Patterns in the brain. Neuronal population coding in the somatosensory system.

The aim of this article is to review some basic principles of neural coding, with an emphasis on mechanisms of stimulus representation in ensembles of neurons. The theory of "across-neuron response patterns" (ANRPs), first suggested by Thomas Young (1802) and fully developed by Robert Erickson (1963-2000), is summarized and applied to the problem of coding in primary afferent fibers and cortical neurons of the somatosensory system. The basic premise of the theory is that precise information about stimulus features cannot be encoded by single neurons, but is encoded by patterns of activity across populations of neurons. Different stimuli produce uniquely different patterns of ensemble activity (ANRPs)-discrimination between two stimuli is based on the absolute difference in total amount of activity (neural mass difference) of the ANRPs for those stimuli. Review of the literature shows that ANRPs and related population codes can accurately represent and differentiate among various stimulus parameters that cannot be distinguished by single neurons alone. Finally, the behavior of neuronal ensembles can be used to account for the sensory-perceptual changes associated with plasticity of thalamocortical circuits following selective sensorimotor deprivation or experience.

Perceptual significance of somatosensory cortical reorganization following peripheral denervation.

The functional significance of reorganization in somatosensory cortex following peripheral denervation has not been thoroughly addressed. In this paper, two distinct hypotheses dealing with this issue are discussed. The first is the hypothesis of functional respecification. This influential view suggests that sets of partially deafferented cortical neurons, which respond to new peripheral inputs and acquire new receptive fields, undergo corresponding changes in perceptual meaning. Excitation of these neurons by stimulation of their novel receptive fields is thought to result in a change in referral of sensation from the original (now denervated) skin fields to the newly acquired skin fields. The second hypothesis is that of functional conservation. This equally plausible alternative is that sets of partially deprived neurons, although they respond to novel peripheral inputs, retain their original perceptual meaning. Excitation of these neurons by stimulation of their new receptive fields is thought to evoke sensation formerly mediated by those neurons, and hence is still projected to the original, now denervated skin regions or phantom. Behavioral evidence strongly suggests that cortical reorganization after peripheral denervation does not result in major functional respecification, but that the original perceptual function mediated by those neurons is preserved.

Distribution of neurons immunoreactive for parvalbumin and calbindin in the somatosensory thalamus of the raccoon.

The aim of this study was to assess the distribution of neurons immunoreactive for parvalbumin (PV), calbindin (CaBP), glutamic acid decarboxylase (GAD), and gamma-aminobutyric acid (GABA) in the somatosensory thalamus of the raccoon and to compare these features to those of other species, especially primates. Immunohistochemistry was used to study the location of these neurons in the ventroposterior nucleus (VP), ventroposterior inferior nucleus (VPI), posterior group of nuclei (Po), and reticular nucleus (Rt). A consistent differential pattern of PV-positive (PV+) and CaBP-positive (CaBP+) cells was found in the somatosensory thalamus. Many PV+ neurons were observed in VP and Rt, but very few were found in VPI or Po. In contrast, CaBP+ neurons were distributed throughout VP, VPI, and Po but were very sparse or absent in Rt. In the VP nucleus, PV+ cells were distributed in clusters separated by interclusteral regions with a sparse distribution of PV+ cell bodies. The distributions of PV+ and CaBP+ cells tended to be complementary to each other in VP; regions with a high density of PV+ neurons had a low density of CaBP+ cell bodies. Double-labeling experiments revealed very few neurons in which PV and CaBP immunoreactivities were colocalized. Cells immunoreactive for GAD or GABA were found in PV-dense clusters of VP; fewer GABAergic neurons were present in the CaBP-dense interclusteral regions of VP and in VPI and Po. GAD+ and GABA+ neurons were most prominently distributed in Rt. We conclude that the distributions of PV+ and CaBP+ cell bodies in the raccoon somatosensory thalamus are very similar to those in primates. The density of GABAergic neurons in the somatosensory thalamus of the raccoon is less than that in the cat and monkey, but the relative distribution of GABAergic neurons in the different somatosensory nuclei is very similar to that in the cat and monkey. These results are discussed in relation to findings in other species and are related to the functions of lemniscal and nonlemniscal somatosensory pathways.

Progressive changes in cutaneous trigger zones for sensation referred to a phantom hand: a case report and review with implications for cortical reorganization.

The dominant model of cortical plasticity induced by peripheral denervation suggests that a physiologically-reorganized cortical area can acquire new perceptual meaning, including a change in the peripheral referral of sensation. An alternative view is that such an area may retain its former perceptual significance, even though it becomes responsive to new peripheral inputs. To examine evidence related to this issue, a clinical case is presented documenting the time course of changes in phantom limb sensation in a patient with accidental amputation of a hand. About 24 h after injury, a vivid phantom hand was present; tactile stimulation revealed cutaneous trigger zones on the arm, stimulation of which elicited sensation referred to specific fingers of the phantom. While the phantom hand percept remained fairly stable over time, the trigger zones expanded progressively in size during the next 1-8 weeks but had contracted and changed location considerably about one year later. At all times studied, the trigger zones were topographically related to specific fingers and other parts of the phantom hand. The implications of these and other recent clinical findings for cortical reorganization are discussed, and the following tentative conclusions are drawn. (1) A phantom percept is mediated by central neural networks which remain functionally intact after amputation. (2) Cutaneous trigger-zones mapped in humans correspond to novel receptive fields of cortical neurons mapped in animals following peripheral denervation. (3) Cortical reorganization induced by denervation does not produce a major change in perceptual meaning or peripheral reference. In the present case, stimulation of new trigger zones (receptive fields) on the patient's arm presumably activated a reorganized cortical hand area but evoked sensation still referred to the (now missing) hand. Hence, physiological cortical remapping is not necessarily accompanied by functional respecification.

Short-term plasticity in primary somatosensory cortex of the rat: rapid changes in magnitudes and latencies of neuronal responses following digit denervation.

Recordings were made from neurons in primary somatosensory (SmI) forepaw cortex of rats to study the time course of changes in responses beginning immediately following denervation (ligation) of a single digit. Before denervation, neuronal receptive fields (RFs) defined by tactile stimulation varied in size from small regions of one digit to larger areas covering several digits and palmar pads. With electrical stimulation, most neurons responded best to one (on-focus) digit and less to other (off-focus) digits; on-focus stimulation yielded more spikes per stimulus and shorter spike latencies (Lmin) than did off-focus stimulation. After ligation of the on-focus digit, most neurons showed increased responsiveness to stimulating one or several off-focus digits and palmar regions of the forepaw: (1) tactile stimulation showed that the RFs of all but one neuron expanded to include previously "ineffective" skin regions, such as digits or palmar pads adjoining the original RF; (2) electrical stimulation usually evoked stronger responses from neighboring off-focus digits and sometimes elicited novel responses from previously ineffective digits--seven of ten neurons showed increases in spikes per stimulus, which tended to approach stable values within 60-90 min after denervation; three of ten neurons showed decreases in Lmin with time, but most revealed no significant changes. These results suggest that dynamic response properties, as well as RFs, of SmI cortical neurons can be modified rapidly by blocking afferent input from dominant on-focus skin regions. RFs expand and novel responses appear, with concomitant increases in response magnitude and, in some cases, decreases in response latency over time. These findings seem to reflect a rapid increase in synaptic efficacy of weak or previously ineffective inputs from cutaneous afferent nerve fibers.

Immunoreactivity for GAD and three peptides in somatosensory cortex and thalamus of the raccoon.

Immunocytochemical methods were used to determine the distributions of glutamic acid decarboxylase (GAD), vasoactive intestinal polypeptide (VIP), cholecystokinin (CCK), and somatostatin (SOM) in the primary somatosensory cortex and somatosensory thalamus of adult raccoons. The cortex showed extensive immunoreactivity for GAD, revealing a large population of GABAergic neurons. GAD-labeled cells were numerous in all cortical layers, but were most concentrated in laminae II-IV. The cells were nonpyramidal and of varying morphology, typically with somata of small or medium size. GAD-immunoreactive puncta, presumably synaptic terminals, were widespread and often appeared to end on both GAD-negative and GAD-positive neurons. Immunoreactivity for the peptides was much less extensive than that for GAD, with the number of labeled neurons for VIP > CCK > SOM. Peptidergic cells were preferentially located in the upper and middle cortical layers, especially laminae II and III. The cells were nonpyramidal, often bitufted or bipolar in morphology, and small to medium in size. Their processes formed diffuse plexuses of fibers with terminal-like varicosities that occasionally surrounded nonpeptidergic neurons. The thalamus showed a clearly differentiated pattern of immunoreactivity for GAD, but little or no labeling for the three peptides. Nuclei adjoining the ventral posterior lateral (VPL)/ventral posterior medial (VPM) complex--including the reticular nucleus--contained many GAD-positive neurons and fibers. In contrast, the VPL and VPM nuclei displayed considerably less GAD immunoreactivity, somewhat surprising given the raccoon's highly developed somatosensory system. However, the ventral posterior inferior (VPI) nucleus revealed rather dense GAD labeling, perhaps related to a specialized role in sensory information processing. Thus, the primary somatosensory cortex of the raccoon showed patterns of immunoreactivity for GAD and peptides that were similar to those of other species; the somatosensory thalamus revealed a distinctive profile of GAD immunoreactivity, with labeling that was light to moderate in the VPL/VPM complex and relatively extensive in VPL.

Convergent inputs to single neurons in two different subdivisions of somatosensory forepaw digit cortex of the raccoon.

The aim of this study was to compare the physiological properties of single neurons in the glabrous (G) and heterogeneous (H) subdivisions of primary somatosensory digit 3 cortex of adult raccoons. Extracellular recordings were obtained from 50 G neurons whose receptive fields (RFs) were confined to the glabrous skin of a digit, and 41 H neurons whose RFs were located on hairy skin, claws, or mixtures of skin types. Both electrical and mechanical stimulation of the digits were used to assess excitatory neuronal responsiveness. The two sets of neurons, which had nearly identical depth distributions, differed considerably in their input convergence: (i) the percentage of neurons (%N) responding to electrical or mechanical stimulation of each off-focus digit and (ii) the number of digits from which individual cells could be driven were significantly greater for H neurons. Those G and H cells which could be excited by off-focus inputs were examined for probability of response (P), number of spikes per response (S/R), and latency of response (L) to digit stimulation. Surprisingly, for input from any one digit, there were no significant differences in these response properties between the two sets of neurons. However, inputs from different (on-focus versus off-focus) digits varied significantly and revealed patterns of response properties that were qualitatively similar for both G and H neurons. Specifically, %N and P decreased while L increased symmetrically with distance of each off-focus digit from the central on-focus digit 3, reflecting corresponding variations in the synaptic accessibility and conduction time of off-focus excitatory inputs. In contrast, S/R values were very similar for all digits, suggesting that the synaptic strength of off-focus inputs is regulated independently of accessibility. Finally, preliminary findings indicated that denervation of the third digit caused a decrease in off-focus response latencies, while the normal latency profile across digits was retained. This suggests that the previously existing pattern of off-focus inputs to G and H neurons provides a template for denervation-induced cortical reorganization, whereby the synaptic efficacy of off-focus inputs is increased by disinhibition or facilitation.

Immunoreactivity for glutamic acid decarboxylase and several neuropeptides in the spinal cord of the raccoon.

The peroxidase-antiperoxidase method was used to examine major immunohistochemical features of the spinal cord of adult raccoons. The lateral portions of the ventral horn contained many large multipolar neurons that showed cholecystokinin-like immunoreactivity, suggesting the coexistence of cholecystokinin with acetylcholine in a subset of motoneurons. The dorsal horn revealed unique but overlapping patterns of immunoreactivity for glutamic acid decarboxylase, somatostatin, substance P, vasoactive intestinal polypeptide and cholecystokinin. The data imply that some of the peptides may coexist within the same dorsal root ganglion cells and their spinal cord processes.

Coding of stimulus location and intensity in populations of mechanosensitive nerve fibers of the raccoon: I. Single fiber response properties.

The primary aim of this study was to determine and compare the receptive field (RF) characteristics and response properties of single mechanosensitive nerve fibers innervating the glabrous skin of the forepaw and hindpaw of the raccoon. The action potentials of 129 median nerve fibers and 61 posterior tibial nerve fibers were recorded in response to punctate mechanical stimuli varying in location and intensity. The stimuli were delivered to six standard test sites on digit 1 and the contiguous pads of each paw. Attempts were made to classify each fiber according to its rate of adaptation to sustained stimulation; the RF of each fiber was mapped using a standard series of stimulus intensities. The results indicated that the response properties of individual fibers were highly complex and depended on the location and intensity of stimulation. 1) The distributions of absolute threshold were not different for the median or tibial nerve fibers or for different classes of fibers based on adaptation rate. A distal to proximal increase in threshold was found for each paw, suggesting a corresponding gradient of sensitivity across the glabrous skin. 2) Threshold RF areas did not vary across either paw nor did they differ between the two paws. Suprathreshold RFs were quite large relative to expected tactile acuity and displayed complex features. 3) Response properties such as adaptation rate, on- and off-responses, were found to vary with both stimulus location and intensity. It was concluded that the responses of individual nerve fibers could not uniquely encode any stimulus parameter tested, and that the properties of single fibers could not account for apparent differences in tactile acuity across each paw or between the two paws.

Coding of stimulus location and intensity in populations of mechanosensitive nerve fibers of the raccoon: II. Across-fiber response patterns.

The major aim of this study was to determine whether information about stimulus location and intensity can be precisely encoded in the population response profiles of mechanosensitive nerve fibers. Across-fiber patterns (AFPs) were reconstructed using data obtained previously (58) from 129 single median nerve fibers and 61 posterior tibial nerve fibers of the raccoon. Punctate mechanical stimuli of varying intensities were presented to six standard test sites on the glabrous skin of the forepaw and hindpaw, and the AFP produced by each unique stimulus was determined. Innervation density (ID) at each of the test sites was estimated from the integrated voltage of compound action potentials evoked by constant current stimulation of the skin at each site. Differences in ID between corresponding regions of the forepaw and hindpaw were estimated by counts of myelinated fibers and measurements of skin areas supplied by the median and tibial nerves, respectively. All AFPs were adjusted to take into account the differences in ID among the various skin sites on the two paws. The data indicated that different stimulus locations and intensities on each paw could be precisely and unambiguously encoded by AFPs; location was represented primarily by the shape of the AFP, whereas intensity was represented by the height of the AFP. The discriminability of location and intensity was shown to be directly related to the magnitude of the difference between the AFPs elicited by any two stimuli. Consistent with expectations, AFPs for stimuli delivered distally on each paw showed greater differences as a function of stimulus location and intensity than AFPs for stimuli applied proximally. Furthermore, AFPs produced by stimuli to the forepaw showed greater changes than AFPs for stimuli to comparable sites on the hindpaw. It was concluded that systematic differences in the population response profiles of median and tibial nerve fibers could account for regional variations in tactile acuity across each paw and differences between the two paws.

Physiological changes in the somatosensory forepaw cerebral cortex of adult raccoons following lesions of a single cortical digit representation.

The purpose of this study was to determine whether restricted lesions within primary somatosensory (SmI) cortex cause changes in the functional organization of cortical areas bordering on the site of injury. Focal ablations of cortical tissue were made in the representational area for digit 3 within the SmI forepaw cortex of adult raccoons. Electrophysiological mapping experiments done 15-17 weeks later showed that significant alterations had occurred in the response properties of clusters of neurons within those representational zones adjoining the lesion--the zones for digit 2, digit 4, and the palmar pads. These three cortical areas were modified by the appearance of new, usually weaker secondary inputs and changes in some properties of the normal primary inputs from the forepaw. (i) Many neurons responded to stimulation of previously ineffective skin regions; the new inputs often originated from digit 3 but frequently involved other digits or the pads as well. (ii) Neuronal receptive fields (RFs), mapped at a standard suprathreshold stimulus intensity, were larger than normal. (iii) Skin type and submodality sensitivity typically were less specific than normal; more neurons had RFs that included both glabrous and hairy skin or claws and displayed mixtures of responsiveness to skin touch, hair deflection, or claw touch. (iv) The representation of RF location, skin type, and submodality sensitivity was more variable as a function of horizontal and vertical distance through the cortex. In general, the physiological changes were found to degrade the somatotopic order and response specificity of the intact cortical areas adjoining the lesion.(ABSTRACT TRUNCATED AT 250 WORDS)

Projections from cortical area SmII and claustrum to two functional subdivisions of SmI forepaw digit cortex of the raccoon.

The retrograde HRP tracer method was used to study the projections from cortical area SmII and the claustrum to two electrophysiologically defined, functionally distinct subdivisions of the SmI forepaw cortex of the raccoon. Individual SmI cortical digit zones were found to receive ipsilateral projections from SmII and immediately adjoining cortical regions; the projections to the "heterogeneous" (hairy skin and claw) functional subdivision of a given digit zone were considerably more extensive than those to the glabrous skin functional subdivision of that zone. HRP-filled neurons within SmII were located primarily in layers VI and V, often formed clusters, and were distributed antero-posteriorly in a manner consistent with a loosely topographic representation of the digits. The SmI cortical digit zones received ipsilateral projections from approximately the middle 1/3 of the anterior-posterior extent of the insular claustrum; no clear difference was found in the projections to the two functional subdivisions of a given digit zone. Labeled neurons were typically scattered throughout much of the claustrum but were more numerous in its dorsal regions, tended to aggregate in clusters, and were distributed antero-posteriorly in an overlapping but roughly topographic fashion.

Thalamic connections of two functional subdivisions of the somatosensory forepaw cerebral cortex of the raccoon.

The purpose of this study was to compare the thalamic interconnectivities of 2 functionally distinct subdivisions of the somatosensory (Sml) forepaw cortex of the raccoon--the somatotopic subdivision representing the glabrous skin of the digits and the more heterogeneous subdivision representing the hairy skin and claws. Injections of HRP were made into one or the other functional subdivision of a specific digit subgyrus of Sml cortex in 10 adult raccoons. The distribution of HRP-labeled neurons and axon terminals in the thalamus revealed that the 2 sectors have different patterns of thalamic projections. The glabrous skin region of each cortical digit zone was interconnected with a specific crescent-shaped lamella of neurons that extended rostrocaudally through the ventral posterior lateral (VPL) nucleus and typically was separated from adjacent lamellae by small bundles of myelinated fibers. The VPL lamellae constituted relatively distinct digit subnuclei that were connected somatotopically with the glabrous subdivisions of the corresponding cortical digit areas. The projections were dense, topographic, and reciprocal; labeled neurons and axon terminals within a particular lamella overlapped considerably and tended to be arranged in clusters. In contrast, the heterogeneous region of each cortical digit zone was reciprocally connected with the somatotopically appropriate VPL digit subnucleus and with adjoining subnuclei as well. The projections were comparatively sparse, less topographic, and more widely distributed than those of the glabrous skin sectors; groups of HRP-positive neurons and terminals in VPL tended to straddle the borders of the appropriate lamella and extended into adjacent lamellae. Furthermore, small clusters of labeling were found in the dorsal, presumed kinesthetic region of VPL and in portions of the ventral posterior inferior nucleus and the posterior nucleus. These results indicate that the glabrous cortical subdivisions have precise, somatotopically organized connections with specific VPL subnuclei, whereas the heterogeneous cortical subdivisions have more diffuse and scattered connections with several subregions of VPL and other thalamic nuclei as well. These 2 thalamocortical projection patterns may account for many of the differing functional properties of neurons residing within the 2 cortical sectors. Finally, the convergent thalamic projections to the heterogeneous cortical regions could contribute, at least indirectly, to the functional reactivation that occurs within Sml cortex of the raccoon following peripheral nerve transection (Kelahan and Doetsch, 1984).

Intracortical connections of two functional subdivisions of the somatosensory forepaw cerebral cortex of the raccoon.

The aim of this study was to compare the intrinsic intracortical connectivities of 2 functionally distinct subdivisions of the somatosensory (Sml) forepaw cortex of the raccoon--the somatotopic glabrous skin representation and the more heterogeneous, hairy skin and claw representation of the digits. HRP was injected into one or the other functional subdivision of a particular digit subgyrus of Sml cortex in 10 adult raccoons. The distribution of HRP-labeled neurons and axon terminals in the cortex showed that intrinsic "horizontal" connections exist within and between individual cortical digit zones; the labeling tended to have an oval-shaped configuration that was longer in the mediolateral than in the anteroposterior curvilinear plane. The 2 cortical sectors were found to have different patterns of intracortical projections. The connections of the glabrous skin region of each cortical digit zone were primarily local and confined to that same digit representation. HRP-filled neurons were concentrated near the injection site and decreased in density within the banks and fundi demarcating the injected digit subgyrus; few labeled cells were found in adjoining digit zones. Longer projections to the glabrous subdivision of a particular digit area typically originated from neurons in the heterogeneous subdivision of that same digit area. In contrast, the connections of the heterogeneous region of each digit zone were much more extensive and usually included projections from nonadjacent, as well as neighboring digit zones. The density of HRP-positive neurons declined more gradually with distance from the injection site, and considerable labeling was present in the heterogeneous sectors of adjacent digit zones. The intracortical projections of both functional subdivisions were often, but not always, reciprocal, and the cells of origin tended to be distributed in clusters. The laminar distributions of labeled neurons were similar for both sectors; HRP-filled cells were concentrated more in the supragranular layers, especially in layer III; fewer were found in the infragranular layers, mainly in layer VI and rarely in layer V. These results show that the intrinsic connections of the glabrous cortical subdivisions are fairly localized, whereas those of the heterogeneous cortical subdivisions are more diffuse and highly convergent. The differing intracortical connectional patterns of the 2 sectors are consistent with their contrasting thalamocortical projection patterns and may contribute to the unique functional properties of neurons located within each sector.(ABSTRACT TRUNCATED AT 400 WORDS)

Time-dependent changes in the functional organization of somatosensory cerebral cortex following digit amputation in adult raccoons.

Surgical removal of the third forepaw digit in raccoons causes both long-term and short-term changes in functional organization within the digit 3 primary somatosensory (SmI) cortex. Previous studies have shown that 36-52 weeks following amputation in infant raccoons, neurons within the digit 3 cortical territory had become responsive to cutaneous stimulation of "new" forepaw regions adjoining the digit stump (Carson et al., 1981; Kelahan et al., 1980, 1981); the "novel" receptive fields (RFs) were often larger than normal and revealed no orderly somatotopic organization. In the present study, the cortical effects of digit 3 removal were examined in adult raccoons. Within 36 weeks after amputation, the digit 3 zone was also found to be reactivated by "novel" inputs from the forepaw, with no strictly topographic representation of the "new" skin fields. The basic features of cortical reactivation were very similar in animals amputated as adults and as infants, except that the former typically had larger neuronal RFs than the latter. Short-term cortical changes were studied in adult raccoons within 1 day and between 1 and 4 weeks after amputation: Significant time-dependent differences were found in the reactivated digit 3 territory. Within 1 hr following amputation, some cells in the digit 3 zone began to respond to low-intensity cutaneous stimulation of "new" forepaw regions, limited almost exclusively to digits 2 and 4. Neuronal RFs tended to be larger than normal and showed no strictly topographic organization. One to 4 weeks following amputation, the condition of the digit 3 zone differed dramatically from that found immediately and long after amputation--the majority of responsive neurons could be excited only by high-intensity stimulation of small RFs on the digit 3 stump; relatively few cells were sensitive to low-intensity stimulation of adjacent, intact skin regions. Again, no true somatotopic organization was evident. The combined results of these experiments indicate that within 36 weeks following removal of a digit in raccoons, the deprived SmI cortical sector undergoes a dynamic sequence of changes in functional organization: Neurons that are normally excited by stimulation of digit 3 first become responsive primarily to stimulation of digits 2 and 4 (within 1 day after amputation), then to the digit 3 stump (from at least 1-4 weeks after amputation), and finally again to digits 2 and 4 (within at least 36 weeks after amputation).(ABSTRACT TRUNCATED AT 400 WORDS)

Functional reorganization of adult raccoon somatosensory cerebral cortex following neonatal digit amputation.

Amputation of a forepaw digit in raccoons 2-8 weeks of age produced dramatic changes in the functional organization of somatosensory cortex examined electrophysiologically 9-12 months later. The cortical region normally representing the digit that was amputated received widely overlapping input from the entire forepaw, with local disruption of somatotopic organization. Compared with normal animals, the receptive fields of cortical neurons in amputated animals were larger, often included both glabrous and hairy skin, sometimes involved discontinuous skin regions, and were much more variable in peripheral location as a function of recording distance across the cortex and of depth within the cortex.

The pyramidal system of the woodchuck.

The pyramidal system of the woodchuck was examined anatomically and electrophysiologically. The pyramidal tract was found to originate entirely within the anterior half of the cerebral hemispheres and to follow a course typical of most rodents, decussating almost entirely and descending the length of the spinal cord in the ventral part of the dorsal funiculi. It decreased in size uniformly with distance along the spinal cord; most of its fibers terminated in the medial half of the dorsal horn, though they scattered widely and even appeared to terminate on motoneurons. Each tract contained 140,000 +/- 20,000 fibers, with 60-80% of the fibers being about 1 micrometer and 90% being less than 3 micrometer in diameter. Stimulation of the medullary pyramid evoked a minute antidromic potential (alpha wave) which was generally obscured by a large surface-positive response that reversed polarity deep in the cortex and that appeared to be synaptic in origin (r wave). It is proposed that the r wave results from intracortical pyramidal cell collateral activity. Though largest in the apparent region of origin of the pyramidal tract, the r wave also showed local maxima in the forepaw and hindpaw foci of somatosensory cortex. The somatosensory cortex was organized in a manner similar to other rodents, but an "association" area lacking topographical organization was found near the anterior pole of the hemispheres. In an allometric sense, the woodchuck was found to be a "normal" rodent and a "normal" mammal.

Neuron and stimulus typologies in the rat gustatory system.

Although gustatory neurons may be categorized in terms of one or a few characteristics (e.g. 'best stimulus'), such typologies are essentialistic and inconsistent with modern taxonomic methods. If olythetic taxonomic criteria are used and the variability among neuronal responses is closely analyzed, neuronal 'types' are found to disappear, at least within the acid-salt range. This applies to both the primary nerve level (chorda tympani nerve) and secondary level (nucleus tractus solitarius) of the taste system in the rat. In the same context, taste stimuli may fall into different groups if several very similar stimuli are used (e.g. sodium and lithium salts). This is not surprising, and may depend on the choice of stimulus arrays rather than a differentiation of a few stimulus types by the taste system. Finally, it should be noted that the arguments regarding neuron and stimulus typologies presented here for the taste system are also valid for other sensory systems, although the conclusions may be different.

Three-dimensional plotting of neuronal population response patterns.

A computer technique is described for generating three-dimensional plots of the distribution of spike activity in neuronal populations. Data obtained from single neurons in the sensorimotor cerebral cortex of cats are used to plot spike density with respect to depth in the cortex and time following a cutaneous stimulus. The computer programs include a masking (hidden line) subroutine and produce a cross-hatched plot showing the rise and fall of neuronal activity in both space and time. By manipulation of the data matrix and by axis rotations, the plot can be viewed from any desired perspective. This method constitutes a powerful technique for analyzing the spatio-temporal response patterns of neuronal populations. The significance of neuronal population responses for encoding stimulus information and predicting behavior is discussed. Note: Details concerning the computer programs developed for this plotting technique may be obtained from the Division of Systems and Computer Services, Medical College of Georgia, Augusta, GA 30901.