Brain Advance Access originally published online on May 30, 2007
Brain 2007 130(9):2470-2473; doi:10.1093/brain/awm124
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Book Review |
The thalamus revisited: where do we go from here?
In my academic lifetime, two outstanding books on the thalamus have been published. Now both have been reissued as expanded, updated and improved second editions. The most massive of these, the second edition of The Thalamus by E.G. Jones, follows his 900-page, 1985 edition. While the early edition was truly remarkable, the two-volume, second edition of 1679 pages of expanded size represents an effort reminiscent of Polyak's 1390-page tome, The Vertebrate Visual System, published in 1957. Jones retains one of my favorite quotations from the first to the second volume, one attributed to the great neuroanatomist and neuroscientist, Jerzy Rose: ‘The thalamus is like the Flying Dutchman: many have heard of it, some believe in it, but few have actually seen it’ (c. 1940). Today, I am not sure how many have heard of the Flying Dutchman, a mythical sailing ship, but Jones’ first edition did much to correct uncertainties about the thalamus, as The Thalamus became one of the most referenced books in neuroscience. The basic organization and many features of the first edition are carried over to the second, but this is bigger and better in every way. Most impressive, the second edition is profusely illustrated with drawings, photographs of investigators, and especially, photomicrographs of brain sections through the thalamus. The photomicrographs are not just of a brain section here and there, but of series of sections from the same cases, and not just from the laboratory species that we can view in the many stereotaxic atlases that are now available, but also of species such as tree shrews, galagos and the egg-laying monotremes. On page 1210, for example, one can see how well differentiated the anterior nuclei of the thalamus are in rabbits compared to rats, or even most primates. No wonder rabbits played such a significant role in the determination of the cortical connections of these nuclei.
Another strength of this edition is the extensive discussion of early to current research on specific nuclei and divisions of the thalamus. If you have worked on the thalamus, your research is included in these volumes, although you may not always like the discussion of that work, since Jones’ opinions are not hidden. There are detailed chapters on the functional properties of thalamic neurons, the neurochemistry of the thalamus and the development of the thalamus. The first chapter on the history of the thalamus has been expanded from the 1985 edition to include a score of more recent investigators, including my PhD advisor, Irving Diamond, but leaves out some of those who are perhaps still too active. The complete second volume of the current edition is perhaps the most valuable, as this is where individual nuclei are considered. The comparative section includes an extensive discussion of the anatomy of the thalamus in non-mammalian vertebrates, followed by a complete chapter on the human thalamus. Here, as well as in other chapters, Jones helps the reader by comparing the different terminologies in current use by various authors. I hope that this publication, as well as related efforts, will help standardize the names of thalamic nuclei to provide a more encouraging environment for researchers.
|
280.00 / $480.00 ISBN-13: 9780521858816
|
The tack taken by Murray Sherman and Ray Guillery in their shorter, 2001 312-page volume, Exploring the Thalamus, was much different. Their focus was on the cellular anatomy and the response properties of neurons in the thalamus, and neurons projecting to the thalamus. These investigators described two that are anatomically and functionally distinct classes of afferents terminating in the thalamus as ‘drivers’ and ‘modulators’. The drivers are the inputs that activate thalamic neurons at above threshold levels. They are proportionately few in number, but are powerful activators. Thus, the messages sent to cortex by thalamic relay neurons that are also drivers closely reflect those of the drivers that activate them. The drivers of the thalamic neurons determine the content of the relayed messages. The well-known drivers are those of sensory systems. For example, ganglion cells of the retina project to the dorsal lateral geniculate of the thalamus where they drive the neurons that project to primary visual cortex, and provide the geniculate neurons with the content of the messages they bring to cortex. Other sources of driving afferents include the superior colliculus, the hippocampus via the fornix and the mammilary body, the amygdala and the deep nuclei of the cerebellum. But, most important for Sherman and Guillery, driving afferents also constitute a subpopulation of afferents originating in the neocortex. Cortical projections that are drivers can be anatomically identified since they originate in the layer 5 pyramidal neurons and terminate with distinctive types of large terminals on thalamic neurons. It is our neglect of the functional roles of these cortical drivers of thalamic neurons that concerns Sherman and Guillery. The textbook picture of the role of the thalamus has been to relay information from lower centres to cortex, where streams of cortical processing take place and cortical outputs govern our lives. In their 2001 volume, Sherman and Guillery stressed the significance of the cortical projections that drive thalamic neurons, which, in turn, project back to cortex with further messages. The selection and modification of the messages depend on the modulators that project to the thalamus as well as the intrinsic circuitry of the thalamus. The modulators modify messages sent to the cortex by amplifying or suppressing, as in different states of awareness and attention. Since the modulators come from many sources, they undoubtedly have many functional roles. Sherman and Guillery pointed out that an important source of modulation is cortex itself. These modulators can be identified since they originate from layer 6 cortical cells and they terminate with characteristic types of terminals on thalamic neurons.
The specific functions of cortical modulators and drivers of thalamic neurons must be as varied as the functions of the cortical areas themselves. In their recent (2005) expanded second edition, Exploring the Thalamus and its Role in Cortical Function, Sherman and Guillery more fully discuss the interactive functioning of the thalamus and the cortex. Both Jones, and Sherman and Guillery lament that neuroscientists have almost exclusively focused on corticocortical processing while neglecting or ignoring the roles that corticothalamocortical circuits might play. They agree that we need to know more about the roles of the thalamus in cortical function. The question is, how do we do this. Sherman and Guillery suggest that we need to determine the response properties of the drivers of each thalamic nucleus and each cortical area, and compare these response properties to the outputs of the drivers of each thalamic nucleus and each cortical area. Thus, we would discover what sorts of transformations of messages take place within each nucleus and area, and thereby understand the contributions of each nucleus and area to the processing within a system. Of course, this research goal is technically challenging and it is of considerable magnitude. One complication is that neural circuits are not necessarily stable over time, as they can be structurally and functionally modified by experience. In addition, any comprehensive study of nuclei and areas is presently limited by our incomplete understanding of how the cortex of various mammals is subdivided into cortical areas, and perhaps equally limited by our partial knowledge of how the thalamus is subdivided into nuclei. We should not take the view that all or even most areas and nuclei have already been adequately identified.
Cortical areas were defined by Brodmann (1909
) as the ‘organs of the brain’. Areas are the larger functional subdivisions of cortex, and are thought to be divided into smaller, repeating functional units called columns or modules. A cortical area can be most reliably identified when many distinctive features of the area are revealed and they are all congruent with the proposed borders. These features might include many different architectonic (histological) characteristics, patterns of connections with other structures, neural response properties, and patterns of representation of activating inputs. Thalamic nuclei are the conceptual equivalent of the cortical areas, and can be identified using the same criteria. Historically, both areas of the cortex and nuclei of the thalamus were identified by their cytoarchitectonic features. This led to a high level of agreement among investigators for a few cortical areas, and perhaps even more so for some thalamic nuclei. However, the limits of a purely cytoarchitectonic approach resulted in the misidentification of cortical areas and boundaries of areas within studied species, and to the misidentification of homologues of areas across species. As homologous areas often differ in histological distinctiveness among species, the same cortical area is sometimes identified as different across species. For example, area 3b, or primary somatosensory cortex, was identified by Brodmann as area 1 in lemurs, area 1 plus 3 in marmosets and area 5 plus 7 in flying foxes and hedgehogs. In squirrels, Brodmann misidentified the less developed monocular section of area 17 (primary visual cortex) as area 18 (secondary visual cortex, V2), resulting in the misleading legacy of an area 18 placed immediately medial to area 17 in many of the contemporary portraits of cortical areas in rats and mice. Cortical organization now has been largely reinterpreted in the well-studied mammals, such as rats, cats, New and Old World monkeys and humans based on the results of many experimental studies that were designed to help identify areas as functionally distinct subdivisions of cortex. Nevertheless, of the 35 or so areas of visual cortex that have been proposed for macaque monkeys, only three or four have been so well identified that there is nearly unanimous agreement about their validity and borders.
Errors of identification have occurred in the thalamus as well. As a long-standing example, we commonly speak of a ventroposterior medial nucleus and a ventroposterior lateral nucleus. As these two cell masses are largely separated by a narrow cell-poor septum, they were easily distinguished in early studies of the primate thalamus. But now we know that together they constitute a single functional unit, the ventroposterior nucleus, which systematically represents cutaneous receptors from mainly the opposite side of the body, and relays to primary somatosensory cortex, area 3b. The ventroposterior superior nucleus (VPS) that relays proprioceptive information to cortex is often not recognized and is included instead in the ventroposterior nucleus. Jones goes halfway by placing VPS of anthropoid primates in the dorsal ‘shell’ of the ventroposterior nucleus. One might argue that fewer such mistakes have been made in subdividing the thalamus, as most nuclei are more histologically distinct than are most cortical areas. However, this is uncertain because the thalamus has been less studied with techniques that could reveal further or different subdivisions as nuclei.
We now generally recognize that species of mammals vary greatly in numbers of cortical areas, and this is likely to be true in the less investigated thalamus (Kaas, 2007a). Comparative studies of cortical organization support the conclusion that early mammals had relatively few cortical areas, on the order of 15–20 or so, and these were rather poorly differentiated (Kaas, 2007b). The number of cortical areas did not change very much in some lines of mammalian evolution, and some extant species with very small brains may have even lost some areas. But for the most part, the number of cortical areas increased in at least many of the successful lines of descent. Thus, from an ancestral state of four or five visual areas, present-day macaque monkeys have something like 30–40 visual areas. Although few of these areas have been adequately identified, there is widespread agreement that the estimated number is approximately correct. Because of the power of functional magnetic resonance imaging, progress in understanding how the human brain is divided into functional areas has proceeded rapidly, and some investigators estimate that the total number of cortical areas in humans exceeds 200. Studies on domestic cats indicate these and probably other carnivores have more visual areas than their early mammal ancestors did. This increase would have been independent from the increase that occurred in primates, and thus most of the visual areas in monkeys do not exist in cats, although they may have independently evolved visual areas with similar features and functions.
Although comparative studies have led us to re-evaluate early views on cortical organization, it seems that we are more reluctant to abandon the view that one scheme of thalamic organization fits all. If all cortical areas interconnect with thalamic nuclei, does it not follow then, that cortical complexity will be reflected in thalamic complexity, such that species with more cortical areas will have more thalamic nuclei?
To be fair, Jones does not argue that all mammals have the same thalamic nuclei, and he does a beautiful job of showing us histological sections through the thalamus of various species, and discussing the relevant data on known species differences in thalamic connections and other features. However, the research emphasis has been on those distinctive nuclei that can be identified across species, and appear to be homologous. Nevertheless, when thalamic cytoarchitecture alone is considered, difficulties can exist in the consistent identification of even the most distinctive of nuclei. In the confusing thalamus of monotremes, Jones notes that the identity of the usually distinctive lateral geniculate nucleus is uncertain, although studies on retinal projections to the thalamus have located the probable nucleus. Clearly, more experimental studies are needed across a variety of species in order to reveal patterns of cortical connections with the thalamus. The many different histological preparations that are now available can be extremely helpful in efforts to relate patterns of connections to structural and histochemical distinctions in the thalamus.
One thalamic region where a promising start has occurred is that part of the visual thalamus known as the pulvinar. As it was once thought that only primates have a pulvinar, the pulvinar is typically called the lateral posterior nucleus in mammals such as rats, or it is divided into lateral posterior and pulvinar regions as in cats. The visual pulvinar of primates is usually divided into inferior and lateral ‘nuclei’. To reduce confusion in the future, hopefully investigators will start to call the extrageniculate visual thalamus of all mammals, the pulvinar complex. In any case, evidence from comparative studies suggests that early mammals had a pulvinar complex of two or three nuclei, differing in connections with visual cortical areas, and having or lacking visual inputs from the superior colliculus. This is the number now apparent in rats, and even in highly visual rodents such as squirrels, where the pulvinar complex is rather large. Cats have another arrangement of four or more visual nuclei, while tree shrews, the closest studied relative of primates, have four differently ordered nuclei of the pulvinar complex. Primates also have a different number and arrangement of nuclei, with four recently identified in the territory of the traditional inferior pulvinar, and at least two nuclei in the territory of the traditional lateral pulvinar. These nuclei connect differently with various visual areas (Kaas and Lyon, 2007), some of which (primary and secondary visual areas, V1 and V2) are shared with other mammals, and some of which are not (the middle temporal visual area, MT, and adjoining areas). Some pulvinar nuclei in these highly visual mammals are likely homologous and retained, but elaborated, from those of early mammal ancestors. Others may be relatively new additions. The main point here is that the extrageniculate visual thalamus has evolved in different ways. Highly visual mammals of separate lines of descent have both ancient and newer visual nuclei and visual cortical areas. The newer areas and nuclei were differently derived across taxanomic groups, while the older nuclei and areas differentiated structurally and functionally in both diverse and similar ways. Undoubtedly, other regions of the thalamus and cortex have done the same thing, but we know more about the visual thalamus and cortex because more relevant research has been published.
The demonstrated and probable variability in the complexity of the thalamus across species raises the issue of how to proceed in studying thalamic nuclei in order to see what functions they mediate. Comparative studies are valuable because they tell us which species share the same nuclei or cortical areas, and how these homologous structures are similar or different in their numerous characteristics. Since only primates have the middle temporal visual area, MT, and the medial nucleus of the inferior pulvinar, IPm, with major interconnections with MT, the functional inter-relationships of MT and IPm can be studied only in primates. In addition, since some of the characteristics of IPm and MT likely vary across primates, some aspects of these structures might be accessible for study only in some primates. When we are particularly interested in human brain function, those primates available for study that are most closely related to humans are likely to share with humans more features of homologous thalamic nuclei, but such sharing is not necessarily the case for any specific feature. Thus, the goal of understanding how the human brain works, a brain dominated by a massive thalamus and an even more massive cortex, is a daunting enterprise. Appropriate comparative studies, especially on primates, and those that can be done in humans, are needed. However valuable the many ongoing studies on mice are, they provide types of understandings that apply broadly to mammals, but do not capture the significance of the complexity of the human brain.
Department of Psychology
Vanderbilt University
References
Brodmann K. Vergleichende Lokalisationslehre der Grosshirnrhinde. (1909) Leipzig: Barth.
Kaas JH. The evolution of the dorsal thalamus in mammals. In: Evolution of nervous systems—Kaas JH, Krubitzer LA, eds. (2007a) 3. Mammals, London: Elsevier. 499–516.
Kaas JH. Reconstructing the organization of neocortex of the first mammals and subsequent modifications. In: Evolution of nervous systems, Vol. 3—Kaas JH, Krubitzer LA, eds. (2007b) Mammals, London: Elsevier. 499–516.
Kaas J, Lyon DC. Pulvinar contributions to the dorsal and ventral streams of visual processing in primates. Brain Res Rev (2007) doi:10.1016/j.brainresrev.2007.02.008.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||