JN Ad Instruments
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

J Neurophysiol 78: 2924-2936, 1997;
0022-3077/97 $5.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (22)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Rasmusson, D. D.
Right arrow Articles by Northgrave, S. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Rasmusson, D. D.
Right arrow Articles by Northgrave, S. A.

The Journal of Neurophysiology Vol. 78 No. 6 December 1997, pp. 2924-2936
Copyright ©1997 by the American Physiological Society

Reorganization of the Raccoon Cuneate Nucleus After Peripheral Denervation

Douglas D. Rasmusson and Stacey A. Northgrave

Department of Physiology and Biophysics, Dalhousie University, Halifax, Nova Scotia B3H 4H7, Canada

    ABSTRACT
Abstract
Introduction
Methods
Results
Discussion
References

Rasmusson, Douglas D. and Stacey A. Northgrave. Reorganization of the raccoon cuneate nucleus after peripheral denervation. J. Neurophysiol. 78: 2924-2936, 1997. The effects of peripheral nerve transection on the cuneate nucleus were studied in anesthetized raccoons using extracellular, single-unit recordings. The somatotopic organization of the cuneate nucleus first was examined in intact, control animals. The cuneate nucleus in the raccoon is organized with the digits represented in separate cell clusters. The dorsal cap region of the cuneate nucleus contains a representation of the claws and hairy skin of the digits. Within the representation of the glabrous skin, neurons with rapidly adapting properties tended to be segregated from those with slowly adapting properties. The representations of the distal and proximal pads on a digit also were segregated. Electrical stimulation of two adjacent digits provided a detailed description of the responses originating from the digit that contains the tactile receptive field (the on-focus digit) and from the adjacent (off-focus) digit. Stimulation of the on-focus digit produced a short latency excitation in all 99 neurons tested, with a mean of 10.5 ms. These responses had a low threshold (426 µA). Stimulation of an off-focus digit activated 65% of these neurons. These responses had a significantly longer latency (15.3 ms) than on-focus responses and the threshold was more than twice as large. Two to five months after amputation of digit 4, 97 cells were tested with stimulation of digits 3 and 5. A total of 44 were in the intact regions of the cuneate nucleus. They had small receptive fields on intact digits and their responses to electrical stimulation did not differ from the control neurons. The remaining 53 neurons were judged to be deafferented and in the fourth digit region on the basis of their location with respect to intact neurons. All but two of these cells had receptive fields that were much larger than normal, often including more than one digit and part of the palm. When compared with the off-focus control neurons, their responses to electrical stimulation had lower thresholds and an increased response probability and magnitude. The latencies of these cells did not decrease, however, and were the same as the off-focus control values. The enhanced responses of the deafferented neurons to adjacent digit stimulation indicate that there is a strengthening of synapses that were previously ineffective. The increased proportion of neurons that could be activated after amputation suggests that there is also a growth of new connections. This experiment demonstrates that reorganization in the adult somatotopic system does occur at the level of the dorsal column nuclei. As a consequence, many of the changes reported at the cortex and thalamus may be due to the changes occurring at this first synapse in the somatosensory pathway.

    INTRODUCTION
Abstract
Introduction
Methods
Results
Discussion
References

The responses of central somatosensory neurons can be modified in adult mammals by altering the inputs from the periphery (reviewed by Kaas 1991; Wilson and Snow 1990). Either disconnection or enhanced use of a body part can alter the topographical organization of the cortex. The discovery that this plasticity can occur in the adult nervous system has renewed investigation into human conditions such as phantom limb sensations (e.g., Mogilner et al. 1993; Ramachandran 1993). The physiological and possibly structural mechanisms responsible for this plasticity are poorly understood. Before these mechanisms can be fully explored, it is necessary to localize the site at which the changes occur. Most studies on this type of plasticity have concentrated on the primary somatosensory cortex, largely because it is relatively easy to access the cerebral cortex. The dominant, and fastest, pathway carrying detailed information from the skin is the dorsal column-medial lemniscal system, which synapses in the dorsal column (gracile and cuneate) nuclei and the ventroposterior (VPL) thalamus. Changes at one of these earlier levels would have a major impact on the somatosensory cortex and may account for some or all of the changes seen in the cortex.

Few studies have examined plasticity in the dorsal column nuclei. Experiments on the gracile nucleus, which receives input from the hindlimb, are limited by the small cross-sectional area of the gracile nucleus (Dostrovsky et al. 1976; LaMotte and Kapadia 1993; Millar et al. 1976). The only major study showing reorganization of the cuneate nucleus was carried out not in adult animals but in kittens (Kalaska and Pomeranz 1982). Experiments on the immediate consequences of peripheral nerve damage found new receptive fields (RFs) for some neurons in the dorsal column nuclei (Dostrovsky et al. 1976; Millar et al. 1976; Panetsos et al. 1995; Pettit and Schwark 1993, 1996). These results show that some inputs to the dorsal column nuclei are masked normally. Although unmasking may provide the baseline from which further plasticity can proceed, it is the long-term changes that are described more accurately by the term "reorganization." The purpose of the present study was to determine if long-term reorganization can be demonstrated in the cuneate nucleus. The digit amputation model in the raccoon was used to permit direct comparison with previous studies in this species (Kelahan and Doetsch 1984; Rasmusson 1982, 1996a,b; Turnbull and Rasmusson 1991).

The cuneate nucleus can be divided into rostral, middle, and caudal regions on cytoarchitectonic grounds (Cheema et al. 1983). The digit regions in the raccoon cuneate nucleus are most detailed in the middle (MCu), or cluster region, which extends for several millimeters caudal to the obex (Johnson et al. 1968). The digit representations, as defined by natural or mechanical stimulation of the skin, correspond to different cell clusters with little or no apparent overlap of excitatory inputs from adjacent digits (Johnson et al. 1968). This specificity provides an opportunity for exploring the changes produced by digit amputation: amputation should completely denervate a single digit region in MCu, with little effect on the adjacent digit regions. If plasticity occurs at this level, the deafferented digit region should become responsive to new inputs.

After digit amputation, the cortical and thalamic regions become responsive to inputs from adjacent digits. This sets some limits on what might be expected in the cuneate nucleus. The greatest degree of reorganization that might be expected in the cuneate would yield new cuneate responses that were comparable with those seen in VPL thalamus (Rasmusson 1996a,b). If, on the other hand, reorganization does not occur at the cuneate level but begins in the thalamus, or is due to changes in the spino-thalamic projection, the deafferented region of the cuneate nucleus should not respond to peripheral inputs. Using this rationale, we have examined the cuneate nucleus after digit amputation and found neurons with response characteristics similar to those of reorganized thalamic neurons. This indicates that the initial site of reorganization in the somatosensory system responsible for fine discriminative touch is the dorsal column nuclei.

    METHODS
Abstract
Introduction
Methods
Results
Discussion
References

Animals

Recordings were made from 25 adult raccoons of either sex. They were trapped in the wild and maintained in a communal room with ad lib food. At the time of recording they weighed between 4 and 12 kg.

Seven raccoons were studied after amputation of the fourth digit of one forepaw. Amputation at the metacarpal-phalangeal joint was carried out under halothane anesthesia, using sterile operating conditions as described previously (Rasmusson 1982). The animals received intramuscular injections of a long-acting antibiotic (cesazolin, 15 mg/kg) and an analgesic (buprenorphin, 0.3 mg/kg) at the end of the surgical procedure. The interval between amputation and recording ranged from 2 to 4.5 mo (60-135 days).

Recording procedure

Recordings were performed while the animal was anesthetized with alpha -chloralose. Induction was carried out using ketamine (100 mg im), and a catheter was inserted into the radial vein of the forearm opposite to the forepaw being studied. Administration of alpha -chloralose (5% iv) was sufficient to maintain an areflexic state, usually consisting of 2 ml initially, followed by 0.5 ml/h thereafter.

To stabilize the preparation, the animal was placed in a Kopf stereotaxic and the neck was ventroflexed. The skull was secured in this position by cementing a bar to the top of the skull using dental caulk. The spine was supported by clamping to one of the thoracic vertebrae and securing this clamp to the stereotaxic frame. The skin and muscles overlying the medulla and cervical spinal cord then were reflected and parts of the atlas, axis, and occipital bone were removed. A dam was created around the opening using gauze soaked in 3% agar before opening the dura. Penetrations were concentrated in the region caudal to the obex where the cuneate nucleus produces a visible bulge on the dorsal surface of the medulla.

Recordings were made using either carbon-fiber electrodes or Parylene-C insulated tungsten electrodes (2 MOmega impedance, A-M Systems, Everett, WA). The signals were amplified using a CWE amplifier (Ardmore, PA) and monitored with an oscilloscope and an audio speaker. Seven raccoons were used to examine the somatotopic organization of the cuneate nucleus. In these animals, the RFs were examined at 50- or 100-µm intervals within each penetration by tapping the skin and deep tissues of the forearm or by touching the skin with fine-tipped glass probes and Semmes-Weinstein von Frey monofilaments (Stoelting, Wood Dale, IL). RFs were drawn on schematized outlines of the forepaw and digits. Thresholds were assessed as being either low or high, and maintained stimulation was used to classify the responses as slowly or rapidly adapting, as described previously (Rasmusson et al. 1991). In each case, several rows of penetrations were made across the cuneate nucleus with 200-300 µm separation. The dorsoventral (DV) height of the representations of distinct forepaw regions was estimated from the number of recording sites with RFs on each part of the forepaw. As this gives only a rough indication of the size of the various representations, statistical analysis was not carried out on these measurements.

In 18 animals, the responses also were tested using electrical stimulation. In these experiments, the electrical signals were saved on a microcomputer using the DataWave A/D and software package (Thornton, CO). Activity was sampled at 11.4 kHz in trials that were 350 ms in duration with the stimulus delivered at 50 ms. The timing of individual spikes was measured with 0.1-ms resolution. Electrical stimulation of the digits was accomplished by inserting two Teflon-coated stainless steel wires (0.003 in) subcutaneously into the distal pad of each of two digits---the same digit that contained the RF for the unit under study (called the on-focus digit) and one of the immediately adjacent digits (the off-focus digit). There is no a priori reason to expect a difference in the on-focus responses for different digits or between the two off-focus digits. In all of the animals with an amputation, both of the adjacent digits (3 and 5) were stimulated to ensure that the most effective new input was tested. Stimulation of nonadjacent digits was not tested in the control animals because reorganization at other levels rarely includes a nonadjacent digit and when it does so has a much higher threshold than on the adjacent digit.

A Maser-8 stimulator (AMPI, Jerusalem, Israel) was used to deliver 0.2-ms constant-current pulses to the digit. Trials were organized in groups of three with 2 s between trials, stimulating first one digit, then the second digit, and then recording with no stimulation to monitor spontaneous activity. At least 20 trials were repeated at a given intensity. All cells were tested at either 500 and 600 µA. If the neuron was activated, the intensity was decreased by 100-200 µA for the next block of trials. Conversely, if it was not activated the intensity was increased by 200-500 µA. In this manner, the threshold for each digit was bracketed. The maximum intensity tested was 3,000 µA. Recording of compound action potentials from the ulnar nerve in the raccoon using the same stimulation technique revealed that the thresholds for Adelta fibers and C fibers were 600 and 2,000 µA, respectively. Their conduction velocities were ~15 and 1.5 m/s. With a distance of 40 cm from digit tip to cuneate nucleus in the smallest animals, the earliest possible influence of Adelta fibers on cuneate neurons would be at a latency of ~25 ms and of C fibers at ~250 ms.

Two to five neurons usually were studied in detail during each penetration. Based on the location of the RFs, subsequent penetrations were made 200-300 µm medially or laterally with the goal of localizing and concentrating on the third, fourth, and fifth digit regions. In preliminary experiments, it was found that marking lesions interfered with subsequent recordings. As a result, lesions were only placed at the end of the last penetration in subsequent experiments. These lesions were used to confirm the rostrocaudal level of the cuneate at which recordings were made.

Data analysis

All data were analyzed off-line using the DataWave software. Separation of two or three units at a single recording site was often possible on the basis of their amplitudes or waveforms using the cluster-cutting feature of this software. Spontaneous firing rates were calculated from all of the trials without stimulation. Responses to digit stimulation were analyzed for each block of 20 trials as detailed previously for thalamic neurons (Rasmusson 1996b). Peristimulus time histograms (PSTHs) were constructed using 1-ms bin size. The latency of the excitatory response was first estimated from a suprathreshold PSTH. The latencies of the first spike in each trial that occurred within a 10-ms window after this estimate then were averaged at each stimulus intensity. The variability of the response was calculated as the standard deviation of these first spikes (SDL). The probability of a response, P(r), was calculated for each intensity as the proportion of trials with at least one spike within this 10-ms window. Response magnitude was estimated by the number of action potentials within this window and averaged only across the trials with a response. Threshold was considered to be the lowest intensity with a P(r) of 0.5. If P(r) was not exactly 0.5 at any intensity, threshold was taken to be halfway between intensities that yielded P(r) values greater and less than 0.5. When comparing the thresholds between groups, cells responding well at <= 200 µA were assigned a threshold of 200 µA. Similarly, cells that responded to >= 2,000 µA with a P(r) of >= 0.2, but did not reach a P(r) of 0.5 were assigned a threshold of 2,000 µA. Neurons that had P(r) <=  0.2 at the highest intensity tested were considered to be nonresponsive. In comparing groups on the basis of all response properties except threshold, the values at 600 µA were used. If the cell was nonresponsive at 600 µA, the values at threshold were used. Inhibition of activity was noted in a few cells, but, for the most part, the spontaneous activity of the neurons was too low to permit detailed analysis of inhibition.

The responses to on-focus and off-focus stimulation in the control animals were compared using a paired Student's t-test. In the amputated animals, the neurons were separated into "deafferented" and on-focus groups as described in RESULTS. Comparisons between cell groups in amputation and control animals and between the deafferented and on-focus amputation groups were made using unpaired Student's t-tests. Group histograms also were compared using the nonparametric Komolgorov-Smirnov test.

    RESULTS
Abstract
Introduction
Methods
Results
Discussion
References

Somatotopic organization of the normal cuneate nucleus

Preliminary mapping experiments were carried out on seven normal raccoons to examine the somatotopy and sequence of RFs in the cuneate nucleus. A total of 41 histologically confirmed penetrations were made in the region caudal to the obex in these animals. The RFs and response characteristics were defined at a total of 1,051 sites.

The top and bottom of the cuneate nucleus could be easily defined from these recordings (cf. Fig. 1). The response characteristics distinguishing the overlying fasciculus from the nucleus were as detailed by Johnson et al. (1968); in particular, the shift from thin, isolated spikes to clusters of initially negative spikes with variable amplitudes was considered to be the dorsal surface of the cuneate nucleus. The ventral surface of the MCu was distinguished by a shift in the RFs from skin to deep RFs that are characteristic of the basal cuneate nucleus (bCu) (Ostapoff and Johnson 1988). Neurons in bCu had high-thresholds and were activated only by strong tapping of the muscles or tendons of the forearm. These upper and lower boundaries were used to delineate the cutaneous representation of the forepaw in MCu. The DV height of the region with cutaneous RFs ranged from 1,000 to 1,500 µm and was occupied primarily by the representation of the glabrous skin of forepaw. The mediolateral organization of the cutaneous representation in MCu was readily apparent in these penetrations, with the ulnar side of the forepaw (i.e., toward the 5th digit) represented medially and the radial side represented laterally.


View larger version (61K):
[in this window]
[in a new window]
  		FIG. 1. Photomicrograph and drawing of the middle cuneate nucleus. Cluster region, which contains the representation of glabrous skin, is labeled MCu. Neurons in the dorsal cap region (CAP) are larger and more widely spread than in the underlying cluster region. Neurons in the basal cuneate nucleus (bCu) are small and diffusely organized. CC, central canal; Gr, gracile nucleus.

DORSAL CAP REGION. The first neurons encountered in MCu within a penetration had RFs either on the dorsal, hairy surface of the digits (in 33 of 41 penetrations) or on one of the claws (7 penetrations). Both hair and claw RFs were seen in 17 penetrations and in each case hair-responsive neurons were encountered dorsal to claw-responsive neurons. This region of hair and/or claw-responsive cells averaged 350 µm in height (range 200-500 µm) and was characterized by action potentials that were consistently larger and more clearly isolated from background noise than in more ventral regions.

The RFs on hairy skin usually included the distal part of the digit. The responses at almost all of these sites (86%) were rapidly adapting; those cells that were slowly adapting (SA) responded well to lateral stretching of the skin, which is indicative of inputs from the SA type II receptors. The thresholds for claw-responsive neurons were very low in almost all (96%) penetrations and the responses were rapidly adapting in 73% of the penetrations. The neurons could be activated best by slight movement of the claw in a dorsiflexion direction and were also very sensitive to light touch on the skin at the sides of the claw. These neurons also could be activated by a light tap of the digit tendons at the wrist, indicating their extreme sensitivity to movement of the claw. In slightly more than half of the penetrations (13 of 24) the RF was on more than one claw. This always included the digit the glabrous representation of which was encountered in the underlying cluster region. This indicates that this dorsal region is organized topographically and is basically in register with the representation of the glabrous skin.

CLUSTER REGION. RFs on the glabrous skin of the digits were found immediately ventral to the hair/claw representation. The glabrous skin of the raccoon forepaw consists of two pads on each digit (distal and proximal pads) that are ~5 mm wide and 10 mm long and six large pads on the palm. Examples of typical RF sizes as well as the progression of RFs within two penetrations are illustrated in Fig. 2. Digit RFs always were restricted to a single digit, and the largest digit RFs, which occupied about one-quarter of the pad, were always on the proximal digit pad. RFs on the palm were much larger, sometimes extending onto more than one pad.


View larger version (23K):
[in this window]
[in a new window]
  		FIG. 2. Progression of cutaneous receptive fields (RFs) in MCu. A and B: data from 2 penetrations in different animals. Numbers beside the RFs are the depths (in µm) below the surface of cuneate nucleus. Shaded region on digit 4 indicates a claw-responsive neuron. Note that on the distal pads the RFs show small movement over several hundred micrometers, whereas on the proximal pads and the palm, there are more rapid shifts in RF location. Also the distal pad is represented over a larger dorsoventral distance than the proximal pad.

Within a single penetration, the RFs always progressed from distal to proximal on a digit and then onto the palm. The progression of RFs on any digit pad was quite gradual (Fig. 2), but large jumps in RF sequence often occurred from the distal pad of the digit to the proximal pad (as in Fig. 2B) or from the digit to the palm (Fig. 2A). Occasionally the RF jumped from the distal pad of one digit to the proximal pad of an adjacent digit, depending on the orientation of the cell clusters relative to the electrode. These jumps usually occurred after a short region with low background activity and no discernible action potentials, suggestive of the fiber bundles separating cell clusters. These RF sequences and the fact that RFs for individual neurons never included more than one digit are consistent with the conclusion that the anatomic clustering of cells corresponds to segregation of information processing from different digits (Johnson et al. 1968) and perhaps from the two pads for each digit. The amount of the cuneate nucleus devoted to the distal digit pad was greater than the proximal digit pad: the median DV height was 300 µm for the distal digit pad and only 200 µm for the proximal digit pad. This and the more gradual progression of RFs on the distal digit, as in Fig. 2, are consistent with the much larger magnification factor for the distal digit at the cortical level (Rasmusson et al. 1991; Welker and Seidenstein 1959).

In the mediolateral axis, the ulnar to radial progression of RFs was apparent across the whole nucleus (from digit 5 medially to digit 1 laterally) as well as within a single digit region. Thus two adjacent penetrations in the same row could have RFs on the same digit, but the more medial penetration always had RFs located more toward the ulnar side of the digit (e.g., Fig. 2B) than the lateral penetration. The palm region also was organized topographically and in register with the overlying digit representations, as illustrated in both penetrations in Fig. 2.

Within the glabrous skin representation, the responses almost always had low thresholds. The number of sites that were slowly and rapidly adapting were approximately equal, and within a single penetration this submodality usually remained the same. Submodality changed within the glabrous zone in only nine penetrations, and in six of these instances, the change occurred at the same point as a large jump in RF that might correspond to a different cell cluster. Thus in only 3 of 41 penetrations was there evidence of mixing of slowly and rapidly adapting neurons within a cell cluster.

Characterization of on- and off-focus responses in control animals

A total of 99 neurons with RFs on the glabrous skin of a digit were analyzed in the control animals. The response properties (group means and SEs) for these cells are presented in the first two columns of Table 1.

 
View this table:
[in this window] [in a new window]
  		TABLE 1. Properties of excitatory responses

All 99 neurons responded to electrical stimulation of the on-focus digit. These neurons characteristically responded at a short latency (mean 10.46 ms) and with a high probability of firing (0.8). Responses to off-focus digit stimulation were tested in 74 of these 99 neurons. Approximately one-third of the neurons tested (26 of 74, 35%) did not respond even when using a stimulation intensity of >= 1,000 µA. The 48 neurons that did show an excitatory response did so at a longer latency (mean 15.33 ms) and with lower probability (0.51) than they did to on-focus stimulation. The results of a representative neuron are illustrated in Fig. 3. This neuron responded to tactile stimulation of digit 5, with a RF similar to that labeled 450 in Fig. 2A. Its threshold for digit 5 (on-focus) stimulation was 450 µA, whereas for digit 4 (off-focus) stimulation threshold was 950 µA. The PSTHs show the large differences in responsiveness to suprathreshold on- and off-focus stimulation. This neuron responded to on-focus stimulation on every trial [i.e., P(r) = 1.0] at a short latency (9.59 ms). The timing of the initial spike was very consistent over trials as seen in the dot rasters (SDL = 0.76 ms). The response to off-focus stimulation (Fig. 3B) was much weaker, with a P(r) of only 0.65 at this higher intensity. The latency was ~7 ms longer than with on-focus stimulation and the variability of the initial spike was much greater (SDL = 2.17 ms). The PSTHs of most control neurons were comparable with this example.


View larger version (14K):
[in this window]
[in a new window]
  		FIG. 3. Peristimulus time histograms (PSTHs) and dot rasters for a control neuron that had a RF on the distal part of digit 5. Stimulation is at time 0 and bin size is 1 ms. Each PSTH is the summation of the 20 trials shown in the dot rasters below. A: stimulation of digit 5 (on-focus) at 600 µA. Threshold for this neuron was 450 µA. Mean latency was 9.59 ms, and the neuron responded on every trial. Small variability in the first spike is reflected in a low standard deviation (SDL), which was 0.76. This stimulation produced an average of 3.8 spikes/response. B: stimulation of digit 4, the off-focus digit. At 1,000 µA, the probability of response[P(r)] was 0.65, but at 900 µA P(r) was 0.4. Threshold for the off-focus response therefore was considered to be 950 µA. Latency for off-focus stimulation was considerably longer (16.68 ms) than for on-focus, and the variability of the response was also much greater (SDL = 2.17 ms). Response magnitude was only 1.7 spikes/response. This pattern of a longer and more variable latency, a higher threshold, and a weaker response to off-focus stimulation was typical of the control neurons.

The distributions of latencies for the control neurons are shown in Fig. 4. The range for responses to on-focus stimulation (Fig. 4A) was 7.6-15.6 ms. The vast majority of these cells (87%) had a latency <12 ms. For off-focus stimulation (Fig. 4B), the histogram is shifted clearly to the right. The mean latency for off-focus stimulation (15.3 ms) was significantly greater than that for on-focus stimulation (paired t-test, t = 4.30, P < 0.001). The few long latency off-focus responses would skew this distribution (the maximum was 37 ms), but the shorter on-focus responses were apparent in both the minimum latencies (7.6 vs. 9.6 ms) and the median latencies (10.5 vs. 13.1 ms). Only two responses had a latency >25 ms and therefore might have received a contribution from Adelta fibers.


View larger version (29K):
[in this window]
[in a new window]
  		FIG. 4. Latency histograms for stimulation of the on-focus digit in control animals (A), the off-focus digit in control animals (B), and the digit yielding the best response in deafferented neurons (C).

There was a distinct shortening of latency with an increase in intensity of on-focus stimulation (Fig. 5, circles). At low intensities, the average latency was ~12-13 ms, and it decreased to 8-9 ms at higher intensities. Thus the intensity used for the majority of comparisons in this study (600 µA) yielded a response that was near, but not at, the minimum latency obtainable with suprathreshold stimulation. The latency for off-focus stimulation also decreased as stimulus intensity increased (Fig. 5, triangle ), however, the off-focus responses were consistently 2-5 ms longer than the on-focus responses at all intensities. Only two neurons had a shorter latency to off-focus stimulation than they did to on-focus stimulation at threshold, but in both cases this was due to the higher intensities that were necessary to reach threshold for off-focus stimulation.


View larger version (13K):
[in this window]
[in a new window]
  		FIG. 5. Relationship between mean latency and stimulus intensity obtained with on-focus digit stimulation (open circle ) and with off-focus stimulation (triangle ) in control animals.

The thresholds for on-focus responses were grouped tightly (Fig. 6A). The average was 426 µA. Only one cell had a threshold >1,000 µA. The distribution of thresholds for off-focus stimulation was much broader, although some had quite low thresholds (Fig. 6B). The mean threshold for off-focus stimulation (905 µA) was more than twice as high as the on-focus threshold. Less than half of the responsive neurons (20 of 48) had thresholds <= 600 µA compared with 90% of the on-focus responses. The differences in threshold obtained with these two types of stimulation were statistically significant (t = 6.87, P < 0.001). Thresholds for individual neurons also were compared as the ratio of off-focus threshold to on-focus threshold. This ratio averaged 1.98, with only six neurons having a lower threshold to off-focus than to on-focus stimulation.


View larger version (19K):
[in this window]
[in a new window]
  		FIG. 6. Histograms of thresholds for the on-focus (A) and the off-focus (B) stimulation in control animals.

On-focus excitatory responses could be elicited very reliably, as indicated by the high probability of a response on each trial [mean P(r) = 0.80] and a low variability in the latency of the first spike (mean SDL = 1.2 ms). The majority of the neurons (53/99) responded on >= 90% of the trials at 600 µA. These neurons also tended to fire multiple spikes, with an average of 3.1 spikes per response. The responses to off-focus stimulation were much less reliable. The first-spike variability was almost twice as high (2.1 ms) as with on-focus stimulation (t = 5.72, P < 0.001). Only five neurons (10%) had an SDL that was <1 ms compared with 46% of the on-focus responses. The probability of responding to off-focus stimulation (0.51) at 600 µA was also much lower than on-focus stimulation (t = 4.76, P < 0.001). The magnitude of the off-focus responses (2.3 spikes/response) was also significantly less than the on-focus response (t = 5.67, P < 0.001).

Most of the neurons showed differences between on- and off-focus stimulation similar to that seen in Fig. 3. The PSTHs for a neuron with more subtle differences are shown in Fig. 7, with the responses to on-focus stimulation shown on the left and to off-focus stimulation on the right. This neuron had only a small difference between on- and off-focus thresholds (250 vs. 350 µA), but there was a clear difference in the magnitude of the responses. It had a much weaker, longer latency response to off-focus stimulation, even when comparing the off-focus response at 700 µA to the on-focus response at 400 µA. In a few cases, one response measure, such as a lower threshold, suggested a stronger response to the off-focus stimulus; however, in such cases, consideration of the other measures made it clear that, overall, input from the on-focus digit was more effective.


View larger version (15K):
[in this window]
[in a new window]
  		FIG. 7. PSTHs for a neuron in the cuneate nucleus that had a RF on digit 2. Off-focus stimulation was tested using digit 1. Response properties for on-focus stimulation at 600 µA were latency = 8.12 ms, SDL = 0.44, P(r) = 1, spikes/response = 6.7; for off-focus stimulation: latency = 13.28 ms, SDL = 1.85, P(r) = 0.65, spikes/response = 1.7. Threshold for on-focus stimulation was 250 µA and for off-focus 350 µA.

The subgroup of neurons in which digit 4 constituted the on-focus digit (n = 32) were virtually identical to the overall group means for all measures except threshold, which was slightly higher (541 µA). The differences between on-focus digit groups for each of these measures were not statistically significant (analysis of variance).

In summary, cuneate neurons in the control animals typically were activated at a short and consistent latency from the digit containing the tactile RF, whereas activation from an adjacent digit was weaker and occurred at a longer latency. The differences between on- and off-focus stimulation were large and consistent in terms of latency and threshold but were more subtle with respect to variability and magnitude of the response. There was also a significant percentage of neurons (35%) that did not respond to off-focus stimulation even at very high intensities, whereas all of the neurons responded to on-focus stimulation.

Amputation animals

A total of 125 neurons were sampled in the animals with amputation of the fourth digit. Twenty-eight of these neurons were in the dorsal cap region (n = 15), with clear claw or hairy skin responses, or were ventral to the digit regions (n = 13), with RFs on the palm or proximal digit. The remaining 97 neurons were analyzed with respect to electrical stimulation as in the control animals.

Neurons were separated into on-focus and deafferented groups on the basis of the location of the neuron within the nucleus using the RFs of neurons within the same and adjacent penetrations as a reference. The stereotyped sequence of RFs within a penetration (Fig. 2) provides one criterion. The denervated region should begin ~500 µm below the surface of the cuneate nucleus with the possibility of some multiclaw cells in the dorsal cap responding to stimulation of the claws of digits 3 and/or 5. Because the palm innervation was intact, as a penetration leaves the denervated region, it should encounter neurons with RFs on palmar pads C or D. A second criterion was created by stepping across the cuneate nucleus in closely spaced penetrations. Penetrations passing through the denervated zone would be medial to penetrations with small RFs on the third digit and lateral to those with RFs on the fifth digit. Only neurons with an appropriate DV depth and mediolateral location were included in the deafferented group. This is a conservative approach to the question of whether reorganization occurs in this nucleus because any neurons that had changed so much that they were indistinguishable from the intact third or fifth digit neurons would not be included in the deafferented group.

Using these criteria, 44 of the 97 neurons were considered to be within intact regions. These neurons had small RFs that were located on digit 2 (n = 5), 3 (n = 20), or 5 (n = 19). All of these RFs were similar in size and shape to those illustrated in Fig. 2. The response properties of these neurons to on-focus stimulation were virtually identical to those of the control animals (Table 1).

The remaining 53 neurons were assigned to the deafferented group. The RFs for these neurons were much larger than any seen in the normal animal. The RFs of six representative neurons are shown in Fig. 8, with darker shading indicating lower thresholds to tactile stimulation. In two-thirds of the cells, the RF included parts of both digits 3 and 5. Most of these RFs extended onto the palm adjacent to the amputation wound (e.g., Fig. 8, D and E). Several neurons had split RFs on the two digits with clear nonresponsive regions in between (Fig. 8, A-C). In those cases with large, continuous RFs, there were often regions of lower threshold within the high-threshold RF as in Fig. 8D. Only 2 of the 53 cells had RFs that were restricted to a single digit pad and were as small as any seen in control animals. These two neurons were included in the deafferented group because cells above and below them clearly were assigned to the deafferented group. The deafferented neurons rarely responded throughout sustained stimulation and therefore would be classified as rapidly adapting. However, even deafferented cortical neurons show a poor ability to respond to repeated activation (Rasmusson 1982). Consequently, this finding does not prove the absence of input from peripheral slowly adapting receptors.


View larger version (41K):
[in this window]
[in a new window]
  		FIG. 8. Representative tactile RFs obtained from animals with a digit amputation. Each figure shows the RF for a single neuron that was classified as in the deafferented group. A-C: split RFs on both digits 3 and 5. D-F: large diffuse RFs encompassing part of the palm in addition to most or all of digits 3 and 5. Darker stippling indicates lower threshold.

When tested with electrical stimulation, all 53 neurons in the deafferented group responded to stimulation of either digit 3 or digit 5 and the majority (72%) responded to stimulation of both. The digit that produced the strongest response at the shortest latency was considered to be the "best digit" for that cell. The mean response properties to electrical stimulation of the best digit are presented in the last column of Table 1. The responses of these cells differed from control on-focus neurons by having a longer latency (15.3 vs. 10.5 ms, t = 6.45, P < 0.001). The latency distribution was in fact not different from that for off-focus stimulation in control animals (Fig. 3C). These differences between the three groups are seen more readily in cumulative frequency plots (Fig. 9A). The Komolgorov-Smirnov test showed that the latency of the on-focus control group was significantly shorter than both of the other groups (P < 0.001). Only 28% of the deafferented neurons and 23% of the off-focus control group had a latency of <= 12 ms compared with 85% of the on-focus control group. The variability of the response in the deafferented group was intermediate between the two control groups (Fig. 9B). It was significantly greater than with on-focus stimulation (P < 0.02), but not significantly different from the off-focus group (P > 0.05).


View larger version (30K):
[in this window]
[in a new window]
  		FIG. 9. Comparison of response properties for on- and off-focus responses in control animals and the responses to the best digit in the deafferented group of neurons. Relative cumulative frequencies plots for latency (A), variability of the initial spike (B), probability of a response (C), and threshold (D).

The other response measures were similar to those of control on-focus responses and significantly different from the control off-focus responses, indicating an improvement in synaptic efficacy (Fig. 9, C and D). The mean threshold for activation was decreased from 905 to 431 µA, the probability of responding was increased from 0.51 to 0.79, and the response magnitude was increased from 2.3 to 3.0 spikes/response. In each of the measurements, on- and off-focus stimulation were significantly different from each other (P < 0.001), but the deafferented group was not significantly different from the on-focus control group (P > 0.05).

Examples of PSTHs for two cells in the deafferented group are shown in Figs. 10 and 11. The neuron illustrated in Fig. 10 had a relatively short latency response (14-15 ms) that was relatively long lasting. The responses to the two digits were similar for all of the response parameters. Figure 11 shows an example of the results from another deafferented cell that had a long latency response (>19 ms) after stimulation of each digit with similar thresholds but with a clear difference in effectiveness. The response to digit 3 stimulation (right) had a shorter latency and a greater magnitude than the digit 5 response.


View larger version (16K):
[in this window]
[in a new window]
  		FIG. 10. PSTHs of a neuron in the deafferented group that had strong and roughly equal responses to stimulation of both digits 3 (left) and 5 (right). Response measures for digit 3 and digit 5 are latency = 14.52 and 14.91 ms; SDL = 1.5 and 0.86; P(r) = 1.0 and 1.0; threshold, 350 and 350 µA; spikes/response = 5.1 and 4.7.


View larger version (11K):
[in this window]
[in a new window]
  		FIG. 11. PSTHs of a deafferented neuron with long latencies and low thresholds to stimulation of both digits 5 (left) and 3 (right). Response measures for digit 5 and digit 3 are latency = 26.4 and 19.5 ms; SDL = 1.63 and 1.33; Pr = 0.65 and 0.8; threshold = 650 and 550 µA; spikes/response = 1.67 and 1.77.

The differences between this group of neurons and the control groups in terms of their RFs and response properties to electrical stimulation indicate that these cells are located in the denervated fourth digit region and have undergone considerable change in their afferent drive. It is also important to note that in the numerous penetrations through the medial part of MCu of the amputation animals, there were no instances of large gaps with nonresponsive neurons, which would be expected if the denervated fourth digit region were silent.

Spontaneous activity

The mean spontaneous rates for the control and amputation on-focus neurons were almost identical (7.9 and 7.0 spikes/s, respectively). Although the mean for the deafferented neurons was slightly higher (9.5), the differences between the three groups were not statistically significant (F = 0.793). Approximately the same percentage of each group could be considered to be silent, with a spontaneous rate <2 spikes/s (Pubols et al. 1989): 28% of control neurons, 30% of intact neurons in amputation animals, and 23% of deafferented neurons. Thus there was no indication that amputation resulted in a permanent change in spontaneous activity.

    DISCUSSION
Abstract
Introduction
Methods
Results
Discussion
References

Organization of cuneate

The overall organization of the cuneate nucleus in mammals has been demonstrated repeatedly using both physiological mapping and anatomic tracing methods. The forepaw is represented with the radial side laterally and the ulnar side medially in all species examined (rat: Beck 1981; Nord 1967; cat: Dykes et al. 1982; Kruger et al. 1961; Kuhn 1949; Nyberg and Blomqvist 1982; raccoon: Johnson et al. 1968; Rasmusson 1988, 1989; squirrel: Ostapoff et al. 1983; primate: Florence et al. 1989). In most of these species, the digits are represented dorsal to the palm projection, the exception being the primate in which the digits are represented ventrally. The most detailed representation of the forepaw is in the MCu, which also has been described as the "cell nest" or cluster region because of the grouping of neurons in this region (Keller and Hand 1970). These cells in MCu are the major relay cells in the projection of cutaneous information to the VPL thalamus (Cheema et al. 1983; Ellis and Rustioni 1981).

The cluster region is enlarged greatly in the raccoon in comparison with other species (Welker et al. 1964), and the digits of the raccoon forepaw are represented in different clusters of neurons (Johnson et al. 1968). This segregation, or parcellation, of digit information in the cuneate was evident in the present study in that tactile RFs on a digit always were restricted to a single digit. However, electrical stimulation of an adjacent, off-focus digit was able to activate 65% of the neurons with digit RFs. This discrepancy between electrical and tactile stimulation is likely due to the high degree of synchronization of the afferent input produced by the electrical pulse as well as the fact that high intensities of electrical stimulation were used. These responses to electrical stimulation of a off-focus digit were characterized by longer latencies and higher thresholds compared with on-focus stimulation, indicating weaker, perhaps multisynaptic connections from off-focus afferents. Anatomic tracing studies showed that the afferents from a single digit terminate only in the appropriate cell clusters and not in adjacent cell clusters (Rasmusson 1988), which makes it unlikely that direct, monosynaptic connections occur within the digit cluster of the on-focus digit. Monosynaptic connections might occur if the dendrites of the cuneate digit neurons extend outside their own cluster and into neighboring clusters. Such monosynaptic connections would be expected to be much less dense than those occurring within the appropriate clusters. The longer space constants of synapses on any such distal dendrites also would help account for the longer latency and weaker responses to off-focus stimulation.

Information from off-focus digits also might arrive at the cuneate nucleus indirectly, via a multisynaptic pathway. This might involve recurrent collaterals from the cuneothalamic neurons ending in adjacent cells clusters. Another possible pathway is the dorsal column postsynaptic pathway (Uddenberg 1968), which might have larger convergence from digit afferents than the direct input to the cuneate nucleus. If either of these pathways is involved, the off-focus connections must normally be subthreshold or else the precise somatotopy of the cuneate nucleus would not be seen when using tactile stimulation. The off-focus inputs could be suppressed actively by local inhibitory interneurons or may not produce sufficient synaptic current to bring the cuneate neuron to threshold in the normal animal. A third possible multisynaptic pathway could involve excitatory interneurons within the cuneate nucleus that cross between clusters. This seems to be the least likely explanation, as the neurons in the cluster regions appear to either be cuneothalamic neurons or inhibitory neurons (Rustioni et al. 1984).

Several aspects of the organization of the raccoon cuneate nucleus that were noted in the present study deserve emphasis. One is that input from the two major subdivisions of the digit, namely the distal and proximal digital pads, may be segregated in different clusters. The major evidence in support of this interpretation is that within a penetration there were consistently breaks in the progression of RFs at the crease between these two digit pads. These breaks coincided with regions of greatly reduced background activity, which would correspond to the cell-poor regions between clusters. A theoretical advantage of parcellation is that it would permit greater interaction within a cluster than between clusters. Separation of inputs from the two parts of the digit into different clusters would suggest that distal and proximal digit regions may have different functions. This is consistent with the known differences between the distal and proximal pads of raccoon digits in receptor density (Pubols et al. 1971; Rice and Rasmusson 1997; Turnbull and Rasmusson 1986), as well as the increased magnification factor for the distal pad at all levels of the somatosensory system (Rasmusson et al. 1991; Welker and Seidenstein 1959). There was also a clear tendency for neurons with slowly adapting or rapidly adapting properties to be grouped together. This spatial segregation of the two submodalities is consistent with the data obtained in the cat cuneate nucleus (Dykes et al. 1982).

The present study also indicates that the dorsal cap region should be considered a distinct subnucleus. The histological distinctiveness of the dorsal region in the cat was noted by Cheema et al. (1983) and in the raccoon by Ostapoff and Johnson (1988). Its neurons are larger and more sparsely distributed than those in the cluster region (Haring and Rowinski 1982; Johnson et al. 1968; Ostapoff and Johnson 1988). The recordings in present study were consistent with these histological differences, as the action potentials in the dorsal cap region were larger and more easily isolated than in the cluster regions. The responsiveness of the cells in this region to hair and claw stimulation is consistent with its innervation by the radial and dorsal ulnar nerves, which arise from the dorsal surface of the forepaw (Rasmusson 1989). The dorsal cap region also has a different projection pattern than the underlying cluster region. Only the dorsal cap neurons project to the cerebellum as well as the thalamus (Haring and Rowinski 1982; Ostapoff and Johnson 1988). For these reasons, the dorsal cap region should be considered to be a separate subnucleus, distinct from the cluster region.

Plasticity in the cuneate nucleus

IMMEDIATE UNMASKING. A number of studies have shown that new RFs can be unmasked immediately after removal of afferent input to the dorsal column nuclei. Blocking the spinal cord or a peripheral nerve during the recording is advantageous because one can observe the same neuron before, during, and after the block. Using cold block of the spinal cord, Dostrovsky et al. (1976) reported that one-quarter of their sample of neurons in the gracile nucleus began to respond to new RFs when their dominant input was silenced. Changes also have been described in cuneate and gracile neurons immediately after deafferentation by injection of lidocaine or capsaicin directly into the RF (Panetsos et al. 1995; Pettit and Schwark 1993, 1996). The appearance of unusual receptive fields also has been reported in some of these experiments. Cutting all but one dorsal root, for example, resulted in a small number of cells with split or double RFs that were never seen in control animals (Millar et al. 1976). Unmasking also can be inferred from studies in which the nucleus is mapped before and immediately after deafferentation and the areas of intact versus deafferented representations are estimated. For example, an enlargement of the area of the gracile nucleus representing the abdomen was seen after cutting the dorsal roots innervating the hindlimb (Dostrovsky et al. 1976). The difficulty with this type of experiment is that it may take several hours to map the nucleus, during which time changes may be occurring. In kitten visual cortex, for example, clear changes in response properties are seen within 6 h of visual deprivation (Mioche and Singer 1989). It is difficult therefore to determine if changes in the map are due to unmasking or to very rapid reorganization.

LONG-TERM PLASTICITY: REORGANIZATION. Before the present study, few experiments have shown any long-term changes in the dorsal column nuclei. Cutting the dorsal roots for the hindlimb in adult cats produced an increase in the amount of the gracile nucleus that is occupied by input from the abdomen (Dostrovsky et al. 1976; Millar et al. 1976). The extensive denervation caused substantial shrinkage of the nucleus, however, so that it is difficult to make firm conclusions based on the topographic maps. Shrinkage also was seen in the cuneate nucleus after complete denervation of the forepaw (Avendaño and Dykes 1995). Forepaw denervation has been shown to produce reorganizational changes in the cuneate nucleus of kittens (Kalaska and Pomeranz 1982), but normal developmental changes may have contributed to the results in that study.

The major challenge in demonstrating reorganization is in the identification of the deafferented region. This was facilitated in the VPL thalamus of the raccoon by the organization of digit regions as a series of nested lamellae (Rasmusson 1996a,b). An electrode therefore could be angled to pass through several successive digit regions, in effect bracketing the deafferented zone by regions with normal RFs and responsiveness. Unfortunately the digit regions in the cuneate nucleus are not organized in such a convenient fashion. The dorsoventral position of the neurons within the cuneate nucleus was useful in determining whether a neuron might be in a digit region, i.e., the digit region must be several hundred micrometers below the surface of the nucleus and any claw-responsive neurons and above the palm representation. The likelihood that a penetration was in the deafferented region also could be based on the responses in adjacent penetrations. These anatomic considerations as well as the distinctive responses that were seen in the group of neurons that we identified as deafferented strongly support the conclusion that cuneate neurons undergo marked changes in their inputs after digit amputation. Finally, the absence of silent "holes" in any of the penetrations also indicates that reorganization does occur in the cuneate nucleus.

It might be argued that the deafferented neurons were in fact in the intact digit 3 or digit 5 regions. Two aspects of the data argue against this conclusion. One is that the size and shape of the RFs for 96% of the deafferented neurons were significantly different from any RFs seen in control animals (compare Figs. 2 and 8). The RFs of deafferented neurons usually encompassed a whole digit or extended beyond a single digit, whereas control RFs were much smaller than a single digit pad. Another argument against this interpretation is that the latency histogram was similar to that for the off-focus control neurons and quite different from the on-focus latency histogram (Figs. 4 and 9A).

Examination of the other response properties in the deafferented group demonstrated a great increase in the efficiency of information transmission from the previous off-focus digit to the deafferented neuron. All of the deafferented neurons responded to stimulation of one off-focus digit (compared with 65% in controls), and a high percentage (72%) responded to both digits 3 and 5. This latter point also argues against the possibility that this sample contained intact digit 3 or digit 5 neurons because these digits are not adjacent to one another. Testing of multiple digit inputs to the cortex of normal raccoons revealed that only one-quarter of the cells that responded to the adjacent digit also responded to a nonadjacent digit (Doetsch et al. 1992). A similar difference in the cuneate nucleus would mean that we would expect only 16% of these neurons to respond to both digits 3 and 5 instead of the 72% that was found.

The deafferented neurons also showed a much lower threshold than off-focus control neurons (431 vs. 905 µA). This is consistent with the low thresholds to tactile stimulation and probably is related to the increase in RF size. There was also an improvement in the measures related to response strength. The increase in the number of spikes/response, the increase in the probability of firing, and the decrease in first-spike variability are all consistent with the idea that the synaptic efficiency from the off-focus digits increases during reorganization.

The latencies of the responses in the cuneate nucleus after reorganization indicate that the new inputs are not carried by Adelta fibers, which likely cannot reach the cuneate nucleus within 25 ms. Similarly C fibers are far too slow to be involved in the new responses, although a role for C fiber input in reorganization has been proposed in the sense of providing a tonic inhibitory control (Calford and Tweedale 1991). Thus the new RFs after reorganization are likely due to activity in the large, Abeta afferents.

Mechanisms of cuneate reorganization

Two general mechanisms have been proposed to explain the type of reorganization seen in the somatosensory system: the growth of new connections and the strengthening of preexisting connections. The strongest evidence for new growth or sprouting in the cuneate nucleus comes from a study by Florence and Kaas (1995) in which they labeled the forearm afferents in primates that had undergone therapeutic amputation of the entire hand. These afferents were found to innervate a more extensive region of both the dorsal horn and the cuneate nucleus compared with normal animals. The encroachment into the digit areas of the cuneate was quite limited, however, possibly due to the extensive amount of the periphery that was denervated. A study looking for sprouting in the raccoon cuneate nucleus after digit amputation did not find any new projections from the adjacent intact digits to the deafferented digit region (Rasmusson 1988).

As mentioned previously, there is evidence in the cat for immediate unmasking of some connections in the cuneate nucleus. Any direct relevance of these unmasked synapses to long-term reorganization depends on the assumption that a RF that is unmasked immediately after deafferentation will be the same RF that is observed several months later. This assumption is difficult to test, particularly when the extent of chronic denervation and temporary inactivation are different. These conditions have been compared more directly in the studies on the raccoon, in which temporary block of the complete digit was used to mimic chronic digit amputation. With long-term reorganization after digit amputation, new RFs appear on the adjacent digits (Kelahan and Doetsch 1984; Rasmusson 1982, 1996a; present study). The only way that unmasking of ineffective synapses could account for this reorganization would be if temporary digit block also revealed new RFs on adjacent digits. However, this was not seen in either the cuneate nucleus or VPL thalamus (Northgrave and Rasmusson 1996; Rasmusson et al. 1993). Ineffective inputs from these digits do exist before denervation (as revealed by electrical stimulation of adjacent digits), but the strength of these ineffective inputs does not change in the cuneate nucleus during complete digit block (Northgrave and Rasmusson 1996). These findings indicate that any ineffective synapses are not suddenly released from tonic inhibition with removal of the afferents from the denervated digit. Clearly, there must be a change in the effectiveness of these weak synapses during reorganization, perhaps via mechanisms similar to those responsible for long-term potentiation in the hippocampus (Bliss and Collingridge 1993). In addition, the greater percentage of neurons that respond to the adjacent digits after reorganization suggests that there is also a growth of new connections. The cuneate nucleus provides a more favorable spatial framework for such interaction to take place than does the cerebral cortex. In the raccoon, cortical digit representations are separated by many millimeters, whereas in the cuneate nucleus the representations of adjacent digits are within several hundred micrometers of each other.

The data presented here indicate that reorganization does occur within the cuneate nucleus after digit amputation. A comparison of some of these changes with those seen in the raccoon VP thalamus and primary somatosensory (SI) cortex is presented in Table 2. Each of these studies used electrical stimulation, but the data in Zarzecki et al. (1993) were based on intracellular recordings and the timing of EPSP onset. At each level, amputation results in a substantial increase in the proportion of neurons activated by off-focus digits. Mean latency was found to decrease in the cortical and thalamus studies, but not in the cuneate nucleus. However, this may result from the fact that the off-focus groups in both the thalamic and cortical studies contained numerous long latency responses that skewed the latency distribution. In the thalamic study, median latency for the off-focus group was much lower than mean latency and was not different from the median latency for the deafferented neurons. An interesting point in Table 2 is that the difference between the cuneate and the thalamic latencies was approximately the same for both the on-focus control and deafferented neurons (2.5 and 2.8 ms, respectively). This suggests that the conduction time from cuneate to thalamus was not increased after reorganization; this might have been the case if additional changes occurred at the thalamic level. The increased probability of firing and the lower thresholds in both the cuneate nucleus and the thalamus after amputation are consistent with the appearance of new RFs that are seen with mechanical stimulation. The similar changes at both the thalamic and cuneate levels for these two measures are consistent with the hypothesis that the major changes are occurring at the cuneate level. Another similarity between these two levels is that almost all of the deafferented neurons in the cuneate nucleus and the cuneate have abnormally large RFs (Fig. 8) (Rasmusson 1996a). This is not the case in the cortex, where many new RFs are as small as control RFs (Kelehan and Doetsch 1984; Rasmusson 1982; Turnbull and Rasmusson 1991). This difference between the reorganized thalamus and cortex indicates that further changes likely occur at the cortical level via either cortico-cortical changes as proposed by Zarzecki et al. (1993) or local inhibitory connections.

 
View this table:
[in this window] [in a new window]
  		TABLE 2. Comparison of responses at cortical, thalamic, and cuneate levels of the raccoon before and after digit amputation

    ACKNOWLEDGEMENTS

  We appreciate the assistance of J. Jordan, L. Tremere, and S. Dick in various aspects of this experiment.

  This research was funded by the Medical Research Council of Canada and the Scottish Rite Charitable Foundation.

    FOOTNOTES

  Address reprint requests to: D. Rasmusson.

  Received 4 June 1997; accepted in final form 5 August 1997.

    REFERENCES
Abstract
Introduction
Methods
Results
Discussion
References
0022-3077/97 $5.00 Copyright ©1997 The American Physiological Society



This article has been cited by other articles:


Home page
 J. Neurophysiol.Home page
R. D. Lane, C. P. Pluto, C. L. Kenmuir, N. L. Chiaia, and R. D. Mooney
Does Reorganization in the Cuneate Nucleus Following Neonatal Forelimb Amputation Influence Development of Anomalous Circuits Within the Somatosensory Cortex?
J Neurophysiol, February 1, 2008; 99(2): 866 - 875.
[Abstract] [Full Text] [PDF]

Home page
 J. Neurophysiol.Home page
R. Hanajima, R. Chen, P. Ashby, A. M. Lozano, W. D. Hutchison, K. D. Davis, and J. O. Dostrovsky
Very Fast Oscillations Evoked by Median Nerve Stimulation in the Human Thalamus and Subthalamic Nucleus
J Neurophysiol, December 1, 2004; 92(6): 3171 - 3182.
[Abstract] [Full Text] [PDF]

Home page
 Proc. Natl. Acad. Sci. USAHome page
S. A. Chowdhury, K. A. Greek, and D. D. Rasmusson
Changes in corticothalamic modulation of receptive fields during peripheral injury-induced reorganization
PNAS, May 4, 2004; 101(18): 7135 - 7140.
[Abstract] [Full Text] [PDF]

Home page
 J. Neurophysiol.Home page
S. A. Chowdhury and D. D. Rasmusson
Corticocortical Inhibition of Peripheral Inputs Within Primary Somatosensory Cortex: The Role of GABAA and GABAB Receptors
J Neurophysiol, August 1, 2003; 90(2): 851 - 856.
[Abstract] [Full Text] [PDF]

Home page
 J. Neurophysiol.Home page
A. S. Stojic, R. D. Lane, and R. W. Rhoades
Intracortical Pathway Involving Dysgranular Cortex Conveys Hindlimb Inputs to S-I Forelimb-Stump Representation of Neonatally Amputated Rats
J Neurophysiol, January 1, 2001; 85(1): 407 - 413.
[Abstract] [Full Text] [PDF]

Home page
 J. Neurosci.Home page
T. M. Woods, C. G. Cusick, T. P. Pons, E. Taub, and E. G. Jones
Progressive Transneuronal Changes in the Brainstem and Thalamus after Long-Term Dorsal Rhizotomies in Adult Macaque Monkeys
J. Neurosci., May 15, 2000; 20(10): 3884 - 3899.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (22)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Rasmusson, D. D.
Right arrow Articles by Northgrave, S. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Rasmusson, D. D.
Right arrow Articles by Northgrave, S. A.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Visit Other APS Journals Online