Figure 132

FIGURE 13.3 Schematic illustrating the potential for and limitations on receptive field reorganization based on divergent afferent inputs. To the far left is a cartoon of the glabrous hand showing the region on the distal phalange of D2 that is magnified in the panels to the right. Sensory stimuli applied to this skin region (illustrated by series of small circles within the shaded skin region) activate central neurons in a manner that produces partially overlapping receptive fields (shown as larger circles to the top). Because of the divergence of the inputs, stimulation to the skin site indicated by a plus (+) produces excitatory potentials (indicated by a + sign in the receptive fields) over a larger extent than would be predicted if there was a simple one-to-one relationship between skin and the central targets. Thus even if sensory inputs are deafferented, indicated by large X in the panel to the right, central neurons with alternate sources of excitatory inputs will not be functionally silenced. However, if the denervation eliminates all sensory inputs, as for the central neuron in the panel on the right, the neuron will be deactivated permanently or until new inputs are formed.

FIGURE 13.3 Schematic illustrating the potential for and limitations on receptive field reorganization based on divergent afferent inputs. To the far left is a cartoon of the glabrous hand showing the region on the distal phalange of D2 that is magnified in the panels to the right. Sensory stimuli applied to this skin region (illustrated by series of small circles within the shaded skin region) activate central neurons in a manner that produces partially overlapping receptive fields (shown as larger circles to the top). Because of the divergence of the inputs, stimulation to the skin site indicated by a plus (+) produces excitatory potentials (indicated by a + sign in the receptive fields) over a larger extent than would be predicted if there was a simple one-to-one relationship between skin and the central targets. Thus even if sensory inputs are deafferented, indicated by large X in the panel to the right, central neurons with alternate sources of excitatory inputs will not be functionally silenced. However, if the denervation eliminates all sensory inputs, as for the central neuron in the panel on the right, the neuron will be deactivated permanently or until new inputs are formed.

cortical representation is reactivated by inputs predominantly from the wrist and forearm (Figure 13.2D, see also Reference 26). In this series of examples, the pecking order for reactivation in area 3b is dictated by the extent of the denervation, so that as progressively more and more inputs are removed, the next nearest neighbor in the map provides the major source of new activation.

A common explanation for the dominance of near neighbors as alternate influences on deprived neurons in area 3b is that the thalamocortical afferent terminals are divergent so there is topographic overlap between adjacent sensory representations (Figure 13.3, see also References 27 and 28). Under normal circumstances, the overlapping inputs are not apparent, and only one sensory representation is dominant. However, if the dominant source of activation is deprived, the latent inputs gain strength and become the new governing influence, presumably through disinhibition of the latent inputs (see below).

Some of the topographic hierarchies that emerge in cortex after peripheral denervation cannot be accounted for simply by topographic overlap among the thalamocortical afferents. In squirrel monkeys, if all of the glabrous surface of the hand is deactivated, through transection of both the median and ulnar nerves, the dorsal hand representation takes over the large zone of affected neurons in area 3b (Figure 13.4; see also Reference 17). The dominance by dorsal hand inputs is surprising because the dorsal hand has a minor presence in normal hand maps in primates. For example, less than 10% of the map in normal squirrel monkeys is devoted to the representation of the dorsal hand.29 Thus, a takeover of the full extent of the hand representation by inputs from the dorsal hand would constitute a manifold enlargement in the dorsal hand representation. If such an expansion was subserved

FIGURE 13.4 Patterns of topographic organization in somatosensory cortical area 3b after transection of nerves to the hand in owl or squirrel monkeys. A. The normal somatotopy of the hand representation. The representations of the hairy, dorsal skin of the digits and hand typically lie lateral and medial to the digit representations as shown. B. After transection of both the median and ulnar nerves to denervate the glabrous representations of the digits and palm, the full extent of the deprived zone is reactivated by remaining inputs from the hairy skin. Adapted from Reference 17. C. After transection of the median and radial nerves to eliminate both the glabrous and dorsal inputs from the lateral hand, the deprived cortical zone is only partially reactivated and a large non-responsive region (black) persists (from Reference 57). The changes in cortical map organization have been overlaid on a summary drawing of the structural isomorph of the hand (see Reference 24) so that the extent of change can be evaluated in the context of the normal hand representation. M, medial; R, rostral.

FIGURE 13.4 Patterns of topographic organization in somatosensory cortical area 3b after transection of nerves to the hand in owl or squirrel monkeys. A. The normal somatotopy of the hand representation. The representations of the hairy, dorsal skin of the digits and hand typically lie lateral and medial to the digit representations as shown. B. After transection of both the median and ulnar nerves to denervate the glabrous representations of the digits and palm, the full extent of the deprived zone is reactivated by remaining inputs from the hairy skin. Adapted from Reference 17. C. After transection of the median and radial nerves to eliminate both the glabrous and dorsal inputs from the lateral hand, the deprived cortical zone is only partially reactivated and a large non-responsive region (black) persists (from Reference 57). The changes in cortical map organization have been overlaid on a summary drawing of the structural isomorph of the hand (see Reference 24) so that the extent of change can be evaluated in the context of the normal hand representation. M, medial; R, rostral.

by divergent thalamocortical afferents, it would require that afferents relaying information about the dorsal hand project to all neurons in the hand representation. Since there are considerably fewer neurons in the dorsal hand representation of VP compared to those that relay information from the glabrous hand, each afferent carrying dorsal hand information would need to span large distances in area 3b to cover the full extent of the hand representation. However, no evidence of such expansive afferents have been found at the single axon level,30-31 except for one report of a zone of terminal labeling in somatosensory cortex of a macaque monkey that spanned

8-9 mm mediolaterally across the face representation after a bulk injection in VPM.28 Since bulk injections can label projections from multiple neurons, the widespread label does not support the contention that individual arbors have such massive divergence. Moreover, no other studies have obtained similar results.30-32

The alternative and more plausible explanation for the dramatic enlargement of the dorsal hand representation after denervation of glabrous hand is that reactivation takes place at earlier stations in the afferent relay. Xu and Wall33-34 have found that neurons in the glabrous hand representation of the cuneate nucleus acquire new receptive fields on the dorsal skin immediately after median nerve transection. The rapid reactivation likely results from the near-neighbor arrangement of afferent terminals from the skin of the hand in the cuneate nucleus; inputs from the dorsal hand are directly adjacent to those from the corresponding part of the glabrous hand.7'8 Afferents from glabrous D1 terminate adjacent to, and perhaps partially overlap afferents from, dorsal D1 in the cuneate nucleus of monkeys. Thus, because of the convergence of inputs from the two surfaces of the hand, cuneate neurons probably have the potential to be driven by either dorsal or glabrous inputs. Presumably, the glabrous inputs dominate normally; however, when the glabrous inputs are placed at a competitive disadvantage, such as occurs after median nerve transec-tion, the afferents from the dorsal hand become potent enough to activate the target neurons. This shift in functional efficacy of the dorsal hand inputs would yield a new pattern of representation in the cuneate nucleus, consisting of a dorsal hand takeover of the entire hand representation. The relay neurons would transfer the new pattern of representation to VP, and after local adjustments that may occur in VP, the dorsally dominant hand map would be relayed to area 3b.

13.2.3 Other Inputs

Other types of reorganization are not readily explained by existing connections, at any level of the pathway. The most dramatic example is that the face representation takes over hand cortex when all inputs from the forelimb are denervated. Such large-scale expansion was first shown by Pons and colleagues.35 Somatosensory cortex was mapped in macaque monkeys that had long-standing loss of all sensory inputs from the upper extremity as a result of dorsal rhizotomy at the cervical level. Throughout the large extent of cortex where the forelimb representation normally is situated, neurons had acquired new receptive fields. The vast majority of receptive fields were on the chin of the face. Thus, the reorganized map consisted of a greatly expanded representation of the chin. More evidence of the capacity of face inputs to reactivate hand cortex came from studies in monkeys that suffered injuries to the forelimb that ultimately resulted in surgical amputation of some portion of the limb.25,26 The level of the amputations varied considerably from animal to animal, so that in some monkeys the amputation occurred distally, below the elbow, and in others, nearly the entire arm had to be removed. If only the hand was amputated, much of the deprived cortex was taken over by wrist and forearm inputs (Figure 13.2D); however, there may have been some expansion of the face at the lateral border of the deprived zone. At the other extreme, if most of the arm was amputated, much of the denervated zone was reoccupied by an expanded representation of the face (Figure 13.2E). A similar takeover of forelimb cortex by the face also has been reported in monkeys that had dorsal column transection.5

What is the mechanism for face expansion? Such large-scale takeover by face inputs cannot be accounted for simply by divergent thalamocortical afferents. In the macaque monkeys that had dorsal rhizotomy, the face representation expanded to take over a mediolateral extent in cortex of 11 millimeters or more.35 The largest thalamocortical afferent arborization observed in area 3b of a macaque monkey spanned a total extent of about 2.5 mm, and most others were much smaller.31 Thus, the extents of thalamocortical afferents are too limited to accommodate changes that occupy many millimeters of cortical territory. Similarly, there is no explanation for the massive takeover of hand cortex by face inputs, based on existing subcortical connections. The sensory inputs from the face terminate in the trigeminal nucleus, which is lateral to the cuneate nucleus, and there is no overlap among face and hand inputs normally.36 Since there is no evidence of a normally existing, structural framework that could allow for the widespread activation of hand cortex by face stimulation, the alternative explanation had to be considered, that new connections formed somewhere in the pathway so that face inputs had access to neurons normally activated by the hand.

To explain the takeover of forelimb neurons by the face representation, the potential for sprouting of primary sensory afferents from the face into the deprived hand representation of the cuneate nucleus was considered.36 As will be discussed in more detail in a subsequent section, sprouting has been described at multiple levels of the pathway in monkeys after peripheral sensory manipulations. In monkeys that had either forelimb amputation or dorsal column section, bilateral injections into the face to reveal the primary sensory afferent projection patterns produced label contralateral to the limb denervation that was restricted to the trigeminal nucleus. This is the normal projection target for face inputs. However, the labeled face projection extended into the cuneate nucleus on the side ipsilateral to the denervation, and labeled areas that normally only receive inputs from the hand.6-8 The expanded projection to the cuneate nucleus was sparse, but might have had enough functional potency to activate the deprived neurons in the cuneate nucleus. If the new connections were effective, the representation of the face would expand to take over the deprived zone in the cuneate nucleus. Although no studies yet have evaluated the electrophysiological effects of limb denervation at the level of the dorsal column nuclei in monkeys, the data from recent studies of VP of monkeys with long-standing forelimb denervation support that claim that the redistributed peripheral afferents have physiological impact. In VP of monkeys that had arm amputation37 and in monkeys with dorsal rhizotomy,38 the face representation extended beyond the cell-free septa that normally separates the face and hand subnuclei into the deprived hand representation. Presumably, the expanded projection from the trigeminal nucleus to the cuneate nucleus generated a functional map of the face that was relayed to the hand subnucleus of VP. In turn, the new map in VP would be relayed to somatosensory cortex, and the consequence would be widespread reactivation of hand cortex by face inputs.

Another question that is considerably more difficult to address is why inputs from the face are favored in the deprived hand representation of the primary somatosensory relay over other inputs? The gracile nucleus, which receives primary sensory afferents from the lower trunk and hindlimb is just medial to the cuneate nucleus, and just as closely adjacent to the cuneate nucleus as the trigeminal nucleus. In fact, the gracile and cuneate are separated by a less distinctive border than the cuneate and trigeminal nuclei. Thus, sprouting of gracile neurons into the deprived cuneate nucleus would seem to be equally as likely as sprouting of face neurons. Because there is no functional evidence of hindlimb expansion, the possibility of new growth from the gracile nucleus into the cuneate nucleus has not been explored in monkeys. If such sprouting occurs, the inputs presumably have no physiological impact. Indeed, in rats that had forelimb amputation early in life, hindlimb inputs sprout into the cuneate nucleus, but their impact is largely suppressed by GABA-mediated inhibition in S1.39-40 At present, the explanation for the apparent preference for face inputs as new sources of activation after forelimb deafferentation in the relay to area 3b remains a mystery.

There is less extensive information about hierarchies among sensory representations at other levels of the pathway in monkeys, but the limited data suggest that the propensity for takeover by the face representation after forelimb deafferentation is not shared by some higher order cortical areas. Pons and colleagues41 mapped the second somatosensory cortical area, SII, in macaque monkeys that had a lesion throughout the forelimb representation of cortical SI. Six to eight weeks after the lesion, neurons in the deprived forelimb representation responded to stimulation of the foot and lower leg, not the face. In SII of macaques, the foot and leg representation is adjacent to the forelimb on the anterolateral side, so it is not entirely surprising that the foot inputs reactivated the deprived zone. However, other adjacent representations did not expand similarly. For example, the face and head representations are medially adjacent to the forelimb in SII, yet there appeared to be only minimal enlargement of the head representation. This suggests that hindlimb inputs may be the preferred alternate in hand cortex of SII.

0 0

Post a comment