Case series evidence for changed interhemispheric relationships in cortical structure in some amputees.

Limb amputation and related changes in body feelings are associated with cortical functional reorganization that is reflected by increased interhemispheric asymmetry of body maps in the postcentral somatosensory cortex (PCS). As a pilot test to determine if limb amputation affects interhemispheric symmetry in PCS structure, we used MRI and computational morphometry to examine interhemispheric relationships of PCS thicknesses in a case series of eight lower limb amputees compared with 11 control subjects. As a further control, the same relationships were compared in the lateral occipital visual cortex (LOV) which, by nature of its visual connectivity, would be expected to be less related to amputation. The PCS thicknesses in the left and right hemispheres were positively related in control subjects, but not in amputees. The range of the PCS interhemispheric thickness differences (ID) in amputees was larger than the range in control subjects, and four of eight amputees had PCS ID that were at or above the maximal control subject ID. In contrast, LOV thicknesses in the two hemispheres were positively related and LOV ID ranges were similar in both amputees and control subjects. The results from this case series suggest the hypothesis that amputation alters PCS interhemispheric thickness relationships in some amputees. Further tests of this hypothesis would be useful to determine whether changes in structural symmetry contribute to known post-amputation alterations in PCS functional map symmetry and body feeling.

Human brain plasticity: an emerging view of the multiple substrates and mechanisms that cause cortical changes and related sensory dysfunctions after injuries of sensory inputs from the body.

Injuries of peripheral inputs from the body cause sensory dysfunctions that are thought to be attributable to functional changes in cerebral cortical maps of the body. Prevalent theories propose that these cortical changes are explained by mechanisms that preeminently operate within cortex. This paper reviews findings from humans and other primates that point to a very different explanation, i.e. that injury triggers an immediately initiated, and subsequently continuing, progression of mechanisms that alter substrates at multiple subcortical as well as cortical locations. As part of this progression, peripheral injuries cause surprisingly rapid neurochemical/molecular, functional, and structural changes in peripheral, spinal, and brainstem substrates. Moreover, recent comparisons of extents of subcortical and cortical map changes indicate that initial subcortical changes can be more extensive than cortical changes, and that over time cortical and subcortical extents of change reach new balances. Mechanisms for these changes are ubiquitous in subcortical and cortical substrates and include neurochemical/molecular changes that cause functional alterations of normal excitation and inhibition, atrophy and degeneration of normal substrates, and sprouting of new connections. The result is that injuries that begin in the body become rapidly further embodied in reorganizational make-overs of the entire core of the somatosensory brain, from peripheral sensory neurons to cortex. We suggest that sensory dysfunctions after nerve, root, dorsal column (spinal), and amputation injuries can be viewed as diseases of reorganization in this core.

Evidence for brainstem and supra-brainstem contributions to rapid cortical plasticity in adult monkeys.

Cortical maps can undergo amazingly rapid changes after injury of the body. These changes involve functional alterations in normal substrates, but the cortical and/or subcortical location(s) of these alterations, and the relationships of alterations in different substrates, remain controversial. The present study used neurophysiological approaches in adult monkeys to evaluate how brainstem organization of tactile inputs in the cuneate nucleus (CN) changes after acute injury of hand nerves. These data were then compared with analogous data from our earlier cortical area 3b studies, which used the same approaches and acute injury, to assess relationships of cuneate and cortical changes. The results indicate that cuneate tactile responsiveness, receptive field locations, somatotopic organization, and spatial properties of representations (i.e., location, continuity, size) change during the first minutes to hours after injury. The comparisons of cuneate and area 3b organization further show that some cuneate changes are preserved in area 3b, whereas other cuneate changes are transformed before being expressed in area 3b. The findings provide evidence that rapid reorganization in area 3b, in part, reflects mechanisms that operate from a distance in the cuneate nucleus and, in part, reflects supracuneate mechanisms that modify brainstem changes.

Functional organization of tactile inputs from the hand in the cuneate nucleus and its relationship to organization in the somatosensory cortex.

Central processing of tactile inputs from the hand begins in the main cuneate nucleus and continues in the thalamus and area 3b cortex. Little is known about cuneate functional organization in primates or about how cuneate and area 3b organization are related. In this study, neurophysiologic approaches were used to evaluate how tactile inputs from the hand and adjacent body are organized in the cuneate nucleus of squirrel monkeys. Cuneate data on the organization of hand inputs were then compared with analogous area 3b data from our earlier cortical studies that used the same approaches. Evaluations of several cuneate properties, including (1) responsiveness to tactile stimulation, (2) incidences and sizes of receptive fields, (3) somatotopic progressions, (4) properties of representations, and (5) relationships between functional inputs and cytochrome oxidase staining, suggest that tactile afferents from the hand form consistently organized cuneate representations that, in turn, relate to the parcellated organization of cuneate structural substrates. Comparisons of cuneate and area 3b organization indicate that tactile processing from the brainstem to cortex involves a preservation of tactile responsiveness and somatotopic organization but, in addition, involves transformations that produce receptive field sharpening, suppression of hairy hand inputs, amplification and refinement of glabrous inputs, and relocations of representations. Ascending lemniscal substrates are characterized by cascading excitatory convergence/divergence that increments at successively higher levels between sensory afferents and area 3b. It is suggested that the observed preservations and transformations reflect this organization but, in addition, reflect mechanisms that cause counterbalancing sharpening and suppressions of hand inputs.

Rapid changes in brainstem maps of adult primates after peripheral injury.

Cerebral cortical maps in adult primates reorganize within minutes-hours after peripheral injuries, but subcortical versus intracortical contributions to this rapid reorganization remain controversial. The present results show that injury of nerves to the hands of adult monkeys triggers rapid (minutes-hours) changes in maps of the hand in the brainstem main cuneate nucleus. These findings suggest that peripheral injury causes an initial concurrent reorganization of brainstem and cortical substrates and that early sensory changes emerge from reorganization involving multiple central levels.

Initial cortical reactions to injury of the median and radial nerves to the hands of adult primates.

The area 3b hand cortex of adult squirrel monkeys was mapped during the first minutes to hours after transecting the radial and median nerves to the hand. The objective was to evaluate initial cortical reactions to this injury and to determine whether patterns and extents of cortical change are similar in different individuals. There are 5 main findings. First, cortical aggregates related to ulnar nerve inputs from the hand rapidly expanded to occupy an additional 21% of the cortical hand map. Second, face and forearm inputs, which normally activate areas adjacent to hand cortex, rapidly expanded into areas of 4% and 1% of the hand cortex respectively. Third, cortical changes involved shifts in receptive field locations that were initiated within minutes after injury. Fourth, the spatial patterns and extents of cortical change were similar in different individuals. Finally, the pattern of cortical change produced after this injury differed from the pattern seen after injury of the median and ulnar nerves. These rapid expansions are a beginning point from which further changes must progress; however, in contrast to changes accompanying chronic hand injuries, these initial cortical reactions do not appear dictated by use of uninjured inputs.

Cutaneous representations of the hand and other body parts in the cuneate nucleus of a primate, and some relationships to previously described cortical representations.

Dynamic properties of primate somatosensory maps are dependent on normal central adjacencies of cutaneous representations. The cuneate nucleus is an important brainstem processing center of cutaneous information. Surprisingly, there are no descriptions of functional representations of the skin in the primate cuneate nucleus; as a result, the relationships of functional representations at the brainstem level and other levels of the somatosensory neuraxis remain obscure. The present neurophysiological study indicates that the main cuneate nucleus of marmoset monkeys (Callithrix jacchus) contains organized representations of cutaneous inputs from the hand, forelimb, and adjacent body between the lateral face and proximal hindlimb. Inputs from the glabrous hand are represented continuously across transverse planes in the cuneate, whereas inputs from the hairy hand are represented discontinuously. Inputs from distal to proximal, and radial to ulnar, parts of the hand are mapped in an organized manner. At rostrocaudal levels where the cuneate nucleus is near its largest transverse area, the map of the hand is about 2600 times smaller than the hand skin area it represents. Cuneate representations of the forelimb and trunk are represented both medial and lateral to the hand representation, and interface with representations in the adjacent gracile and trigeminal nuclei. These findings provide a starting point for understanding functional representations of the skin in the cuneate nucleus of primates. Furthermore, they provide a basis for understanding relationships of cutaneous representations at different levels of the neuraxis. In this regard, comparisons of the present results to previously defined representations in the somatosensory (area 3b) cortex indicate that cuneate hand representations are several times smaller than cortical representations, and that there are similarities and differences in adjacencies of cuneate and cortical representations.

The consistency, extent, and locations of early-onset changes in cortical nerve dominance aggregates following injury of nerves to primate hands.

The somatosensory cortex of primates contains patch- and bandlike aggregates of neurons that are dominantly activated by cutaneous inputs from the radial, median, and ulnar nerves to the hand. In the present study, the area 3b hand cortex of adult monkeys was mapped immediately before and after combined median and ulnar nerve transection to evaluate the consistency, extent, and location of early post-injury alterations in the deprived median and ulnar nerve cortical bands. Several alterations were observed acutely after injury. (1) The patchlike cortical aggregates of intact radial nerve inputs from the hand underwent a two- to three-fold expansion. This expansion was not related to peripheral changes in the radial nerve skin territory, but was due to rapid central decompression of radial nerve dominance patches. (2) The largest changes involved patches in lateral to central locations of the hand map. (3) The expanded patches occupied cortical zones that were activated by inputs from the digits, palm, and posterior hand prior to injury. These receptive field shifts were initiated within minutes after injury. (4) Receptive fields of neurons within expanded radial nerve patches were normal in size. (5) Besides changes involving radial nerve inputs from the hand, there was a small expansion of forelimb inputs into the preinjury hand cortex; however, the representation of face inputs did not expand into this cortex. (6) Finally, neurons across 50-69% of the hand cortex were unresponsive to tactile stimuli acutely after this injury. These findings indicate that the distribution patterns of nerve dominance aggregates in adult primates begin changing within minutes after nerve injury. Cortical changes involving specific inputs occupy similar extents and locations of cortex, and are arranged in highly consistent patterns, in different individuals. It is suggested that this consistency reflects specific patterns of central sensitization or disinhibition that are triggered by the injury.

Sensory afferent projections and area 3b somatotopy following median nerve cut and repair in macaque monkeys.

The fidelity of median nerve regeneration and the consequent effects of regeneration errors on cortical organization were determined in combined anatomical and electrophysiological studies. In three adult macaque monkeys, the median nerve was cut, sutured, and allowed to regenerate for 7-13 months. After regeneration, distributions of afferents to the dorsal horn of the spinal cord and the cuneate nucleus of the brainstem were determined by making injections of horseradish peroxidase conjugates into the distal phalanges of digit 1 or 2. While label from a single digit on the normal hand was confined to the appropriate locations in the median nerve territories of the dorsal horn and cuneate nucleus, label from the reinnervated digits spread out to cover most of the median nerve territories in those structures. These results are consistent with the interpretation that some proportion of primary sensory fibers normally innervating other digits and pads of median nerve skin erroneously reinnervated the skin of the injected digits. In the same monkeys, microelectrodes were used to record from an array of closely spaced sites across the representation of the hand in area 3b of somatosensory cortex. The reactivation pattern was abnormal, with neurons at many recording sites having more than one receptive field, larger than normal receptive fields, or receptive fields at abnormal skin locations. Thus, there is somatotopic disorder both in the regenerated median nerve and in reactivated cortex, indicating that primary somatosensory cortex does not reorganize to compensate fully for peripheral reinnervation errors in these adult primates. Nevertheless, the organization of receptive fields in area 3b suggests the existence of some central selection of synapses.

Nerve innervation of the hand and associated nerve dominance aggregates in the somatosensory cortex of a primate (squirrel monkey).

The cutaneous innervation territories of the median, ulnar, and radial nerves to the hand were determined from neurophysiological recordings of peripheral mechanoreceptor axons in adult squirrel monkeys. These territories were then related to cutaneous receptive fields of cortical area 3b neurons to determine how low-threshold inputs from each hand nerve map onto the primary somatosensory cortex. The results indicate that mechanoreceptor axons in each nerve innervate a continuous skin territory covering about 40% of the hand surface. The total territory of each nerve contains subregions of skin that are either autonomously innervated by that nerve or that receive overlapping innervation from more than one nerve. The autonomous, overlap, and total territories of each nerve are relatively constant from hand to hand. In the area 3b cortex, low-threshold afferents from each nerve provide inputs to aggregates of cortical neurons. The cortical aggregates relating to the median and ulnar nerves are arranged as continuous, rostrocaudally oriented bands, whereas aggregates relating to the radial nerve are discontinuous and more patch-like. Similar patterns of bands and patches, and similar compression ratios of skin/cortical area, are seen across different monkeys. These findings demonstrate that the primary somatosensory cortex of normal adult primates contains bands or patches of neurons that are dominantly activated by low-threshold inputs from specific hand nerves. This approach of delineating nerve territories and their related cortical dominance aggregates provides a useful means of analyzing cortical images of nerves and of quantitating peripheral and central patterns of deprivation after nerve injury.

Lesion-induced changes in the central terminal distribution of galanin-immunoreactive axons in the dorsal column nuclei.

Rats that sustained forelimb removal on either embryonic day (E) 16, on the day of birth (P-0), or transection of the brachial plexus in adulthood had brainstem sections stained for galanin, calcitonin gene-related peptide (CGRP), or substance P (SP) at various intervals after these lesions were made. In normal adult rats, only a few galanin-immunoreactive fibers are present in the cuneate nucleus and most are located in its caudal portion. CGRP-positive axons are also sparse in the cuneate and are distributed mainly in the periphery of the nucleus. SP-positive axons are seen throughout the cuneate nucleus. In rats that sustained forelimb removals at birth or transection of the brachial plexus in adulthood, dense galanin immunoreactivity was present throughout the cuneate nucleus at all rostrocaudal levels on the side of the brainstem ipsilateral to the lesion. The changes after lesions that were made in the adult animals were apparent within 1 week, the earliest time analyzed. Increases in galanin immunoreactivity in the cuneate of animals that sustained forelimb removals on P-0 were first visible on P-2. Neither forelimb removal at birth nor brachial plexus lesions in adulthood had any qualitative effect upon the distribution or density of CGRP- or SP-immunoreactivity in the cuneate nucleus. Removal of a forelimb on E-16 did not increase the density of galanin-immunoreactive fibers in the cuneate nucleus. Such lesions also failed to produce any appreciable change in the density of either CGRP- or SP-positive fibers in the cuneate nucleus. The present data raise the possibility that large caliber, non-peptidergic primary afferent axons which innervate the cuneate nucleus may express galanin after damage at birth or in adulthood.

Anatomical and functional changes in the organization of the cuneate nucleus of adult rats after fetal forelimb amputation.

A previous study has shown that fetal forelimb removal in the rat results in an increase in the size of the hindlimb representation in primary somatosensory cortex and suggested that this anomalous cortical organization may have resulted from alterations in the primary afferent innervation of the dorsal column nuclei (Killackey and Dawson, 1989). The present study used both anatomical and electrophysiological techniques to examine the effects of fetal forelimb amputation on the dorsal column nuclei. Rats sustained forelimb removals on embryonic day 16 and were used in terminal experiments when they reached adulthood (> 60 d of age). Analysis of cytochrome oxidase-stained sections demonstrated that the cuneate nucleus ipsilateral to the lesion decreased in volume by an average of 36.7% (N = 7, p < 0.001, paired t test), but there was no corresponding increase in the volume of the gracile fasciculus and nucleus. Bilateral application of HRP to the sciatic nerves demonstrated that axons that innervate only the gracile nucleus on the intact side of the brainstem were present in the cuneate nucleus on the deafferented side. Injection of HRP into the skin overlying the point of the amputation (the stump) indicated that axons innervating this region filled most of the dorsal one-half of the shrunken cuneate nucleus and overlapped with the sciatic nerve afferents innervating the cuneate on this side. Mapping the receptive fields of multiple unit clusters demonstrated that most recording sites in the shrunken cuneate nucleus were activated by inputs from the stump and adjacent skin. In addition, 9.1% (N = 30) of such unit clusters (N = 328) could also be excited by stimulation of the hindlimb. These were observed in only three of the nine experiments. Unit clusters with split receptive fields including the skin overlying the stump and the hindlimb were located throughout the rostrocaudal extent of the cuneate nucleus. These results indicate that fetal forelimb amputation results in anatomical expansion of the central projections of hindlimb afferents into the cuneate nucleus. This anatomical organization appears weakly expressed in the receptive fields of cuneate neurons.

Changes in the cortical map of the hand following postnatal median nerve injury in monkeys: modification of somatotopic aggregates.

Median nerves to the hands of 8-15-d-old marmoset monkeys were transected and precluded from regeneration by ligation. Following periods of 0.4-1.5 years, features of organization in the cortical area 3b hand map were assessed neurophysiologically, and compared to features in normally reared monkeys. Cortical features in monkeys with both histories were similar in certain respects. (1) Receptive field organization was similar in terms of tactile thresholds and receptive field size, continuity, and glabrous-hairy specificity. (2) Somatotopic organization was similar in terms of the continuity of the glabrous representation, and progressions of receptive field shifts across some parts of the hand map. (3) Finally, the overall size of the hand map did not change. In contrast, other cortical features clearly differed following these developmental histories. (1) Neurons at virtually all recording sites in normal hand maps responded to light mechanical stimulation, whereas, following injury, neurons at about 8% of the recording sites responded only to high-intensity stimuli. (2) Somatotopic organization differed in terms of the presence or absence of the representation of skin autonomously innervated by the median nerve, the number and continuity of representations of hairy skin, and the spatial interfacing of representations. (3) Finally, there were differences in the areas and widths of representations of parts of the hand. The overall impression is that there is a correspondence between the cortical features that changed most after injury, and the features that varied most in individual normal monkeys: in both circumstances the most variable features involved properties of spatial patterning across large aggregates of neurons as reflected by the size, shape, continuity, and interfacing of representations. A hypothesis is proposed that suggests that the cortical hand map normally consists of a number of representations that are capable of developing and surviving somewhat autonomously of each other. The features of spatial patterning in the mosaiclike map of these representations are influenced by postnatal availability of inputs from intact hand nerves.

Changes in the cortical map of the hand following postnatal ulnar and radial nerve injury in monkeys: organization and modification of nerve dominance aggregates.

The ulnar and radial nerves to the hands of 12-31-d-old marmoset monkeys were transected and ligated, and the monkeys were subsequently reared for periods of 1.4-1.6 years with only median nerve innervation to the hand. Features of organization in the cortical area 3b hand map were then assessed with neurophysiological mapping procedures, and compared to features in monkeys that had undergone either a normal postnatal development with three intact hand nerves, or an abnormal development with two intact nerves due to postnatal injury of the median nerve. A systematic comparison of cortical organization in these monkeys led to three main findings. First, some features of organization show little or no change when monkeys are reared with one, two, or three hand nerves. These features include receptive field size and the overall size of the hand map. Second, other features are, in contrast, clearly altered in an injury-dependent manner. These features include cortical neuronal thresholds to light tactile stimuli, and the spatial location, size, shape, continuity, and somatotopic interfacing of representations of the parts of the hand. Finally, estimates of the peripheral innervation territories of the hand nerves, and of the corresponding distributions of cortical neurons activated by inputs from these territories, indicate that the normal hand map contains bandlike aggregates of neurons that are dominantly activated by inputs from each nerve. Postnatal nerve injuries alter the size of these nerve dominance aggregates.

Current hypotheses of structural pattern formation in the somatosensory system and their potential relevance to humans.

Current hypotheses of structural pattern formation in the mammalian somatosensory system are modeled on experimental findings from the trigeminal system of rodents. The present results show that, like rodents, the trigeminal nucleus principalis of humans contains a parcellated pattern of cytochrome oxidase dense patches. These results provide an indication of the potential usefulness of rodent-based hypotheses for understanding pattern formation in human somatosensory connections.

Parcellated organization in the trigeminal and dorsal column nuclei of primates.

The subdivisions of the brainstem trigeminal complex in non-primate mammals are characterized by aggregated or parcellated patterns of neural organization. The present studies used cytochrome oxidase histochemistry to test if parcellated organization patterns also occur in the brainstems of primates. The results demonstrate that a parcellated pattern of neural organization exists in the trigeminal nucleus principalis, but not in the spinal trigeminal nuclei, of macaque and squirrel monkeys. The results further suggest that parcellation in the nucleus principalis qualitatively resembles the aggregated organization in dorsal column nuclei. Taken together with previous findings from non-primates, these results indicate that central parcellation is an organizational feature of specific ascending somatosensory projections in many mammals including primates.

Central projections from the skin of the hand in squirrel monkeys.

Central termination patterns of afferents from the hands of squirrel monkeys were studied after subdermal injections of wheat germ agglutinin conjugated with horseradish peroxidase (WGA-HRP) or cholera toxin subunit B conjugated to HRP (BHRP). WGA-HRP more effectively labeled axons terminating in the superficial dorsal horn of the spinal cord, while BHRP more effectively labeled axons terminating in the deeper layers. Injections of both tracers, when restricted to parts of glabrous digits, palm, or dorsal hand, revealed somatotopic patterns in the spinal cord and pars rotunda of the cuneate nucleus that were, in some respects, similar and, in other respects, quite different from those previously reported for macaque monkey (Florence et al., J. Comp. Neurol. 286:48-70, '89). As in macaques, injections in digits 1-5 produced a rostrocaudal sequence of foci of terminations in the cervical spinal cord. However, inputs from the palm were located medial to those from the digits, whereas the palm is represented lateral to the digits in macaque monkeys. Since inputs from the palm is also medial in the dorsal horn in cats (Nyberg and Blomqvist, J. Comp. Neurol. 242:28-39, '85), the condition in squirrel monkeys may be similar to the generalized state. In the cuneate nucleus, single injections in the hand produced dense label in the pars rotunda, and sparse label in the rostral and caudal poles. As in macaque monkeys, inputs from specific parts of the hand related to rostrocaudal clusters of cells that are cytochrome oxidase dense. The representation of the digits differed from macaques in that the digits were represented dorsal to the palm, rather that ventral to the palm as in macaques. Again, comparisons with cats suggest that squirrel monkeys have the more generalized pattern. Finally, inputs from the hair, dorsal surfaces of the digits terminated on the same clusters as the inputs from the glabrous, ventral surfaces, apparently overlapping somewhat. The proximity of these terminations from dorsal and ventral surfaces of the digits may be related to observations that cortical representations of the glabrous surfaces of digits become responsive to dorsal surfaces of the same digits when inputs from glabrous skin are chronically deactivated (e.g., Merzenich et al., Neuroscience 3:33-55, '83).

Temporal progression of cortical reorganization following nerve injury.

Damage to peripheral nerves of adult mammals causes reorganization of somatosensory maps in the cerebral cortex. An understanding of the temporal progression of cortical changes is important for understanding the underlying mechanisms. The present experiments utilized neurophysiological recordings to analyze the time course of reorganization in the S-I cortical hindpaw area in adult rats. Following loss of sciatic inputs, the cortical area responding to low threshold inputs from the hindpaw saphenous nerve expands. A brief, early onset period of rapid expansion is followed by a prolonged period of slow increase. The temporal progression suggests that early onset changes condition the central nervous system for later changes.

Somatotopic organization of inputs from the hand to the spinal gray and cuneate nucleus of monkeys with observations on the cuneate nucleus of humans.

Central termination patterns of primary afferents from the hand and forelimb were studied following subdermal injections of HRP conjugates in macaque monkeys. In the middle layers of the dorsal horn of the spinal cord, afferents from digits 1-5 terminated in a rostrocaudal sequence in separate, elongated columns at cervical levels 5-7. Afferents from the glabrous digits extended to the medial margin of the dorsal gray, while afferents from the dorsal skin of the digits terminated more laterally. Afferents from the dorsal hand and palm terminated lateral to those from the digits, while inputs from the forearm occupied tissue rostral and caudal to the representation of the hand. In the cuneate nucleus, terminations from each digit formed an elongated column that was densely labelled in the central pars rotunda and sparsely labelled in both the rostral and caudal reticular poles. Within the pars rotunda, digits 1-5 were represented in order from lateral to medial. Inputs from the digit tips terminated ventral to inputs from the proximal digits. Afferents from the dorsal skin of the digits terminated in an even more dorsal position, while the most dorsal portion of the pars rotunda related to the glabrous and dorsal hand. Within the pars rotunda, terminations from specific parts of the hand overlapped parcellated clusters of neurons. These clusters were densely reactive for cytochrome oxidase (CO) and were surrounded by myelinated fibers. Much sparser label in the reticular poles was found consistently only after injections in the glabrous digits. Inputs to the poles appeared diffuse and overlapping while preserving some somatotopic order. When treated for CO or stained for Nissl substance or myelin, the pars rotunda of humans showed parcellation patterns that closely resembled the patterns seen in monkeys. From the relationship of inputs to the CO dense cell clusters in monkeys, it was possible to postulate in detail the somatotopic organization of inputs to pars rotunda of humans. The present results provide a comprehensive description of the somatotopic patterns of termination of afferents from the skin of the hand and forearm in the spinal cord and cuneate nucleus of macaque monkeys. A direct relationship of afferent somatotopy and identifiable cell clusters in the pars rotunda of the cuneate nucleus is further demonstrated. Finally, the patterns of cell clusters in the pars rotunda of macaque monkeys and humans suggest that the somatotopic organization of the cuneate nucleus may be very similar in human and nonhuman primates.

Somatotopic organization of the lateral sulcus of owl monkeys: area 3b, S-II, and a ventral somatosensory area.

Multiunit microelectrode recordings and injections of horseradish peroxidase (HRP) were used to reveal neuron response properties, somatotopic organization, and interconnections of somatosensory cortex in the lateral sulcus (sylvian fissure) of New World owl monkeys. There were a number of main findings. 1) Representations of the face and head in areas 3b, 1, and S-II are found on the upper bank of the lateral sulcus. Most of the mouth and lip representations of area 3b were found in a rostral extension along the lip of the lateral sulcus. Adjacent cortex deeper in the lateral sulcus represented the nose, eye, ear, and scalp. 2) S-II was located on the upper bank of the lateral sulcus and extended past the fundus onto the deepest part of the lower bank. The face was represented most superficially in the sulcus, with the hand, foot, and trunk located in a rostrocaudal sequence deeper in the sulcus. The orientation of S-II is "erect," with the limbs pointing away from area 3b. 3) Neurons in S-II were activated by light tactile stimulation of the contralateral body surface. Receptive fields were several times larger than for area 3b neurons. 4) A 1-2-mm strip of cortex separating the face and hand representations in S-II was consistently responsive to the stimulation of deep receptors but was unresponsive to light cutaneous stimulation. 5) Injections of horseradish peroxidase in the electrophysiologically identified hand or foot representations of area 3b revealed somatotopically matched interconnections with mapped hand and foot representations in S-II. 6) A systematic representation of the body, termed the "ventral somatic" area, VS, was found extending laterally from S-II on the lower bank of the lateral sulcus. Within VS, the hand and foot were represented deep in the sulcus along the hand and foot regions of S-II, and the face was lateral near the ventral lip of the sulcus. 7) Neurons at most recording sites in the VS region were activated by contralateral cutaneous stimuli. However, a few sites had neurons with bilateral receptive fields. Receptive field sizes were comparable to those in S-II. In addition, neurons in islands of cortex in the VS region had properties that suggested that they were activated by pacinian receptors, while other regions were difficult to activate by light tactile stimuli but responded to stimuli that would activate deep receptors. 8) A few recording sites caudal to S-II on the upper bank of the lateral sulcus were responsive to somatic stimuli.(ABSTRACT TRUNCATED AT 400 WORDS)