THE EVOLUTION OF THE CEREBRAL CORTEX (4,300 WORDS)© John E. LaMuth 2011
The parallel evolution of the human telencephalon is most meaningfully investigated in terms of a developmental style of analysis. The primitive vertebrate telencephalon is dominated exclusively by its own distinctive brand of olfactory sensibility. Unlike the visual sense, however, the sense of smell is exclusively restricted to only the very most ancient cerebral levels of all higher vertebrates including man. The olfactory sense in humans arises predominantly from the approximately 100 million smell receptors concentrated in the olfactory mucosa of each nostril. The axons of the receptor cells, in turn, pass through the first cranial nerve to terminate in the bilaterally paired complement of human olfactory bulbs.
In many other vertebrate species such as the rabbit, a separate (accessory) nasal mucosa gives rise to a parallel olfactory projection in the vomeronasal nerve to separately to terminate in the bilaterally paired accessory olfactory bulbs. According to Humphrey (1940) the accessory bulbs in humans atrophy after a rather cursory peak development in mid-fetal life, whereas the main olfactory bulbs continue to develop to modest adult proportions. The accessory bulb is actually larger than the main bulb in some species of snakes (Carey, 1966). This dual complement of main and accessory olfactory bulbs in the telencephalon is structurally suggestive of the dual pattern of optic input to the diencephalon. The olfactory bulb actually appears to be an extension of the cerebral cortex just as layers in the retina of the eye can be developmentally traced back to a respective complement of laminae within the diencephalon (Sarnat and Netsky, 1974). The central olfactory tract is actually the developmental homologue of the diencephalic optic tract. Each pair of olfactory bulbs relays the input from its respective set of olfactory mucosa to a specific projection site within the cerebral hemisphere just as the parietal and lateral eyes project to the epithalamic and subthalamic precursors of the thalamus. The neocortical mantle of each cerebral hemisphere is surrounded by two phylogenetically ancient types of cortex, just as the dorsal thalamus is bounded by the epithalamus and the subthalamus. The midline (medial) boundary of the neocortex is occupied by the archaecortex, while the outer (lateral) border is taken up by the paleocortex. A comprehensive survey of the relevant literature indicates that the paleocortex is specific to the main olfactory bulb input, whereas the archaecortex is alternately restricted to the accessory bulb input.
The paleocortex of humans primarily occupies the ancient pyriform lobe of the lateral aspect of the hemisphere. Together with the archaecortex, the pyriform cortex comprises the most primitive type of cortex termed the allocortex. Input fibers from the main olfactory tract enter the pyriform cortex via the surface (molecular) layer for processing within the immediately subjacent layer of gray matter. This intermediate gray layer is studded with a layer of pyramidal cells, so named for the roughly pyramidal shape of the typical nerve cell body. The apex of the pyramidal cell emits an elongated apical dendrite that ascends to the outermost layer for synaptic contact with the incoming olfactory input. The base of the pyramidal cell emits a downwardly directed efferent axon that relays the output of each pyramidal cell through the innermost (subcortical) layer of white matter to the secondary olfactory areas within the hemisphere.
The mammalian archaecortex also exhibits this same outer-to-inner sequence of input processing. With only slight variation, a similar pattern is retained in the human archaecortical formations of the hippocampus and the dentate gyrus. In many higher primate species, the hippocampal cortex becomes displaced by the neocortical growth pressures into a highly convoluted configuration deep within the lateral temporal lobes. This loss of continuity with the frontally situated olfactory bulbs, along with the ultimate atrophy of the accessory bulbs in humans, appears to contribute to the experimental notion that the mammalian archaecortex remains outside the olfactory realm of influence.
The preceding analysis, in large measure, has convincingly demonstrated a dual style of olfactory specificity for both the archaecortical and paleocortical precursors of the neocortex. This basic determination raises a further related consideration; namely, how does this pair of allocortical moieties combine to form the neocortex. Each type of allocortex was previously specified to contain only a single pyramidal layer, whereas a cross-section of the neocortex exhibits a dual pairing of pyramidal cell layers. This fundamental observation indicates that the two allocortical substrates layered one upon another to form the more complex neocortical pattern of organization.
Direct proof of this primordial type of allocortical lamination is still evident in reptilian stages of cortical evolution immediately presaging the emergence of the more advanced mammalian neocortex. In the cerebral hemisphere of turtles, the hippocampal formation directly underlies the paleocortex at a common zone of juncture (Sanides, 1970). The cerebral hemisphere of the most primitive living placental mammal, the European hedgehog, exists as an expansive ring of allocortex surrounding a modest neocortical core. The neocortical regions adjacent to the allocortex have been designated periallocortex by Sanides (1970). In a cross-section, the periallocortex differs fundamentally from the more basic allocortical pattern in terms of its advanced arrangement of two pyramidal cell bands separated by the cell-free lamina dessicans. The outer cell band is continuous with the paleocortex, whereas the inner band of cells blends with the archaecortex (Sanides, 1970). This periallocortical configuration is clearly suggestive of the lamination of the paleocortex over the archaecortex. The periallocortex of the hedgehog represents only the very most primitive stage of neocortical differentiation. Through intervening stages of evolution, however, the hedgehog hemisphere has further developed the beginnings of an even more advanced form of cortex known as the isocortex.
The hedgehog hemisphere is centrally crowned by the most initial stage of isocortical development termed proisocortex by Sanides. This proisocortical core does not represent an additional lamination phase, but rather is identified as a more advanced modification of the more basic periallocortical format. The core structure of the proisocortex differentiates concentrically within the periallocortex, turning the latter into a growth ring, just as the preceding differentiation of the periallocortex modified the allocortex into a ring within the hedgehog hemisphere. According to Sanides, the segment of the newly formed periallocortical growth ring adjacent to the paleocortex is termed the peripaleocortex, whereas the band paralleling the archaecortex is termed the periarchaecortex.
The proisocortex differs fundamentally from the periallocortex in terms of the neocortical feature of granularization. In the proisocortex of the hedgehog, and more generally in higher mammals, the cell-free layer known as the lamina dessicans is invaded by a short-axoned interneuron cell known as the granule cell. In contrast to the surface-bound course taken by the olfactory and thalamic inputs to the periallocortex, the thalamic input fibers to the proisocortex predominantly ascend via the subcortical white matter to synapse directly upon the granule cells within the newly formed granular layer. The short axon of the granule cell, in turn, reaches into the adjacent pyramidal cell layers, whereby relaying the processed thalamic input data to the efferent pyramidal cells. This phasing-out of surface-bound inputs in favor of a more centralized granule cell contingent represents a more advanced pattern that is more completely consolidated in terms of the even more recent age levels within the isocortex.
The cerebral hemisphere of some related insectivorous species of bats exhibit one additional age-level of isocortical differentiation termed the para-proisocortex (para = beside). The para-proisocortical growth core of the bat hemisphere modifies the original proisocortex into yet one further concentric growth ring within a pre-existing sequence of periallocortical and allocortical growth rings. This pattern of para-proisocortical differentiation is clearly characterized in higher mammals by a marked expansion of the granule cell layer and the phasing out of proisocortical features related to the no longer prevalent occurrence of surface-bound inputs. From the preceding comparative analysis of a broad number of representative insectivore and reptilian species, a general pattern of neocortical evolution clearly emerges. The various stages of neocortical evolution are evident as individual age levels within a growth ring pattern spanning the evolution of the most primitive of mammalian species. According to Sanides, the neocortex of the remainder of mammalian species more advanced than insectivores display (in addition to the insectivore complement of age levels) a fourth and final growth core of differentiation unique to the more advanced mammalian forms.
THE SIX AGE LEVELS OF THE HUMAN NEOCORTEX
A comprehensive survey of the growth ring patterns for representative primate species by Sanides (1969) predominately focuses upon the novel characteristics of the fourth and final growth core of neocortical evolution. In most species of apes (including man) this final growth core expands so dramatically as to displace the other three neocortical growth rings completely off the exposed convexity of the hemisphere. These three more ancient neocortical growth rings are forced either onto the medial (limbic) aspect of the hemisphere, as is the case for the archaecortical gradient, or displaced deeply within the insular lobe of the lateral hemisphere (paleocortical gradient). The para-proisocortical growth core initially seen in the bat hemisphere is respectively modified into a circumferential growth ring that surrounds the final concentric growth core of primates. The growth ring segment continuing the paleocortical gradient is termed the parainsular zone according to Sanides, whereas the alternate segment continuing the archaecortical gradient is designated the paralimbic zone. Although topographically separated by the final growth core, the limbic and parainsular segments of the last surrounding growth ring exhibit a similar cytoarchitectonic structure throughout their entire peripheral range. The same homogeneity also appears to be the case for the proisocortical and periallocortical growth rings, which (with the exception of small variations around their circumference) appear as large rings of uniform laminar characteristics. Such is not the case for the final primate growth core, which is structured as an uneven admixture of basic types of cortex, alternately designated as homotypical and heterotypical neocortex.
In the human cerebral hemisphere, the homotypic cortex represents the more basic neocortical pattern, whereby all six characteristic cellular laminae are clearly distinguishable. The heterotypical type of cortex, in turn, represents a more advanced modification of the general homotypical pattern, where all six laminae are still evident, but due to functional modification are not all clearly demarcated. The most basic type of heterocortex is termed the koniocortex (Greek – konio – dust), designated for the powdery appearance of the outer four layers due to the extensive accumulation of granule cells. The superabundant inner granule layer (IV) is effectively fused with the outer granule (II) and pyramidal (III) layers, making this type of isocortex highly specialized for processing sensory input. Different subdivisions of koniocortex have been designated for the various optic, auditory, and somatosensory input classifications that ascend to terminate in the neocortex.
The remaining class of heterotypical cortex is agranular rather than granular in structure. The agranular neocortex is represented in the human hemisphere by thought classical motor area “gigantopyramidalis.” This area is designated for the giant inner pyramidal cells of lamina (V), the axons of which descend through the brain stem and spinal white matter to synapse in the vicinity of the spinal motor-neurons. This pronounced functional emphasis upon motor control rather than sensory processing is paralleled by the manifest lack of an inner granule cell layer within the motor cortex. This alternate agranular variety of hetereotypical cortex is, nevertheless, considered to be of the same high level of differentiation that was previously established for the granular form as well.
Both the granular and agranular forms of koniocortex examined so far have further exhibited a generalized type of distended elongation that extends into continuity with either the parainsular or paralimbic segments of the precursor growth ring. This phenomenon appears to have served as the grounds for Sanides (1969) contention that the koniocortical age levels evolved directly as the result of differentiation within the paralimbic/parainsular growth ring. The massively distributed homotypical areas of the growth core, however, also come into an extensive approximation with the paralimbic/parainsular growth ring, exhibiting extremely uniform zones of transition across the growth core boundary. Furthermore, this homotypical cortex actually surrounds all of the remaining borders of the heterotypical cortex not already accounted for by the paralimbic/parainsular junctures.
In this latter type of role homotypical cortex formally meets the conditions set down by Sanides for what he terms the “belt areas” surrounding the koniocortex. For instance, the primary visual (striate) koniocortex is surrounded by two bands of homotypical cortex designated the parastriate and peristriate regions. In similar fashion, the closely adjacent sensorimotor representations surrounding the central sulcus each exhibits its own complement of belt areas, some of which are actually shared between these two different types of heterotypical cortex. Electrophysiological studies have demonstrated that the surrounding belt areas are functionally specific to the same type of sensory input processed within the koniocortical core the, however, with less of the precise topographical organization that characterizes the primary sensory representations.
In piecing together each of these individual clues, a radically new idea emerges that is contrary to what originally was proposed by Sanides; namely, and that the homotypical belt areas actually represent a separate growth ring of isocortical evolution that is intermediate along the evolutionary continuum between the primitive paralimbic/parainsular growth ring and the highly differentiated heterotypical growth core. According to this revised interpretation, the homotypical cortex is of a more ancient origin than the heterotypical cortex, the former essentially being displaced into a belt-like configuration by the final growth core of differentiation characteristic of the koniocortex proper. The ideal concentric configuration of this intermediate homotypical growth ring is only camouflaged by the highly distended pattern exhibited by the koniocortical representations.
There is a well-investigated body of further evidence to support the claim for the addition proposed to the already established growth ring complement. Until now, the growth ring gradient of the neocortex has been examined according to cytoarchitectonic criteria; i.e., the variations in form and structure of the nerve cell bodies in the various cortical laminae. Some histological stains, however, demonstrate a specific affinity for the myelinated fibers that inundate much of the open space within the cortical laminae. Through corresponding myelographic techniques, the precise degree of myelination within a specific cortical area can be measured for a calculation of their relative age of each region. Faster conducting, and more heavily myelinated fibers represent a more recent stage of the neocortical evolution than those areas containing thinner fibers.
Both the granular and agranular heterotypical cortex represent focal maxima of the evolutionary trend for an increase in myelination within the human cerebral hemisphere. In contrast, the myelination factor within the homotypical cortex does not reach that degree of accentuation achieved in the heterotypical cortex, suggested in a clear difference in evolutionary time frames for these two basic types of cortex. According to both cytoarchitectonic and myelographic criteria, A radical modification of Sanides’ final growth core region is clearly warranted. This final age level, in re-evaluation, is necessarily split into an evolutionary pair-sequence of age levels characterized as an additional growth ring surrounding a considerably more narrow growth core. Retaining the format previously introduced by Sanides, is proposed that this distinction be recognized by restricting the designation koniocortex (including the agranular variety) to the sixth wave of development, while coining the term “prekoniocortex” to define the ring-like fifth neocortical wave.
Koniocortex is unique among cortical waves, in that it appears as a central core surrounded concentrically by all of the older growth rings. Those portions of the growth rings, positioned between the medial koniocortical border and the archaecortex are termed the medial ur-trend, while the growth ring segments, intermediate to the paleo- cortex and the lateral koniocortical boundary, are termed the lateral ur-trend (Sanides, 1970). Using cytoarchitectonic and myelographic techniques, Sanides demonstrated that the classical sensorimotor representations had developed via continuity across both ur-trends; however, with one ur-trend being generally more accentuated than its counterpart. The parameter grid shown in Figure l portrays an attempt by the author to apply this principle to all representative areas within the final koniocortical growth core. It is this precise insistence on the existence of an unaccented ur-trend, which accounted for the paired redundancies between unit squares of adjacent ur-trends. Figure 3 shows the appropriately revised version of the parameter grid, showing each koniocortical representation as a product of a single dominant ur-trend.
The jagged line, depicted at the interface between the medial and lateral ur-trends of Fig. 3 represents what is termed the ur-trend limiting sulcus (Sanides, 1969). This limiting sulcus in man is represented rostrally as the inferior frontal sulcus, intermediately as the sensorimotor sulcus between the arm and head representations and caudally as the interparietal sulcus (Sanides, (1970). The position of each koniocortical representation relative to this limiting sulcus, is a valuable criterion for determining its ur-trend of origin. In the frontal lobe, for instance, the inferior frontal gyrus is composed of a sequence of three highly differentiated areas, denoted as pars opercularis (#44), pars triangularis (#45), and pars orbitalis (#12) (Sanides, 1964). Each of these areas characteristically displays giant pyramidal cells in lamina IIIC: an essential property of areal maximums derived across the lateral ur-trend. The middle frontal gyrus is host to an analogous sequence of ur-trend maxima; areas #46, #8δ, and #6aα. Each of these areas displays a size accentuation of pyra¬midal cells in lamina V characteristic of their medial ur-trend origin.
A close inspection of the posterior association region reveals the presence of a pair of conspicuously myelinated koniocortical bands. The visuo-auditory band is situated in area #39 of the visual association region, while the corresponding visuo-sensory band is located between areas #7 and #40 in the somatosensory association region. These bands were first detected during gross dissection as regions of dramatic myelin accentuation (Smith, 1907). These same cortical strips were subsequently shown to commence myelination much earlier than the surrounding associa¬tion regions (Flechsig, 1920). The medially derived visuo-sensory band is located on the dorsal wall of the interparietal sulcus, while the visuo-euditory band is situated lateral to the occipital continuation of this sulcus. The classical visual area #17 was cited by Sanides (1970) as derived solely along a medial ur-trend gradient. Accordingly, area #17 exhibits the giant pyramidal cells of Meynert in cortical lamina V. In the same article, Sanides proposed that the classical auditory representation developed solely along a lateral ur-trend gradient; at least in the case of lamina IIIc accentuated area #42. Auditory area #41 does not display giant lamina IIIc pyramidal cells, accordingly being derived by way of a medial ur-trend gradient span¬ning the superior temporal gyrus.
The remaining intermediate segment of the koniocortical core is composed of sensorimotor areas #4γ, #3a, and #3b of the pre- and post-central gyri. The distended parallel orientation of all three areas perpendicular to the ur-trend limiting sulcus, not only promotes somatotopic cross-modal continuity, but unfortunately invalidates the limiting sulcus as an ur-trend determining criterion. Unlike the classi¬cal somatosensory area #3b, which displays giant pyramidal cells in lamina IIIc, the cortical motor areas #4γ and #3a exhibit large pyramidal cells in both inner and outer laminae (Bailey & von Bonin, 1951). These latter two areas will be provisionally included as fitting the pattern of strict unit alternation for medially and laterally derived core areas. Here, only single alternate ur-trends are prerequisite for koniocortical differentiation, so that the few remaining cortical duplications occur in the primitively differentiated and experimentally inaccessible insular and cingulate gyri.
The preceding ur-trend analysis of the human koniocortical growth core clearly establishes a pattern of equipotential growth core development relative to the ur-trend limiting sulcus. Seven koniocortical subdivisions evolve by way of the medial ur-trend, whereas the remaining seven are derived across the extent of the lateral ur-trend. In addition to this strictly quantitative type of equivalence, the medially derived koniocortical subdivisions exhibit a strict unit alternation relative to the laterally derived representations across the entire extent of the ur-trend limiting sulcus. This staggered medial/lateral pattern of koniocortical differentiation results in the characteristic interlocking pattern of growth core areas across the ur-trend limiting sulcus.
SUMMARY
The neocortical manifestation of the parameter of phylogenetical age is most clearly defined in terms of Sanides’ circumferential growth rings of neocortical differentiation. Reiterating Sanides (1972) there occur two distinct stages of structural differentiation in the cortex. The first stage is demonstrated as the lamination of the paleocortex onto the archaecortex forming the periallocortical growth wave. Periallocortex, in turn, gives rise to an additional sequence of neocortica! growth waves characterizing the second stage of structural differentiation. These subsequent neocortical growth waves, collectively termed isocortex, are cytoarchitectonic variations of the fundamental periallocortica! format rather than additiona! lamination phases. Each successive isocortica! wave develops concentrically, turning its precursor into a circumferentially situated cortical growth ring.
Isocortex initially occurs as the second cortical growth wave, termed proisocortex by Sanides. Proisocortex is characteristically distinguished from periallocortex by the appearance of an inner granular layer, composed of true stellate interneurons. Proisocortex is subdivided into insular and cingular belt regions which totally encircle the remainder of the isocortex.
The third growth wave consists of a pair of paralimbic and parainsular belt regions that collectively comprise a third-order growth ring. This tertiary growth ring is distinguished from the proisocortex by an accentuation of the outer pyramidal layer, and a diminution of proisocortical (limbic) features. The third growth ring has been electrophysiologically demonstrated to encompass regions mapped as the secondary and supplementary sensorimotor representations (Sanides, 1972).
Sanides’ fourth and final koniocortical growth wave, covering virtually the entire hemisphere convexity, was in reevaluation found actually to be composed of a pair of developmentally distinct neocortical growth waves. According to cytoarchitectonic criteria, it appears highly inconsistent to group the heterotypical classical koniocortex and the homotypica! association cortex in the same cortical growth wave. By myelographic standards, koniocortex repre¬sents the focal maximum of the hemisphere myelination trend (Sanides, 1969) signifying a chronologically later development than adjacent association regions. In accordance with these basic criteria, Sanides’ fourth and final koniocortical growth wave is essentially modified into a fifth-order koniocortical maximum, flanked on either side by a pair of belt zones corresponding to a fourth-order growth ring. In keeping with the format introduced by Sanides, it is proposed that this distinction be recognized by restricting the term koniocortex to the fifth wave of differentiation while coining the term prekoniocortex to define the ringlike fourth cortical wave. The close proximity of the paralimb!c ring to the koniocortex must be interpreted as an artificial continuity brought about by cortical growth pressures.
Koniocortex is unique among the cortical waves in that it exists as an unpaired core region surrounded concentrically by the older growth rings. The belt areas positioned between the medial koniocortical border and the archaecortex are termed the medial ur-trend, whereas those belt areas intermediate to the paleocortex and the lateral koniocortical boundary ace termed the lateral ur-trend (Sanides, 1972). Koniocortex frequently lines the limiting axial sulcus located at the interface between the medial and lateral ur-trends This limiting sulcus in man represented rostrally as the inferior frontal sulcus, intermediately as the sensorirnotor sulcus between the arm and head representations, and caudally as the interparietel sulcus (Sanides, l972). Koniocortex is frequently positioned on only a single side of this limiting sulcus suggesting a developmental emphasis from the respectively adjacent ur-trend.
A cursory examination indicates that the cortical association areas outnumber the classical cortical areas by an order of magnitude. A closer inspection of the posterior association region, however, reveals the presence of a pair of conspicuously myelinated koniocortical-like bands. The visuo-auditory band is situated between areas #37 and #39 in the visual association region, while a corresponding visuo-sensory band is located between areas #7 and #40 in the somatosensory associational zone. These bands were first detected during naked eye dissection as regions of dramatic myelin accentuation (Smith, 1907). These same cortical strips were subsequently shown to commence myelination much earlier than the surrounding association regions (Flechsig, 1920). A pronounced myelin content, as well as an early sequence of myelination, are specifically indicated as criteria which are also characteristic of the classical variety of koniocortex. The term associational koniocortex is introduced to emphasize the inclusion of these bandlike associational maximas into the fifth order koniocortical growth wave.
The visuo-sensory band is located on the dorsal wall of the interparietal limiting sulcus, while the visuo-auditory band is situated lateral to the occipital continuation of ths suicus. In relation to this limiting sulcus, the somatosensory associational koniocortex reaches its developmental maximum from the medial ur-trend, whereas the visual associational koniocortex attains its emphasis from the lateral ur-trend.
The visual and somatosensory classical koniocortices display a similar developmental disparity of individual ur-trend components. Sanides (1972) demonstrated that the classical somatosensory koniocortex is developmentally derived across both ur-trends, however, with the major differential emphasis from the lateral ur-trend. The classical visual koniocortex was proposed to have developed solely from the medial ur-trend, although a gradient from the entorhinal area was also demonstrated. (Sanides, l972). Designation of the entorhinal area as peripaleo- rather than periarchaecortex (Valvarde, 1965) suggests the presence of a lateral ur-trend leading to the striate area #17. This lateral ur-trend is nevertheless less accentuated than the medial ur-trend from the parahippocampal gyrus. Thus it appears that for the visual and somatosensory modalities the ur-trend disparity shown for the associational koniocortex is reciprocally counterbalanced by the conversely accentuated ur-trend complement of the classical koniocortex. In the discussion of the parameter of imput specificity this dichotomous ur-trend pattern is shown also to hold true for all remaining constituents of the koniocortical growth wave as well.
Comprehensive analysis of the first cortical stage of structural differentiation provides the mechanistic rationale behind this reciprocal interplay of the paired koniocortical ur-trend complements. This initial neocortical growth wave is composed of peripaleocortical and periarchaecortical belt regions forming a ring encircling the entirety of the isocortex. The paleocortex and archaecortex are primordial precursors of the periallocortex by means of a complex structural lamination of the two respective moieties.