THE JOURNAL OF COMPARATIVE NEUROLOGY 313~151-161(1991)

Functional Regions Within the Map of a Single Digit in Raccoon Primary Somatosensory Cortex D.D. RASMUSSON, H.H. WEBSTER, R.W. DYKES, AND D. BIESOLD’ Department of Physiology and Biophysics, Dalhousie University, Halifax, Nova Scotia, Canada B3H 4H7 (D.D.R.), Department de Physiologie, Universite de Montreal, Montreal, Quebec, Canada H3C 357 (H.H.W., R.W.D.), and Paul-Flechsig-Institut fur Hirnforschung, University of Leipzig 0-7039, Leipzig, Germany (D.B.)

ABSTRACT Electrophysiological recordings were made at a large number of sites in the primary somatosensory cortex of six anesthetized raccoons. A high density of penetrations (110-229 per animal), within or near the representation of the fourth digit, allowed identification of three cortical regions with different physiological properties: a glabrous zone, containing a highly detailed, somatotopically ordered representation of the glabrous surface of the digit; rostral to this a claw-dominant zone, in which the neurons at most penetrations respond to stimulation of the claw of the fourth digit, but may also receive input from the hairy skin or surrounding glabrous skin; and a more rostral multidigit zone, in which the neurons respond to stimulation of two to five digits, with the dominant digit usually being the one represented caudally (i.e., the fourth digit at most of the sites sampled here). Claw-dominant zones with receptive fields restricted to digit three or five are also found rostral to the representations of the glabrous skin of the corresponding digit. The glabrous and claw-dominant zones constitute a complete map of the fourth digit. The multidigit region presumably is a separate map, since its neurons have different spatial convergence, higher thresholds, and a lower incidence of slowly adapting inputs than those in the claw-dominant and glabrous zones. A comparison between animals with lesions of the basal forebrain and intact animals found no differences in the organization of these zones or in the responses to peripheral input, suggesting that cholinergic inputs to the cortex are not essential to these properties. The detailed description of these regions and the proposed terminology should resolve some inconsistencies in the use of the term “heterogeneous zone” in this species. Key words: cortical organization,cutaneous, claw, spatial convergence

Since the pioneering study of Welker and Seidenstein (’591, the North American raccoon (Procyon lotor) has been used in a variety of studies on the processing of somatosensory information (e.g., Welker and Johnson, ’65; Johnson et al., ’68; Pubols et al., ’71) and on somatosensory cortical organization (Johnson et al., ’82). This interest is due to the enlargement of primary somatosensory cortex and the extensive glabrous skin on their forepaws. This enlargement has also proved useful in the study of reorganization of the somatotopic map that occurs when peripheral input is disrupted (Kelahan et al., ’81;Rasmusson, ’82). One persisting difficulty with the raccoon as a model is the extent to which its somatosensory cortex is different from more commonly used mammals, such as cats or monkeys. Efforts to classify raccoon somatosensory cortex histologically on the basis of cytoarchitectonic criteria (i.e., areas 3a, 3b, 1, and 2) have not been successful. This is

o 1991 WILEY-LISS, INC.

partly due to the complex infoldings of the numerous sulci in the forepaw representation. It has also been argued that subprimates do not have all of these regions, but only an “SI proper” corresponding to area 3b or sensory “koniocortex” (Kaas, ’83). Kaas suggested five criteria that would support the definition of separate cortical regions: cytoarchitectonic differences, a complete representation of the body surface in each region, unique physiological properties at the single-cell level, unique anatomical connections, and different behavioral impairments after selective lesions. In the raccoon, a rostral kinesthetic region, responding to inputs from muscle fierents and possessing distinct connections with the thalamus, has fulfilled many of these criteria Accepted August 7,1991. ‘Deceased, May 29,1991.

152

D.D. RASMUSSON ET AL.

(Feldman and Johnson, ’88; Gugino et al., ’90). Another region of raccoon primary somatosensory cortex, located caudal to this kinesthetic region and rostral to the representation of the glabrous skin of the digits, has been suggested as being distinct on physiological grounds (Johnson et al., ’82). This region, called the “heterogeneous zone,” was defined as having inputs from multiple modalities (skin, claw, and deep inputs) and/or from more than one digit. A more recent study (Feldman and Johnson, ’88) restricted the term “heterogeneous zone” to the region that receives inputs from several digits. The “heterogeneous zone” has been proposed to play an important role in cortical reorganization (Doetsch et al., ’88a,b). However, evaluation of this hypothesis requires a clarification of how the “heterogeneous zone” is defined. The purpose of the present experiment is to provide a detailed map of the representation of the fourth digit in raccoon somatosensory cortex. It has become apparent in recent years that, due to interanimal variability, the details of a map of primary somatosensory cortex can best be appreciated if a very large number of penetrations are made in each animal (Paul et al., ’72; Merzenich et al., ’83, ’87). Therefore we have recorded from several hundred penetrations in each animal. The resulting map will provide a framework for future studies on cortical reorganization following deafferentation.

MATERIALS AND METHODS This study presents data from six adult raccoons that were trapped from the wild by a licensed trapper. Three were intact control animals and three had received extensive lesions of the basal forebrain (BF) 14-17 days earlier. These animals served as controls for a study of the effect of BF lesions on cortical reorganization following digit amputation. However, since the BF lesioned group did not differ from the intact animals in terms of any of the data presented in this paper, the lesioning technique and the extent of cholinergic denervation will not be presented here.

Mapping procedure The animal was given 1-2 ml ketamine (100 mg/ml, i.m.) so that an intravenous catheter could be introduced. Anesthesia was maintained with a-chloralose (5% in propylene glycol; 2 ml initially and 0.5 ml supplementary doses). The animal was placed in a stereotaxic apparatus and a large craniotomy was made over the somatosensory cortex. The cortex was mapped according to standard electrophysiological procedures (cf. Turnbull and Rasmusson, ’90). Multiunit activity was recorded with carbon-fiber electrodes (Armstrong-Jamesand Millar, ’79). The activity was monitored with an audio amplifier and speaker and an oscilloscope. Penetrations were oriented approximately perpendicular to the surface of the brain and the electrode was lowered to the middle depths of the cortex ( 800 pm), where evoked activity was strongest. The skin of the forepaw was searched with glass probes or brushes to determine the region that activated the cells at the recording site, The threshold was classified as low or high (cf. Rasmusson ’82). The receptive field was drawn on schematic outlines of the forepaw or digits and various characteristics of the response were noted. If the cells maintained an increased response throughout several seconds of sustained indentation, the site was termed slowly adapting (SA). Those that did not were classified either as rapidly adapting (RA) or as tap, the

-

latter having much higher thresholds for activation than RA sites. A second categorization was the probable location of receptors responsible for the response. Neurons with glabrous skin inputs were for the most part easily detectable because of their low thresholds and small, discrete receptive fields. Those sites with claw inputs were defined by their response to light taps on the claw or to flexion or extension of the claw. In most cases these neurons were also responsive to light touch at the base of the claw on either the glabrous or hairy surface. Hairy skin inputs were identified by flicking the hairs or by light touch of the dorsal surface of the digits or the forepaw. Finally, sites with “deep” inputs were defined as those with cells responding to pressure in the palm or flexion or extension at the wrist. While such sites could often be activated by tapping the tips of the digits, they were also activated by tapping the muscles or tendons in the paw or forearm. In the present study we concentrated on the fourth digit (D4) region of somatosensory cortex because it is relatively flat and easily accessible. Penetrations in the surrounding digit and palm regions were necessary to delimit the borders of the D4 region and the collection of data from these areas supplemented our conclusions about the rules of organization in raccoon somatosensory cortex. Penetrations were performed in rows aligned either mediolaterally or anteroposteriorly and spaced at intervals of less than 0.5 mm except when necessary to avoid blood vessels or sulci. The resulting sampling density ranged from approximately 5 to 8 penetrations/mm2.

RESULTS The region of the primary somatosensory cortex that was studied in these experiments is illustrated in a drawing of the left hemisphere of the raccoon brain (Fig. 1). A total of 1,094 sites were sampled, 474 in the three intact animals and 620 in the three animals with BF lesions (Table 1). Since there were no obvious differences between the control and the BF-lesioned animals, either in terms of somatotopic organization or adaptation properties, they will be presented together. In eight penetrations, primarily near sulci, the neurons could not be driven from the periphery. In 48 penetrations the neurons received input from deep receptors; these were consistently found in the most rostral parts of the sampled region and are identified by “D” in the following figures. However, since they were not studied extensively, and our findings were consistent with those of Feldman and Johnson (’88), they will not be discussed further.

Description of three distinct zones Glabrous zone. The predominant input to raccoon primary somatosensory cortex is from the glabrous skin of the forepaw. The extent of the glabrous representation in each animal is illustrated in Figure 2, in which the location of each penetration with a glabrous receptive field is indicated by a number for the corresponding digit or “P” for the palm. Penetrations whose neurons did not have receptive fields restricted to glabrous skin are indicated by a dot in this figure to illustrate the overall sampling density. Although the general pattern of the forepaw representation is similar in each case, there was considerable variability among animals, in particular with regard to the precise shape of the triradiate sulcus (Tr), the number and length

REGIONS WITHIN RACCOON DIGIT CORTEX

153

had receptive fields near the tip of the digit. It was difficult to localize these receptive fields due to their very low thresholds and to the mechanical problems introduced by the claw. Claw-dominant zone. Rostral to the glabrous zone of D4 most of the recording sites were responsive to stimulation of the claw of D4, either alone or in combination with glabrous or hairy skin. The locations of different combinations of these inputs are illustrated for two animals (Figs. 4, 5). At most sites directly rostral to the glabrous region the receptive fields were located on the claw and the glabrous skin at the base of the claw (Figs. 4A, 5 A “claw glabrous”). Neurons at these sites were extremely sensitive to light taps of the claw and to light pressure at the base of the claw. In many cases, however, the effective glabrous skin extended many millimeters away from the claw so that it was unlikely that the response was due only to movement of the claw. Rostral to the “claw glabrous” sites, many sites were activated by stimulation of the claw and hairy skin (Figs. 4B and 5B). The effective hairs were often the long guard hairs that extend over the claw. Few sites were activated by stimulation of hairy skin alone (Figs. 4C, 5C) or claw alone (Figs. 4C, 5D). Combining all of the penetrations that included inputs from claw and/or hair from a single digit produces the maps shown in Figures 4E and 5E. We have called the resulting area the “claw-dominant zone.” The claw-dominant areas for the other four animals are shown in the right side of Figure 6. The D4 claw-dominant zone is a large, continuous area rostral to the D4 glabrous digit region; however, no mediolateral somatotopy was seen within it. While claw-dominant regions for the third and fifth digits are shown in these figures, they are probably not complete, due to the less extensive sampling in the D3 region and the apparent extension of the D5 claw-dominant region into the walls of the postcruciate sulcus. The similarities in location and extent of the clawdominant region between animals and between digits, and the large variability of subregions as shown in parts A-D of Figures 4 and 5, support the idea that the claw-dominant map is a useful construct. Multidigit zone. In contrast to all of the penetrations in the glabrous and claw-dominant zones, sites that were rostral to (and to some extent between) the claw-dominant digit regions contained neurons that responded to stimulation of more than one digit. The extent of this “multidigit” region is illustrated in Figures 4F and 5F for two animals and on the left side of Figure 6 for the other four. Most of the penetrations in this region rostral to D4 cortex had inputs from two digits and there was a rough topographic organization: cells in the medial penetrations tended to respond to D4 and D5 stimulation, whereas cells in lateral penetrations tended to respond to D4 and D3 stimulation. It was not uncommon, however, for neurons in a penetration to respond to three (25 of 121 sites), four (3 sites), or even all five (4 sites) digits. In 78 of 121 sites there was a clear difference in threshold or strength of response between the digits, with the threshold being lowest on the digit that was represented caudal to the recording site. Nevertheless, in 59 sites the receptive fields on both on- and off-focus digits were judged to have low thresholds. The tendency for higher thresholds for “off-focus” digits in the multidigit zone means that the most likely experimenter error, not detecting a high threshold input from an adjacent digit, would result in that penetration being

+

Fig. 1. Surface diagram of the left hemisphere of the typical raccoon brain, indicating the sulci near the primary somatosensory cortex and (stippled) the region of primary somatosensory cortex studied in the present experiment. A, ansate sulcus; C, coronal sulcus; Cr, cruciate sulcus; PC, postcruciate complex; Tr, triradiate sulcus. Bar = 10 mm.

TABLE 1. Number of Penetrations in the Defined Zones of Each Animal

Animal Intact BFlesion’

1 2 3 4

5 6

Total

Glabrous Claw-dominant Multidigit zone zone zone Deepzone Total 81 118 58 118 147 122

51 35 37 30 41 86

36 23 6 23 18 16

13

5

181 183 110 185 206 229

644

280

122

48

1,094

7 9 14 0

‘BF, basal forebrain.

of branches from Tr, and the exact location of the fifth digit (D5) representation with respect to the medial end of Tr. The largest part of the D4 representation was rostral to Tr, but part of the proximal digit was represented caudal to Tr in some animals (Fig. 2A,C,E). Neurons with receptive fields on proximal digits 2 or 3 were also found caudal to Tr in a few instances. The extent of exposed cortex rostral to T r covered by the glabrous D4 representation ranged from 2 mm (Fig. 2A) to 7 mm (Fig. 2F). Previous experiments suggest that those animals with little D4 representation on the exposed surface have a significant part of the map down the rostral bank of Tr (Turnbull and Rasmusson, ’90).The palm representation was always located caudal to Tr. The boundary between D3 and D4 cortex was a well-defined sulcus in four of the six animals and in all animals the border between D4 and D5 was identifiable from a dimple and a large vein (not shown in this figure; cf. Rasmusson, ’82). The internal somatotopy of D4 cortex is shown for the two animals with the most extensive D4 region on the exposed surface (Fig. 3). Each penetration is labeled by U, M, or R (for ulnar, midline, and radial, respectively) tn indicate the location of the center of the receptive field LSI D4. The region representing the proximal digit is shaded. Despite some difficulty in assigning penetrations to these categories due to the narrowness of the digit (approximately 5 mm), it is clear that an ulnar-to-radial progression of receptive fields is found in moving from medial to lateral over the cortex and that it is a reliable rule of organization. The few inconsistencies in this figure were for the rostralmost penetrations (towards the left in each map), which

+

D.D. RASMUSSON ET AL.

154

A

D

PC

' 5 5 5 5 '55 55

...

DD D

. . .-. - . . . ... ...I .....

DDDD' ~ D .D

" 5

' 5 5 555 5 5 55 55 5 5 5 5 5 5 5 p p p

::..y/ . . . . . . . .-. . -

f

-

D . .

......

' " 4 " " 4 4 4 4 4 4

..

......

. ...

3 3 3

T-

E

A

P

PP P P P P P P P 33 3P PP PPP PP Pp

. . - 4 4 3

- *

PPP P p P

-

PP

. . .33. 3" ;33{/ .. - 3 333

4 4 p p p p p p 4PPPPPppp

p p PPPP

' 3

..

2

2 2 2

r

. . ..5

5-K

"j.

4-4.444 4 5 55 -5 1.44 4 4 4 ':'4'4 442 3 4 4 24 4. 4 44442444442 4t4%4444 : . 4 444 &

.-:- -

bp p

Fbpp

. -.: ;y $444 4$F ;! ....... .: : . - .. .33;333 33 PPP P P . . .. 3 3333333 * * *

........ .." ..... ml m 1 3 3

C

F

Medial Rostra1 Fig. 2. The maps of sites responding to glabrous and deep (kinesthetic) inputs are shown for each animal. A-C Intact animals. D-F Animals with BF lesions. Each symbol represents the location of a single penetration: numbers (2-5) indicate receptive fields on the

glabrous skin of the corresponding digit; P, glabrous skin of the palm; D, deep. Dots indicate penetrations with other inputs. Solid lines indicate sulci using the same abbreviations as in Figure 1.

REGIONS WITHIN RACCOON DIGIT CORTEX

155

8; Fig. 3. Somatotopy within the fourth digit (D4) area in the two animals with the largest D4 representation on the exposed cortex. Heavy lines indicate sulci. Each penetration is indicated by a letter representing the mediolateral quadrant of the fourth digit on which the

receptive field center was located as shown in the inset diagram of the digit: U, ulnar side; M, middle; R, radial side. The shaded portion represents the proximal half of the digit. Calibration bar = 1mm.

included in the claw-dominant zone. Thus, the clawdominant zones shown in Figures 3-5 may be overestimated and the multidigit zones underestimated.

glabrous skin. It has been argued that conclusions about the somatotopic organization of sensory cortex can best be made from a large number of closely spaced penetrations made in a few animals (Merzenich et al., '87). Maps derived in this manner give a finer grain with which to evaluate the rules of organization that are common across individual animals, and also provide a clearer indication of the variability among animals. Using this approach in the present experiments on the raccoon, we identified three zones that differ in terms of their peripheral input, the degree of sensory convergence,and the predominant adaptation characteristics. The glabrous zone of raccoon somatosensory cortex has been studied most extensively and its unusually large surface area accounts for most of the expansion of the somatosensory cortex in this species. Welker and Seidenstein ('59) estimated that the total forepaw cortex of the raccoon is larger in absolute terms than in primates and that 95%of this cortical region is devoted to the glabrous skin. It is characterized by small and discrete receptive fields that are restricted to the glabrous surface of the digit and by a 50430% incidence of penetrations with SA characteristics in the middle layers (Turnbull and Rasmusson, '90). The internal somatotopy of the glabrous zone is highly ordered in both the mediolateral and rostrocaudal dimensions (Welker and Seidenstein, '59; Johnson et al., '82; Kelahan and Doetsch, '84; Turnbull and Rasmusson, '90; and Fig. 3, above). The second region is characterized by neurons having inputs from a single claw and/or the hairy skin on the dorsal surface of the same digit. We have called this region the "claw-dominant zone," but it is important to recognize that this region also receives inputs from hairy and glabrous skin. However, the exact receptor location for the sites that were classified as having a glabrous component is

Response properties in each zone In accord with previous studies (Rasmusson, '82; Turnbull and Rasmusson, 'go), the glabrous region in both the control and BF-lesioned animals had a high proportion of penetrations with SA characteristics in the middle layers (>50%;Fig. 7A). In the claw-dominant zone this proportion was lower (approximately 30%; Fig. 7B) and in the multidigit region much lower (10-15%, Fig. 7C). In both the claw-dominant and multidigit zones, the incidence of RA penetrations and penetrations that required highthreshold taps was greater than in the glabrous zone. A statistical comparison of intact vs. BF-lesioned animals on the basis of the three categories shown in Figure 7 indicated that BF lesions did not alter this property in any region (xz = 1.96, 0.04, and 0.29 for glabrous, claw-dominant,and multidigit regions, respectively; df = 1 and P > 0.05 in each case). Therefore the data from all six animals were combined and the three functional zones were compared in terms of the number of SA vs. non-SA penetrations. The zones were significantly different (x2 = 49.74, df = 2, P -=K 0.01). This analysis indicates that, in addition to differing in terms of submodality inputs and the extent of spatial convergence,the glabrous, claw-dominant,and multidigit zones also differ in terms of the adaptation properties of the constituent neurons.

DISCUSSION This paper presents a detailed description of a digit representation in raccoon somatosensory cortex, and in particular those regions rostra1 to the representation of the

D.D. RASMUSSON ET AL.

156

claw

+

B claw

glabrous

+ hair

2%: ...... 5 .:.:::::.-

$34

- 3

l l

mm

C hair

Fig. 4. Localization of regions rostral to the glabrous zone that responded to different combinations of inputs in an intact animal (same one as Fig. 2A). The extent of the glabrous and deep zones are demarcated by thin lines and labelled with a number (corresponding to digits 2-5) or letter (P, palm; D, deep). A-D Subregions in which the cells responded to stimulation of the claw and glabrous skin (A);claw and hairy skin (B); hairy skin only (C); and claw only (D). E: Entire

“claw-dominant region,” being a summation of the regions shown in

A-D, in which all penetrations had inputs from nonglabrous receptors on a single digit. F: “Multidigit” region in which cells responded to stimulation of more than one digit. The shading is consistent throughout the figure: stipple, D5; diagonal hatching, D4; horizontal hatching, D3.

REGIONS WITHIN RACCOON DIGIT CORTEX

A+

claw

157

claw

E claw

+ hair

Clawdominant

C hair

multidig it

D

Fig. 5. Localization of regions in a BF-lesioned animal (same as Fig. 2D). A-D: The individual subregions. E: The combined claw-dominant region. F The multidigit region. Same conventions as in Figure 4.

difficult to prove, since the glabrous receptive fields were usually at the base of the claw and might be the result of activation of claw receptors via the mechanical coupling between claw and skin.

Sites with receptive fields that include or are restricted to hairy skin were not common and their location was not consistent in different animals. They were therefore included in the claw-dominant region. It is interesting that

D.D. RASMUSSON ET AL.

158

Multidigit

Claw-dominant

A..:

P

P

Fig. 6. Left column: the multidigit region; right column: the claw-dominant regions for the remaining four animals. Same convention as in Figure 4.

REGIONS WITHIN RACCOON DIGIT CORTEX

A

159

related to their common embryological origin, i.e., from the epidermis, in contrast to the separate representation of deep, mesodermally derived structures (muscle and ten8o Glabrous zone dons). The somatotopic organization of the claw-dominant zone was unclear, perhaps partly due to the anatomical 60 arrangement of receptors at the base of the claw. The displacement of this very stiff structure would likely activate receptors on both sides of the claw. YO 40 The third, “multidigit” zone was obviously different from the other two in that it received much wider spatial convergence, responding to stimulation of more than one 20 digit and sometimes of all five. This was apparent for single units within this region as well as in multiunit recordings. The dominant input to a multidigit site was always from 0 the digit that was represented caudal to it, and multidigit SA RA Tap responses were seen rostral to the D3 and D5 representations. Within the multidigit zone in front of D4 cortex there was a rough somatotopy,with D5 being the usual secondary input in medial sites and D3 in lateral sites. The maps of the B multidigit region, however, showed greater variability 8o fz-dominant 0 Intact among animals than the claw-dominant zones in terms of BF lesion both size and location. This may be due to greater difficulty 60 in eliciting responses from the “off-focus’’digit, with their higher thresholds, and to the possibility that the anesthetic may have more of a depressive effect on the higher threshold inputs. YO 40 The differences between areas in the occurrence of SA inputs may be a consequence of the relevant peripheral receptors. The high proportion of SA sites in the glabrous 20 zone (50-80%) is similar to the proportion of SA afferent fibers innervating the glabrous skin of the raccoon forepaw (69%, Turnbull and Rasmusson, ’86). The afferents from 0 the base of the claw are also predominantly SA (81%), SA RA Tap whereas those afferents responsive to hair movement are mostly RA (61%).A convergence of input from these two populations would account for the lower proportion of SA responses in the claw-dominantarea. In the multidigit area, 80 the higher thresholds and lower proportion of penetrations with SA properties may be the result of the central processing in nonlemniscal relay nuclei (Wiener et al., ’87; Doetsch 60 et al., ’88a). Raccoons are judged to be the equal of many primates in terms of fine texture discrimination and tactile acuity 40 (Rensch and Ducker, ’63) and their high manual dexterity YO despite the lack of an opposable thumb (Lohmer, ’75; Nudo and Masterton, ’90). Several authors have commented that, 20 during the exploration of objects, the sense of touch predominates over vision in raccoons (Lyall-Watson, ’63; Lohmer, ’75). These behavioral features correlate well with the enlarged representation of the glabrous skin (Welker and 0 Seidenstein, ’59), but the large claw-dominant region deSA RA scribed here may also be related to their manual dexterity Fig. 7. Percentage of penetrations within the glabrous (A); claw- since they use their claws extensively in the manipulation dominant (B);and multidigit zones (C) that were classified as slowly of locks and latches and in the retrieval of objects (Davis, adapting (SA), rapidly adapting (MI and , tap. The similarity between ’07; Cole, ’12; Lohmer, ’75). The tamarin (Saguinus), intact animals (stipple)and basal forebrain (BF)-lesionanimals (hatch- another species with long, nonretractable claws, exhibits a ing) is apparent in this figure. similar claw-dominatedregion that is located rostral to the glabrous representation (Carlson et al., ’86). As a result of these findings we would suggest that claw receptors, which many of the hair-responsive sites were activated by the dominate in this large cortical region, play a tactile role in large guard hairs projecting over the claw. These guard the raccoon rather than the kinesthetic role that is implied hairs would likely be activated concurrently with the claws by the usual classification of claw receptors as “deep.” The in many situations, a condition which may strengthen high proportion of sites with SA characteristics in both connections in the cortex (Jenkins et al., ’90). The conver- glabrous and claw-dominant zones is also probably related gence of information from claws and hairs may also be to the highly developed manipulative ability, a property

1

]

C

160

shared with primates, which also display large amounts of SA activity in somatosensory cortex (30-60%, Paul et al., ’72;Sur et al., ’80). The lack of a separate region representing hairy skin in the raccoon suggests that these inputs are not very important for the raccoon. It should be noted that the situation is quite different in monkeys (Merzenich et al., ’87), in which the representation of the hairy skin of the hand is consistently larger than in the raccoon even though its location is highly variable. The general caudal-to-rostra1 organization described here, in which the receptive field progression moves up the glabrous side of the digit and then over the tip onto the dorsal surface, has also been described in the galago (Sur et al., ’80).In most primates the representation of the dorsum of the digits is intermingled with that of the glabrous skin (e.g., Paul et al., ’72; Merzenich et al., ’87). A major reason for describing these three regions in such detail is to clarify some confusion that has been entering the literature. When Johnson et al. (’82)first described the region rostral to the main glabrous representation, they identified many units that responded to stimulation of claw and hairy skin and some that responded to input from more than one digit. They called this entire rostral region the “zone of heterogeneous input .” It would, however, include both the multidigit and claw-dominant zones as defined here. More recently, Feldman and Johnson (’88) have used the term “heterogeneous zone” to refer only to the region that receives input from more than one digit, which would correspond to our “multidigit zone.” An example of the confusion that can result is seen in two important anatomical papers that presented data on the thalamic and cortical connections of these areas: Doetsch et al. (’88a,b) identified the “heterogeneous zone” using the early definition (Johnson et al., ’82) and made horseradish peroxidase (HRP) injections into the region that contained claw and other nonglabrous inputs. Although they noted that the heterogeneous sector might be better viewed as several subregions, it is not clear whether their injections included both the claw-dominant and multidigit zones as described here. Thus their conclusions about the unique anatomical connections for the “heterogeneous zone” may apply to the claw-dominant zone, the multidigit zone, or both. In contrast, a recent examination of corticocortical connections using intracortical microstimulation and intracellular recording adopted the more restricted definition of “heterogeneous zone” (Feldman and Johnson, ’88) and studied the projection from the multidigit zone to the glabrous zone (Smits et al., ’91). An underlying question in this and similar studies concerns how different cortical areas are defined and what differences are necessary or sufficient to delimit a separate map of the body. The way in which a map is defined is a function not only of the properties of the source and target spaces and the mapping function between them, but also of the variables that the experimenter actually measures or concentrates on (Dykes and Ruest, ’86). In terms of suggested criteria for defining cortical regions (Kaas, ’83), there does not appear to be any compelling reason to regard the claw-dominant and glabrous zones as separate maps: 1) they complement one another to provide a complete map of the digits; 2) both are within area 3b as defined by cytoarchitectonic criteria (cf. photomicrographs in Johnson et al., ’82); 3) there is inconclusive evidence of different anatomical connections, since the injections made by Doet-

D.D. RASMUSSON ET AL. sch et al. (’88a,b) likely encroached on the multidigit region; and 4)the different physiological properties described here are similar to the properties of the associated peripheral receptors. On the basis of these similarities, we would suggest that the glabrous and claw-dominant zones not be considered as separate maps. In contrast, the multidigit zone is clearly different from the other two areas. 1. It has distinct physiological properties, with greater spatial and submodality convergence, higher thresholds, and lower incidence of slowly adapting characteristics. 2. It is a remapping of the digits. Given the existence of similar properties rostral to the D3 and D5 single-digit areas, there is likely a complete map of the forepaw. Regions yielding similar responses from the proximal forelimb and the hindlimb have also been described, largely in the walls of the postcruciate sulcus (Johnson et al., ’82), so it is possible that the multidigit region is part of a map of the complete body. 3. The distinct anatomical connections described by Doetsch et al. (’88a,b) probably apply to the multidigit region, given the similarities with other studies (e.g., Wiener et al., ’87). Any cytoarchitectural distinctions between these areas will require new techniques, since even the transition from area 3a to 3b is difficult to determine in the raccoon using Nissl staining techniques (Johnson et al., ’82; Feldman and Johnson, ’88;Gugino et al., ’90). The criterion of selective behavioral impairments after cortical lesions has not been examined in the raccoon. The present study in large part confirms the interpretation and summary map of Feldman and Johnson (’88).The additional information about differing response properties and somatotopy in the glabrous, claw-dominant, and multidigit areas will hopefully facilitate further study into the anatomical and physiological organization of raccoon somatosensory cortex. Also, the observation that the properties that are directly related to sensory processing, namely thresholds, adaptation properties, or somatotopic organization, are not altered by BF lesions will be useful in evaluating the role of cholinergic inputs on other cortical functions.

ACKNOWLEDGMENTS This work was supported by grants from the Medical Research Council of Canada to D.D.R. and R.W.D. and from the Health Ministry of the German Democratic Republic to D.B. H.H.W. was supported by an IBROKJNESCO postdoctoral fellowship from the German Democratic Republic. We wish to thank Dr. Ch. Wilke of the Central Animal House in Leipzig for his assistance.

LITERATURE CITED Armstrong-James, M., and J. Millar (1979) Carbon fiber microelectrodes.J. Neurosci. Methods, 1.279-281. Carlson, M., M.F. Huerta, C.G. Cusick, and J.H. Kaas (1986) Studies on the evolution of multiple somatosensory representations in primates: The organization of anterior parietal cortex in the new world callitrichid, Saguinus. J. Comp. Neural. 246:409426. Cole, L.W. (1912) Observations of the senses and instincts of the raccoon. J. h i m . Behav. 2299-309. Davis, H.B. (1907) The raccoon: A study in animal intelligence. Am. J. Psychol. 18:447487.

REGIONS WITHIN RACCOON DIGIT CORTEX Doetsch, G.S., G.P. Standage, K.W. Johnston, and C.-S. Lin (1988a) Thalamic connections of two functional subdivisions of the somatosensory forepaw cerebral cortex of the raccoon. J. Neurosci., 8~1873-1886. Doetsch, G.S., G.P. Standage, K.W. Johnston, and C.-S. Lin (1988b) Intracortical connections of two functional subdivisions of the somatosensory forepaw cerebral cortex of the raccoon. J. Neurosci., 8r1887-1900. Dykes, R.W., and A. Ruest (1986) What makes a map in somatosensory cortex? I n E.G. Jones and A. Peters (eds.): Cerebral Cortex, Vol. 5, Sensory-Motor Areas and Aspects of Cortical Connectivity. New York: Plenum, pp. 1-29. Feldman, S.H., and J.I. Johnson (1988) Kinesthetic cortical area anterior to primary somatic sensory cortex in the raccoon (Procyon lotor). J. Comp. Neurol. 277330-95. Gugino, L.D., M.J. Rowinski, and S.D. Stoney (1990) Raccoon forelimb motorsensory cortex: I. Somatic afferent inputs to different cytoarchitectonic areas. Brain Res. Bull. 24~819-825. Jenkins, W.M., M.M. Merzenich, M.T. Ochs, T. Allard, and E. Guic-Robles (1990) Functional reorganization of primary somatosensory cortex in adult owl monkeys after behaviorally controlled tactile stimulation. J. Neurophysiol. 63%-104. Johnson, J.I., E.-M. Ostapoff, and S. Warach (1982) The anterior border zones of primary somatic sensory (S,) neocortex and their relation to cerebral convolutions, shown by micromapping of peripheral projections to the region of the fourth forepaw digit representation in raccoons. Neuroscience 7:915-936. Johnson, J.I., W.I. Welker, and B.H. Pubols (1968) Somatotopic organization of raccoon dorsal column nuclei. J. Comp. Neurol. 132:l-44. Kaas, J.H. (1983) What, if anything, is SI? Organization of first somatosensory area of cortex. Physiol. Rev. 63206-231. Kelahan,A.M., R.H. Ray, L.V. Carson, C.E. Massey, andG.S. Doetsch (1981) Functional reorganization of adult raccoon somatosensory cerebral cortex followingneonatal digit amputation. Brain Res. 223:152-159. Kelahan, A.M., and G.S. Doetsch (1984) Time-dependent changes in the functional organization of somatosensory cerebral cortex following digit amputation in adult raccoons. Somatosens. Res. 2:49-81. Lohmer, R. (1975) Vergleichende Untersuchungen iiber die Manipulier- und Lernflihigkeiten von Waschbaren (Procyon lotor) und Krabbenwaschb a e n (Procyon cancriuorus cancriuorus). 2. Saugetier. 40:36-49. Lyall-Watson, M. (1963) A critical re-examination of food “washing” behaviour in the raccoon (Procyon lotor linn.). Proc. 2001. SOC.Lond. 141:371395. Merzenich, M.M., J.H. Kaas, J. Wall, R.J. Nelson, M. Sur, and D. Felleman

161 (1983) Topographic reorganization of somatosensory cortical areas 3b and 1 in adult monkeys following restricted deaerentation. Neuroscience 8:33-55. Merzenich, M.M., R.J. Nelson, J.H. Kaas, M.P. Stryker, W.M. Jenkins, J.M. Zook, M.S. Cynader, and A. Schoppmann (1987) Variability in hand surface representations in areas 3b and 1 in adult owl and squirrel monkeys. J. Comp. Neurol. 258:281-296. Nudo, R.J., and R.B. Masterton (1990) Descending pathways to the spinal cord, IV Some factors related to the amount of cortex devoted to the corticospinal tract. J. Comp. Neurol. 296t58P597. Paul, R.L., M. Merzenich, and H. Goodman (1972) Representation of slowly and rapidly adapting cutaneous mechanoreceptors of the hand in Brodmann’s areas 3 and 1 ofMucaca mulatta. Brain Res. 36:229-249. Pubols, L.M., B.H. Pubols, and B.L. Munger (1971) Functional properties of mechanoreceptors in glabrous skin of the raccoon’s forepaw. Exp. Neurol. 31:165-182. Rasmusson, D.D. (1982) Reorganization of raccoon somatosensory cortex followingremoval of the fifth digit. J. Comp. Neurol. 205313-326. Rensch, B., and G. Ducker (1963) Haptisches Lern- und Unterscheidungsvermogen bei einem Waschbliren. Z. Tierpsychol. 20:608-615. Smits, E., D.C. Gordon, S. Witte, D.D. Rasmusson, and P. Zarzecki (1991) Synaptic potentials evoked by convergent somatosensory and corticocortical inputs in raccoon somatosensory cortex: Substrates for plasticity. J. Neurophysiol., 66:688-695. Sur, M., R.J. Nelson, and J.H. Kaas (1980) Representation of the body surface in somatic koniocortex in the prosimian galago. J. Comp. Neurol. 189:381-402. Turnbull, B.G., and D.D. Rasmusson (1986) Sensory innervation of the raccoon forepaw: 1. Receptor types in glabrous and hairy skin and deep tissue. Somatosens. Res. 4:43-62. Turnbull, B.G., and D.D. Rasmusson (1990) Acute effects of total or partial digit denervation on raccoon somatosensory cortex. Somatosens. Motor Res. 7:365-389. Welker, W.I., and J.I. Johnson (1965) Correlation between nuclear morphology and somatotopic organization in ventro-basal complex of the raccoon’s thalamus. J. Anat. 99~761-790. Welker, W.I., and S. Seidenstein (1959) Somatic sensory representation in the cerebral cortex of the raccoon (Procyon lotor). J. Comp. Neurol. 111t469-501. Wiener, S.I., J.I. Johnson, and E.-M. Ostapoff (1987) Organization of postcranial kinesthetic projections to the ventrobasal thalamus in raccoons. J. Comp. Neurol. 258:49&508.