SEARCH

SEARCH BY CITATION

Keywords:

  • striatum;
  • pallidum;
  • pallium;
  • thalamus;
  • evolution

Abstract

  1. Top of page
  2. Abstract
  3. AVIAN BASAL GANGLIA ORGANIZATION AND EVOLUTION
  4. AVIAN PALLIAL ORGANIZATION AND EVOLUTION
  5. AVIAN ARCOPALLIUM AND AMYGDALA
  6. MEDIAL TELENCEPHALIC WALL ORGANIZATION AND EVOLUTION IN BIRDS
  7. AVIAN THALAMIC ORGANIZATION AND EVOLUTION
  8. CONCLUSION
  9. Acknowledgements
  10. LITERATURE CITED

Early 20th-century comparative anatomists regarded the avian telencephalon as largely consisting of a hypertrophied basal ganglia, with thalamotelencephalic circuitry thus being taken to be akin to thalamostriatal circuitry in mammals. Although this view has been disproved for more than 40 years, only with the recent replacement of the old telencephalic terminology that perpetuated this view by a new terminology reflecting more accurate understanding of avian brain organization has the modern view of avian forebrain organization begun to become more widely appreciated. The modern view, reviewed in the present article, recognizes that the avian basal ganglia occupies no more of the telencephalon than is typically the case in mammals, and that it plays a role in motor control and motor learning as in mammals. Moreover, the vast majority of the telencephalon in birds is pallial in nature and, as true of cerebral cortex in mammals, provides the substrate for the substantial perceptual and cognitive abilities evident among birds. While the evolutionary relationship of the pallium of the avian telencephalon and its thalamic input to mammalian cerebral cortex and its thalamic input remains a topic of intense interest, the evidence currently favors the view that they had a common origin from forerunners in the stem amniotes ancestral to birds and mammals. © 2005 Wiley-Liss, Inc.

Several prominent investigators in the early 20th century formulated a theory of avian brain organization and evolution based on studies of living species (Edinger et al., 1903; Edinger, 1908; Ariëns-Kappers et al., 1936). Since techniques for histological study of the nervous system were at that time limited to stains that detected all cell bodies or fiber tracts, or defined the morphology of individual neurons (Northcutt, 2001), their conclusion about the identity of any given brain region was based on the general shape of that brain region and on the appearance and distribution of its cells. For example, the telencephalon of mammals possesses two major, cytoarchitectonically distinct subdivisions: a region forming the outer rind of the telencephalon, whose cells are arrayed in six layers that are spanned by the ascending dendrites of the neurons making up the layers, and a region more centrally and basally located whose neurons are more uniformly distributed and possess radially symmetrical dendritic trees (Fig. 1A). Because of its apparent uniqueness to mammals and because of its histological configuration, the former telencephalic region came to be called the neocortex, while due to its more basal position and nuclear configuration, the latter came to be called the basal ganglia (Edinger et al., 1903; Edinger, 1908; Ariëns-Kappers et al., 1936). Examination of avian telencephalon revealed that a structure that histologically resembled mammalian neocortex was not evident in birds, and that most of the avian telencephalon, instead, resembled basal ganglia in that its neurons were not laminar in their distribution and columnar in their dendritic distribution, but rather were relatively uniformly distributed and typically possessed radially symmetrical dendritic trees (Fig. 1B). Given that most of the avian telencephalon then resembled mammalian basal ganglia in its histological appearance, it was concluded that birds lacked a neocortex and instead possessed a basal ganglia-dominated telencephalon (Fig. 2A and B) (Edinger et al., 1903; Edinger, 1908; Ariëns-Kappers et al., 1936).

thumbnail image

Figure 1. A series of low-power images of transverse sections through the telencephalon of adult rat (A and C) and adult pigeon (B and D), labeled for various markers. Comparison of the Nissl-stained sections for rat and pigeon (A and B) shows that bird telencephalon conspicuously lacks a laminated structure resembling mammalian cerebral cortex along its outer surface. Rather, much of the telencephalon of birds resembles the striatal part of mammalian basal ganglia in its histological appearance. These features of the telencephalon in birds led comparative neuroanatomists in the early part of the 20th century to conclude that telencephalon in birds largely consists of a hypertrophied basal ganglia (Ariëns-Kappers et al., 1936). The striatum in mammals is characterized by enrichment in dopaminergic terminals (C) arising from neurons in the midbrain tegmentum, as can be visualized by immunolabeling for tyrosine hydroxylase (TH). Immunolabeling of pigeon striatum for TH (D) reveals that only the basal part of the telencephalon in birds is enriched in dopaminergic terminals and is therefore the only part of telencephalon in birds that is striatal. Scale bar in D = 2 mm (magnification in A–D the same).

Download figure to PowerPoint

thumbnail image

Figure 2. A series of schematic line drawings of midtelencephalic transverse brain sections of pigeon and rat illustrating. A: The outdated interpretation of the organization of the telencephalon in birds (Ariëns-Kappers et al., 1936) and the outdated nomenclature that view engendered for the telencephalon of birds. B: The long-standing interpretation of mammalian telencephalic organization and the established nomenclature consistent with that view. C: The current interpretation of the organization of avian telencephalon and the new avian telencephalic nomenclature (Reiner et al., 2004b). In each schematic interpretation of telencephalic organization, the speckled region represents pallium, the striped region represents striatum, and the checked region represents globus pallidus. As indicated by a comparison of A and B, the terms used for specific telencephalic regions in birds were seemingly appropriate at the time that most of the telencephalon of birds was thought to represent a hypertrophied basal ganglia (Ariëns-Kappers et al., 1936). Now that it is clear that the telencephalon of birds includes both a basal ganglia and a large pallial territory comparable in their relative extents to those in mammals (B compared to C), the names for pallial regions implying striatal homology were clearly inappropriate and misleading and were thus in need of revision. Hp, hippocampus.

Download figure to PowerPoint

Edinger, Ariëns-Kappers, and their colleagues used a similar approach to examine the brains of other living vertebrates in an effort to identify homologous (i.e., retained) structures and novel (i.e., newly evolved) structures across vertebrate phylogeny. They concluded that the telencephalon had gradually expanded in the vertebrate lineage by the successive addition of new parts, beginning with a telencephalon devoted exclusively to olfaction in the ancestral jawless vertebrates. They further concluded a globus pallidus was added in the jawed fish lineage, a striatum in amphibians, and a primitive cerebral cortex in reptiles. Telencephalic evolution was thought by them to culminate in a divergence between birds and mammals in the relative elaboration of basal ganglia versus neocortex, with birds favoring basal ganglia and mammals neocortex (Fig. 2A and B). This framework became established as the dominant view by major neuroanatomy books by Ariëns-Kappers et al. (1936) and Herrick (1948, 1956). Although neuroanatomical, molecular biological, and neurochemical studies of the past 40 years have refuted this view of telencephalic evolution, vestiges of it have survived in the continued use of the terms “paleostriatum” and “neostriatum” to refer to globus pallidus and caudate-putamen, respectively, in mammals. Moreover, the Edinger-Ariëns-Kappers view of telencephalic evolution became the basis of the neuroanatomical terminology most widely employed for the avian telencephalon throughout the 20th century, which was not replaced by a more appropriate terminology until recently (Reiner et al., 2004b; Jarvis et al., 2005). Modern studies show that the basal ganglia occupies only a small part of the avian telencephalon, with the remainder occupied by a large pallial territory that functions much like cerebral cortex in mammals (Fig. 2C). In the present article, we review the modern understanding of the organization and function of the avian telencephalon and thalamus, and we also review current thinking about the evolutionary relationships of these to the cerebral cortex, basal ganglia, and thalamus of mammals.

AVIAN BASAL GANGLIA ORGANIZATION AND EVOLUTION

  1. Top of page
  2. Abstract
  3. AVIAN BASAL GANGLIA ORGANIZATION AND EVOLUTION
  4. AVIAN PALLIAL ORGANIZATION AND EVOLUTION
  5. AVIAN ARCOPALLIUM AND AMYGDALA
  6. MEDIAL TELENCEPHALIC WALL ORGANIZATION AND EVOLUTION IN BIRDS
  7. AVIAN THALAMIC ORGANIZATION AND EVOLUTION
  8. CONCLUSION
  9. Acknowledgements
  10. LITERATURE CITED

Avian Basal Ganglia Organization

Evidence rebutting the classic view of avian telencephalic evolution began to accrue in the 1960s and 1970s with the advent of new neuroanatomical methods, such as histochemical methods for determining regional variation in brain neurochemistry, and the Nauta-Gygax and later the Fink-Heimer methods for studying degenerating fibers to ascertain nervous system connectivity experimentally. This made it possible to compare brain regions between species by more diverse criteria than mere histological appearance (Northcutt, 2001). By the 1960s, numerous researchers had shown that the striatal part of the basal ganglia in mammals (Koelle, 1954; Butcher and Woolf, 1984), but not the neocortex or globus pallidus, is enriched in acetylcholinesterase (AchE), and Karten (1969) and Parent and Olivier (1970) showed that this was true of only a basal part of the avian telencephalon, a region we now recognize as the striatum in the revised avian forebrain terminology (Reiner et al., 2004b). Similarly, with the development of catecholamine histofluorescence methods in the 1960s, the striatal part of the basal ganglia in mammals (i.e., the caudate-putamen) was also found to be highly enriched in dopaminergic terminals (Fig. 1C) arising from the substantia nigra pars compacta (SNc) of the midbrain (Dahlström and Fuxe, 1964). Juorio and Vogt (1967) and Karten and Dubbeldam (1973) found that only the AchE-rich telencephalic sector was enriched in dopaminergic terminals in birds, and that these terminals arise from a large midbrain field of dopaminergic neurons now termed the SNc in birds (Fig. 1D) (Reiner et al., 1994). These findings clearly established the true location of the striatum in birds (Fig. 2C). Pathway tracing studies in conjunction with neurochemical studies showed that one major projection of the striatal cell group was to a ventromedially located cell field that contains loosely packed large neurons with aspiny dendrites, a field that is poor in AchE and dopaminergic terminals (Karten and Dubbeldam, 1973). This cell field shares these traits with the globus pallidus in mammals and is now known by the same name in birds. As in mammals, this pallidal cell group was shown to give rise to the descending projections by which the basal ganglia influences motor control (Karten and Dubbeldam, 1973). More recently, studies of the regional expression of genes regulating the development of the major telencephalic territories have confirmed the location of the parts of the avian basal ganglia. These studies show that the basal ganglia develops from a ventral telencephalic territory called the subpallium, as it does in mammals as well (Smith-Fernandez et al., 1998; Puelles et al., 2000). Within this territory, pallidal identity is determined by expression of the genes Dlx1, Dlx2, and Nkx2.1, while striatal identity is determined by expression of Dlx1 and Dlx2 but not Nkx2.1.

The advent of single- and double-label immunohistochemical methods and improved pathway tracing methods in the 1980s made it possible to identify specific cell types by their projection targets and the neurotransmitters or neurotransmitter-related chemicals they contain. These approaches confirmed that the relative sizes of the striatum and pallidum are similar in birds and mammals and further revealed that the neuronal cell type composition and functional circuitry of the basal ganglia are similar as well (Graybiel, 1990; Veenman and Reiner, 1994; Marin et al., 1998a, 1998b; Reiner et al., 1998; Reiner, 2002). The most abundant type of neurons in striatum in both birds and mammals are the medium-sized neurons with radially symmetrical spiny dendrites (Reiner et al., 1998; Sun et al., 2005), whose primary neurotransmitter is GABA. In both birds and mammals, about half of these contain the neuropeptide substance P (SP) and the other half contain enkephalin (ENK) neuropeptides. As in mammals, the SP+ neurons of avian striatum typically contain dynorphin as a coneuropeptide transmitter (Reiner and Anderson, 1990; Reiner et al., 1999). The SP+ and ENK+ neurons are the projection neurons of the striatum, and in birds those of medial striatum project to the GABAergic neurons of substantia nigra pars reticulata (SNr) and the dopaminergic neurons of pars compacta (SNc), and those of lateral striatum project to separate sets of GABAergic neurons in globus pallidus (Karten and Dubbeldam, 1973; Veenman and Reiner, 1994; Reiner et al., 1998, 1999; Sun et al., 2005). While mammalian SP+ and ENK+ striatal neurons also project to globus pallidus and nigra, striatopallidal and striatonigral neurons are intermingled in mammals. The avian striatum also contains the three main types of aspiny interneurons characteristic of mammalian basal ganglia, the large cholinergic, the large parvalbuminergic/GABAergic, and the medium-sized somatostatinergic (Reiner and Carraway, 1987; Medina and Reiner, 1994; Reiner et al., 1998). The large cholinergic interneurons in birds and mammals innervate the spiny projection neurons of the striatum, accounting for the striatal enrichment in AchE and muscarinic receptors (Wächtler and Ebinger, 1989; Medina and Reiner, 1994; Bolam and Bennett, 1995; Reiner et al., 1998).

The dopaminergic input from the SNc represents one of the major striatal afferents in both birds and mammals (Reiner et al., 1998). Dopaminergic effects on striatum in both are mediated by D1- and D2-type dopamine receptors, with DA agonists inducing hyperkinesia and DA antagonists yielding hypokinesia (Nistico et al., 1983; Richfield et al., 1987; Dietl and Palacios, 1988; Reiner et al., 1994; Demchyshyn et al., 1995; Yanai et al., 1995; Sun and Reiner, 2000; Reiner, 2002). As in mammals also, nearly the entire pallium is the source of a massive excitatory input to the striatum (Veenman et al., 1995; Veenman and Reiner, 1996; Csillag et al., 1997), with the thalamus also providing excitatory input (Wild, 1987; Reiner et al., 1998). The excitatory input ends on the dendritic spines of the SP+ and ENK+ striatal projection neurons, and it provides these neurons with the information they need to execute their role in movement control (Bolam and Bennett, 1995; Veenman and Reiner, 1996; Csillag et al., 1997). The communication between pallium and striatum in birds is mediated by the same two corticostriatal cell types as in mammals (Cowan and Wilson, 1994; Veenman et al., 1995; Reiner et al., 2001, 2003), a type that projects to brainstem and spinal cord premotor and motor cell groups and a type that only has intratelencephalic connections. Thalamic input to striatum in birds arises from several dorsomedial thalamic nuclei, which also have widespread projections to the pallium (Wild, 1987; Veenman et al., 1997). The distribution of several neuropeptides and transmitters within this thalamic region, as well as topographic and hodological considerations, bolsters the interpretation that dorsomedial thalamic nuclei of birds resemble the intralaminar thalamic nuclear complex in mammals (Veenman et al., 1997). Both the cortical and thalamic excitatory influences on striatum in birds are mediated by the same cell type-specific glutamate receptors as in mammals (Reiner, 2002; Wada et al., 2004; Laverghetta et al., 2005).

In mammals, the SP+ and ENK+ striatal projection neurons are the sources of functionally distinct output pathways by which the basal ganglia execute their role in motor functions, referred to as the direct and indirect, respectively (Fig. 3) (Albin et al., 1989; Crossman, 1990; DeLong, 1990). The SP+ neurons of the direct pathway project to the internal pallidal segment (GPi) and the SNr, which in turn project to the motor thalamic region that projects to premotor and motor cortex. Since both the striatal SP+ neurons and the neurons of GPi and SNr are GABAergic, while those of motor thalamus are excitatory, activation of striatal SP+ neurons has a disinhibitory effect on motor thalamus, thereby promoting the specific set of movements controlled by the particular activated SP+ neurons. ENK+ neurons of the indirect pathway are part of a different output circuit and play a different role in motor control. The ENK+ neurons project to the external pallidal segment (GPe), which itself projects to the subthalamic nucleus (STN), which in turn projects to the GPi and the SNr, with the STN projections utilizing the excitatory amino acid glutamate. Activation of ENK+ striato-GPe neurons produces inhibition of GPe neurons, therefore producing a disinhibition of glutamatergic STN input to GPi and SNr. The end result of this is that activation of ENK+ striato-GPe neurons leads to enhanced inhibition of motor thalamus by GPi and SNr. The indirect pathway circuit thereby promotes suppression of movements by enhancing the activity of GPi and SNr neurons that suppress neuronal activity in motor thalamus. The action of the indirect pathway complements the action of the direct pathway, with the indirect pathway inhibiting movements potentially conflicting with the desired movements promoted by the direct pathway (Albin et al., 1989; Crossman, 1990; DeLong, 1990). The SP+ and ENK+ striatal neurons of the avian basal ganglia are organized into direct and indirect pathway motor control circuits as well (Fig. 3). As in mammals, the SP+ direct pathway neurons have output to a motor thalamus cell group, called the ventrointermediate area (VIA) in birds, and to the SNr (Medina et al., 1997; Reiner et al., 1998; Medina and Reiner, 2000). An additional direct pathway target of SP+ striatal neurons exists in birds, a pretectal nucleus called spiriformis lateralis (SpL) (Reiner et al., 1998). The direct pathway to pretectum exists in reptiles and amphibians as well (Reiner et al., 1998), but is not obvious in mammals. These various striatal output pathways all appear to facilitate desired movements (Reiner et al., 1998). Birds also possess a subthalamic nucleus, and ENK+ neurons of the avian striatum are the source of an indirect pathway looping back to VIA, SpL, and SNr via the subthalamic nucleus (Jiao et al., 2000). Thus, motor control by the basal ganglia in birds is also organized into circuits that simultaneously promote desired movements while inhibiting potentially conflicting movements.

thumbnail image

Figure 3. Circuit diagrams comparing the functional organization of the basal ganglia in mammals and birds. The pluses and minuses indicate whether specific projections use an excitatory (+) or inhibitory (−) neurotransmitter. Striatal projection neuron types, which all use GABA as their primary neurotransmitter, are distinguished by their characteristic neuropeptide. The terminology used for basal ganglia subdivisions in birds is now similar to that in mammals per the recent revision in avian brain nomenclature (Reiner et al., 2004b). As in mammals, the striatal and pallidal output circuitry of birds is organized into direct SP+ striatal outputs to pallidal neurons promoting movement and ENK+ striatal outputs to pallidal neurons inhibiting unwanted movement. The pallidal neurons of the indirect pathway have direct outputs to the targets of the SP+ striatal neurons (i.e., GPi, SNr, and SpL) and indirect outputs to these same targets via the subthalamic nucleus. In mammals, SP+ neurons target two populations of pallidal-type neurons (GPi and SNr), while in birds, three are targeted (GPi, SN, and SpL). It is not yet certain, however, if GPe-like neurons in the avian globus pallidus (where they are intermingled with GPi-type neurons) have a projection to GPi-type neurons of globus pallidus. Such a projection is present in mammals. The striatum in mammals receives extensive dopaminergic input from the midbrain and excitatory glutamatergic input from thalamus and cerebral cortex. Striatal inputs are similar in birds, with a massive dopaminergic input from the midbrain, an excitatory glutamatergic input from thalamus, and an excitatory glutamatergic input from most of the pallium overlying the striatum corresponding to the corticostriatal input of mammals. ENK, enkephalin; GLUT, glutamate; GPe, external segment of globus pallidus; GPi, internal segment of globus pallidus; SNr, substantia nigra pars reticulata; SP, substance P; SpL, nucleus spiriformis lateralis; STN, subthalamic nucleus.

Download figure to PowerPoint

Basal Ganglia Specializations in Vocalizing Birds

Two telencephalic circuits underlie song learning and production in oscine songbirds (Margoliash, 1997; Doupe and Kuhl, 1999). These circuits are well developed in males in those species in which only males sing and in both males and females in those species in which both sing (Langmore, 1998). One of these circuits, called the anterior forebrain pathway, is a circuit routed through a specialized part of the basal ganglia unique to songbirds called area X (Fig. 4). This circuit is a multisynaptic ipsilateral pathway connecting a higher-order pallial song control area called the higher vocal center (HVC) to the robust nucleus of the arcopallium (RA), whose full components are: HVC–area X of the basal ganglia–the dorsolateral medial nucleus of the thalamus (DLM)–the lateral magnocellular nucleus of the anterior nidopallium (LMAN)–RA (Nottebohm et al., 1976; Bottjer, 1997; Nordeen and Nordeen, 1997). The other song control circuit involves a direct projection from HVC to RA and is necessary for song production. By contrast, the X-DLM-LMAN-RA pathway is necessary for initial song learning (Bottjer et al., 1984; Scharff and Nottebohm, 1991) and rehearsal-based adjustments in song during adulthood (Doupe and Solis, 1997; Margoliash, 1997; Jarvis et al., 1998; Mello and Ribeiro, 1998; Brainard and Doupe, 2000). Area X is located in the medial striatum and, consistent with its location, predominantly appears to contain small GABAergic spiny neurons with the anatomical, neurochemical, and physiological characteristics of spiny striatal projection neurons (Grisham and Arnold, 1994; Farries and Perkel, 2002). Area X also contains a less numerous population of larger GABAergic neurons that are the neurons of area X that project to the thalamic nucleus DLM (Luo and Perkel, 1999a, 1999b; Reiner et al., 2004a). These neurons possess pallidal neurochemistry, physiology, and dendritic morphology (Luo and Perkel, 1999a; Farries and Perkel, 2002; Reiner et al., 2004a). The HVC input to area X seems likely to terminate extensively on spiny striatal-type neurons, with these spiny neurons then having their major efferent projection to pallidal-type neurons that project to DLM (Luo and Perkel, 1999a, 1999b; Farries and Perkel, 2002; Reiner et al., 2004a). Given its role in a well-defined behavior, study of area X, and its self-contained striatopallidal elements, could advance understanding of the cellular mechanisms underlying the role of the basal ganglia in motor learning and control (Doupe et al., 2005).

thumbnail image

Figure 4. Two telencephalic circuits underlie song learning and production in male oscine songbirds. The schematic shown in A illustrates these two circuits, as viewed in the sagittal plane. One circuit consists of a direct projection from the HVC to RA. The serially connected components of the other circuit, called the anterior forebrain pathway, are the cortical area called HVC, which receives direct input from avian auditory cortex; area X of the medial striatum (X); the dorsolateral medial nucleus of the thalamus (DLM); a cortical region termed the lateral magnocellular anterior nidopallium (LMAN); and the cortical area called the robust nucleus of the arcopallium (RA), which projects to the vocal motoneurons of the brainstem. B, captured using differential interference contrast microscopy, shows two neurons of area X retrogradely labeled from DLM with biotinylated dextran amine 3 kDa. These neurons have the morphological characteristics of pallidal neurons, which include possessing aspiny dendrites and a relatively large perikaryon (12–14 μm).

Download figure to PowerPoint

Songbirds are not the only group of birds possessing vocal learning ability and complex vocalizations. Parrots and hummingbirds also use learned vocalizations to communicate with conspecifics (Farabaugh et al., 1994; Durand et al., 1997; Pepperberg, 1999; Hile et al., 2000; Jarvis et al., 2000). While the vocalizing abilities of parrots, hummingbirds, and songbirds are likely to have evolved independently, based on the separate evolutionary histories of these avian groups, a similar pair of neural circuits subserve vocalization in all three (Ball, 1994; Striedter, 1994; Durand et al., 1997; Jarvis et al., 2000). The circuit in hummingbirds and parrots resembling the anterior forebrain pathway of songbirds includes a portion of the medial striatum called the magnocellular nucleus of medial striatum (MStm) in parrots and the vocal nucleus of the anterior paleostriatum (VAP) in hummingbirds (Ball, 1994; Striedter, 1994; Durand et al., 1997; Jarvis et al., 2000). While some details of the neurochemistry and circuitry of parrot MStm are known (Ball, 1994; Roberts et al., 2002), it is unknown if parrot MStm and hummingbird VAP possess the unique mix of striatal and pallidal elements found to be the case for area X of songbirds (Luo and Perkel, 1999a, 1999b; Farries and Perkel, 2002; Reiner et al., 2004a).

Avian Basal Ganglia Evolution

Birds evolved from theropod dinosaurs (Fig. 5), and so it is not surprising that there are no noteworthy differences between the basal ganglia in reptiles and birds in size, location, neurochemically defined cell types, inputs, outputs, or functional organization (Reiner, 2002). Thus, the basal ganglia in both birds and reptiles (collectively called sauropsids) show many organizational similarities to the basal ganglia of mammals. For this reason, it seems likely that the major aspects of the cellular composition and output organization of the basal ganglia, as well as the glutamatergic and dopaminergic inputs to striatal projection neurons and the related receptors and signaling mechanisms employed by specific types of striatal neurons in stem amniotes, were already as observed in living amniotes (Reiner, 2002). The size and organization of the basal ganglia in stem amniotes, however, must have been significantly advanced beyond that of the immediate anamniote ancestors, since the basal ganglia in all living anamniotes studied are substantially smaller and less well developed than in living amniotes (Marin et al., 1998a, 1998b; Reiner et al., 1998). The changes that occurred at the anamniote-amniote transition must have included the emergence of prominent modality-specific sensory thalamopallial pathways, pallial information processing, and motor planning areas, and major projections from pallium to striatum. Such changes appear to have been part of an increased involvement of the forebrain in processing sensory information and in behavioral planning in amniotes. As part of this, the pallium appears to have become the major source of the information that the striatum requires for its role in movement control (Reiner et al., 1998). These changes may have been necessary to allow a more complex and adaptable behavioral repertoire of food procurement and/or predator avoidance in the shift from the freshwater semiaquatic habitat of amphibians to the complex and variable terrestrial habitat of stem amniotes. Additionally, refinements and improvements in musculoskeletal anatomy occurred (Hildebrand et al., 1985; Butler and Hodos, 1996) and may have required commensurate enhancements in the neural substrate for movement control. Basal ganglia elaboration may have made it possible to promote effectively the desired and inhibit the nondesired movements from among the now richer repertoire of learned and stereotypic behaviors and movements. Specific elaborations in basal ganglia organization in the avian and mammalian lineages are relatively minor compared to the major modifications at the anamniote-amniote transition and chiefly concern differentiation of the globus pallidus into two sectors and loss of the direct pathway to pretectum in mammals.

thumbnail image

Figure 5. Cladogram showing the evolutionary relationships among the major living tetrapod groups plus dinosaurs. A geological time scale in millions of years before recent times (myr), with the geological eras and periods indicated, is shown at the left to provide a time frame for the major evolutionary lineage divergences and group origins. The cladogram is largely based on paleontological data (Romer, 1945; Gauthier et al., 1988; Lee, 1993, 1997). Carb, Carboniferous period; Cret, Cretaceous period; Dev, Devonian period; Jur, Jurassic period; Perm, Permian period; Tri, Triassic period.

Download figure to PowerPoint

AVIAN PALLIAL ORGANIZATION AND EVOLUTION

  1. Top of page
  2. Abstract
  3. AVIAN BASAL GANGLIA ORGANIZATION AND EVOLUTION
  4. AVIAN PALLIAL ORGANIZATION AND EVOLUTION
  5. AVIAN ARCOPALLIUM AND AMYGDALA
  6. MEDIAL TELENCEPHALIC WALL ORGANIZATION AND EVOLUTION IN BIRDS
  7. AVIAN THALAMIC ORGANIZATION AND EVOLUTION
  8. CONCLUSION
  9. Acknowledgements
  10. LITERATURE CITED

Avian Pallial Organization

Some groups of birds show behavioral and cognitive skills comparable to those of primates (Marler, 1996; Burish et al., 2004; Emery and Clayton, 2004). For example, crows are able to make and use simple tools, while parrots are capable of the rudiments of referential language and abstract categorical reasoning (Hunt, 2000; Pepperberg, 2002; Weir et al., 2002; Hunt and Gray, 2003). Moreover, many species of birds are capable of sophisticated foraging strategies, elaborate parental and social behavior, challenging homing and migratory behavior, and/or complex vocal learning. In light of this, it is not surprising that birds possess a large and complex brain. Brain size-bodyweight correlations in fact show that brain size for birds rivals that of many mammals (Jerison, 1985). Contributing to the similarity in brain size between birds and mammals is the similarity in the size of the cerebrum or telencephalon. The large avian cerebrum reflects expansion of both the basal ganglia as well as the regions above the basal ganglia, referred to as the pallium, which are now known to be functionally comparable to mammalian cerebral cortex, and which provide the substrate for the substantial behavioral and cognitive skills of birds.

The outdated notion that the avian telencephalon is one large basal ganglia went hand in hand with the outdated view that birds lacked the neural substrate for learning and cognition. The latter view began to collapse at the same time that the true extent of the basal ganglia in birds was being ascertained by neurochemical and pathway tracing studies. The pallial regions above what we now know to be the basal ganglia in birds possess two major subdivisions, the dorsal ventricular ridge (DVR) and the Wulst, with the former containing the regions now called the mesopallium and nidopallium, and the latter comprising the region now called the hyperpallium. While the relatively uniform distribution of neurons make the DVR and Wulst superficially resemble the striatal part of the basal ganglia in mammals (Fig. 1A), they are clearly less uniform than the striatum of mammals and contain cytoarchitectonically evident subregions, which are now known to be functionally specialized as well. This began to be revealed by Karten and colleagues (Karten, 1969; Nauta and Karten, 1970) in a series of pathway tracing and behavioral studies in parallel with their studies on the basal ganglia. They showed that the structures that had been called the neostriatum and hyperstriatum in birds (and are now called the nidopallium and hyperpallium) contain specific nuclei that receive direct visual, auditory, and somatosensory input from thalamus (Fig. 6) and play a role in sensory information processing resembling that of layer 4 neurons of sensory neocortex in mammals (Karten, 1969; Nauta and Karten, 1970). Additionally, they showed that the upper hyperpallium and the region now termed the arcopallium give rise to major descending projections to diverse sensory, premotor, and motor cell groups of the thalamus, midbrain, hindbrain, and spinal cord, reminiscent of the outputs of layers 5 and 6 of mammalian neocortex (Zeier and Karten, 1971; Karten et al., 1973). Subsequent work by others further reinforced the conclusion that hyperpallium, nidopallium, and arcopallium play roles in information processing and motor control (Bonke et al., 1979; Güntürkün, 1991, 1997; Heil and Scheich, 1991; Hodos, 1993; Wild et al., 1993; Knudsen and Knudsen, 1996; Cohen et al., 1998; Shimizu and Bowers, 1999). Early in his studies, Karten concluded, bolstered by the embryological studies of Källén (1962), that the hyperpallium and nidopallium were pallial territories (Karten, 1969; Nauta and Karten, 1970). More recent data on the regional expression of genes involved in specifying pallial identity during development, such as Emx-2, Pax6, and Tbr1, reinforce the conclusion that the avian telencephalic subdivisions now called the hyperpallium, mesopallium, nidopallium, and arcopallium (together with hippocampal complex and piriform cortex) derive from the developing telencephalic sector called the pallium (Smith-Fernandez et al., 1998; Medina and Reiner, 2000; Puelles et al., 2000), from which hippocampus, neocortex, claustrum, piriform cortex, and pallial amygdala derive in mammals.

thumbnail image

Figure 6. Schematic diagrams of frontal views through the cerebrum, diencephalon, and midbrain of the pigeon showing the cell groups that make up the major visual circuit and the major auditory circuit in birds. BG, basal ganglia; Ento, entopallium; EW, nucleus of Edinger-Westphal; Hp, hippocampus; M, mesopallium; Hy, hypothalamus; IC, inferior colliculus; N, nidopallium; Ov, nucleus ovoidalis; Rt, nucleus rotundus; SN, substantia nigra.

Download figure to PowerPoint

The most prominent and cytoarchitectonically distinct of the specialized regions of the DVR in birds are a visual region and an auditory region (Fig. 6), which represent the thalamorecipient parts of the major visual and auditory processing streams of the avian cerebrum, and which reside within the nidopallium (or lower DVR). The thalamorecipient visual region, now called the entopallium, receives its input from a distinct round thalamic nucleus called nucleus rotundus, which itself receives a bilateral visual input from a type of neuron in layer 13 of the visual midbrain roof (or tectum) termed a tectal ganglion cell (Karten et al., 1997; Luksch et al., 1998; Major et al., 2000; Hellmann and Güntürkün, 2001). A similar tectothalamopallial pathway is present in mammals. In less visually oriented placental mammals and in marsupials and monotremes, the thalamic component of this circuit is termed the lateral posterior (LP) nucleus, while in more visually oriented placental mammals (e.g., primates, tree shrews, and ground squirrels), the LP appears to have differentiated into a more complex cell group called the pulvinar (Robson and Hall, 1977; Jones, 1985). In either case, the caudal and/or inferior part of this thalamic nucleus resembles the avian nucleus rotundus in that it receives a bilateral input from tectal ganglion cells (Robson and Hall, 1977; Raczkowski and Diamond, 1981; Benevento and Standage, 1983; Huerta and Harting, 1983; Mooney et al., 1984; Lane et al., 1997; Major et al., 2000; Stepniewska et al., 2000). This tectorecipient thalamic region also resembles the avian nucleus rotundus in the topology of its telencephalic projection, since it projects heavily to a visual area of the temporal lobe of neocortex (Mathers et al., 1977; Robson and Hall, 1977; Haight et al., 1980; Raczkowski and Diamond, 1980; Coleman and Clerici, 1981; Luppino et al., 1988; Steele and Weller, 1993; Schobber, 1981; Lyon et al., 2003). We will refer to this visual area of the temporal cortex as the tectothalamic visual area (Vtt).

The auditory region of the DVR in birds is called field L, and it receives its input from a midline auditory thalamic nucleus that is ovoid in shape and consequently called the nucleus ovoidalis (Karten, 1968; Wild et al., 1993). Nucleus ovoidalis itself receives its auditory input from the auditory midbrain (the central nucleus of the inferior colliculus) (Karten, 1968; Wild et al., 1993). Field L is the primary auditory area (A1) of the avian telencephalon, and it consists of three cytoarchitectonic subdivisions called L1, L2, and L3, with L2 being the thalamorecipient zone (Karten, 1968; Bonke et al., 1979; Heil and Scheich, 1991; Wild et al., 1993). In its organizational features, the ovoidalis to field L circuit in birds resembles the pathway in mammals from the ventral part of the medial geniculate nucleus (MGNv) of the thalamus to the primary auditory area in temporal cortex (Karten, 1969; Clerici and Coleman, 1990). The projection from MGNv to primary auditory cortex of temporal lobe is present in monotremes, marsupials, and placental mammals (Kaas, 1980; Rowe, 1990; Beck et al., 1996), thus suggesting it was present in basal mammals as well (Fig. 5).

The Wulst contains two separate sensory areas, a primary somatosensory area (S1) and a primary visual area (V1), which receive their input from specific thalamic nuclei that will be discussed in more detail below. The primary sensory areas of Wulst and DVR give rise to intratelencephalic circuits that participate in information processing, with the circuits ultimately sending input to the areas of the Wulst and arcopallium that project on the brainstem and spinal cord. Thus, the avian hyperpallium, nidopallium and arcopallium, although not organized cytoarchitectonically into layers like mammalian neocortex, perform the same type of neural operations at the cellular level as mammalian neocortex (Karten and Shimizu, 1989; Karten, 1991; Cohen et al., 1998; Jarvis et al., 2005). Increase in the size and regional specialization of the neocortex in mammals and the pallium in birds appears to be the common underpinning of the emergence of complex vocal and cognitive abilities in particular mammalian and avian species (Brenowitz et al., 1997; Margoliash, 1997; Doupe and Kuhl, 1999; Jarvis et al., 2000). As noted above, specialized cell groups that reside in the nidopallium and arcopallium mediate the ability of songbirds to learn and produce species-typical song (Margoliash, 1997; Doupe and Kuhl, 1999). Analogous regions in the nidopallium and arcopallium, plus one or more in the mesopallium, mediate vocalization in parrots and hummingbirds (Jarvis and Mello, 2000; Jarvis et al., 2000; Iwaniuk and Hurd, 2005). The tool-making abilities of corvids and the conceptual abilities of parrots appear to be based in the expansion, in particular, of the mesopallium (Lefebvre et al., 2004; Iwaniuk and Hurd, 2005).

Avian Pallial Evolution

No living nonmammal has a telencephalic structure that possesses the six-layered cytoarchitecture characteristic of mammalian neocortex, and all living mammals do (Allman, 1990; Northcutt and Kaas, 1995). Thus, neocortex evolved uniquely in the mammalian lineage after its divergence from the lineage leading to living reptiles but before the radiation of mammals into monotremes, marsupials, and placentals (Fig. 5). Whether neocortex derived from any antecedent structures within the telencephalon of stem amniotes has been much debated, as has whether birds possess their own unique derivatives of those same antecedent structures. Recent opinions on this issue have divided into two camps. Both accept that stem amniotes possessed a simple cortical structure that was the forerunner of the superior part of mammalian neocortex and of the Wulst in birds (Medina and Reiner, 2000). The two hypotheses diverge on the origin of mammalian temporal neocortex. One viewpoint proposes that temporal neocortex had no antecedent in stem amniotes and arose de novo after the divergence of the lineages leading to modern mammals on the one hand and living reptiles and birds on the other (Bruce and Neary, 1995; Striedter, 1997; Smith-Fernandez et al., 1998; Puelles et al., 2000). This hypothesis further proposes that the DVR of birds, which seems to resemble temporal neocortex in possessing a Vtt and an A1, is derived from a subcortical pallial region in stem amniotes that in the mammalian lineage gave rise to the claustrum, endopiriform region, and/or the pallial amygdala (Fig. 7). This viewpoint we will refer to as the “de novo origin of temporal neocortex” hypothesis, and it has its antecedents in the work of Holmgren (1925). The alternative hypothesis proposes that a region in stem amniotes situated at the ventrolateral edge of dorsal cortex gave rise to temporal neocortex in the mammalian lineage and to at least rostral DVR in the sauropsid (reptile and bird) lineage (Fig. 7) (Karten, 1969, 1991; Nauta and Karten, 1970; Reiner, 1993; Butler, 1994a, 1994b). We will call this the “common origin of DVR and temporal neocortex” hypothesis. In the following sections, we evaluate these hypotheses.

thumbnail image

Figure 7. Proposed pallial homologies between mammals and birds using transverse sections through rat telencephalon and pigeon telencephalon for illustration according to the “de novo origin of temporal neocortex” hypothesis and the “common origin of temporal neocortex and DVR” hypothesis. The key difference between the two hypotheses concerns the proposed mammalian homologue of the DVR of birds. The de novo hypothesis proposes that temporal neocortex has no homologue in reptiles and birds, and that reptilian/avian DVR is homologous to the claustrum, endopiriform region, and/or parts of the basolateral/basomedial amygdala. By contrast, the common origin hypothesis proposes that rostral DVR of reptiles and birds is homologous to temporal neocortex of mammals. While the sauropsid homologue of the claustrum-endopiriform-pallial amygdaloid complex is not entirely clear, it may include a small region just deep to olfactory cortex or to parts of caudal DVR. Shadings are used to highlight the homologous structures according to each of these two hypotheses.

Download figure to PowerPoint

Common origin of DVR and temporal neocortex.

Proponents of the homology of temporal neocortex and DVR have offered several lines of evidence in favor of this view. First, the topological and cytoarchitectonic continuity of rostral dorsal cortex and DVR in reptiles resembles that of superior and temporal neocortex in mammals and has led some authors to propose the homology of rostral DVR and temporal neocortex (Reiner, 1993; Butler, 1994a, 1994b). The similarity and continuity of rostral dorsal cortex and DVR is evident in turtles, but is especially impressive in the primitive lizard-like rhynchocephalian reptile Sphenodon punctatum, the tuatara, in which the DVR consists of a cell plate that is continuous with that of dorsal cortex (Fig. 8) (Elliot-Smith, 1919; Durward, 1930; Reiner and Northcutt, 2000). The evolutionary histories of turtles and Sphenodon suggest that this may resemble the ancestral cytoarchitectural pattern for the DVR (Reiner and Northcutt, 2000). Paleontological data indicate that living reptiles are members of two lineages that diverged from stem amniotes, the anapsid reptiles, a primitive radiation from which turtles appear to have arisen, and the diapsid reptiles, from which lizards, snakes, Sphenodon, and crocodilians arose (Fig. 5) (Romer, 1945; Gauthier et al., 1988; Lee, 1993, 1997). The diapsid lineage itself was characterized by an early divergence into the lepidosaurian and archosaurian lineages. Sphenodon is an early and relatively unchanged offshoot of the former (with lizards and snakes arising later), and crocodilians, dinosaurs, and birds from the latter. Thus, both the turtle and Sphenodon diverged early during reptile evolution, with Sphenodon retaining many primitive lepidosaurian features, among which is likely to be a primitive DVR.

thumbnail image

Figure 8. Image and schematic illustrating the organization of Sphenodon dorsal cortex and dorsal ventricular ridge (DVR). The illustrations show an image of a transverse section through the rostral telencephalon that was stained for succinic dehydrogenase (SDH), juxtaposed to a line drawing of this same section. The major telencephalic subdivisions are identified, as are the tectothalamic visual (Vtt) and primary auditory (A1) areas of the rostral DVR. The dorsal cortex in Sphenodon constitutes a thin tissue overlying the lateral ventricle. Note that the DVR in Sphenodon consists of a cell plate that histologically resembles the dorsal cortex. The two SDH+ zones in Sphenodon DVR are very likely to be Vtt and A1, based on the documented efficacy of SDH histochemistry in identifying these sensory areas in reptile and bird DVR, and the position of these zones in birds and other reptilian species (Reiner and Northcutt, 2000).

Download figure to PowerPoint

The second line of evidence for homology of rostral DVR and temporal neocortex deals with the high similarity in their connectivity. For example, both contain a Vtt and an A1, with the thalamic and midbrain cell groups of origin for these inputs being remarkably similar in their neurochemical, neuroanatomical, and functional organization between mammals on the one hand and reptiles and birds on the other (Karten, 1969, 1991; Pritz, 1980; Balaban and Ulinski, 1981; Bruce and Butler, 1984a, 1984b; Brauth and Reiner, 1991; Reiner, 1993, 1994; Cohen et al., 1998; Bruce et al., 2002; Ferland et al., 2003; Haesler et al., 2004; Teramitsu et al., 2004; Yamamoto and Reiner, 2005). While such similarities could be independently evolved, the similarities are so extreme in the midbrain and thalamic parts of the circuit, as discussed in more detail in the subsequent section on thalamus, as to be unlikely to have evolved separately (Brauth and Reiner, 1991; Luksch et al., 1998; Major et al., 2000; Bruce et al., 2002; Ferland et al., 2003; Haesler et al., 2004; Teramitsu et al., 2004; Yamamoto and Reiner, 2005). A third line of evidence in support of the homology of temporal neocortex and DVR is that the topological arrangement of S1, V1, Vtt, and A1 in reptiles and birds (Fig. 9) is identical to that likely to be the ancestral pattern for mammals (Orrego, 1961; Hall et al., 1977; Kaas, 1980; Coleman and Clerici, 1981; Luppino et al., 1988; Rowe, 1990; Lyon et al., 2003; Reiner, 1993; Beck et al., 1996). This pattern could not have been carried over from the amphibian ancestors of stem amniotes, since there is no evidence from modern amphibians that these areas existed in the stem tetrapods ancestral to amphibians and amniotes (Northcutt and Kicliter, 1980). Thus, the similarity between modern sauropsids and mammals in the topology of these cortical sensory areas is likely to be due to common inheritance from a stem amniote possessing these same areas in this same configuration.

thumbnail image

Figure 9. Schematics illustrating the location and extent of the major thalamorecipient sensory areas in side view in the turtle cortex and DVR, with DVR flattened to provide a schematized side view; and tree shrew. While many sensory areas are present in the neocortex of tree shrew, this image focuses on those sensory areas found in turtle dorsal cortex and rostral DVR, namely, V1, S1, Vtt, and A1 (as defined in the text). Note that V1 and S1 are above the dotted line, which separates dorsal cortex and DVR in turtles and separates superior and temporal neocortex in tree shrew. By contrast, Vtt and A1 are in the DVR of turtles and in the temporal neocortex of tree shrew. In addition, note that S1 in both turtle and tree shrew is rostral to V1, while V1, V2, and A1 are arrayed in a superior to inferior sequence. Placement of these sensory areas is based on the references cited in the text.

Download figure to PowerPoint

Karten (1969, 1991; Nauta and Karten, 1970) on the one hand and Reiner and Butler on the other (Reiner, 1993, 2000; Butler, 1994b) have proposed different scenarios for the common evolution of the DVR of extant sauropsids and the temporal neocortex of mammals from stem amniotes. Both viewpoints assume that stem amniotes possessed a dorsal cortex and a small proto-DVR at the lateral angle of the ventricle, which possessed distinct S1, V1, Vtt, and A1 regions. Karten (1969, 1991; Nauta and Karten, 1970) proposed that this dorsal cortex and proto-DVR possessed the major neuron types making up the different layers of neocortex (Fig. 10). Karten further suggested that in the mammalian lineage the proto-DVR neurons came to migrate to the lateral telencephalic surface and array themselves in the six-layered fashion that characterizes neocortex, while neurons of dorsal cortex arrayed themselves into six-layered cortex dorsally. In the sauropsid lineage, by contrast, Karten suggested that the neurons of the proto-DVR proliferated and formed distinct cell groups in situ, with the DVR then coming to bulge into the lateral ventricle due to its enlargement. The alternative hypothesis (Reiner, 1993, 2000; Butler, 1994b) posits that the cells of the proto-DVR and dorsal cortex in stem amniotes largely consisted of the type of corticocortical projection neuron found in cortical layers 5–6 in living mammals, which receive thalamic input on their apical dendrites (Fig. 11). The view offered by Reiner (2000), in particular, hypothesizes that the cytoarchitecture of the proto-DVR of stem amniotes resembled that of dorsal cortex, as true of DVR in Sphenodon. The changes leading to the therapsid (i.e., early mammal lineage) grade would have consisted mainly in a shift in the final adult positions of the cortical plate, the basal ganglia, and the piriform cortex (Fig. 5). Addition of layer 2–4 cell types and the shift to an inside-out gradient of neurogenesis would have occurred at some point in the transition from therapsids to mammals (Reiner, 1991, 1993; Butler, 1994b). In the sauropsid lineage, the changes in the proto-DVR leading to the Sphenodon grade would have consisted mainly of enlargement of the proto-DVR and its ingrowth into the ventricle. Cell types resembling cortical layer 2–4 cell types in mammals are assumed to have evolved independently in sauropsids in this scenario, perhaps not till dinosaurs or birds (Fig. 5). Both the Karten (1991) and the Reiner-Butler scenarios account for the topological similarities in the cortical sensory areas evident among modern sauropsids and mammals and the prominent similarities in the Vtt and A1 circuits.

thumbnail image

Figure 10. Schematic illustrations depicting the evolutionary derivation of sauropsid DVR (A) and mammalian temporal cortex (B) proposed by Karten (Karten, 1969, 1991; Nauta and Karten, 1970). In the pallium of both lineages, numerous neurons proliferate during development from the subventricular zone (arrows 1 and 2), which give rise to the dorsal cortex in sauropsids and the superior neocortex in mammals. Only short stretches of the continuous dorsal proliferative zone are indicated. Temporal neocortex in mammals is depicted as arising from an augmentation and migration of neuroblasts derived from the proto-DVR of stem amniotes to the lateral telencephalic surface (arrow 3). In sauropsids, proto-DVR also comes to be augmented, but its neurons remain at the lateral angle of the ventricle and come to bulge into the lateral ventricle, forming the DVR.

Download figure to PowerPoint

thumbnail image

Figure 11. Schematic illustration of evolution of the pallium proposed by Reiner (2000) and Butler (1994b), who hypothesize a proto-DVR in stem amniotes. The evolutionary changes in the proto-DVR leading to early mammals involve a shift in the positions of the cortical plate, the basal ganglia, and the piriform cortex, followed by a thickening and differentiation of the cortical plate, presumably accompanied by the addition of the layers 2–4 cell types. Finally, the changes in the neocortex from the early mammal to the early placental grade are assumed to consist mainly in the expansion and further laminar differentiation of the neocortex, and the further ventralward shift of the olfactory cortex. The evolutionary changes in the proto-DVR leading to the Sphenodon grade consist mainly of enlargement of the DVR and its ingrowth into the ventricle. The changes in the DVR from the Sphenodon to the turtle grade consist of cytodifferentiation of the DVR into cell groups. Evolution to the avian grade consisted of further differentiation of the DVR. Note that all schematic drawings are of transverse sections through the telencephalon at approximately the level of the anterior commissure. DP, dorsal pallidum; S, septum; St, striatum.

Download figure to PowerPoint

De novo origin of temporal neocortex.

Proponents of the de novo origin of temporal neocortex have offered three lines of evidence to argue for the homology of the DVR of reptiles and birds to the claustrum, endopiriform region, and/or pallial amygdala of mammals. Because the DVR largely lies deep to piriform (olfactory) cortex in reptiles and birds (Fig. 7), as do the claustrum, the endopiriform region, and the pallial amygdala in mammals (Fig. 7), several investigators have proposed a homology of the DVR of birds and reptiles to claustrum, endopiriform region, and/or pallial amygdala of mammals based on topological similarity (Bruce and Neary, 1995; Striedter, 1997). In general, however, topological relationships of homologous structures are not so invariant as to make an argument for homology solely on this basis secure (Swanson and Petrovich, 1998). A second line of evidence for the homology of DVR of birds and reptiles to claustrum, endopiriform region, and/or the pallial amygdala of mammals concerns a claimed similarity in embryological derivation from the basolateral edge of the pallial proliferative zone during neurogenesis (Striedter, 1997). The data for this interpretation are, however, not unambiguous, and Källén (1962) claimed that DVR of birds and reptiles has a different spatial and temporal derivation from the embryonic pallial proliferative zone than do claustrum, endopiriform region, and pallial amygdala of mammals.

The third basis for claiming homology of DVR to subcortical pallial regions in mammals comes from homeobox gene mapping studies (Smith-Fernandez et al., 1998; Puelles et al., 2000). Among the genes expressed preferentially in pallial regions, Tbr-1, Emx-2, and Pax-6 are expressed throughout the entire pallium in mammals (Smith-Fernandez et al., 1998; Puelles et al., 2000). In contrast, expression of Emx-1 was initially claimed to be restricted to the developing hippocampal cortex, neocortex, olfactory cortex, claustrum, and more lateral parts of the pallial amygdala, but absent from the endopiriform region and more medioventral parts of the pallial amygdala of mammals (Puelles et al., 2000). Puelles et al. (2000) termed the Emx-1-negative region the “ventral pallium” and suggested it was a phylogenetically conserved pallial sector that all amniotes possess as part of their pallial bauplan. The genes Tbr-1, Emx-2, and Pax-6 are also expressed throughout the pallium in reptiles and birds (Smith-Fernandez et al., 1998; Puelles et al., 2000). The ventral DVR in reptiles and birds, however, does not appear to express Emx-1, and Puelles et al. (2000) suggested this territory was the ventral pallium of sauropsid telencephalon and homologous to the endopiriform region and parts of pallial amygdala of mammals. Medina et al. (2004) recently proposed that the parts of pallial amygdala comprising the ventral pallium are the basomedial and lateral anterior nuclei. Puelles et al. (2000) also suggested that the dorsal DVR (i.e., the mesopallium in birds) is homologous to the claustrum, largely for topographic reasons.

Several lines of evidence rebut a homology of the DVR of reptiles and birds to the claustrum, endopiriform region, and/or pallial amygdala of mammals. First, histologically distinct claustral and endopiriform nuclei cytoarchitectonically resembling those in placental and marsupial mammals have not been unambiguously demonstrated in monotremes (Abbie, 1940; Divac et al., 1987; Butler et al., 2002; Ashwell et al., 2004). This raises the possibility that the claustrum and endopiriform may have arisen with the common ancestor of placental and marsupial mammals (Fig. 5). Under these circumstances, no part of the DVR of birds and reptiles could be homologous to the claustrum or endopiriform region. Second, recent fate-mapping studies more sensitive than those of Puelles et al. (2000) have revealed that among the putative ventral pallial nuclei, only the ventralmost part of the endopiriform nucleus is entirely Emx-1-negative (Guo et al., 2000; Gorski et al., 2002). In contrast, the remainder of the endopiriform nucleus and nearly all pallial amygdaloid nuclei are rich in Emx-1-expressing neurons. Thus, the evidence does not favor the view that a distinct ventral pallial territory persists during development and gives rise to specific Emx-1-negative ventral pallial nuclei in mammals, rendering problematic claims of homology for ventral DVR (i.e., nidopallium) of birds to specific claustroamygdaloid nuclei in mammals. Moreover, Emx-1 would be expected to be absent from the nidopallium according to Karten's view (1991) that the nidopallium contains the same granule cell neuron type as found in layer 4 of mammalian neocortex, because cortical layer 4 neurons do not express Emx-1 (Guo et al., 2000; Gorski et al., 2002). Consistent with this interpretation, neurons of both layer 4 of neocortex and the thalamorecipient parts of the nidopallium express the genes Rorb/Nr1f2 and EAG2 (Dugas-Ford and Ragsdale, 2003). Finally, as discussed below, the de novo hypothesis leaves to coincidence the extreme similarities in the mesothalamocortical circuits for the A1 and Vtt of mammals and extant sauropsids (Brauth and Reiner, 1991; Reiner, 1993; Luksch et al., 1998; Major et al., 2000; Bruce et al., 2002; Ferland et al., 2003; Haesler et al., 2004; Teramitsu et al., 2004; Yamamoto et al., 2005) and does not deal with the fact that the thalamic projections to the claustrum and endopiriform region are meager and do not resemble those to DVR (Sloníewski et al., 1986; Dinopoulos et al., 1992). As discussed in the subsequent section on the avian thalamus, thalamic projections to the pallial amygdala also do not resemble those to the avian DVR.

AVIAN ARCOPALLIUM AND AMYGDALA

  1. Top of page
  2. Abstract
  3. AVIAN BASAL GANGLIA ORGANIZATION AND EVOLUTION
  4. AVIAN PALLIAL ORGANIZATION AND EVOLUTION
  5. AVIAN ARCOPALLIUM AND AMYGDALA
  6. MEDIAL TELENCEPHALIC WALL ORGANIZATION AND EVOLUTION IN BIRDS
  7. AVIAN THALAMIC ORGANIZATION AND EVOLUTION
  8. CONCLUSION
  9. Acknowledgements
  10. LITERATURE CITED

Numerous authors have proposed that the region of the avian telencephalon formerly known as the archistriatum is at least partly comparable to mammalian amygdala (Edinger et al., 1903; Edinger, 1908; Ariëns-Kappers, 1922; Ariëns-Kappers et al., 1936; Zeier and Karten, 1971; Bruce and Neary, 1995; Puelles et al., 2000), a structure that itself possesses both pallial and subpallial subdivisions (Swanson and Petrovich, 1998; Puelles et al., 2000). Based on neurochemical and developmental data, it is clear that all parts of what was called the archistriatum in birds are pallial (Puelles et al., 2000; Reiner et al., 2002, 2004b; Wada et al., 2004; Sun et al., 2005). Moreover, the taenia (also called nucleus taeniae) has typically been regarded as a part of the archistriatal complex, although this was not reflected in its name (Ariëns-Kappers et al., 1936; Zeier and Karten, 1971; Thompson et al., 1998; Cheng et al., 1999; Absil et al., 2002), and that much or all of taenia is subpallial (Foidart et al., 1999; Cobos et al., 2001; Absil et al., 2002; Sun et al., 2005). Hodological, developmental, neurochemical, and behavioral evidence support the amygdaloid nature of the taenia and the posterior part of what was called the archistriatum (Zeier and Karten, 1971; Veenman et al., 1995; Lanuza et al., 2000; Puelles et al., 2000; Absil et al., 2002; Roberts et al., 2002; Yamamoto et al., 2005). By contrast, what were called anterior archistriatum, intermediate archistriatum, and at least lateral parts of medial archistriatum have largely somatic connectivity and neurochemistry, making them unlike amygdala in mammals (Zeier and Karten, 1971; Veenman et al., 1995; Davies et al., 1997; Mello et al., 1998; Reiner et al., 2002, 2004; Wada et al., 2004; Yamamoto and Reiner, 2005; Yamamoto et al., 2005). For these reasons, in the newly revised avian telencephalic nomenclature, the posterior part of what was called the archistriatum was renamed as the posterior amygdala, while the taenia was renamed the nucleus taenia of the amygdala. The posterior amygdala in birds may be at least partly comparable to the cortical amygdala of mammals, while the nucleus taenia appears homologous to part of the medial amygdala of mammals (Cheng et al., 1999; Absil et al., 2002; Reiner et al., 2004b; Yamamoto et al., 2005). For the remaining parts of what was the archistriatum, the term “arcopallium” has replaced the term “archistriatum” (Reiner et al., 2004b). The arcopallium receives input from diverse areas of the DVR and gives rise to descending projections to sensory, premotor, and motor areas of the diencephalon, midbrain, and hindbrain (Zeier and Karten, 1971; Wild et al., 1993; Davies et al., 1997). In this regard, its neurons are comparable to layer 5 and 6 neurons of the cerebral cortex in mammals, as also evidenced by the preferential expression of the ETS transcription factor ER81 in both the intermediate arcopallium and layer 5 neurons of neocortex (Dugas-Ford and Ragsdale, 2003). Finally, a subpallial region inferior to what is now called globus pallidus in birds has been recognized in the nomenclature revision as a subpallial part of the avian amygdala comparable to the sublenticular part of the mammalian extended amygdala and named the subpallial amygdala (Reiner et al., 2004b; Yamamoto et al., 2005).

MEDIAL TELENCEPHALIC WALL ORGANIZATION AND EVOLUTION IN BIRDS

  1. Top of page
  2. Abstract
  3. AVIAN BASAL GANGLIA ORGANIZATION AND EVOLUTION
  4. AVIAN PALLIAL ORGANIZATION AND EVOLUTION
  5. AVIAN ARCOPALLIUM AND AMYGDALA
  6. MEDIAL TELENCEPHALIC WALL ORGANIZATION AND EVOLUTION IN BIRDS
  7. AVIAN THALAMIC ORGANIZATION AND EVOLUTION
  8. CONCLUSION
  9. Acknowledgements
  10. LITERATURE CITED

Hippocampal Complex

The hippocampal complex constitutes the bulk of what is called the medial pallium in mammals. The mammalian hippocampal complex possesses several histologically distinct three-layered subdivisions, such as the dorsomedially situated dentate gyrus and Ammon's horn (or cornu ammonis), as well as the subiculum, by which the cornu ammonis grades into the posterior ventral pallium (Columbo and Broadbent, 2000; Day, 2003). The hippocampus in mammals is important for processing and retaining spatial information. Birds also possess a medial pallium, which is thought to be homologous as a field to the medial pallium of mammals (Columbo and Broadbent, 2000; Puelles et al., 2000). Nonetheless, the cytoarchitecture and neurochemistry of this region in birds does not lend itself to identification of simple one-to-one equivalents for the dentate gyrus, cornu ammonis, and subiculum of mammals (Erichsen et al., 1991; Krebs et al., 1991; Columbo and Broadbent, 2000; Day, 2003). Despite the prominent cytoarchitectural and neurochemical differences between the medial pallia of mammals and birds, this region in birds appears to contain a homologue of the mammalian hippocampal complex, since its afferent and efferent connections are highly similar to those of the hippocampal complex in mammals (inputs from thalamus, hypothalamus, diagonal band, locus ceruleus, raphe, and contralateral hippocampus and outputs to septum, diagonal band, and hypothalamus) and because the avian medial pallium too appears to play a major role in such spatial tasks as navigation, spatial learning, and spatial memory (Casini et al., 1986; Columbo and Broadbent, 2000; Atoji et al., 2002; Day, 2003). Notably, the size of the avian medial pallium is correlated with the extent to which birds use spatial abilities in such activities as food caching and nocturnal nesting in complex habitats (Krebs et al., 1989; Hampton et al., 1995; Abbott et al., 1999). Thus, despite the prominent difference between the avian medial pallium and mammalian medial pallium in cytoarchitecture and neurochemistry, the two regions are homologous and carry out remarkably similar functions (Columbo and Broadbent, 2000; Day, 2003). Moreover, the role of the medial pallium in spatial tasks appears to have been conserved throughout vertebrate evolution to an extent that might not be readily inferred from the cytoarchitectural variation of this region among living vertebrates (Bingman et al., 2003; Day, 2003).

Septum

The septum constitutes most of the subpallial part of the medial wall of the telencephalon in mammals (Jakab and Leranth, 1995; Puelles et al., 2000). In mammals, the septum receives a prominent input from the hippocampus and projects to a variety of limbic brain regions involved in regulating responses to stress and in regulating social behaviors, including sexual and aggressive behaviors (Goodson et al., 2004). Projection targets of the septum by which it mediates these functions include the medial amygdala, preoptic area, anterior hypothalamus, ventromedial hypothalamus, and central gray (Jakab and Leranth, 1995; Risold and Swanson, 1997a, 1997b; Goodson et al., 2004). A septal region resembling that of mammals in cytoarchitecture and shape is present in the telencephalon of most jawed vertebrate groups, the most notable exception being in the case of the everted telencephalon of ray-finned fish (Northcutt, 1995). Detailed analysis of the neurochemistry and connectivity of the septum in songbirds has identified the four main subdivisions of the avian septum and shown that they are homologous to the main subdivisions of the septum in mammals (Casini et al., 1986; Atoji et al., 2002; Goodson et al., 2004; Montagnese et al., 2004). The avian septal nuclei are the medial septal nucleus, the lateral septal nucleus, the septohippocampal nucleus, and the caudocentral septum (Goodson et al., 2004). The first three of these are homologous to the similarly named mammalian septal nuclei, and the caudocentral nucleus in birds appears to be homologous to the septofimbrial nucleus of mammals according to Goodson et al. (2004). These septal nuclei must have already been present in the stem amniotes for birds and mammals to both possess them (Fig. 5), and the similarities between birds and mammals for this region suggest birds to be a useful model system in which to elucidate more details of their function.

AVIAN THALAMIC ORGANIZATION AND EVOLUTION

  1. Top of page
  2. Abstract
  3. AVIAN BASAL GANGLIA ORGANIZATION AND EVOLUTION
  4. AVIAN PALLIAL ORGANIZATION AND EVOLUTION
  5. AVIAN ARCOPALLIUM AND AMYGDALA
  6. MEDIAL TELENCEPHALIC WALL ORGANIZATION AND EVOLUTION IN BIRDS
  7. AVIAN THALAMIC ORGANIZATION AND EVOLUTION
  8. CONCLUSION
  9. Acknowledgements
  10. LITERATURE CITED

Avian Thalamic Organization

As in mammals, the avian thalamus is a multinucleated region that receives sensory inputs and projects to the pallial and striatal sectors of the telencephalon. Three major regions can be recognized, a dorsomedially situated intralaminar thalamus that receives diverse inputs and has polysensory projections to both striatum and pallium, a dorsally and laterally situated lemnothalamus that receives specific lemniscal sensory inputs and projects to modality-specific parts of the Wulst (hyperpallium), and a more ventrally and medially situated collothalamus, which receives visual and auditory input from midbrain and projects to the lower DVR (nidopallium) (Butler, 1994a, 1994b; Gonzalez et al., 2002). The dorsomedial thalamic nuclei in birds (also called the dorsomedial thalamic zone, or DTZ) are comparable to the intralaminar, midline, and mediodorsal thalamic nuclear complex (IMMC) of mammals (Fig. 12) based on neurochemical and topographic considerations (Veenman et al., 1997). The diverse sources of input to the DTZ also resemble those of the IMMC in mammals (Zeier and Karten, 1971; Wild, 1989; Korzeniewska and Güntürkün, 1990; Medina and Reiner, 1997; Veenman et al., 1997). The common expression of ErbB4 in neurons of the DTZ of birds and IMMC of mammals further supports the view they are homologous (Bruce et al., 2002). The lemnothalamus includes the dorsointermediate ventral area (DIVA), which transmits somatosensory information to the rostral S1 of Wulst, the ventrointermediate area (VIA), which transmits basal ganglia outflow to the M1 of Wulst, and the dorsal lateral geniculate nucleus, which transmits retinal information to V1 of the Wulst (Karten et al., 1973; Medina et al., 1997; Medina and Reiner, 2000; Reiner, 2000). The major collothalamic nuclei are the nucleus ovoidalis, which receives tonotopic auditory input from the inferior colliculus and projects to field L (the primary auditory cortex), and nucleus rotundus, which receives visual input from the tectum (much of it motion-related) and projects to the entopallial nucleus of the nidopallium (Karten et al., 1997; Luksch et al., 1998; Reiner, 2000). These various ascending thalamopallial sensory projections provide the telencephalon of birds with the information it needs to execute its role in higher-order information processing, learning, and cognition.

thumbnail image

Figure 12. Drawings of cross-sections illustrating the comparability of the avian dorsal thalamic zone (DTZ) and the mammalian intralaminar, mediodorsal, and midline thalamic nuclear complex (IMMC). The schematics to the left represent transverse views through pigeon DTZ at a rostral (A) and a caudal level (C), while the schematics to the right represent transverse views through rat IMMC at a rostral (B) and a caudal level (D). The three different shading patterns indicate comparable parts of the avian DTZ and the mammalian IMMC. Note that Korzeniewska and Güntürkün (1990) have suggested that the DLP corresponds to the mammalian posterior thalamic nucleus. CL, centrolateral thalamic nucleus; CM, central medial thalamic nucleus; DIP, nucleus dorsointermedius posterior thalami; DLM, nucleus dorsolateralis anterior thalami, pars medialis; DLP, nucleus dorsolateralis posterior thalami; DMA, nucleus dorsomedialis anterior thalami; DMP, nucleus dorsomedialis posterior thalami; FRf, fasciculus retroflexus; Hb, habenular nuclei; IMD, intermediodorsal nucleus; LHb, lateral habenular nucleus; MD, mediodorsal nucleus; MHb, medial habenular nucleus; PC, paracentral thalamic nucleus; PF, parafascicular thalamic nucleus; PV, nucleus paraventricularis; Re, reuniens thalamic nucleus; Rh, rhomboid thalamic nucleus; SHL, lateral subhabenular nucleus; SPC, nucleus superficialis parvocellularis.

Download figure to PowerPoint

Avian Thalamic Evolution

Since the target of the lemnothalamus of birds (the hyperpallium) is accepted as homologous to the superior sector of mammalian neocortex, it is not surprising that the lemnothalamic nuclei are accepted as homologous to the comparable mammalian nuclei. Thus, the DIVA is considered homologous to the mammalian ventrobasal thalamic complex, VIA to mammalian motor thalamus, and the avian dorsal lateral geniculate nucleus to the mammalian dorsal lateral geniculate nucleus (Medina and Reiner, 2000). The homology of the avian collothalamic nuclei to specific mammalian thalamic nuclei, however, has been the subject of disagreement. Since this dispute is germane to interpretations of the mammalian homologue of the DVR, we will discuss the opposing viewpoints for the mammalian homologues of nucleus rotundus and nucleus ovoidalis held by proponents of the two major competing hypotheses on the evolutionary relationship of DVR and cerebral cortex.

Nucleus rotundus.

Proponents of the de novo origin of temporal cortex reject the hodological evidence for DVR and temporal neocortex homology presented by Karten and others because they claim that nucleus rotundus in birds is homologous to the posterior/suprageniculate nuclear complex in mammals, rather than to the lateral posterior/caudal pulvinar complex as in the “common origin of DVR and temporal cortex” view. Part of the evidence for the notion that nucleus rotundus is homologous to the posterior/suprageniculate nuclear complex in mammals relates to an interpretation of homology of the tectal layers that project to these thalamic regions. Karten and his coworkers have noted the extreme morphological and functional similarity of the neurons that project to nucleus rotundus and posterior lateral posterior/caudal pulvinar (LP/cPUL), which reside in layer 13 (called the stratum griseum centrale) of avian tectum and the deepest sublayer of stratum griseum superficiale (SGS3) of the mammalian superior colliculus, respectively (Karten et al., 1997; Luksch et al., 1998; Major et al., 2000). Davila et al. (2000) argue that layer 13 of birds is a deep tectal layer and akin to a deep layer of the mammalian superior colliculus termed the “stratum griseum intermedium,” which projects to the posterior/suprageniculate nuclear complex. This view overlooks the profound morphological and functional similarity of the avian layer 13 neurons to mammal SGS3 neurons (Karten et al., 1997; Luksch et al., 1998; Major et al., 2000) and stems from the claims of Puelles and coworkers that stratum griseum intermedium (SGI) neurons in mammals develop early, as do tectal layer 13 neurons in birds, but that layer SGS3 neurons develop relatively later (compared to other layers) than do layer 13 neurons in birds (Davila et al., 2000). This claim is, however, based on an inaccurate simplification of published data on the laminar histogenesis of mammalian superior colliculus (Güntürkün, 2003). In brief, Davila et al. (2000) claimed that the published data of Altman and Bayer (1981) show that the neurons of the collicular layer projecting to LP/cPUL (in SGS3) are generated later in development than are neurons in deeper colliculus, and that the tectal layer projecting to avian nucleus rotundus arises earlier than other tectal layers. By this reasoning, avian tectal layer 13 is more like mammalian SGI than it is like mammalian SGS3, and thus its target more like posterior/suprageniculate nuclear complex than like LP/cPUL. The claim for deep colliculus in mammals is, however, based on only one early-born minority large neuron type in a sublayer of SGI. In fact, neurons of the superficial gray layer (SGS) in mammals otherwise have birthdates overlapping those of the neurons in other layers. Thus, the developmental data are inconclusive, while morphological evidence favors the view that layer 13 of birds is like SGS3 of mammals, and that nucleus rotundus is thus homologous to mammalian LP/cPUL. Recent data on ErbB4, LAMP, and FoxP2 localization in avian and mammalian thalamus further supports this interpretation, since nucleus rotundus and LP/cPUL are rich in LAMP and FoxP2 and poor in ErbB4, while the suprageniculate/posterior nuclear region is poor in LAMP and FoxP2 but rich in ErbB4 (Bruce et al., 2002; Ferland et al., 2003; Haesler et al., 2004; Teramitsu et al., 2004; Yamamoto and Reiner, 2005).

Nucleus ovoidalis.

One of the proponents of the de novo hypothesis has suggested that the projection of nucleus ovoidalis to L2 of field L resembles the projection of posterior intralaminar thalamus (PIN) to the mammalian lateral amygdalar nucleus (Bruce and Neary, 1995). A number of lines of evidence indicate this interpretation to be incorrect and support the alternative view that ovoidalis is homologous to ventral MGN (MGNv) in mammals. First, the developmentally regulated marker ErbB4 and the neuropeptide CGRP are both found in neurons of PIN in mammals but rarely in MGNv neurons (Fig. 13) (Brauth and Reiner, 1991; Yasui et al., 1991; Bruce et al., 2002). In birds, these markers are found in a nucleus that surrounds nucleus ovoidalis (Ov) called periovoidalis (pOv), but not notably in Ov itself (Brauth and Reiner, 1991; Lanuza et al., 2000; Durand et al., 2001; Bruce et al., 2002). Moreover, both the avian peri-Ov and the mammalian PIN receive auditory input from the external nucleus of the inferior colliculus, while the avian Ov and mammalian MGNv receive their auditory input from the core of the inferior colliculus (Winer, 1991; Durand et al., 1992; Wild et al., 1993; Wang and Karten, 2004). The nucleus ovoidalis displays an orderly tonotopic map (Biedermann-Thorsen, 1970; Zaretsky and Konishi, 1976) similar to that observed in the mammalian MGNv, but not observed in the medial and dorsal subnuclei of the mammalian MGN (Winer, 1991). In the case of both avian Ov and mammalian MGNv, the tonotopy derives from the input from the core of the inferior colliculus, which is itself tonotopically ordered (Winer, 1991; Durand et al., 1992; Wild et al., 1993). Curiously, a specific part of the mammalian MGNv has also been called the “nucleus ovoideus” by Cajal (1966) or the “pars ovoidea” (Winer, 1991). Thus, the projection of Ov to L2 does not resemble that of mammalian PIN to the lateral anterior amygdala, since Ov does not resemble PIN in either its neurochemistry or source of midbrain input. Rather, Ov resembles MGNv, and its projection to L2 of field L thus resembles that of MGNv to layer 4 of A1 of temporal cortex in mammals.

thumbnail image

Figure 13. Images of CGRP-immunostained neurons in the peripeduncular-intralaminar nuclei (PIN) of rat thalamus (A) and the periovoidalis (pOv) region of pigeon thalamus (B) in transverse view. Note that both PIN and pOv lie inferior to and partly surround the main auditory thalamic nuclei, the ventral subdivision of the medial geniculate nucleus (MGNv) in rats, and the nucleus ovoidalis (Ov) in pigeons. By their neurochemistry and input from the midbrain, these CGRP+ populations of neurons appear to be homologous. Medial lies to the left and dorsal to the top in both images.

Download figure to PowerPoint

CONCLUSION

  1. Top of page
  2. Abstract
  3. AVIAN BASAL GANGLIA ORGANIZATION AND EVOLUTION
  4. AVIAN PALLIAL ORGANIZATION AND EVOLUTION
  5. AVIAN ARCOPALLIUM AND AMYGDALA
  6. MEDIAL TELENCEPHALIC WALL ORGANIZATION AND EVOLUTION IN BIRDS
  7. AVIAN THALAMIC ORGANIZATION AND EVOLUTION
  8. CONCLUSION
  9. Acknowledgements
  10. LITERATURE CITED

During the early 20th century, birds were thought to possess an inflexible, innate behavioral repertoire. Accordingly, the forebrain of birds was thought to be dominated by a hypertrophied basal ganglia, which at that time was thought to mediate reflexive behavior. It is now clear that many birds possess a complex and malleable behavioral repertoire mediated by a brain that rivals that of many mammals in size and sophistication. The cerebral hemispheres of birds are especially enlarged, and the parts of them devoted to higher-order processes such as learning and cognition, namely the Wulst and DVR, are comparable to the cerebral cortex of mammals and subserve similar functions. The recognition that the avian telencephalon superficially looks very different than the mammalian telencephalon but performs similar functions provides challenges and opportunities for further elucidating the evolutionary and neurobiological bases of the difference in appearance and similarity in function.

The morphological differences are least prominent in the case of the basal ganglia, consisting mainly in the greater segregation of some neuronal populations than in mammals, and in the specialization of part of the basal ganglia for a role in vocal learning and vocalization. These differences could be profitably exploited to ascertain aspects of basal ganglia function shared in common by birds and mammals. For example, the segregation of the two main types of spiny projection neurons in different parts of the avian striatum and of the two main types of corticostriatal neurons in different parts of the telencephalic pallium (Veenman et al., 1995; Reiner et al., 1998, 2001) potentially makes birds convenient for comparing the physiology of these different striatal and corticostriatal neuron types. In the case of songbirds, the dedication of a specific part of the basal ganglia (i.e., area X) to a well-defined behavior (i.e., vocal learning) potentially provides a model system for elucidating the cellular mechanisms underlying the broader role of the striatum and pallidum in motor learning and motor plasticity (Doupe et al., 2005). Similarly, the now evident evolutionary conservatism in the septum, medial pallium, and parts of the amygdala among amniotes indicates the utility of birds for addressing fundamental structure-function questions about these regions. Of course, birds and mammals are not identical in their social, appetitive, and reproductive behaviors, and investigation of the neurobiological bases of any differences will be valuable for understanding the changes in otherwise similar structures that allow for species-typical functional differences.

Further study of the avian pallium offers the opportunity to address a particularly intriguing issue concerning the relationship between neural design and function: how higher-order cognition can be carried out without a laminated cerebral cortex, which earlier generations of comparative neuroanatomists had regarded as not possible. In part, birds are able to do this because their telencephalic pallium contains the same types of neurons as does the cerebral cortex in mammals, except they are arrayed into nuclei rather than layers (Karten, 1991). In some regions, however, the neurons of the different types are organized into flattened juxtaposed cell groups that have the appearance of layers. A noteworthy example of this is field L, the primary auditory region of birds (Karten, 1968; Bonke et al., 1979; Heil and Scheich, 1991; Wild et al., 1993). Field L contains a flattened disk-like collection of neurons called L2 that receives the auditory thalamic input, and a medially juxtaposed L1 and laterally juxtaposed L3. The L1 and L3 resemble layers 2/3 of primary auditory cortex in their connectivity. It would be interesting to know to what extent the functional interactions of L2 with L1/L3 resemble those of layer 4 with layers 2/3 in mammals. Moreover, the evolutionary basis of any such resemblances is unresolved. While all studied reptiles possess a homologue of field L, this region is not obviously organized into cytoarchitectonically distinct L1, L2, and L3 subfields (Reiner and Northcutt, 2000). Nonetheless, the cell types characteristic of L1-3 could be present in reptiles, but covert. Further study of this issue in birds and reptiles will clarify whether birds have independently evolved a pseudolaminar organization in the primary auditory pallium resembling that in mammals to allow their refined auditory abilities, or whether they have built on cell types already present in stem amniotes. Along these lines, further information on the genes regulating the identity and development of specific parts of the Wulst and DVR in birds and their apparent homologues in mammalian cerebral cortex will be valuable for further characterizing the evolution of the avian pallium and its evolutionary relationship to the mammalian neocortex (Puelles et al., 2000; Yamamoto et al., 2005). Such information will help provide insight into, for example, the precise regional and cellular homologue of the avian entopallium in mammalian temporal cortex. Finally, regardless of which features of avian pallial organization are newly evolved and which ancestral, study of structure-function relationships in avian and mammalian pallia will help reveal which neural traits are and which are not essential for allowing higher-order cognition. Such studies in birds will also help clarify the basis of the difference in cognitive abilities found among birds, for example, the changes in the size and complexity of the mesopallium accounting for the substantial cognitive differences between chickens and pigeon on the one hand and parrots and corvids on the other.

Acknowledgements

  1. Top of page
  2. Abstract
  3. AVIAN BASAL GANGLIA ORGANIZATION AND EVOLUTION
  4. AVIAN PALLIAL ORGANIZATION AND EVOLUTION
  5. AVIAN ARCOPALLIUM AND AMYGDALA
  6. MEDIAL TELENCEPHALIC WALL ORGANIZATION AND EVOLUTION IN BIRDS
  7. AVIAN THALAMIC ORGANIZATION AND EVOLUTION
  8. CONCLUSION
  9. Acknowledgements
  10. LITERATURE CITED

The authors thank their various collaborators over the years who have made the parts of their research presented here possible. They particularly thank Drs. Glenn Northcutt, Steve Brauth, Loreta Medina, and Ann Butler for their many stimulating conversations and ideas about basal ganglia evolution.

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. AVIAN BASAL GANGLIA ORGANIZATION AND EVOLUTION
  4. AVIAN PALLIAL ORGANIZATION AND EVOLUTION
  5. AVIAN ARCOPALLIUM AND AMYGDALA
  6. MEDIAL TELENCEPHALIC WALL ORGANIZATION AND EVOLUTION IN BIRDS
  7. AVIAN THALAMIC ORGANIZATION AND EVOLUTION
  8. CONCLUSION
  9. Acknowledgements
  10. LITERATURE CITED