THE EVOLUTION OF THE THALAMUS (4,700 WORDS)
© John E. LaMuth 2011
The immediate proximity of the diencephalon to the brainstem gives this subdivision of the forebrain the greatest potential for a exhibiting a cranial style of sensory function. The strictly sensory portion of the diencephalon, the dorsal thalamus, lies dorsal to the basal hypothalamic region. The thalamus is flanked by two subregions, the epithalamus (above-thalamus) and the subthalamus (or ventral thalamus). In the diencephalon of the developing human embryo, the epithalamus, dorsal thalamus, and subthalamus are approximately the same size range separated from each other by the sulcus dorsalis and medius. This configuration is similar to that observed in the fully mature diencephalon of more primitive lower vertebrates. At this more primitive stage of forebrain development, the diencephalon is primarily dominated by visual inputs consistent with its intimate interrelationship with the second cranial (optic) nerve. At more advanced stages of diencephalic evolution, the thalamus undergoes a massive expansion relative to the other remaining subdivisions. In the adult human diencephalon, the enormous expansion of the dorsal thalamus both dwarfs and displaces the more rudimentary epithalamic and subthalamic subdivisions. This expansion of the thalamus is paralleled by the invasion of related inputs relayed from lower levels of the neuraxis. True to their more primitive origins, however, the epithalamic and subthalamic subdivisions still demonstrate a characteristic functional specificity to strictly visual inputs.
The human subthalamus, for instance, is dominated primarily by input from the dual pair of optical nerves. The subthalamus of man is distorted into pieces by the downward growth component of thalamic expansion. Systematic embryological investigations, however, permit the tracing of the migration course for each of these subthalamic fragments. According to Reinoso-Suarez (1966) the pregeniculate nucleus of the human diencephalon represents one such component of the subthalamic fragmentation. This pregeniculate nucleus significantly represents the most phylogenetically ancient site of termination for direct optic tract fibers to the diencephalon. The more recent fiber components of the optic nerve carry across through evolution into the more recently evolved visual areas of the thalamus such as the lateral geniculate nucleus.
The epithalamus, on the other hand, it exhibits a radically different style of visual specificity. The human epithalamus is composed of a symmetrically paired set of habenular nuclei that attach by way of a twin set of stalks to the peripherally situated pineal gland. The uncertain functional characteristics of the human pineal gland do not offer any clear indication of its true origins; namely, the atrophied remnant of the primordial parietal eye of lower vertebrates. The parietal eye is an accessory visual organ centrally situated in the forehead of a wide variety of fish, frog, and lizard species. In terrestrial vertebrates, this third eye is invariably located beneath a transparent area of the skin, usually within a special recess upon the medio-dorsal aspect of the skull. Only a very primitive shadow-type sensibility is possible for this primative style of lighting arrangement, implicating the parietal eye in the control of seasonal or diurnal rhythms. The parietal eye is embryologically derived from the outermost extension of the pineal or parapineal evagination of the epithalamic ependyma. This extension further develops a retina, a lens, and a cornea in a sequence of differentiation analogous to that followed by the paired lateral eyes (Sarnat and Netsky, 1974). Axons from the retinal nerve cells of the third eye grow into the habenula similar to the retinal ganglion cell projection into the pregeniculate nucleus. In a few living species of primitive jawless fishes, both pineal and parapineal bodies develop ocular structures. In some jawless fossil forms, the cranium exhibits paired frontal eye orbits in addition to the more pronounced lateral orbits (Eakin, 1962). This evidence appears to suggest that a remote ancestral vertebrate possessed both paired dorsal eyes in addition to the standard lateral eyes. The paired parietal eyes eventually fused to form the single parietal eye of living vertebrates, which, in turn, degenerated into the pineal gland at the mammalian stage of evolution. The preceding accounts seem to establish the primitive existence of a dual optic projection to the primordial subdivisions of the thalamus.
The orderly expansion of the dorsal thalamus between the epithalamic and subthalamic moieties acquires a much greater significance in the context of ongoing development of inputs. For instance, in the subthalamus example, a large number of optic tract fibers from the main lateral eyes carry beyond the human pregeniculate nucleus into the more recent visual areas of the thalamus proper. The reticular nucleus of the dorsal thalamus is just one of many new regions to receive extensions of sensory input from the optic track. This observation is corroborated by cytological zones of transition with the pregeniculate nucleus (Hassler, 1971). The reticular nucleus, in actuality, forms the outer-most boundary of the thalamus, encasing it in a sort of enveloping, shell-like sheath. The peripheral, circumferential orientation of the reticular nucleus is reminiscent of the mature, outer growth rings of a tree stump that surround the newer heartwood at the core. As suggested by this model, the reticular nucleus is viewed as the most ancient age level of the human thalamus, occupying the zone of transition between the primordial subthalamus and the more recent age levels of the thalamus. This entire series of thalamic age levels (including the reticu1ar nucleus) comprise what is termed the sub¬thalamic gradient of thalamic evolution.
A roughly equivalent portion of the human thalamus is derived across the corresponding epithalamic gradient of thalamic evolution. The habenula shows similar zones of transition to the human thalamus in the ancient region of the posterior (limitans) group of nuclei (Hassler, 1959). The un¬precedented atrophy of parietal eye in mammals makes tracing the epithalamic gradient into the thalamus more difficult than was encountered in the case of the subthalamic gradient. This timely phasing out of parietal inputs of the mammalian stage of evolution have left the door open for the incorporation of replacement inputs from lower levels of the neuraxis. This introduction of alternate inputs crucial to mammalian survival, in theory, sparked the massive thalamic expansion at the human stage of forebrain evolution.
In line with these considerations, the mammalian thalamus evolved in stages away from both epithalamic and subthalamic moieties. Each newly develop¬ed age ievels arose as a growth core of differentiation at the juncture of the epithalamic and subthalamic gradients, turning each successive¬ly older age level into a concentric growth shell. In this manner the entire series of concentrated growth shells was built up in the thalamus by the time the human stage of development is first apparent.
Each growth shell is theoretically divided into two hemi-regions, denoting either the epithalamic or subthalamic evolutionary gradient. Growth pressures from the development of more recent age levels, compressed each thalamic hemi-region into a compact complex of nuclei within its respective gradient. Although the ideal concentric configuration becomes distorted, the continuities be¬tween each complex of nuclei within the respective epithalamic or subthalamic gradients depicts the correct temporal order of evolutionary development.
The actual existence of a growth shell series in the human thalamus has never before been postulated in the literature, undoubtedly due to the gross distortion of the growth shell configuration in higher mammalian forms. Along similar lines, the German researcher, Rolf Hassler, proposed a related hexapartitionment paradigm for the human thalamus, upon which the current growth shell theory was only recently based (LaMuth, 1977). Indeed, the parallel evolution of the neocortex and the dorsal thalarnus is the basis for postulation of a chronological gradient OF diencephalic differentiation homologous to the corresponding age gradation in the cortex. The three dimensional configuration of the dorsal thalamus renders the diencephalic gradient more difficu1t to detect than the age gradation shown for the planar pallium. The chronological transition From non-specific to specific thalamic organization should by reason of concurrent cortical differentiation exhibit discrete levels of diencephalic differentiation quantitatively equivalent to the set of neocortical growth waves. Indeed, Hassler’s theory of the hexapartition of the dorsal thalamus represents the diencephalic counterpart of the six distinct parameter levels demonstrated for neocortical age gradation. In reiteration of Hassler (1972) there are six cytoarchitectonically distinct thalamic levels representing the aggregate terminus for a specific thalamic input. Listed in order of increasing specificity, these thalamic levels are denoted as (1) relay (2) first integrative level (3) second integrative level (4) composite multisensory (5) reticulate feedback and (6) unspecific protopathic.
The highly differentiated ventral relay nuclei of the lateral thalamic complex are the primary terminus for the ascending specific sensory and motor inputs. Immediately dorsal to the ventral nuclei are the zentrolateral nuclei representing the first integrative level. According to Hassler, this level receives finer caliber input fibers, presumably the collaterals of ventral layer input. Situated over the zentrolateral level are the dorsolateral nuclei corresponding to the second integrative level. A sparse complement of direct sensory terminals is distributed to the dorsolataral layer from lower input levels. The magnocelluar portions of the medial geniculate, ventro-anterior, and dorso-medial thalamic nuclei collectively comprise the composite multi-sensory level. The terminal overlap of related specific inputs at this level has been verified in electrophysiological studies. The reticulate feedback age level is portrayed by the thalamic reticular nucleus, which forms a shell surrounding the entirety of the dorsal thalamus. The reticular nucleus has been demonstrated to receive the specific thalamic input identical to that of the immediately subjacent lateral thalamic subdivision. The unspecific protopathic age level is composed of the nonspecific midline and intralaminar thalamic subdivisions. This nonspecific thalamic level projects to the ancient corticoid ring rather than to the more recent neocortex proper.
The anterior portion of this ancient thalamic grey consisting of the nucleus fasciculous and reuniens, pars ventralis of nucleus medialis, and centralis and the rostral parts of nucleus parafascicularis and centro-median, has been shown to be of subthalamic origins (Reinoso-Suarez, 1966). The reticular nucleus of the next most recent age level is also derived from the subthalamus (Christ 1969) suggesting the existence of a subthalamic gradient of dorsal thalamic differentiation. The subthalamus borders the dorsal thalamus from below dictating that this gradation be termed the “ventral ur-trend” of dorsal thalamic differentiation.
The remaining posterior portion of the unspecific protopathic age level, includes such subdivisions as the nucleus limitans, suprageniculatus, peripeduncularis, and posterior centro-median. This series of caudal nuclei collectively displays a distinctively intense cholinesterasic staining activity (Poirier, 1974), denoting a common developmental origin. The close topographical proximity of this posterior series of nuclei to the habenula (Hassler, l959) suggests that the epithalamus represents the other basic moiety of dorsal thalamic differentiation. The epithalamus borders the dorsal thalamus from above, dictating that the corresponding developmental gradient be termed the “dorsal ur-trend” of dorsal thalamic differentiation.
Hassler’s unspecific protopathic thalamic level is composed of a pair of primordial moieties, each of which is the precursor of a distinct thalamic ur-trend. The five more recent thalamic age levels differentiated concentrically between the epithalamic and subthalamic moieties, apparently without the initial lamination phase characteristic of neocortical differentiation. In relation to the three dimensional configuration of the thalamus, these concentric age levels are termed the “growth shells” of thalamic differentiation, emphasizing their homology to the growth rings of the planar pallium. Each thalamic growth shell is subdivided into two hemi-regions structurally denoting dorsal and ventral ur-trend components. Growth pressures from the development of more recent age levels compressed each thalamic hemi-region into a compact nucleus within the corresponding ur-trend. Although the ideal concentrical configuration becomes distorted, the continuities between nuclei along each thalamic ur-trend depict the correct order of chronological differentiation. Each thalamic growth shell is correspondingly denoted by a distinct pair of thalamic nuclei analogous to the paired belt areas contained within each cortical growth ring.
The thalamic relay age level, similar to the koniocortical growth wave, represents the developmentally most recent core of diencephalic differentiation. This unpaired age level is characteristically situated at the junction between the dorsal and ventral thalamic ur-trends. The respective pair of ur-trends continues across a contiguous series of progressively older nuclei into close proximity with either the epithalamc or subthalamic moieties.
In summary, according to Hassler (1971), the human thalamus is subdivided into six distinct subregions that show clear differences in the exact manner which incoming input fibers branch and terminate. The most primitive regions of the thalamus are grouped into the unspecific-protopathic classification, in ref¬erence to its non-specific mode of input termination. The reticulate-feedback subdivision comprises the next higher level of refinement, where inputs terminate in passage to the newer, and more centrally situated, segments of the thalamus. The next more recent thalamic level is termed the com¬posite-multisensory classification: designated in reference to the extensive terminal overlap of inputs entering this age level.
The three re¬maining growth shells were recognized by Hassler as the second integrative level, the first integrative level, and the relay level; in reference to the integrative relationship the former two classifications exhibit with respect to the latter. The second and first integrative levels actually represent the final two concentric growth shells surrounding the core-like relay level. According to Hassler, the final relay age level is the pri¬mary terminus for sensory and motor inputs to the thalamus, whereas the first and second integrative levels receive collateral (accessory offshoot) branches of these fibers. Hassler never directly specified that his series of hexapartition levels were, in actuality, a gradient of distinct age levels. Although in the qualitative sense, the relay age level was recognized as being of more recent evolutionary origin than the unspecific protopathic level, it was their re-evaluation in the growth shell context that added the extra measure of functional significance.
The hexapartition format, as originally proposed, was not very suggestive of an age level gradient because it did not accurately account for the presence of both the epithalamic and subthalamic gradations in each growth shell. According to this type of revision the epithalamic and subthalamic age gradients actually intersect within the final relay growth core as an inter-locking array of input relay nuclei. The relay growth core is divided into fourteen different subdivisions (or nuclei), each of which are collectively specific to the equivalent number of input classifications that terminate in the thalamus (see Fig. 1). Each of the fourteen input classifications evolved exclusively across either the epithalamic or subthalamic gradient, which taken together span across parts of the entire progression of growth shells to reach into respective proximity with either the epithalamic or subthalamic moieties.
While not nearly as well defined, the pattern of ur-trend alternation is also indicated in the human thalamus as well. A homologue of the ur-trend limiting sulcus in the thalamus is lacking due to the fact that the thalamus is not a surface structure. Through the use of alternate criteria, such as continuity, however, an analogous pattern of interlocking epithalamic and sub¬thalamic ur-trends becomes readily apparent across the entire rostro-caudal extent of the thalamus. Of the fourteen functional subdivisions in the thalamic growth core, half are derived across the epithalamic gradient, while the other half evolve within the subthalamic gradation.
By modification to include both epithalamic and subthalamic component age levels, Hassler’s hexapartition terminology is completely applicable for designating each constituent thalamic growth shell. These hypothesized growth shells in the thalamus are further corroborated by the corresponding demonstration of an equivalent complement of six distinct growth rings in the planar cortex. The six hexapartition levels of the thalamus are analogously equivalent to the five neocortical age levels plus the allocortical age level. In theory, each correspondingly numbered growth rings and growth shell evolved in a synchronous fashion throughout the scale of mammalian evolution. This unified complement of six forebrain age levels (in a group quantitative sense) comprise what is designated the parameter of phylogenetic age. This temporal parameter serves to quantitatively unite the set of thalamic growth shells and cortical growth rings into a globally coherent pattern altogether lacking in the two separate theories.
The theoretical establishment of this unprecedented type of all-encompassing age parameter shows additional applications relative to the overall pattern of forebrain. For a given unit input, only one thalamic nucleus of the growth shell pair was mentioned by Hassler in documentation of the individual levels of his hexapartition thalamic scheme. Apparently the thalamic nucleus of the less obvious ur-trend was not recognized as being of the same age and input characteristics as the more pronounced component of the pair. Many of the overlooked thalamic subdivisions such as the anterior nuclear complex complement the unmatched components of the individual hexapartition levels, By modification to include both ur-trend components, Hassler’s nomenclature is completely applicable for defining each constituent growth shell of dorsal thalamic differentiation.
Similar to the dual pattern of olfactory inputs for the telencephalon, the primordial diencephalic precursors of dorsal thalamus are shown to receive a dual optic projection analogous to the double olfactory bulb projection to the allocortex. More specifically, the epithalamus (habenula) receives direct visual input from the dorsal parietal eye of lower vertebrates, while the subthalamus appears to be similarly related to the main lateral eyes.
The fact that the diencephalon of lower vertebrates is concerned almost entirely with vision has been well documented through electrophysiological means. Only in mammals do other modalities (such as auditory and tactile stimuli) find such extensive terminations within the thalamus. This sudden influx of new types of input seemingly corresponds to the loss of the light sensing function of the parietal eye/pineal gland in mammals. It might be proposed that this the emphasis of the dorsal visual input to the epithalamus occurs precisely at the period where the dorsal thalamus begins to differentiate in waves similar to the growth rings of the neocortex. Apparently inputs from the brainstem structures are incorporated into the dorsal thalamus by taking over the vacancies left by the lost epithalamic inputs.
The ancient projections from the paired lateral eyes to the ventral thalamus have been mentioned only sparingly in the literature, apparently masked by the more massive projection to more recent forebrain structures. The ventral thalamus of lower vertebrates has also become somewhat dispersed by growth pressures within the mammalian brain, with contradictory viewpoints in the literature as to which nuclei are actually fragments of this division. Further confusion in the literature also occurs in that the term ventral thalamus is sometimes used interchangeably with the subthalamus, whereas other researchers strictly maintain the distinction. Ventral thalamic nuclei such as the substantia nigra and the corpus Luysii have been cited in the literature to receive optic fibers over the more ancient accessory optic tract. Other ventral thalamic nuclei such as the zona incerta have been shown to receive collaterals from the visual cortex, wherein emphasizing the visual nature of this thalamic subdivision.
This intrinsic modification of Sanides’ final growth core has even broader evolutionary consequences when both the cortex and the thalamus are collectively considered. The hypothetical existence of an extra age level within the cortex effectively evens out the total number of age levels within the thalamus and the cortex to the common denominator of “six.” The six hexapartition levels of the thalamus appear to be functionally equivalent to the five neocortical age levels (plus the allocortex). In theory, each correspondingly numbered growth ring and growth shell evolved in a synchronous fashion within the span of mammalian evolution. This unified complement of six forebrain age levels in a group quantitative sense comprise what must be theoretically be termed the parameter of phylogenetic age. This temporal parameter serves to quantitatively unite the set of thalamic growth shells with the set of cortical growth rings, resulting in a globally coherent pattern altogether lacking with respect to the two separate theories.
The theoretical establishment of such an all encompassing age parameter format has further ramifications with respect to the overall pattern of forebrain connectivity. The identical number of age levels in both the thalamus and the cortex indicates that only correspondingly numbered growth shells and growth rings are reciprocally connected by the thalamic radiations. In theory, a thalamic cell on a discrete point within the time-differentiation continuum directs its main projection to cells of the cortex derived at the same phylogenetic age. Similarly, cortical cells that emit feedback projections destined for the thalamus analogously respect this age level restriction. These intersegmental considerations restrict the intrinsic interconnectivity of the thalamo-cortical loop the strictly to identical sets of age levels. Yakolev’s (et al, 1966) documentation of the limbic and insular projections of the three most ancient thalamic age levels appears to corroborate these theoretical age level restrictions the. Cortical citations for the newer age levels of the thalamus by Hassler (1959) similarly seem to verify this contention, even though Hassler did not technically considered his hexapartition levels as separate age levels. In addition to this age levels type of consideration, the thalamo-cortical loop is further subject to an alternate style of input functional restriction. As stated previously, the thalamic radiatiod serve to relay the entire complement of forebrain inputs to each of the age levels of the cerebral cortex. The individual fibers within the thalamo-cortical loop, however, all appear more or less indistinguishable, in giving no indication of the type of input that is being relayed. This observation raises the related questions; namely, how the different varieties of forebrain input have differentiated across the entire growth ring age series to reach a koniocortical stage of specialization.
The appropriate answer is not actually a simple as the growth ring/shell theory would lead to believe. In the cortex, for instance, the final koniocortical growth core exhibits a type of topographic dualism that is not accountable simply in terms of the growth ring paradigm. This final cortical growth core is not homogenous, but rather is subdivided into two parallel bands by means of a longitudinal fissure, the ur-trend limiting sulcus. This sulcus is designated for its role in separating the medial and lateral ur-trends of the hemisphere (German ur = old). According to Sanides (1969) the growth ring segments interposed between the archaecortex and the ur-trend limiting sulcus are termed the medial ur-trend, whereas the comparable gradient reaching to this sulcus from the paleocortex is designated the lateral ur-trend. The ur-trend limiting sulcus is depicted schematically as the jagged line interposed between the paired characterizations of the medial and lateral ur-trends. From this diagram, it would appear that the two parallel bands of the koniocortical growth core have been independently derived across both the medial and lateral gradients, respectively. The position of each koniocortical area relative to the ur-trend limiting sulcus serves as a valuable criterion for determining its direct ur-trend of origin. For instance, and the limiting sulcus in humans is rostrally represented as the inferior frontal sulcus that separates the middle and inferior frontal gyri (Sanides, 1970). The inferior frontal gyrus is composed of a sequence of three highly differentiated areas designated as pars opercularis (area #44), pars triangularis (area #45), and pars orbitalis (area #12). Each of these areas characteristically displays an abundance of giant pyramidal cells in the outer pyramidal layer #IIIc. As cited previously, the paleocortex primordially contributes to the formation of the outer pyramidal layer consistent with the corresponding emphasis in the koniocortical maximum derived across the lateral ur-trend. The middle frontal gyrus is host to an alternate sequence of ur-trend maxima: areas #46, #8delta, and #6aalpha. Each of these medial ur-trend maxima exhibit a size accentuation of pyramidal cells in lamina V consistent with the archaecortical origins of the inner pyramidal layer. A close inspection of the posterior association region reveals the presence of a pair of conspicuously myelinated koniocortical areas. The visuo-auditory band is situated in area 39 of the visual association region, whereas the corresponding areas #7 and #40 in the somatosensory area comprise the visuo-sensory band.
This inherent dualism in thalamic development shows a special significance with respect to the evolution of forebrain imports, since each major forebrain input initially enters at the diencephalic level of the neuraxis. According to this format, half of these inputs arise from, or are incorporated into the epithalamic age gradient, whereas the remainder are correspondingly derived across the subthalamic age gradient. This general composite pattern is schematically depicted in Fig. 1 as the symmetrical separation of the 14 varieties of forebrain input into the two complementary subgroups shown on the upper and lower margins of the diagram. Irrespective of its epithalamic or subthalamic affiliation, each input classification evolves through its own respective gradient within the growth shell series to reach its final culmination in the relay growth core. Different subdivisions of nuclei within each thalamic growth shell deal individually with different classes of input. These nuclei are distinguished from one another by the cellular differences that inevitably result from the processing of different types of input. Cytoarchitectonic differences are also observed between adjacent growth shells of different ages processing the same type of input. When these two factors controlling cellular variability are combined into a common system, the result is a comprehensive parcellation of the thalamic mass much as was proposed by Hassler (1959). It furthermore is possible to show that each of Hassler’s experimental subdivisions respects the growth shell and input restrictions embodied in the master schematic grid. Each unit-square of this chart graphically represents the coordinate set of unique age and input parameter levels. The individual unit-squares of this theoretical grid systematically and compass each of the experimentally determined thalamic subdivisions on a one-to-one basis. This point-for-point correspondence accounts for essentially every thalamic subdivision, yet is virtually free of any duplication across adjacent unit-squares. The minimal redundancy observed in the older thalamic growth shells appears to be the result of the hesitancy on the part of some researchers to subdivide regions of already diminutive proportions. Besides being listed by name within each unit-square, the thalamic classifications of Hassler’s system are designated through their exact position with the schematic grid. This Cartesian coordinate paradigm of the human thalamus reduces the great complex of this elaborately structured region to an elementary level of quantitative simplicity. This precise coordinate style of system for the human thalamus had not been achieved prior to the original proposal of the dual parameter grid (LaMuth, 1977). This original model has undergone a number of further modifications, achieving its final format in Fig. 1.
Hassler published the details of his thalamic parcellation scheme a full two decades before the conceptualization of the parameter grid, therefore it is necessary to sift through the bulk of his documented observations for clues to the correct coordinate designation for each thalamic subdivision. By piecing together details dealing with the termination of inputs and the continuity between adjacent nuclei, it proved possible to arrive at a unified and topographically correct model of the human thalamus fitted exactly within the rigidly defined age and input restrictions. In a similar fashion, cortical subdivisions within the human forebrain prove equally applicable for a parallel topographical correlation with the dual parameter grid. Being essentially a uniform surface structure, the cortex is amenable to a similar style parcellation format.
The human cerebral cortex has traditionally been subdivided according to a pair of rival, yet parallel, systems of parcellation, both of which have endured due to their proven experimental accuracy. The German parcellation traditional was introduced by Broadmann in 1909 with the unveiling of his cytoarchitectonic cortical parcellation scheme encompassing 50+ numerically coded subdivisions. A rival French system was later proposed by von Economo in the 1920’s, the subdivisions of which are designated by short sequences of letters: the first letter representing a shorthand for the corresponding lobe of the cortex (P = parietal, F = frontal, etc.). Both German and French parcellation schemes accurately reflect the differences in structural organization between adjacent cortical growth rings. They further take into account the associated functional differences observed within a single cortical growth ring. By utilizing both age and input parameter criteria, the cortical parcellation schemes of Broadman and von Economo correlate on a one-to-one basis with the dual parameter grid. This strict unit-square correspondence accounts for virtually every major subdivision within both parcellation schemes. Here, only a minimal amount of redundancy is basically restricted to only the most ancient insular and cingulate regions.