., Scarnati, E., Ljungberg, T. (1992) Neuronal activity in monkeyventral striatum related to the expectation of reward. J. Neurosci. 12:4595-4610.

Schultz, W., Romo, R., Ljungberg, T., Mirenowicz, J., Hollerman, J.R., Dickinson, A. (1995)Reward-related signals carried by dopamine neurons. In: Models of InformationProcessing in the Basal Ganglia, pp 233-248. Eds J.C. Houk, J.L. Davis, D.G. Beiser.MIT Press: Cambridge, Mass.

Schwartz, M.L. and Goldman-Rakic, P.S. (1984) Callosal and intrahemispheric connectivity ofthe prefrontal association cortex in rhesus monkey: Relation between intraparietal andprincipal sulcal cortex. J. Comp. Neurol. 226:403-420.

Schwindt, P. and Crill, W. (1999) Mechanisms underlying burst and regular spiking evoked bydendritic depolarization in layer 5 cortical pyramidal neurons. J. Neurophysiol. 81:1341-1354.

Schwindt, P.C. (1992) Ionic currents governing input-output relations of Betz cells. In: Singleneuron computation (McKenna T, Davis J, Zornetzer, SF eds.) pp 235-258, New York:Academic Press.

Schwindt, P.C., Spain, W.J., Foehring, R.C., Stafstrom, C.E., Chubb, M.C., Crill, W.E. (1988)Multiple potassium conductances and their functions in neurons from cat sensorimotorcortex in vitro. J. Neurophysiol. 59:424-449.

Seamans, J.K., Floresco, S.B., Phillips, A.G. (1998) D1 receptor modulation of hippocampal-prefrontal cortical circuits integrating spatial memory with executive functions in the rat.J. Neurosci. 18:1613-1621.

Seamans, J.K., Gorelova, N.A., Yang, C.R. (1997) Contributions of voltage-gated Ca2+

channels in the proximal versus distal dendrites to synaptic integration in prefrontalcortical neurons. J. Neurosci. 17:5936-5948.

Seeman, P. and van-Tol, H.H. (1993) Dopamine receptor pharmacology. Curr. Opin. Neurol.Neurosurg. 6:602-608.

Séguéla, P., Watkins, K.C., Descarries, L. (1988) Ultrastructural features of dopamine axonterminals in the anteromedial and the suprarhinal cortex of adult rat. Brain Res. 442:11-22.

Selemon, L.D. and Goldman-Rakic, P.S. (1988) Common cortical and subcortical target areasof the dorsolateral prefrontal and posterior parietal cortices in the rhesus monkey. Adouble label study of distributed networks. J. Neurosci. 8:4049-4068.

Seroogy, K., Ceccatelli, S., Schalling, M., Hökfelt, T., Frey, P., Walsh, J., Dockray, G., Brown,J., Buchan, A., Goldstein, M. (1988a) A subpopulation of dopaminergic neurons in ratventral mesencephalon contains both neurotensin and cholecystokinin. Brain Res. 455:88-98.

Seroogy, K., Tsuruo, Y., Hökfelt, T., Walsh, J., Fahrenkrug, J., Emson, P.C., Goldstein, M.(1988b) Further analysis of presence of peptides in dopamine neurons. Cholecystokinin,peptide histidine-isoleucine/vasoactive intestinal polypeptide and substance P in ratsupramammillary region and mesencephalon. Exp. Brain Res. 72:523-534.

Sesack, S.R. and Bunney, B.S. (1989) Pharmacological characterization of the receptormediating electrophysiological responses to dopamine in the rat medial prefrontal cortex:A microiontophoretic study. J. Pharmacol. Exp. Therap. 248:1323-1333.

Literature 230

Sesack, S.R., Deutch, A.Y., Roth, R.H., Bunney, B.S. (1989) Topographical organization of theefferent projections of the medial prefrontal cortex in the rat: An anterograde tract-tracingstudy with Phaseolus vulgaris Leucoagglutinin. J. Comp. Neurol. 290:213-242.

Shettleworth, S.J. (1983) Memory in food-hoarding birds. Sci. Am. 248:102-110.

Shi, W.X., Zheng, P., Liang, X.F., Bunney, B.S. (1997) Characterization of dopamine-induceddepolarization of prefrontal cortical neurons. Synapse 26:415-422.

Shimizu, T., Cox, K., Karten, H.J. (1995) Intratelencephalic projections of the visual wulst inpigeons (Columba livia). J. Comp. Neurol. 359:551-572.

Shu, S., Ju, G., Fan, L. (1988) The glucose oxidase-DAB-nickel method in peroxidasehistochemistry of the nervous system. Neurosci. Lett. 85:169-171.

Siegel, A., Fukushima, T., Meibach, R., Burke, L., Edinger, H., Weiner, S. (1977) The origin ofthe afferent supply to the mediodorsal thalamic nucleus: enhancemet of HRP transport byselective lesions. Brain Res. 135:11-23.

Sik, A., Penttonen, M., Ylinen, A., Buzsaki, G. (1995) Hippocampal CA1 interneurons: An invivo intracellular labeling study. J. Neurosci. 15:6651-6665.

Sillito, A.M. (1984) Functional consideration of the operation of GABAergic inhibitoryprocess in the visual cortex. In: Cerebral Cortex Vol. 2 Functional Properties ofCortical Cells (Peters A, Jones EG eds) pp 91-117. New York. Plenum.

Simon, H., Scatton, B., Le Moal, M. (1980) Dopaminergic A10 neurones are involved incognitive functions. Nature 286:150-151.

Singer, W. (1993) Synchronization of cortical activity and its putative role in informationprocessing and learning. Ann. Rev. Physiol. 55:349-374.

Smeets, W.J.A.J. (1991) Comparative aspects of the distribution of substance P and dopamineimmunoreactivity in the substantia nigra of amniotes. Brain Behav. Evol. 37:179-188.

Smeets, W.J.A.J. and Gonzales, A. (1990) Are putative dopamine-accumulating cell bodies inthe hyperthalamic periventricular organ a primitive brain character of non-mammalianvertebrates? Neurosci. Lett. 114:248-252.

Smiley, J.F. and Goldman-Rakic, P.S. (1993) Heterogeneous targets of dopamine synapses inmonkey prefrontal cortex demonstrated by serial section electron microscopy: A laminaranalysis using the silver-enhanced diaminobenzidine sulfide (SEDS) immunolabelingtechnique. Cereb. Cortex 3:223-238.

Smith, P.K., Krohn, R.I., Hermanson, G.T., Mallia, A.K., Gartner, F.H., Provenzano, M.D.,Fulimoto, E.K., Goeke, N.M., Klenk, D.C. (1985) Measurement of protein usingbicinchoninic acid. Anal. Biochem. 150:76-85.

Snider, R.K., Kabara, J.F., Roig, B.R., Bonds, A.B. (1998) Burst firing and modulation offunctional connectivity in cat striate cortex. J. Neurophysiol. 80:730-744.

Soha, J.A., Shimizu, T., Doupe, A.J. (1996) Development of the catecholaminergic innervationof the song system of the male zebra finch. J. Neurobiol. 29:473-489.

Sokolowski, J.D. and Salamone, J.D. (1994) Effects of dopamine depletions in the medialprefrontal cortex on DRL performance and motor activity in the rat. Brain Res. 642:20-28.

Sokolowski, J.D., McCullough, L.D., Salamone, J.D. (1994) Effects of dopamine depletions inthe medial prefrontal cortex on active avoidance and escape in the rat. Brain Res.651:293-299.

Song, D.D. and Harlan, R.E. (1994) The development of enkephalin and substance P neurons inthe basal ganglia: insights into neostriatal compartments and the extended amygdala.Brain Res. Dev. Brain Res. 83:247-61.

Literature 231

Spruston, N., Schiller, Y., Stuart, G., Sakmann, B. (1995) Acitivity-dependent action potentialinvasion and calcium influx into hippocampal CA1 dendrites. Science 268:297-300.

Stafstrom, CE., Schwindt, P.C., Chubb, M.C., Crill, W.E. (1985) Properties of persistentsodium conductance and calcium conductance of layer V neurons from cat sensorimotorcortex in vitro. J. Neurophysiol. 53:153-170.

Stewart, M.G., Kabai, P., Harrison, E., Steele, R.J., Kossut, M., Gierdalski, M., Csillag, A.(1996) The involvement of dopamine in the striatum in passive avoidance training in thechick. Neuroscience 70:7-14.

Storm, JF. (1990) Potassium currents in hippocampal pyramidal cells. Prog. Brain Res.83:161-187.

Striedter, G.F. (1994) The vocal control pathways in budgerigars differ from those insongbirds. J. Comp. Neurol. 343:35-56.

Striedter, G.F., Marchant, T.A., Beydler, S. (1998) The „neostriatum“ develops as part of thelateral pallium in birds. J. Neurosci. 18:5839–5849.

Stuart, G.J. and Sakmann, B. (1994): Active propagation of somatic action potentials intoneocortical pyramidal cell dendrites. Nature 367:69-72.

Stuart, G.J. and Sakmann, B. (1995) Amplification of EPSPs by axo-somatic sodium channelsin neocortical pyramidal neurons. Neuron 15:1065-1076.

Stuss, D.T. and Benson, D.F. (1986) The Frontal Lobes. Raven Press. New York.

Surmeier, D.J., Bargas, J., Hemming, H.C. Jr., Nairn, A.C., Greengard, P. (1995) Modulation ofcalcium currents by a D1 dopaminergic protein kinase/phosphatase cascade in ratneostriatal neurons. Neuron 14:385-397.

Surmeier, D.J., Kita, H., Kitai, S.T. (1988) The expression of gamma-aminobutyric acid andLeu-enkephalin immunoreactivity in primary monolayer cultures of rat striatum. BrainRes. 470:265-82.

Swanson, L.W. (1982) The projections of the ventral tegmental area and adjacent regions: Acombined fluorescent retrograde tracer and immunofluorescence study in the rat. BrainRes. Bull. 9:321-53.

Székely, A.D., Boxer, M.I., Stewart, M.G., Csillag, A. (1994) Connectivity of the lobusparolfactorius of the domestic chicken (Gallus domesticus): An anterograde andretrograde pathway tracing study. J. Comp. Neurol. 348:374-393.

Takagi, H. (1986) Distribution of peptides in basal ganglia. Prog. Brain Res. 66:35-72.

Takatsuki, K., Shiosaka, S., Inagaki, S., Sakanak, M. (1981) Topographic atlas of somatostatin-containing neurons system in the avian brain in relation to catecholamine-containingneurons system I. Telencephalon and diencephalon. J. Comp. Neurol. 202:103-113.

Tassin, J.P., Cheramy, A., Blanc, G., Thierry, A.M., Glowinski, J. (1976) Topographicaldistribution of dopaminergic innervation and of dopaminergic receptors in the ratstriatum. I. Microestimation of [3H] dopamine uptake and dopamine content inmicrodiscs. Brain Res. 107:291-301.

Taylor, A.E. and Saint-Cyr, J.A. (1995) The neuropsychology of Parkinson's disease. BrainCogn. 28:281-96.

Terry, R.D., Masliah, E., Salmon, D.P. (1991) Physical basis of cognitive alterations inAlzheimer disease: Synapse loss is the major correlate of cognitive impairment. Ann.Neurol. 30:572-596.

Tohyama, M. and Takatsuji, K. (1998) Atlas of Neuroactive Substances and Their Receptorsin the Rat. Oxford: Oxford University Press.

Literature 232

Tombaugh, T.N. (1981) Effects of pimozide on nondiscriminated and discriminatedperformance in the pigeon. Psychopharmacol. 73:137-41.

van Eden, C.G., Hoorneman, E.M., Buijs, R.M., Matthijssen, M.A., Geffard, M., Uylings, H.B.(1987) Immunocytochemical localization of dopamine in the prefrontal cortex of the rat atthe light and electron microscopical level. Neuroscience 22:849-862.

Vates, G.E, Broome, B.M., Mello, C.V., Nottebohm, F. (1996) Auditory pathways of caudaltelencephalon and their relation to the song system of adult male zebra finches. J. Comp.Neurol. 366:613-642.

Vates, G.E., Vicario, D.S., Nottebohm, F. (1997) Reafferent thalamo- "cortical" loops in thesong system of oscine songbirds. J. Comp. Neurol. 380:275-290.

Veenman, C.L. and Reiner, A. (1994) The distribution of GABA-containing perikarya, fibers,and terminals in the forebrain and midbrain of pigeons, with particular reference to thebasal ganglia and its projection targets. J. Comp. Neurol. 339:209-250.

Veenman, C.L. and Reiner, A. (1996) Ultrastructrual morphology of synapses formed bycorticostriatal terminals in the avian striatum. Brain Res. 707:1-12.

Veenman, C.L., Karle, E.J., Anderson, K.D., Reiner, A. (1995a) Thalamostriatal projectionneurons in birds utilize LANT6 and neurotensin: A light and electron microscopicdouble-labeling study. J. Chem. Neuroanat. 9:1-16.

Veenman, C.L., Medina, L., Reiner, A. (1997) Avian homologue of mammalian intralaminar,mediodorsal and midline thalamic nuclei: Immunohistochemical and hodologicalevidence. Brain Behav. Evol. 49:78-98.

Veenman, C.L., Wild, J.M., Reiner, A. (1995b) Organization of the avian "corticostriatal"projection system: A retrograde and anterograde pathway tracing study in pigeons. J.Comp. Neurol. 354:87-126.

Velayos, J.L. and Reinoso-Suarez, F. (1982) Topographic organization of the brainstemafferents to the mediodorsal thalamic nucleus. J. Comp. Neurol. 206:17-27.

Verma, A. and Moghaddam, B. (1996) NMDA receptor antagonists impair prefrontal cortexfunction as assessed via spatial delayed alternation performance in rats: Modulation bydopamine. J. Neurosci. 16:373-379.

von Fersen, L., Wynne, C.D.L., Delius, J.D. (1990) Deductive reasoning in pigeons.Naturwissenschaften 77:548-549.

Waddington, J.L. and Daly, S.A. (1993) Regulation of unconditioned motor behaviour byD1:D2 interactions. In: D1:D2 Dopamine Receptor Interactions pp 51-78. Ed J.L.Waddington. Academic Press: Toronto.

Walaas, S.I. and Greengard, P. (1984) DARPP-32, a dopamine - and adenosine 3': 5'-monophosphate-regulated phosphoprotein enriched in dopamine innervated brain regions.I. Regional and cellular distribution in rat brain. J. Neurosci. 4:84-98.

Waldmann, C. and Güntürkün, O. (1993) The dopaminergic innervation of the pigeoncaudolateral forebrain: Immunocytochemical evidence for a "prefrontal cortex" in birds?Brain Res. 600:225-234.

Walker, A.E. (1940) A cytoarchitectural study of the prefrontal area of the macaque monkey. J.Comp. Neurol. 73:59-86.

Wallenberg A. (1903) Der Ursprung des Tractus isthmo-opticus (oder bulbo-striatus) derTaube. Neurol. Zentralbl. 22:98-101.

Watanabe, M. (1992) Frontal units of the monkey coding the associative significance of visualand auditory stimuli. Exp. Brain Res. 89:233-47.

Literature 233

Watanabe, M., Ito, H., Ikushima, M. (1985) Cytoarchitecture and ultrastructure of the avianectostriatum: Afferent terminals from the dorsal telencephalon and some nuclei in thethalamus. J. Comp. Neurol. 236:241-257.

Watanabe, M., Kodama, T., Hikosaka, K. (1997) Increase of extracellular dopamine in primateprefrontal cortex during a working memory task. J. Neurophysiol. 78:2795-2798.

Watson, J.E. and Blampied, N.M. (1989) Quantification of the effects of chlorpromazine onperformance under delayed matching to sample in pigeons. J. Exp. Anal. Behav. 51:317-28.

Wernicke, C. (1874) Der aphasische Symptomemkomplex (Cohn and Weigart, Breslau,Germany).

Westenbroek, R.E., Ahlijhanian, M.K., Catterall, W.A. (1990) Clustering of L-type Ca++channels at the base of major dendrites in hippocampal pyramidal neurons. Nature347:281–284.

Wickens, J. (1990) Striatal dopamine in motor activation and reward-mediated learning: stepstowards a unifying model. J. Neural. Transm. Gen. Sect. 80:9-31.

Wickens, J. and Kötter, R. (1995) Cellular models of reinforcment. In: Models of InformationProcessing in the Basal Ganglia, pp 187-214. Eds J.C. Houk, J.L. Davis, D.G. Beiser.MIT Press: Cambridge, Mass.

Wild, J.M. (1987a) Thalamic projections to the paleostriatum and neostriatum in the pigeon(Columba livia). Neuroscience 20:305-327.

Wild, J.M. (1987b) The avian somatosensory system: connections of regions of bodyrepresentation in the forebrain of the pigeon. Brain Res. 412:205-223.

Wild, J.M. (1988) Vestibular projections to the thalamus of the pigeon: an anatomical study. J.Comp. Neurol. 271:451-460.

Wild, J.M. (1989) Avian somatosensory system: II. Ascending projections of the dorsal columnand external cuneate nuclei in the pigeon. J. Comp. Neurol. 287:1-18.

Wild, J.M. (1993) Descending projections of the songbird nucleus robustus archistriatalis. J.Comp. Neurol. 338:225-241.

Wild, J.M. (1994) Visual and somatosensory inputs to the avian song system via nucleusuvaeformis (Uva) and a comparison with the projections of a similar thalamic nucleus ina nonsongbird (Columba livia). J. Comp. Neurol. 349:512-535.

Wild, J.M. (1997) The avian somatosensory system: The pathway from wing to Wulst in apasserine (Chloris Chloris). Brain Res. 759:122-134.

Wild, J.M. and Farabaugh, S.M. (1996) Organization of afferent and efferent projections of thenucleus basalis prosencephali in a passerine (Taeniopygia guttata). J. Comp. Neurol.365:306-328.

Wild, J.M., Arends, J.J.A., Zeigler, H.P. (1985) Telencephalic connections of the trigeminalsystem in the pigeon (Columba livia): A trigeminal sensorimotor circuit. J. Comp.Neurol. 234:441-464.

Wild, J.M., Arends, J.J.A., Zeigler, H.P. (1990) Projections of the parabrachial nucleus in thepigeon (Columba livia). J. Comp. Neurol. 293:499-523.

Wild, J.M., Karten H.J., Frost, B.J. (1993) Connections of the auditory forebrain in the pigeon(Columba livia). J. Comp. Neurol. 337:32-62.

Wild, J.M., Reinke, H., Farabaugh, S.M. (1997) A non-thalamic pathway contributes to a wholebody map in the brain of the budgerigar. Brain Res. 755:137-141.

Literature 234

Williams, G.V. and Goldman-Rakic, P.S. (1995) Modulation of memory fields by dopamine D1receptors in prefrontal cortex. Nature 376:572-575.

Williams, S.M. and Goldman-Rakic, P.S. (1993) Characterization of the dopaminergicinnervation of the primate frontal cortex using a dopamine-specific antibody. Cereb.Cortex 3:199-222.

Williams, S.M. and Goldman-Rakic, P.S. (1998) Widespread origin of the primate mesofrontaldopamine system. Cereb. Cortex 8:321-345.

Wilson, C., Nomikos, G.G., Collu, M., Fibiger, H.C. (1995) Dopaminergic correlates ofmotivated behavior: Importance of drive. J. Neurosci. 15:5169-5178.

Wilson, F.A.W., O´Scalaidhe, S.P., Goldman-Rakic, P.S. (1993) Dissociation of object andspatial processing domains in primate prefrontal cortex. Science 260:1955-1958.

Wilson, F.A.W., O´Scalaidhe, S.P., Goldman-Rakic, P.S. (1994) Functional synergism betweenputative gamma-aminobutyrate-containing neurons and pyramidal neurons in prefrontalcortex. Proc. Nat. Acad. Sci. 91:4009-4013.

Winsauer, P.J. and Thompson, D.M. (1991) Cocaine self-administration in pigeons.Pharmacol. Biochem. Behav. 40:41-52.

Wise, R.A. (1978) Catecholamine theories of reward: A critical review. Brain Res. 152:215-247.

Wong, R.K. and Prince, D.A.. (1981) Afterpotential generation in hippocampal pyramidal cells.J. Neurophysiol. 45:86-97.

Wynne, B. and Delius, J. (1996) Frontal forebrain lesions: Effects on the foraging andapomorphine pecking of pigeons. Physiol. Behav. 59:757-762.

Wynne, B. and Güntürkün, O. (1995) Dopaminergic innervation of the telencephalon of thepigeon (Columba livia): A study with antibodies against tyrosine hydroxylase anddopamine. J. Comp. Neurol. 57:446-464.

Yajeya, J., Quintana, J., Fuster, J.M. (1988) Prefrontal representation of stimulus attributesduring delay tasks. II. The role of behavioral significance. Brain Res. 474:222-230.

Yanai, J., Silverman, W.F., Shamir, D. (1995) An avian model for the reversal of 6-hydroxydopamine induced rotating behaviour by neural grafting. Neurosci. Lett. 187:153-156.

Yang, C.R. and Mogenson, G.J. (1990) Dopaminergic modulation of cholinergic responses inrat medial prefrontal cortex: An electrophysiological study. Brain Res. 524:271–281.

Yang, C.R. and Seamans, J.K. (1996) Dopamine D1 receptor actions in layers V-VI ratprefrontal cortex neurons in vitro: Modulation of dendritic-somatic signal integration. J.Neurosci. 16:1922-1935.

Yang, C.R., Seamans, J.K. and Gorelova, N. (1999) Developing a neuronal model for thepathophysiology of schizophrenia based on the nature of electrophysiological actions ofdopamine in the prefrontal cortex. Neuropsychopharmacol. 21:161-194.

Yang, C.R., Seamans, J.K., Gorelova, N. (1996a) Electrophysiological and morphologicalproperties of layers V-VI principal pyramidal cells in rat prefrontal cortex in vitro. J.Neurosci. 16:1904-1921.

Yang, C.R., Seamans, J.K., Gorelova, N. (1996b) Focal dendritic D1 receptor stimulationdifferentially modulates layers I-II & V-VI glutamate inputs to deep layer prefrontalcortical (PFC) pyramidal neurons in vitro. Soc. Neurosci. Abstr. 22:1770.

Yang, C.R., Seamans, J.K., Gorelova, N. (1997) Mechanisms of dopamine (DA) modulation ofGABAergic inputs to rat layer V-VI pyramidal prefrontal cortical (PFC) neurons in vitro.Soc. Neurosci. Abstr. 23:1771.

Literature 235

Yang, C.R., Seamans, J.K., Gorelova, N. (1998) Dopamine D1/5 receptor activation influencesN- & L-type Ca++ channels to modulate dendritic Ca 21 potentials in pyramidal prefrontalcortex (PFC) neurons. Soc. Neurosci. Abs. 24:854.

Yokel, R.A. and Wise, R.A. (1975) Increased lever pressing for amphetamine after pimozide inrats: Implications for a dopamine theory of reward. Science 187:547-548.

Yung, K.K.L., Bolam, J.P., Smith, A.D., Hersch, S.M., Ciliax, B.J., Levey, A.I. (1995)Immunocytochemical localization of D1 and D2 dopamine receptors in the basal gangliaof the rat: light and electron microscopy. Neuroscience 65:709-730.

Zahrt, J., Taylor, J.R., Mathew, R.G., Arnsten, A.F.T. (1997) Supranormal stimulation of D1

dopamine receptors in the rodent prefrontal cortex impairs spatial working memoryperformance. J. Neurosci. 17:8528-8535.

Zarrindast, M.R. and Amin, R. (1992) Role of D-1 and D-2 receptors in apomorphine-inducedpecking in chicks. Psychopharmacol. (Berl) 106:67-70.

Zarrindast, M.R. and Namdari, S. (1992) Ephedrine and amphetamine induced pecking inchickens: Possible indirect D-1/D-2 dopaminergic mechanism. Eur.Neuropsychopharmacol. 2:135-40.

Zarrindast, M.R., Hajian-Heydari, A., Hoseini-Nia, T. (1992) Characterization of dopaminereceptors involved in apomorphine-induced pecking in pigeons. Gen. Pharmacol.23:427-430.

Zeier, H. (1971) Archistriatal lesions and response inhibition in the pigeon. Brain Res. 31:327-339.

Zeier, H. and Karten, H.J. (1971) The archistriatum of the pigeon: Organization of afferent andefferent connections. Brain Res. 31:313-326.

Zeier, H. and Karten, H.J. (1973) Connections of the anterior commisure in the pigeon(Columba livia). J. Comp. Neurol. 150:201-216.

Zeigler, P.H. (1976) Feeding in the pigeon. Adv. Stud. Behav. 7:283-389.

Zhang, L., Valiante, T.A., Carlen, P.L. (1993) Contribution of the low-threshold T-type calciumcurrent in generating the post-spike depolarizing afterpotential in dentate granule neuronsof immature rats. J. Neurophysiol. 70:223-231.

Zheng, P., Zhang, X.X., Bunney, B.S., Shi, W.X. (1999) Opposite modulation of cortical N-methyl-D-aspartate receptor-mediated responses by low and high concentrations ofdopamine. Neuroscience 91:527-35.

Zhou, F.M., and Hablitz, J.J. (1999) Dopamine modulation of membrane and synapticproperties of interneurons in rat cerebral cortex. J. Neurophysiol. 81: 967–976.

236

Danksagung

Zuallererst möchte ich meinen Eltern und meiner Frau Ingelore danken, die mir immer alles

ermöglicht haben, und die mir nicht nur während der Zeit dieser Doktorarbeit den Rücken

freigehalten haben.

Meinem Betreuer und Chef Prof. Onur Güntürkün möchte ich dafür danken, dass er mehr Freund

als Chef war, und dass auch er mir fast immer alles ermöglicht hat. Ich glaube, dass ich - und

damit auch diese Arbeit - sehr von seiner offenen und kompetenten Art profitiert habe. Sowohl

die persönlichen und wissenschaftlichen Freiheiten die er mir eingeräumt hat, als auch die

gemeinsamen Diskussionen haben meine Arbeit immer wieder beflügelt.

Grossen Anteil am Gelingen dieser Arbeit haben natürlich auch alle Mitarbeiter und

Mitarbeiterinnen der AE Biopsychologie. Von vielen von ihnen habe ich immer wieder

fachliche oder persönliche Unterstützung erhalten. Das gilt vor allem für Markus „Fast Blue“

Hausmann, der mich jeden Tag ertragen musste; Thomas „Mauve Masher“ Kalt, der mich oft zur

Arbeit hin und meistens auch wieder zurück gebracht hat, wobei er auf beiden Wegen so

manchen Unmut abfangen konnte (und in der Zeit dazwischen auch); und Burkhard „Horst“

Hellmann, der mir mit dem einen oder anderen Tip und Trick ausgeholfen hat („aber dafür hat

man ja Post-docs“). Alle anderen von euch haben die Zeit nicht minder geprägt. Vielen Dank

deshalb auch an Bettina Diekamp, Martina Skiba, Martina Manns, Ariane Schwarz, Sabine

Kesch, und Christel Lehnert. Besonderen Dank auch an Daniel Durstewitz, mit dem ich nicht nur

viel Laborarbeit geteilt habe, sondern mit dem ich auch so manch´ anstrengende Diskussion

geschlagen habe. Jede einzelne davon war es wert!

Bei Prof. Hanns Hatt und meinem zweiten Betreuer PD Kurt Gottmann möchte ich mich dafür

bedanken, dass sie mir in ihrer Arbeitsgruppe am Lehrstuhl für Zellphysiologie eine „zweite

Heimat“ gegeben haben. Auch dort habe ich das Glück gehabt, nicht nur etwas lernen zu können,

sondern vor allem es im Umgang mit freundlichen und hilfsbereiten Menschen zu tun. Besonders

hervorgehoben seien dabei Gunnar, Markus, Ralf und Simon!

Dank auch an Prof. Toru Shimizu, dem ich die anatomischen „basics“ und eine tolle Zeit in

Florida zu verdanken habe, sowie Dr. Ulli Schall, der mir ebenfalls sowohl wissenschaftlich

als auch persönlich eine schöne Zeit in Australien verschafft hat.

Dank euch allen!

237

Lebenslauf

Sven Kröner

persönliche Daten und Schulausbildung

04.03.1971 geboren in Bochum (Nordrhein-Westfalen)

1977 - 1981 Besuch der Grundschule Castroper-Hellweg, Bochum

1981 - 1990: Besuch des Heinrich-von-Kleist Gymnasiums, Bochum:

Abitur im Mai 1990, ebenda

Juli 1998 Heirat mit Ehefrau Ingeborg Kröner, Las Vegas, Nevada

Universitäre Ausbildung

1990 - 1996 Studium der Psychologie an der Ruhr-Universität Bochum

Diplomarbeit: Der "präfrontale Cortex" von Tauben - Modell fürdie Chemoarchitektur kognitiver Prozesse

seit Sept. 1996 Wissenschaftlicher Mitarbeiter an der Fakultät für Psychologie,AE Biopsychologie, bei Prof. Onur Güntürkün

Sept. 1996 - Mai 2000 Promotion: „The caudolateral neostriatum of the avian forebrainand its modulation by dopaminergic afferents“Betreuer der Promotion:Prof. Dr. Onur Güntürkün, AE Biospychologie, RUBPD Dr. Kurt Gottmann, Lehrstuhl für Zellphysiologie, RUB

Tag der mündlichen Prüfung 29. 05. 2000

Forschungsvorhaben im Ausland

Juli - Dez. 1994 Forschungspraktikum an der University of South Florida, Tampa,FL, im Department of Experimental Psychology bei Prof. ToruShimizu: Neuroanatomische Untersuchungen zum caudalenVorderhirn von Vögeln

Dez. 1996 - Apr. 1997 Forschungsaufenthalt am Prince of Wales Hospital, Sydney,Australien, in der Biological Schizophrenia Research Unit bei Dr.Phillip Ward: Untersuchungen zum Einfluß von Amphetamin aufPrepulse-Inhibition und auditorische „Gating“-Mechanismen

Stipendien

Okt. 1996 Stipendium der G.A. Lienert Stiftung:

238

Erklärung

Die hier vorgelegte Dissertation wurde von mir selbst und ohne unerlaubte fremde Hilfe

angefertigt. Außer den in den Anmerkungen im Text und im Literaturverzeichnis genannten

Hilfsmitteln wurden keine weiteren benutzt. Bereits im Druck erschienene oder als Manuskripte

eingereichte Teile der Arbeit sind als solche gekennzeichnet.

Neben der „mit Auszeichnung“ bestandenen Diplom-Hauptprüfung im Fach Psychologie an der

Ruhr-Universität Bochum, habe ich bislang keine weiteren Prüfungen abgelegt.

Ich habe diese Dissertation weder in dieser, noch in irgendeiner anderen Form bereits

vorgelegt. Ich habe darüberhinaus bislang auch keine andere Dissertation vorgelegt.

Bochum, den 27. 03. 2000

Sven Kröner

webdoc.sub.gwdg.dewebdoc.sub.gwdg.de/ebook/dissts/Bochum/Kroener2000.pdf · 2005-10-31 · Table of...

Documents

Transcript of webdoc.sub.gwdg.dewebdoc.sub.gwdg.de/ebook/dissts/Bochum/Kroener2000.pdf · 2005-10-31 · Table of...