packed with densely stained neurons. These neurons are highly active metabolically and thus VP is also apparent as the nucleus with the most expression of cytochrome oxidase (CO) in the somatosensory thalamus. The medial subnucleus, ventroposte-rior medial (VPM), representing the face, is separated from the lateral portions, ventroposterior lateral (VPL), representing the rest of the body, by a cell-sparse fiber septum, the arcuate lamina and many early investigators considered them as separate nuclei. A second, less obvious band separates the VPL into a more medial subnu-cleus, representing the hand, from a more lateral subnucleus representing the foot, and a third more lateral fiber band, separating the foot from the tail representation, is sometimes apparent. Such fiber bands separating face, hand, foot, and tail representations are also sometimes seen in non-primate mammals (Welker, 1973). Representations of the trunk and proximal limbs cap the hand and foot subnuclei dorsally.

Within VP, separate clusters of neurons are activated by either SA or RA inputs via the dorsal column-medial lemniscus pathway. The exact arrangement of these clusters is not known, but globally VP forms a systematic representation of the contralateral body surface with some additional representation of the ipsilateral oral cavity (see Bombardiei et al., 1975; Rausell and Jones, 1991a). However, the presence of separate clusters of cells for SA and RA inputs indicates that two maps of the body are interdigitated in VP.

Just dorsal to VP, we distinguish a small ventroposterior superior nucleus, VPS. The neurons in this nucleus express less cytochrome oxidase, and it has less darkly Nissl-stained cells that appear less densely packed. These differences are especially obvious in prosimian galagos and the smaller New World monkeys, but they are less apparent in Nissl preparations of the larger thalamus of macaque monkeys where neurons in general are less densely packed. Perhaps for this reason, the territory of VPS was included in VP in early studies, and the dorsal part of VP was thought to be activated by deep receptors. Now, the evidence for VPS as a separate nucleus is more compelling. Besides the architectonic distinctions, which are obvious in CO and other preparations, the VPS contains a separate representation of the body, and the inputs are largely from muscle-spindle receptors, rather than mechanoreceptors of the skin. The inputs to VPS are from the brain-stem nuclei, devoted to muscle-spindle receptors, such as the external cuneate nucleus. Also, VPS projects in a distinctly different pattern than VP to the cortex (see below).

Just ventral to VP, a ventroposterior inferior nucleus (VPI) has long been recognized in primates, but often not in other mammals (see Krubitzer and Kaas, 1987). VPI is thin dorsoventrally, and has small, lightly stained cells in Nissl preparations. The nucleus expresses little CO. VPI receives spinothalamic inputs and projects widely to the somatosensory cortex, especially areas S2 and PV (Krubitzer and Kaas, 1992). VPI appears to form a crude map of the contralateral body in parallel with the map in VP, and neurons are responsive to both cutaneous and noxious stimuli (Apkarian and Shi, 1994). The projections of VPI are to superficial, rather than middle, cortical layers (Rausell and Jones, 1991b), and transection of afferents in the dorsal columns of the spinal cord, leaving the spinothalamic inputs intact, deactivates neurons throughout areas 3a, 3b, 1, and 2 of the somatosensory cortex (Jain et al., 1997, 1998, 2001a). Thus, the thalamic relay of spinothalamic information to anterior parietal cortex does not seem to be capable of independently activating cortical neurons, and VPI appears to have a role in modulating, rather than evoking, cortical activity. However, spinothalamic terminations are also found in a recently defined nucleus, the posterior ventral medial nucleus (VMpo), which may be specific for pain and temperature sensibilities; VMpo projects to anterior cortex in the lateral sulcus (Craig et al., 1994). Thus, the spinothalamic terminations in VPI probably have functional roles other than mediating the sensations of pain and temperature.

FIGURE 1.5 The representation of the contralateral body surface in area 3b of an owl monkey. This New World monkey has only a shallow central dimple rather than a central fissure, and thus most of area 3b is exposed on the dorsolateral surface of the parietal cortex. A large sector of area 3b represents the face and oral cavity, including both the contralateral and ipsilateral tongue and teeth. This part of the representation curves rostrally to form a rostro-lateral extension of parietal cortex. A mediolateral strip of area 3b represents the body from tail to digits of the hand. Areas 3 a and 1 form additional representations of the body along the rostral and caudal borders of area 3b, and area 2 forms a fourth representation just caudal to area 1. Portions of parietal cortex contain several functional subdivisions of somatosensory cortex, but they have not been well defined. Lateral somatosensory cortex in the lateral sulcus includes the secondary somatosensory area, S2, the parietal ventral area, PV, and the rostral parietal area, PR (all not shown). Primary motor cortex, M1, or area 4, receives inputs from several somatosensory areas, as does the supplementary motor area, SMA. See Merzenich et al., 1978; Jain et al., 1997, and Jain et al., 2001b for the representation in area 3b. Areas are shown on a dorsolateral view of the brain. The encircled portions of the tongue and upper face representations curve around on the orbital surface and the upper bank of lateral sulcus, respectively.

The anterior parietal cortex of monkeys contains four strip-like areas that extend mediolaterally (Figure 1.5). All four areas were originally thought to be within a single cortical representation, S1 (Marshall et al., 1937). But, it is now clear that each of these architectonic fields contains a separate representation of the body and is a distinct area of the somatosensory cortex (for early evidence, see Merzenich et al., 1978; Kaas et al., 1979). In addition, only the representation in area 3b is the homologue of the single representation that is commonly described as S1 in other

FIGURE 1.6 Isomorphs of body parts in area 3b of monkeys. By cutting somatosensory cortex into sections parallel to the surface of the brain and processing for myelin, area 3b can be recognized as a myelin-dense field (A). In addition, it is apparent that the field is subdivided by narrow myelin-light septa, creating a sequence of myelin-dense ovals. Recordings with microelectrode indicate that these ovals correspond to the representations of body parts. In A, the ovals for a owl monkey correspond to digits 1-5 of the hand, the palm, upper face, upper lip, and chin plus lower lip. The hand-face septum and interdigital septa (arrow heads) are marked. These territories are drawn in B for comparison with A. See Jain et al. (1998, 2001b) for details.

FIGURE 1.6 Isomorphs of body parts in area 3b of monkeys. By cutting somatosensory cortex into sections parallel to the surface of the brain and processing for myelin, area 3b can be recognized as a myelin-dense field (A). In addition, it is apparent that the field is subdivided by narrow myelin-light septa, creating a sequence of myelin-dense ovals. Recordings with microelectrode indicate that these ovals correspond to the representations of body parts. In A, the ovals for a owl monkey correspond to digits 1-5 of the hand, the palm, upper face, upper lip, and chin plus lower lip. The hand-face septum and interdigital septa (arrow heads) are marked. These territories are drawn in B for comparison with A. See Jain et al. (1998, 2001b) for details.

mammals (Kaas, 1983). This conclusion is supported by a multitude of observations, including basic similarities of the organization of representations in 3b and S1, their koniocellular nature, similar patterns of intracortical (Jones et al., 1978; Pons and Kaas, 1985; Krubitzer and Kaas, 1990; Huffman and Krubitzer, 2001) and callosal (Killackey et al., 1983) connections, and the dominant projection of VP to 3b and S1 (Jones, 1975). In addition, as for S1, only area 3b responds well to cutaneous stimuli under typical recording conditions in all primates. In the anterior parietal cortex of galagos, only area 3b responds well to cutaneous stimuli (see above). Even in marmoset monkeys, only area 3b consistently responds well to cutaneous stimuli (Carlson et al., 1986; Krubitzer and Kaas, 1990). Thus, there is only one representation that is S 1-like in these primates, which is located in area 3b. As expected for S1, the receptive fields of neurons in area 3b are smaller than for any other soma-tosensory area.

As with other primary sensory areas, only area 3b is densely myelinated and only area 3b expresses a high level of cytochrome oxidase. In such preparations in New World monkeys, area 3b stands out as a narrow mediolateral strip of cortex that curves rostrally near the lateral fissure to extend far into the frontal lobe (Jain et al., 2001b). Furthermore, if cortex is cut parallel to the brain surface, and the sections processed for myelin or CO, area 3b can be seen to be composed of isomorphs of the parts of the contralateral body surface, with narrow septa separating the representations of various body parts (Figure 1.6), much as in the barrel field of mice and rats or the nose field of star-nosed moles. In the hand region of area 3b, separate myelin-dense ovals exist for each digit of the hand, and other ovals represent the pads of the palm (Jain et al., 1998). A major septum, the hand-face septum, separates the hand representation from the face representation. In the more lateral face region, a caudorostral row of three myelin-dense ovals corresponds in turn to the upper face, upper lip, and lower lip plus chin. A sequence of four more rostral ovals successively represents the contralateral teeth, contralateral tongue, ipsilateral teeth, and ipsilateral tongue. These isomorphic ovals make it possible to precisely determine parts of the representation from architecture alone. This can be very meaningful in studies of plasticity where such ovals might become responsive to different parts of the body (see Jain et al., 1998; Kaas and Florence, 2000) and in studies of neuronal connections (Fang et al., 2000). The presence of such isomorphic ovals is harder to detect in the cortex of larger brains with disrupting fissures. Yet, we have seen myelin-poor septa separating the representations of digits in area 3b of macaque monkeys (Jain et al., 1998), and such septa may be detectable in area 3b of humans.

Although modular or columnar organization is probably a common feature of cortical areas, in the somatosensory cortex of primates, modularity, in terms of segregated classes of cells, has only been demonstrated in area 3b. In the hand representation of area 3b of monkeys, separate patches of neurons receive either SA or RA thalamic inputs (Sur et al., 1981; 1984). These inputs may be segregated in other parts of area 3b as well.

Area 3a forms a strip of cortex just rostral to area 3b. Area 3a receives muscle-spindle receptor information (Phillips et al., 1971; Schwarz et al., 1973; Wiesendanger and Miles, 1982) from VPS in the thalamus (Cusick et al., 1985) and cutaneous receptor information from area 3b (i.e., Krubitzer and Kaas, 1990). A head and neck portion of area 3a also receives vestibular information from the thalamus (Akbarian et al., 1992). In microelectrode recording experiments in anesthetized monkeys, area 3a neurons are generally activated by lightly tapping body parts and other moderate levels of stimulation that would activate deep as well as cutaneous receptors (see Huffman and Krubitzer, 2001). Under light anesthesia, neurons responsive to light tactile stimulation are sometimes seen as well. Area 3a represents deep receptors in a somatotopic pattern that parallels that of cutaneous receptors in area 3b (Huffman and Krubitzer, 2001). The receptive fields in area 3a are large and can span multiple digits in the hand area.

Area 1 forms a second representation of the body surface in a strip of cortex just caudal to area 3b (Merzenick et al., 1978; Kaas et al., 1979; Nelson et al., 1980; Sur et al., 1982; Felleman et al., 1983b). Area l receives cutaneous information directly from VP and area 3b. Possibly as many as 20% of the relay neurons in VP project to both area 3b and area 1 (Lin et al., 1979; Cusick et al., 1985). Yet, the projections to area 1 from area 3b terminate in layer 4, while those from VP terminate largely above layer 4 (Jones et al., 1975), suggesting that the 3b inputs have an activating role and the VP inputs have a modulatory role. In support of this conclusion, lesions of area 3b abolish evoked responses in area 1 of squirrel monkeys and owl monkeys (Garraghty et al., 1990b). Area 1 contains a systematic map of cutaneous receptors that is essentially a mirror reversal of the map in area 3b (Merzenich et al., 1978; Nelson et al., 1980). Receptive fields for neurons in area 1 are quite small, although they tend to be larger than in area 3b.

Area 2 forms another strip-like representation of the body just caudal to area 1. While the caudal border of area 2 has been variously defined, correlations of architecture with microelectrode maps have provided a good estimate of the location of this border (see Pons et al., 1985; Lewis et al., 1999). Area 2 receives sparse cutaneous inputs from area 3b, dense cutaneous inputs from area 1 (Pons and Kaas, 1986), and muscle-spindle receptor inputs (Schwarz et al., 1973) from VPS in the thalamus (Pons and Kaas, 1985). Possibly as many as 40% of the VPS neurons that project to area 3a also project to area 2 (Cusick et al., 1985).

Neurons in area 2 generally respond to light touch (Pons et al., 1985), but they are also responsive to manipulations of the muscles and body that would activate deep receptors. Under conditions of deep anesthesia, area 2 neurons may respond poorly or not at all to somatosensory stimuli, or only to more intense stimuli that would activate deep receptors (Merzenich et al., 1978). The receptive fields for neurons in area 2 vary in size, but the receptive fields are larger than for neurons in area 3b or area 1. In the hand region of area 2, neurons with multidigit receptive fields predominate, unlike in areas 3b or 1. Area 2 has been characterized as having a systematic map of the body that is a mirror reversal of the one in area 1, but this is an oversimplification. Reversals of somatotopic organization do occur from area 1 to area 2, but area 2 also has more than one representation of digits of the hand and perhaps other body parts in at least macaque monkeys (Pons et al., 1985). The significance of this more complex somatotopic organization is not clear. Area 2 is interconnected with area 3a and more caudal somatosensory cortex, areas 5 and 7 (Pons and Kaas, 1985). The organization of this more caudal somatosensory cortex is not yet well understood.

In summary, area 3b in monkeys is clearly the homologue of S1 as defined in most mammals (Kaas, 1983). Area 3b activates area 1, which in turn activates area 2. Thus, these three areas are largely serial steps in a processing hierarchy, although area 2 directly receives important proprioceptive information from VPS. Area 3a is part of the group, but, significant area 3a outputs are directed to primary and premotor areas of motor cortex (Huerta and Pons, 1990; Darian-Smith et al., 1993; Huffman and Krubitzer, 2001). All four areas of anterior parietal cortex project to areas PV and S2 in the lateral sulcus (Krubitzer and Kaas, 1990; Qi et al., 2000).

Monkeys have at least three representations of the body surface in lateral parietal cortex of the lateral sulcus, areas S2, PV, and VS. S2 and PV either immediately border the lateral part of 3b or are separated from 3b by a narrow strip of area 1. This issue has been difficult to resolve with microelectrode mapping experiments, since lateral 3b represents the face and the immediately adjoining more lateral cortex also represents the face. This adjoining cortex could be parts of the face representations in S2 and PV, or include part of an area 1 face representation, as well as S2 and PV representations. More distant from the area 3b border and further into the lateral sulcus, S2 and PV represent the limbs and trunk, body parts that would not be represented in adjacent areas 1 or 3b. Thus, this more distant cortex is clearly not area 1. S2 and PV can be distinguished from each other because they form mirror image reversals of each other in somatotopic organization (Krubitzer and Kaas, 1990;

Krubitzer et al., 1995; Disbrow et al., 2000; Coq et al., 1999; Qi et al., 2000). The two areas adjoin along representations of the hand and face, while representations of the trunk and hindlimb are distant from each other.

In monkeys, S2 and PV differ from galagos and non-primate mammals in that they depend on inputs from the areas of anterior parietal cortex for activation (Pons et al., 1987; Garraghty et al., 1990a; Burton et al., 1990; however, see Zhang et al., 1996). Lesions of these areas, or possibly just areas 3a and 3b, abolish evoked activity in S2 and PV. This may be because the large relay neurons in the ventroposterior nucleus project in parallel to both S1 (area 3b) and to S2 and PV in most mammals but do not to project to S2 and PV in monkeys. Instead, S2 and PV receive their major thalamic inputs from VPI (Krubitzer and Kaas, 1992), and this input appears to modulate, rather than activate, S2 and PV neurons. The receptive fields in both areas S2 and PV are quite large and often extend over large parts of the body, although occasionally small receptive fields that are restricted to a single digit of the hand have been noted. Although S2 neurons typically have large receptive fields, information about stimulus location may be retained in population code (Nicolelis et al., 1998).

The ventral somatosensory area, VS, is a proposed subdivision of the soma-tosensory cortex that lies deeper to S2 in the lateral sulcus. In New World monkeys, much of VS is in the fundus of the sulcus and on the lower bank of the sulcus adjacent to the auditory cortex (Cusick et al., 1989; Krubitzer et al., 1995; Qi et al., 2000). Some of the neurons in VS respond to auditory stimuli, but the majority of neurons respond to light tactile stimulation on the contralateral side of the body. The receptive fields in area VS are comparable in size to those in area S2. Overall, VS appears to form a crude mirror reversal of S2, with foot and hand representations bordering S2, and face and trunk representations more distant from their shared border. The connections of VS are not well understood, but S2 and PV have major interconnections (Krubitzer and Kaas, 1990; Qi et al., 2000). S2 and PV also project to a parietal rostral (PR) area just rostral to PV (Krubitzer and Kaas, 1990). Little is known about PR in terms of responsiveness to tactile stimuli and other connections, but it seems likely to be a higher order somatosensory area, possibly relaying to entorhinal cortex and then to the hippocampus (Mishkin, 1979; Friedman et al., 1986) as part of the corticolimbic pathway for touch.

The posterior parietal region of monkeys contains a complex of areas that appear to use somatosensory and visual information for encoding early stages of motor control (Andersen et al., 1990, 1997). The extent of the region varies greatly across simian species, with some New World monkeys having relatively little posterior parietal cortex. Unfortunately, little is known about species differences, and even in the most studied macaque monkeys, only parts of posterior parietal cortex are well understood. Traditionally, posterior parietal cortex has been divided into areas 5 and 7 of Brodmann (1909), and these regions have been further divided into 5a, 5b, 7a, and 7b, as well as even more subdivisions, but these designations have no precise significance. Often more recently proposed subdivisions are based on physiological and anatomical distinctions (see Lewis and Van Essen, 2000), and they are more likely to constitute functionally valid subdivisions of cortex. Most notably, medial (MIP), lateral (LIP), and ventral (VIP) areas of the intraparietal sulcus of macaques have been delimited as distinct visuomotor areas that utilize eye position, proprio-ceptive, vestibular, and visual information to help guide eye, hand, and body movements. More anterior regions of posterior parietal cortex appear to be more directly involved in somatosensory processing, but even neurons in area 5 combine visual and somatosensory signals (Graziano et al., 2000). Comparative studies of posterior parietal cortex organization are clearly needed.

0 0

Post a comment