LE JOURNAL CANADIEN DES SCIENCES NEUROLOGIQUES

Parallel Processing of Cutaneous Information in the Somatosensory System of the Cat

ROBERT W. DYKES

RESUME: Dans le passe le principe de In the past, studies of the somatosen­ beginning (Adrian, 1941) although it base des mecanismes corticaux du systeme sory system have played a major role was not reported in man until much sensitif furent fondes en grande partie sur in developing ideas central to under­ later (Penfield and Rasmussen, 1950). le resultat de Vetude du systeme somato- standing cortical mechanisms ap­ Woolsey and Fairman (1946) found sensitif. Les idees de cartes topographiquesplicabl e to all sensory systems. The SII in a number of primates and sub- de colonnes corticales, de specificite ideas of (i) topographic maps, (ii) cor­ primates and hypothesized that the se­ module, ont toutes originees dans Vetude detica l columns, and (iii) modality cond somatosensory map was an ce systeme, pour etre ensuite appliquees aux systemes auditifs et visuels. Re'cem- specificity originated in this sensory evolutionarily more primitive projec­ ment des changements fondamentaux se system and were later applied to the tion of the body surface that had been sont produits dans notre comprehension auditory and visual systems. Now, superceded by a newer projection to SI. des fonctions somatosensitives et corticales, after several decades of relative con­ These classical experiments il­ ces concepts nouveaux s'appliqueront bien-stancy, our ideas about somatosensory lustrated the regularity of the represen­ tdt aux autres systemes. Le present article cortical function have begun to change tation of the body surface on the cor­ detaille ces developpements. rather rapidly and again the newer tical surface. With the exception of the ideas may find useful applications in hand-face transition, the representation other sensory systems. In this article appeared to follow a sequence dictated the concepts found in current textbooks by the dermatomal sequence of spinal are summarized briefly and the many nerves (Bard, 1938). The maps were elegant experiments which expanded shown to be distorted in a character­ and filled out the details of these istic manner so that the body parts with traditional ideas are noted only briefly the greatest innervation density were so that there is space to present some given a much larger proportion of the of the newer ideas about the somato­ cortical surface. In fact, the size of a sensory system. cortical area devoted to a particular body part is thought to be propor­ tional to, and dictated by, the innerva­ THE TRADITIONAL VIEWS tion density of that part (Mountcastle, Somatosensory Maps: The Homun- 1980). culus A map of the body surface spread on the postcentral gyrus was The Cortical Column and Modality published by Penfield and Boldrey in Specificity In 1957, Mountcastle 1937 (Fig. 1) and eventually became recorded from 685 single units in areas known as the homunculus. Woolsey 3a, 3b, 1, and 2 of cat somatosensory and his colleagues mapped many cortex. When electrodes were inserted species in the next 25 years and drew perpendicular to the surface, all corresponding animal shapes over the neurons encountered were of one postcentral gyrus or its homologue in modality and had overlapping receptive each species studied (Woolsey, 1958; fields. When vertically-oriented 1964). No one questioned the validity electrode penetrations were not exactly of this conceptualization. Instead, dis­ perpendicular to the cortex, abrupt cussions focussed on the details of the shifts in the nature of the adequate organization of the maps. For example, stimulus were observed. As the the relationships between the orienta­ electrode advanced, regions driven by tion of the occiput, hand, and face were cutaneous stimuli gave way to regions From the Microsurgical Research Laboratories. Departments of Neurology and Neurosurgery, discussed in great length (Woolsey, driven by stimulation of deep structures Physiology and Surgery. Royal Victoria Hospital, 1958; Penfield and Jasper, 1954). or conversely responses driven by deep McGill University, Montreal, Quebec. The existence of a second somato­ stimulation changed to responses best Reprint Requests to: Dr. R.W. Dykes, driven by cutaneous stimuli during the Microsurgical Laboratories, Royal Victoria Hospital, sensory representation in the cerebral McGiU University, Montreal, Quebec. cortex (SII) was apparent from the electrode trajectory. From this data he

Vol. 9 No. 1 FEBRUARY 1982—9 Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 25 Sep 2021 at 14:00:48, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0317167100043560 THE CANADIAN JOURNAL OF NEUROLOGICAL SCIENCES

Levitt, 1968; Dreyer et al, 1975) con­ firming the existence of functionally homogenous regions separated by boundaries where the function and receptive field change abruptly. Com­ parable functional units were reported in the visual cortex (Hubel and Wiesel, 1962) soon after their description in the somatosensory cortex and equivalent, vertically-arranged functional units were reported in the auditory cortex (Hindetal., 1960). Modality Specificity of Cortical Neurons Included in the columnar hypothesis, although not stated ex­ plicitly, was the principle of modality specificity: That is, despite the multiple opportunities for the convergence of several classes of sensory receptors onto one cortical neuron, the neurons within one vertically-oriented region of primary somatosensory cortex, when activated by natural stimuli, appeared to receive excitatory inputs from only one class of receptors while adjacent areas were best activated by other clas­ ses. The modality classification employed by Mountcastle (1957) in­ cluded hair, pressure, deep, and joint responses. Several examples of modality-specific neurons for each of these classes were presented in the ear­ ly publications from his laboratory but no evidence of neurons activated by more than one class. (Mountcastle, 1957; Powell and Mountcastle, 1959; Mountcastle and Powell, 1959.) Neurons activated principally by one class of afferent input have now been reported at each level of the neuraxis, including submodality-specific neurons activated by cutaneous slowly adapting receptors, cutaneous rapidly adapting receptors, muscle spindle afferents, joint afferents and afferent fibers from pacinian corpuscles (Dykes, 1978). Neurons activated specifically by nox­ ious stimuli, warm stimuli and cold Figure 1 — (Reproduced with permission from Penfield and Boldrey, 1937) The single map of the stimuli have been identified in the body surface called the homunculus extending over the entire primary somatosensory cortex. The map is bilateral because one side is meant to represent the motor cortex and the other the spinal cord and thalamus (cf Willis and sensory cortex. Coggeshall, 1978). So far the cortical projection of modality specific neurons of these latter classes has eluded the formulated an hypothesis stating that the regions of SI caudal to area 3 while microelectrode. the somatosensory cortex contained a the proportion of cutaneous responses mosaic of functional units consisting of were greatest in the center of the These three concepts; the maps, the vertical columns of neurons serving (i) somatosensory cortex. Subsequent to columns, and the modality specificity one point on the body and (ii) one class these experiments, cortical columns of somesthetic neurons are major parts of sensory receptors. The proportion of have been studied by numerous of the foundation upon which our un­ deep columns were said to be greater in workers (Welt et al, 1967; Levitt and derstanding of the functioning of the

Downloaded10 —FEBRUAR from https://www.cambridge.org/coreY 1982 . IP address: 170.106.33.22, on 25 Sep 2021 at 14:00:48, subject to the Cambridge Core terms of use, availableSomatosensory at Processing https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0317167100043560 LE JOURNAL CANADIEN DES SCIENCES NEUROLOGIQUES

TABLE 1 Peripheral afferent nerve fibers in primates and their roles in sensation*

Type Fiber class, size, and Peripheral Differential sensitivity of skin conduction velocity termination and adaptive propertyt Sensory function

Glabrous A-beta,* 6-12/x, Merkel's cells SA mechanoreceptors Velocity and position detectors; 35-75 ni/sec sense of touch-pressure Meissner's cor- QA mechanoreceptors Velocity and perhaps instanta­ puscles neous position; sense of con­ tact and flutter; best at 30-40 Hz Pacinian corpuscles QA mechanoreceptors Velocity and perhaps higher de­ rivatives of position; sense of contact and vibration: best at 250-300 Hz A-delta, 1-5/z, 5-30 Bare nerve endings SA thermoreceptors Sense of cooling m/sec Bare nerve endings SA nociceptors Sense of pricking pain and, at low frequencies, tickle C fibers, 0.2-1.5 zi, Bare nerve endings SA thermoreceptors Sense of warming 0.5-2 m/sec Bare nerve endings SA nociceptors Sense of burning pain and, at low frequencies, itch Hairy A-bcta,* 6-12 ti, Hair follicle ap­ QA mechanoreceptor Velocity detectors; sense of con­ 35-75 m/sec paratus tact and flutter; best at 30-40 Hz Bare nerve endings, QA mechanoreceptor Velocity detectors; sense of con­ "field" recep­ tact tors Pinkus domes, SA "type I" mech­ Stimulation in human elicits Merkel cells anoreceptor no sensation Ruffini organs SA "type II" mech­ Velocity and position detectors; anoreceptor sense of touch-pressure Pacinian cor­ QA mechanoreceptor Velocity and perhaps higher de­ puscles rivatives of position; sense of vibration; best at 250-300 Hz A-delta, 1-5 p., Bare nerve endings QA mechanoreceptor Velocity; sense of contact 5-30 m/sec Bare nerve endings SA thermoreceptor Sense of cooling Bare nerve endings SA nociceptors Sense of pricking pain and, at low frequencies, tickle C fibers, 0.2-1.5 tt, Bare nerve endings SA thermoreceptors Sense of warming 0.5-2 m/sec Bare nerve endings SA nociceptors Sense of burning pain and, at low frequencies, itch Bare nerve endings SA mechanoreceptors Occur rarely in monkey skin and not present in human skin

*In earlier classifications these fibers were labeled A-alpha. They are, however, similar in size to the second group of afferent fibers in muscle nerves. Thus they are termed A-beta fibers here, which allows a correlation between the systems of classification as follows: group I = A-alpha; group II = A-beta; group III = A-delta; and group IV = C fibers. t SA = slowly adapting; QA = quickly adapting. * From Mountcastle, Vernon B: Sensory receptors and neural encoding: introduction to sensory processes. In Mountcastle, Vernon B, editor: Medical Physiology ed. 14, St. Louis, 1980, C.V. Mosby Co.

Dykes FEBRUARY 1982—11 Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 25 Sep 2021 at 14:00:48, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0317167100043560 THE CANADIAN JOURNAL OF NEUROLOGICAL SCIENCES

somatosensory system rests. Presently, favors this specificity theory. There is Anatomical Tracing Techniques The all of these ideas are undergoing a rapid physiological and anatomical evidence study of neuroanatomy has been re­ evolution. The rest of this article is an for specific thermoreceptors, nocicep­ volutionized by the use of horseradish attempt to restate each idea in a tors, joint afferents, pressure and vibra­ peroxidase and radio-labelled amino modern form following a brief com­ tion detectors (for reviews: cf., Dykes, acids to trace neuronal connections. ment about the techniques and dis­ 1977;Mountcastle, 1980). For the somatosensory pathways, these coveries that prepared the way for techniques have allowed functionally As indicated above, during the last these changes. identified regions to be labelled so that two decades, neurons responding selec­ details of their connections to other tively to only one of these specific areas can be defined. Particularly in CONCEPTUAL AND TECHNICAL receptor classes have been located at this context, the origin of the afferent PRECONDITIONS FOR A each level of the somatosensory inputs to a specific region can be ascer­ MODERN PICTURE OF pathway. Thus, in at least some cases, tained with a precision and detail never SOMATOSENSORY CORTEX specific information about modality before possible, thereby allowing in­ Specificity of Sensory Receptors and submodality is encoded by the ferences about the detailed spatial Over the last 20 years, anatomists receptors and is preserved at successive organization of the various inputs to and physiologists have argued about synapses in the central somesthetic those regions so identified. classifications of cutaneous sensory pathways. Mountcastle (1957) has receptors (these arguments are been shown not only to be correct reviewed by Sinclair, 1967). One about modality specificity, but the CURRENT VIEWS school, believing that the pattern of specificity of synaptic input is even Modern Maps - Multiple Homunculi neural activity presented to the central greater than he originally proposed. Turning now to the newer ways to look nervous system was the most impor­ Perhaps it is valid even to the point that at the somatosensory system, a major tant aspect of sensation, argued that for each class of sensory receptors step was taken by Paul et al (1972) sensory receptors are found in a diver­ there is a unique class of afferent who used the low-impedance micro- sity of sizes and shapes and respond to neurons and a particular sensory ex­ electrode technique to discover that the a diversity of stimuli. This theory held perience. Table I summarizes the major postcentral gyrus of primates con­ that different sensory experiences must classes of sensory receptors found in tained two separate and complete maps arise from the different patterns of the skin of man and other primates and of the hand within what was originally neural activity set up in a heterogene­ lists their putative sensory correlates thought of as the single homunculus. ous receptor population. To support (Mountcastle, 1980). This same microelectrode mapping this theory, evidence was marshalled technique was used effectively by Sur suggesting that distinct classes of Electrophysiological Mapping Tech­ et al (1978) to provide extremely high receptors did not exist. Instead, the niques Another prerequisite for our resolution maps of primary somato­ theory said, light tactile stimuli ac­ modern concepts is the existence of sensory cortex of the grey squirrel. tivated only a few receptors, and the new investigative tools that have pro­ Then Kaas, Merzenich and their col­ sensations elicited by thermal stimuli, vided novel information about the leagues, mapped the entire somato­ for example, were differentiated from somatosensory pathways. The classical sensory cortex of several different tactile stimuli by provoking another maps of the cortex were obtained with species of primates showing in each pattern in this same receptor popula­ evoked potentials recorded from sur­ case that there was at least a double re­ tion. Then, with the addition of a few face electrodes placed at successive presentation of the body on the post­ unmyelinated fiber terminals, this new points one mm or more apart. In man central gyrus (Merzenich et al, 1978; pattern allowed noxious stimuli to be and monkeys, generally only three or Kaas et al, 1979; Sur et al, 1980; Nel­ recognized simply as the intense activa­ four points were recorded across the son et al, 1980). One body map was tion of all large and small fibres. A entire width of the postcentral gyrus. located in cytoarchitectonic area 3b form of this idea still exists today as the Welker (1971) introduced the use of and the other in area 1 (Fig. 2). Thus, gate control theory of pain. low impedance microelectrodes to primary somatosensory cortex of the record multiunit data from points much primate contains at least two complete The other school of thought, believ­ closer together within the somato­ homunculi. In addition there is some ing in specificity, held that there were sensory maps and thereby provided the evidence for additional representations specific classes of receptors each selec­ means to obtain higher resolution infor­ located in adjacent regions (Merzenich tively activated by a different class of mation about the organization of these and Kaas, 1980). This experiment has stimuli and each giving rise to a dif­ regions. The newer maps of both the not been performed in man but the ferent sensation. Thus, tactile stimuli cortical and subcortical structures in likelihood that a dual map exists in were said to activate encapsulated the somatosensory pathways have been areas 3b and 1 of man is overwhelm­ receptors, thermal stimuli to activate obtained with multiunit recordings ing. taken at intervals of 50 to 300/um. This one class of free nerve endings, and If the number of maps of the body method provides more than an order of noxious stimuli to activate another surface reported to exist in the class of free nerve endings. Without magnitude improvement in spatial resolution. somatosensory cortex are enumerated, question, the weight of evidence now the two found in SI can be added to the

Downloaded12 —FEBRUAR from https://www.cambridge.org/coreY 1982 . IP address: 170.106.33.22, on 25 Sep 2021 at 14:00:48, subject to the Cambridge Core terms of use, availableSomatosensory at Processing https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0317167100043560 LE JOURNAL CANADIEN DES SCIENCES NEUROLOGIQUES

answer by carefully recording the nature of the adequate stimulus re­ quired to activate the neurons in each map. From such data it becomes ap­ parent that each map is best activated by a different class of somatic stimuli. Further, these classes of stimuli suggest that the predominant input to each map is one particular class of sensory recep­ tors. Tests were designed to differen­ tially activate rapidly adapting cutaneous receptors, slowly adapting cutaneous receptors, pacinian corpus­ cles and muscle spindles (Dykes et al, 1980b), and were used to identify the receptor classes which best activated cortical neurons. When these tests were applied repeatedly to neurons observed at successive loci within the cat somatosensory cortex, the response classes were not randomly distributed, but were segregated into specific nuclear regions. These regions ap­ peared to contain a complete map of the body. Thus each map consisted of a segregated population of neurons serv­ ing a specific submodality or class of sensory receptors. For example, as has been argued for some time, recordings in area 3a showed that this particular cortical area received input only from deep structures, (Oscarsson and Rosen, 1966) probably muscle spindles, and that when the exploring electrode cros­ sed the cytoarchitectonic boundary between areas 3a and 3b the responses abruptly changed so that the neurons were optimally driven by stimuli which activated cutaneous receptors (Fig. 3) (Rasmusson et al, 1979). Once the electrode crossed into area 3b and the responses were driven by cutaneous stimuli, they never reverted Figure 2 — (Reproduced with permission from Nelson et al, 1980) Two maps in primary to activation by stimuli applied to deep somatosensory cortex of the cynomdogous monkey (Macaca fascicularis) as obtained by the structures. The transition from deep to fine grain microelectrode mapping technique of Welker (1971). cutaneous responses was abrupt and complete; seldom were mixtures of cutaneous and deep inputs observed. map in SII and the more elusive map in that there may be as many as ten; By making many penetrations in the SHI reported only rarely (Tasker, clearly the somatosensory cortex con­ cortex and recording the receptive 1960; Darian-Smith et al, 1966; Tanji tains a multiplicity of body representa­ fields encountered at each cortical et al, 1978). In addition, there is tions. locus, a complete representation of the another representation of the body in This observation prompts the ques­ deep inputs (muscle spindles) was area 3a between area 3b and motor tion of causation — why do multiple found in area 3a and appeared to be cortex (Oscarsson and Rosen, 1966; maps exist? That each is a successive co-extensive with this cytoarchitectonic Dykes et al, 1980b), making a total of evolutionary step in the development of area in cats (Dykes et al, 1980b). five that are well-documented. the neocortex is no longer a satisfac­ Merzenich, Kaas and their colleagues Merzenich and Kaas (1980) suggest tory answer. One can obtain another have shown that there are complete

Dykes Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 25 Sep 2021 at 14:00:48, subject to the Cambridge Core terms of use, availableFEBRUAR at Y 1982— 13 https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0317167100043560 THE CANADIAN JOURNAL OF NEUROLOGICAL SCIENCES

cutaneous maps in cytoarchitectonic

regions of 3b and 1 of the owl monkey NOT (Merzenich et al, 1978) and macaque SKIN DRIVEN DEEP monkey, (Nelson et al, 1980). In cats, by giving careful attention to the sub- • D modality of the cutaneous responses, 100- (ie., whether neurons are best activated by cutaneous slowly adapting or cutaneous rapidly adapting inputs) it is 80- g possible to show that even within area

3b, there are two maps. These two _ 60- response classes are segregated into z UJ distinct bands running mediolaterally O a* within area 3b (Dykes et al, 1980b). £ 40-

Such an analysis can be extended to other receptor classes. It is possible to 20- ii readily identify inputs from pacinian -*-"^, i corpuscles because they are sensitive to very high frequency vibrations and 0 i i 1000 800 600 400 200 4 200 400 600 800 1000 have large receptive fields. Thus, it is MICRONS 1 MICRONS relatively straightforward to ask where .„r. , 3a / 3b their inputs terminate in the sensory AREA 3a BORDER A"EA 3b cortex. Traditionally, input from paci­ nian afferents has been difficult to find Figure 3 — (Reproduced with permission from Rasmusson et al, 1979) Percentage of deep, skin and in SI. Mountcastle et al (1969) found non-drivable sites encountered in progressive lOOum steps during electrode trajectories that cros­ that only 4% of 1099 units could be ac­ sed the area 3a/area 3b border. n< 30. tivated by stimuli thought to selective­ ly drive pacinian corpuscles, and even those cortical neurons so activated sions deep responses were encountered. to one cytoarchitectonic zone, must could not be cyclically entrained. The Mountcastle (1957) and Powell and modify the concept of the cortical reason for this is now apparent; paci­ Mountcastle (1959) reported an in­ column. Mountcastle (1957) and nian input is preferentially directed creasing probability of encountering Powell and Mountcastle (1959) towards SII (Bennett et al, 1980, Fer- regions receiving input from deep struc­ recognized a cortical column by the rington and Rowe, 1980) and it now tures when electrodes were placed at changes that occurred at its boundary. appears likely that SII contains a map progressively posterior regions in areas Either a change in submodality or a of the body devoted to input from paci­ 1 and 2, until in the most posterior shift in the receptive field locus signal­ nian afferent fibres. SII also has been region only 10% of the neurons were led the border. It is apparent now that implicated as a recipient of other clas­ activated by cutaneous receptors. these changes occur predominantly at ses of sensory input (Bennett et al, Thus, it is not yet clear that either areas the edges of each map (Rasmusson et 1980) and eventually may be shown to 1 or 2 serve only one modality, and al, 1979; Dykes et al, 1980b; Dykes contain several other representations in these regions must be explored further. and Gabor, 1980; 1981; Sretavan and addition to the pacinian map. Nevertheless, major portions of Dykes, 1980) and consequently that somatosensory cortex consist of a the functional criteria for recognizing The functions of those portions of series of body representations wherein the edge of a cortical column corres­ primary somatosensory cortex iden­ each map is restricted to a single pond to the transition region between tified cytoarchitectonicaly as areas 1 cytoarchitectonic area and in those two distinct maps in the somatosensory and 2 are still poorly characterized. areas studied most carefully, each cortex. Thus, a change in the nature of However, from the limited data serves a particular class of sensory the adequate stimulus occurs once at available, they seem to contain ad­ receptors. Such an arrangement sug­ the 3a-3b boundary and generally ditional maps of the body surface. gests that the multiple maps exist so again only once within area 3b when Merzenich et al (1978) have drawn a that there will be a distinct cortical the slowly adapting to rapidly-adapting detailed and complete map for area 1 in region to serve each of several discrete transition occurs. monkey and a partial map in area 2. classes of somatic receptors. When long slanting penetrations are Based upon only a few penetrations, Multiple maps and the cortical oriented to run along the mediolateral Dykes et al (1980b) suggested that area column Since the cortical column is length of the somatosensory cortex (in 1 in cats also contains a body map. The recognized electrophysiologically by a direction equivalent to running along penetrations passing through area 1 the nature of the stimulus adequate to the mediolateral length of the post­ generally provided cutaneous rapidly excite the constituent neurons, central gyrus), then the receptive field adapting responses but on some occa- modality-specific maps, each restricted sequences are a smooth gradient of

Downloaded14 —FEBRUAR from https://www.cambridge.org/coreY 1982 . IP address: 170.106.33.22, on 25 Sep 2021 at 14:00:48, subject to the Cambridge Core terms of use, availableSomatosensory at Processing https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0317167100043560 LE JOURNAL CANADIEN DES SCIENCES NEUROLOGIQUES

gradually shifting, partly overlapping Cor. S regions on the body, and in this direc­ tion receptive field jumps that might "\ Ans. S signal the boundary of the cortical ( column are seldom encountered. These Cr. S observations suggest that empirically- demonstrable cortical columns are long • \> _J LATERAL narrow bands running the length of the 3b ANTERIOR v -oW^-V" 0 \ cytoarchitectonic bands in SI and are ;j\vvs^\^^o->*AiK:JW \ w\^^lw^ V* A \X equivalent to the individual functional ° ^^^ * °^^ ' maps which exist within each O A ^° „o o °i*»"00^ o\ D° iPo „6»5Sr^ooa o0 cytoarchitectonic region (Dykes and - > °>° m:ooo\ Gabor, 1981). A 0 3 A A. ° a " a O^ 0 ' ^v\ „l^ 6 a A A 4 A ° iv\^ o °. a oA a i ^vwS' * „ o K The concept of modality-specific a 6 0< 3a «A4" % '&§^ A$y// CAT 180 modality-specific cylinders, the cortex • CUTANEOUS SA \

DownloadedDykes from https://www.cambridge.org/core. IP address: 170.106.33.22, on 25 Sep 2021 at 14:00:48, subject to the Cambridge Core terms of use, availableFEBRUAR at Y 1982— 15 https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0317167100043560 THE CANADIAN JOURNAL OF NEUROLOGICAL SCIENCES

presented above for the cortical areas, have been obtained only recently, and with the exception of two experiments in primates (Friedman and Jones, 1981; Dykes et al, 1981) all of the data have been derived from studies of cats. However, the cat cortex has proven to be an excellent predictor of the or­ ganization of primate cortex and there is little reason to believe that sub­ cortical regions will not follow the same pattern in monkeys and man. Conse­ quently, the following data in cats will be presented with the assumption that in the future an equivalent organization will be found in primates. Functional Organization of the So­ Figure 5b — Loci of successively encountered receptive fields observed during the trajectory shown matosensory Thalamus The ventro- in part A. Letters A through E correspond to the regions marked in part A. (Reproduced with posterior (VP) nucleus of the thalamus permission.) can be divided into five parts. These are the inferior, medial, lateral, oral and posterior group of thalamic nuclei main cuneate nucleus have been known superior divisions, or VPI, VPM, VPL, named nucleus posterior oralis medialis for some time to serve muscle spindles VPO and VPS respectively (Dykes et (POm). This region projects to area 5a (Rosen and Sjolund, 1973) and to pro­ al, 1980a). Using electrophysiological (the third somatic sensory area, SIII) ject preferentially to the muscle spindle mapping techniques combined with in­ (Tanji et al, 1978). Recently a series of region of the thalamus, VPO (Oscars- jections of HRP and radio-labelled pro­ experiments repeating some of the son, 1966; Dykes et al, 1980a). The teins Dykes et al (1980a) confirmed the observations of Poggio and Mountcas- microelectrode maps of the central observations of Oscarsson and Rosen tle (1960) showed that there is a sub- regions of the cuneate and gracile (1966) that VPO receives input from region of Pom which receives input nuclei show that this region receives muscle spindles and projects to area 3a from cutaneous rapidly adapting only cutaneous inputs and that these of the somatosensory cortex. Similar neurons having moderately high inputs are grouped into discrete regions observations concerning VPLc pars thresholds and characteristically large serving cutaneous slowly adapting oralis of monkeys (VPO of cats) receptive fields. These neurons appear receptors and other regions serving (Dykes et al, 1981; Friedman and to receive their inputs from a set of cutaneous rapidly adapting receptors. Jones, 1981) suggest that this region spinal neurons with similar functional The spatial arrangement of these also serves muscle spindles in the properties located in the lateral cervical segregated regions within the central primate. In the monkey, the cortical nucleus (da Costa et al, 1981). The in­ core of the DCN is not clear but all projection of VPLc pars oralis is less ference from these observations is that major body parts have been found to well-defined and may project to several there are a number of discrete be represented in both parts. The data other areas as well as area 3a. (Jones submodality-specific nuclear regions in support the hypothesis that at least two and Porter, 1980). The VPL portion of the somatosensory thalamus. Each complete maps of the body exist here the thalamus receives input from projects preferentially to a specific cor­ — one serving slowly adapting recep­ cutaneous receptors and appears to be tical region and each probably conveys tors and one serving rapidly adapting divided into discrete regions devoted to information from one or a few selected receptors. slowly adapting and rapidly adapting classes of cutaneous sensory receptors. In the caudal portion of the cuneate inputs in the same way area 3b is Obviously the principles of organiza­ and gracile nuclei is a cytoarchitec- divided (Fig. 5) (Dykes et al, 1981). tion of the VP thalamus appear very tonically-distinct region wherein cells Beneath the VPL is the VPI nucleus. similar to those found in the somato­ can be activated by stimuli that selec­ Electrophysiological recordings within sensory cortex where specific regions tively activate pacinian corpuscles. VPI indicate that input from pacinian each serve a specific receptor class. This region appears to receive input corpuscles terminates there. HRP in­ Organization of the Dorsal Column from pacinian corpuscles located in all jections into SII showed that the VPI Nuclei (DCN) The same combined parts of the body except the face and nucleus projects to SII (Herron, 1981; anatomical and electrophysiological thus to contain a representation of the Herron and Dykes, 1981) where the analysis has been extended to the DCN body exclusive of the face. Injections of map of pacinian input is situated. Thus, to show that here again discrete nuclear HRP into the VPI nucleus of the VPI is a relay nucleus for pacinian cor­ volumes serve specific classes of recep­ thalamus label cells in this caudal part puscles. tors and project to specific thalamic of the DCN demonstrating that this A posterior thalamic region just dor­ relay nuclei. The external cuneate region projects to the thalamic area sal to VPL is a rostral extension of the nucleus and its extension under the which receives pacinian input.

Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 25 Sep 2021 at 14:00:48, subject to the Cambridge Core terms of use, availableSomatosensory at Processing https://www.cambridge.org/core/terms16 — FEBRUARY 1982 . https://doi.org/10.1017/S0317167100043560 LE JOURNAL CANADIEN DES SCIENCES NEUROLOGIQUES

SUMMARY At each level of the neuraxis there exist segregated nuclear regions ac­ tivated selectively by a specific class of sensory receptors. Each of the subcor­ Somatosensory tical regions projects preferentially to a cortex selected region at the next higher level of the somatosensory pathway where again a particular nuclear region is ac­ tivated by that same specific class of receptors (Fig. 6). This arrangement can be described as a parallel relay and processing of somatic sensory information. Parallel processing also exists in the visual system (Stone et al, 1979) where it ap­ pears to follow very similar rules of organization. The virtues of this con­ ceptualization in the somatosensory system are: (i) It provides a partial ex­ planation of how modality-specific in­ put can be relayed to the somatosen­ sory cortex despite the obvious Somatosensory divergence and the convergence which thalamus occurs at each synaptic level; (ii) it sug­ gests that the multiple representations of the body located in the somatosen­ sory cortex are there so that each may serve a separate class of inputs; (iii) it predicts where the transition points of the cortical columns will be found; and VPI finally (iv) it suggests a new shape for those cortical units Mountcastle (1957) called columns.

ACKNOWLEDGEMENTS This work was supported by the Medical Research Council of Canada. I am indebted to Mr. Maurice Murphy, graphiste, Mr. Charles Hodge, photo­ grapher, and Mrs. Genevieve Holding, typist for their constant support and cooperation. Dorsal REFERENCES column ADRIAN, E.D. (1941) Double representation nuclei of the feet in the sensory cortex of the cat. J. Physiol. 98: 16p-18p. BARD, P. (1938) Studies on the cortical representation of somatic sensitivity. Bull. N.Y. Acad. Med. 14: 585-607. BENNETT, R.E., FERRINGTON, D.G. and ROWE, M. (1980) Tactile neuron classes within second somatosensory area (SII) of the cat cerebral cortex. J. Neurophysiol. 43: 292-309. da COSTA, D.C.N., METHERATE, R.S., HERRON, P. and DYKES, R.W. (1981) A thalamic terminus for the lateral cervical nucleus. Soc. Neurosci. Abstr. 7. Figure 6 — Summary of modality specific connections in the dorsal column-medial lemniscal DARIAN-SMITH, I., ISBISTER, J., MOK, H. system. Specific classes of sensory receptors converge upon discrete regions of the dorsal column and YOKOTA, T. (1966) Somatic sensory nuclei. These nuclear regions relay their receptor-specific information to separate nuclear regions cortical projection areas excited by tactile in the somatosensory thalamus which, in turn, project to specific cortical areas. This arrange­ stimulation of the cat: a triple representa­ ment gives rise to multiple cortical representations each of which is preferentially activated by a tion. J. Physiol. 182: 671-689. different class of somatic sensory information.

Dykes FEBRUARY 1982 — 17 Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 25 Sep 2021 at 14:00:48, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0317167100043560 THE CANADIAN JOURNAL OF NEUROLOGICAL SCIENCES

DREYER, D.A., LOE, P.R., METZ, C.B. and In: Neural Mechanisms of the Auditory and PAUL, R.L., MERZENICH, M., and WHITSEL, B.L. (1975) Representation of Vestibular Systems, Thomas, Springfield. GOODMAN, H. (1972) Representation of the head and face in the postcentral gyrus of slowly and rapidly adapting cutaneous the Macaque./. Neurophysiol. 38: 714-733. HUBEL, D.H. and WIESEL, T.N. (1962) mechanoreceptors of the hand in Brodman's Receptive fields, binocular interaction and areas 3 and 1 of Macaca mulatto. Brain DYKES, R.W. (1977) Sensory receptors, ch 9: functional architecture in the cat's visual Res. 36: 229-249. pp. 320-333. In: Reconstructive Micro­ cortex. J. Physiol. 160: 106-154. surgery, R. Daniel and J. Terzis. Little, PENFIELD, W. and BOULDREY, E. (1937) JONES, E.G. and PORTER, R. (1980) What is Brown & Co. Boston, Mass. Somatic motor and sensory representation area 3a? Brain Res. Rev. 2: 1 -43. in the cerebral cortex of man as studied by DYKES, R.W. (1978) The anatomy and KAAS, J.H., NELSON, RJ., SUR, M., LIN, electrical stimulation. Brain. 60: 389-443. physiology of the somatic sensory cortical C.S. and MERZENICH, M. (1979) Multiple regions. Prog. Neurobiol. 10: 33-88. PENFIELD, W. and RASMUSSEN, T. (1950) representations of the body within the The cerebral cortex of man. pp. 248. Mac- DYKES, R.W. and GABOR, A. (1980) primary somatosensory cortex of primates. millan, N.Y. Magnification factors and receptive field se­ Science 204: 521-523. quences in submodality-specific bands of cat PENFIELD, W. and JASPER, H. (1954) LEVITT, J. and LEVITT, M. (1968) Sensory SI cortex. (Abstr) Proc. Int. Union. Physiol. Epilepsy and the functional anatomy of the hind-limb representation in the SMI cortex Sci. 14: 392 (Budapest). human brain. In: Functional Localization in of the cat. A unit analysis. Expl. Neurol. 22: the Cerebral Cortex, pp. 41-155. Little, 259-275. DYKES, R.W. and GABOR, A. (1981) Brown and Co., Boston, Mass. Magnification functions and receptive field MERZENICH, M.M. and KAAS, J.H. (1980) sequences for submodality-specific bands in POGGIO, G.F. and MOUNTCASTLE, V.B. Principles of organization of sensory- SI cortex of cats. J. Comp. Neurol. 202: (1960) A study of the functional contribu­ perceptual systems in mammals. In: 597-620. tions of the lemniscal and spinothalamic Progress in Psychobiology and Physiological systems to somatic sensibility. Central ner­ DYKES, R.W., LIN, C.S. and REHMAN, N. Psychology, N.Y. Academic press. Vol. 9, vous mechanisms in pain. Johns Hopkins (1980a) Thalamocortical projection nuclei pp. 1-42, Ed. J.M. Sprague and A.N. Eps­ Hosp. Bull. 106:266-316. for modality-specific regions of somatosen­ tein. sory cortex: demonstration of different pro­ POWELL, T.P.S. and MOUNTCASTLE, V.B. MERZENICH, M.M., KAAS, J.H., SUR, M. jections for areas 3a and 3b of the cat. The (1959) Some aspects of the functional Physiologist 23: 73. and LIN, C.S. (1978) Double representation organization of the cortex of the postcentral of the body surface within cytoarchitectonic gyrus of the monkey: a correlation of DYKES, R.W., RASMUSSON, D.D. and areas 3b and 1 in "SI" in the owl monkey. HOELTZELL, P.B. (1980b) Organization findings obtained in a single unit analysis (Aotus trivirgatus). J. Comp. Neurol. 181: with cytoarchitecture. Johns Hopkins Hosp. of primary somatosensory cortex in the cat. 41-74. J. Neurophysiol. 43: 1527-1546. Bull. 705:133-162. MOUNTCASTLE, V.B. (1957) Modality and RASMUSSON, D.D., DYKES, R.W. and DYKES, R.W. and SRETAVAN, D. (1981) topographic properties of single neurons of HOELTZELL, P.B. (1979) Segregation of Submodality and its relationship to cat's somatic sensory cortex. J. Neuro­ modality and submodality information in SI somatotopy in the forearm region of cat cor­ physiol. 20:408-434. tex. Neurosc. Soc. Abstr. 7: 834. cortex of cat. Brain Res. 166:409-412. MOUNTCASTLE, V.B. (1980) Sensory recep­ ROSEN, I. and SJOLUND, B. (1973) DYKES, R.W., SUR, M., MERZENICH, tors and neuronal encoding: introduction to Organization of group I activated cells in the M.M., KAAS, J.H. and NELSON, R.J. sensory processes, pp. 327-347. ch. 11 In: (1981) Regional segregation of neurons main and external cuneate nuclei of the cat: Medical Physiology Vol. 1. ed. V.B. Mount- Identification of muscle receptors. Exp. responding to quickly adapting, slowly castle. C.V. Mosby Co. St. Louis. adapting, deep and pacinian receptors within Brain Res. 16:221-231. MOUNTCASTLE, V.B. and POWELL, T.P.S. thalamic ventroposterior lateral and ventro- SINCLAIR, D. (1967) Cutaneous Sensation (1959) Central nervous mechanisms sub­ posterior inferior nuclei in the squirrel Oxford University Press, London, pp. 306. monkey (Saimiri sciureus). Neurosci Vol. 6: serving position sense and kinesthesia. John 1687-1692. Hopkins Hosp. Bull. 105: 173-200. STRETAVAN, D. and DYKES, R.W. (1980) Receptive field sequences in the somatosen­ FERRINGTON, D.G. and ROWE, M. (1980) MOUNTCASTLE, V.B., TALBOT, W.H., sory cortex of the cat. Soc. Neurosci. Abst. Differential contributions to coding of SAKATA, H. and HYVARINEN, J. (1969) 6: 763. cutaneous vibratory information by cortical Cortical neuronal mechanisms in flutter- somatosensory areas I and II. J. Neuro­ vibration studied in unanesthetized STONE, J., DREHER, B. and LEVENTHAL, physiol. 43: 310-331. monkeys. Neuronal periodicity and fre­ A. (1979) Hierarchical and parallel quency discrimination. J. Neurophysiol. 32: mechanisms in the organization of visual FRIEDMAN, D.P. and JONES, E.G. (1981) 452-484. cortex. Brain Res. Rev. 1: 345-394. Thalamic input to areas 3a and 2 in monkeys. J. Neurophysiol. 45: 59-85. NELSON, R.J., SUR, M., FELLEMAN, D.J. SUR, M., MERZENICH, M.M., and KAAS, and KAAS, J.H. (1980) The representations J.H. (1980) Some quantitative features of the HERRON, P. (1981) The connections of cor­ of the body surface in postcentral somato­ body surface representations in areas 3b and tical somatosensory areas I and II with sensory cortex of Macaca fascicularis. J. 1 of somatosensory cortex in owl monkeys: separate nuclei in the ventroposterior region Comp. Neurol. 192: 611-644. magnification, receptive field area and of the thalamus in the raccoon. Neurosci. (In hypercolumn size. J. Neurophysiol. 44: 295- press). OSCARSSON, O. (1966) The projection of 311. Group I muscle afferents to the cat cerebral HERRON, P. and DYKES, R.W. (1981) Con­ cortex. In: Nobel Symposium I: Muscle Af­ SUR, M., NELSON, R.J. and KAAS, J.H. nections of the ventroposterior inferior ferents and Motor Control, pp. 307-316. Ed. (1978) The representation of the body sur­ nucleus of the thalamus in cats. Soc. R. Granit. Wiley, N.Y. face in somatosensory area I of the grey Neurosci. Abstr. 7: 396. squirrel. J. Comp. Neurol. 179:425-449. OSCARSSON, O. and ROSEN, I. (1966) Short HIND, J.E., ROSE, J.E., DAVIES, P.W., latency projections to the cat's cerebral cor­ SUR, M., NELSON, R.J. and KAAS, J.H. WOOLSEY, C.N., BENJAMIN, R.M., tex from skin and muscle afferents in the (1980) Representation of the body surface in WELKES, W.S., and THOMPSON, R.F. contralateral forelimb. J. Physiol. 182: 164- somatic koniocortex in the Prosimian (1960) Unit activity in the auditory cortex. 184. Galago. J. Comp. Neurol. 189: 381-402.

Downloaded18 —FEBRUAR from https://www.cambridge.org/coreY 1982 . IP address: 170.106.33.22, on 25 Sep 2021 at 14:00:48, subject to the Cambridge Core terms of use, availableSomatosensory at Processing https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0317167100043560 LE JOURNAL CANADIEN DES SCIENCES NEUROLOGIQUES

TANJI, D.G., WISE, S.P., DYKES, R.W. and In: Neurophysiological Basis of Normal and WOOLSEY, C.N. (1964) Cortical localization JONES, E.G. (1978) Cytoarchitecture and Abnormal Motor Activities, pp. 255-293. as defined by evoked potential and electrical thalamic connectivity of the third somato­ eds. M.D. Yahr, and D.P. Purpura. Raven stimulation studies. In: Cerebral Localiza­ sensory area of the cat cerebral cortex. /. Press, Hewlett, N.Y. tion and Organization, pp. 17-26. eds. G. Neurophysiol. 41: 268-284. Schaltenbrand and C.N. Woolsey, The WILLIS, W.D. and COGGESHALL, R.E. University of Wisconsin Press, Madison. TASKER, R.R. (1960) A third somatic tactile (1978) Sensory Mechanisms of the Spinal sensory area. Physiologist 3: 162. Cord Plenum Press, New York, N.Y. pp. 485. WOOLSEY, C.N. and FAIRMAN, D. (1946) WELKER, C. (1971) Microelectrode delinea­ Contralateral, ipsilateral and bilateral tion of fine grain somatotopic organization WOOLSEY, C.N. (1958) Organization of representation of cutaneous receptors in of SMI cerebral neocortex in albino rat. somatic areas I and II of the cerebral cortex Brain Res. 26: 259-275. somatic sensory and motor areas of the cerebral cortex. In: Biological and of pig, sheep and other mammals. Surgery. WELT, C, ASCHOFF, J.C., KAMEDA, K. Biochemical Bases of Behavior, pp. 63-81. 19: 684-702. and BROOKS, V.B. (1967) Intracortical eds. H.F. Harlow and C.N. Woolsey. The organization of cat's motorsensory neurons. University of Wisconsin Press, Madison.

Dykes FEBRUARY 1982— 19 Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 25 Sep 2021 at 14:00:48, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0317167100043560