SI neuron response variability is stimulus tuned and NMDA receptor dependent.

Skin brushing stimuli were used to evoke spike discharge activity in single skin mechanoreceptive afferents (sMRAs) and anterior parietal cortical (SI) neurons of anesthetized monkeys (Macaca fascicularis). In the initial experiments 10-50 presentations of each of 8 different stimulus velocities were delivered to the linear skin path from which maximal spike discharge activity could be evoked. Mean rate of spike firing evoked by each velocity (MFR) was computed for the time period during which spike discharge activity exceeded background, and an across-presentations estimate of mean firing rate (MFR) was generated for each velocity. The magnitude of the trial-by-trial variation in the response (estimated as CV; where CV = standard deviation in MFR/MFR) was determined for each unit at each velocity. MFR for both sMRAs and SI neurons (MFRsMRA and MFRSI, respectively) increased monotonically with velocity over the range 1-100 cm/s. At all velocities the average estimate of intertrial response variation for SI neurons (CVSI) was substantially larger than the corresponding average for sMRAs (CVsMRA). Whereas CVsMRA increased monotonically over the range 1-100 cm/s, CVSI decreased progressively with velocity over the range 1-10 cm/s, and then increased with velocity over the range 10-100 cm/s. The position of the skin brushing stimulus in the receptive field (RF) was varied in the second series of experiments. It was found that the magnitude of CVSI varied systematically with stimulus position in the RF: that is, CVSI was lowest for a particular velocity and direction of stimulus motion when the skin brushing stimulus traversed the RF center, and CVSI increased progressively as the distance between the stimulus path and the RF center increased. In the third series of experiments, either phencylidine (PCP; 100-500 microg/kg) or ketamine (KET; 0.5-7.5 mg/kg) was administered intravenously (iv) to assess the effect of block of N-methyl-D-aspartate (NMDA) receptors on SI neuron intertrial response variation. The effects of PCP on both CVSI and MFRSI were transient, typically with full recovery occurring in 1-2 h after drug injection. The effects of KET on CVSI and MFRSI were similar to those of PCP, but were shorter in duration (15-30 min). PCP and KET administration consistently was accompanied by a reduction of CVSI. The magnitude of the reduction of CVSI by PCP or KET was associated with the magnitude of CVSI before drug administration: that is, the larger the predrug CVSI, the larger the reduction in CVSI caused by PCP or KET. PCP and KET exerted variable effects on SI neuron mean firing rate that could differ greatly from one neuron to the next. The results are interpreted to indicate that SI neuron intertrial response variation is 1) stimulus tuned (intertrial response variation is lowest when the skin stimulus moves at 10 cm/s and traverses the neuron's RF center) and 2) NMDA receptor dependent (intertrial response variation is least when NMDA receptor activity contributes minimally to the response, and increases as the contribution of NMDA receptors to the response increases).

Effects of spinal dorsal column transection on the response of monkey anterior parietal cortex to repetitive skin stimulation.

The pattern of 14C-2-deoxyglucose (2DG) labeling in anterior parietal cortex was evaluated in three groups of experimental subjects: (1) subjects in which all spinal pathways projecting at short latency to the contralateral hemisphere were intact, (2) subjects with either unilateral or bilateral transection of the dorsal column pathway, and (3) subjects in whom a two-stage tractotomy (dorsal column isolation) restricted short-latency mechanoreceptor drive to that conveyed via the dorsal column pathway. Macaca fascicularis and Macaca arctoides monkeys were studied. When the spinal cord pathways projecting at short latency to contralateral anterior parietal cortex were intact, controlled vibrotactile or skin brushing stimuli evoked one or, more rarely, several loci of maximal 2DG uptake (typically 1.5-2.5 mm in diameter) in the topographically appropriate location(s) within area 3b and/or area 1. The labeling at each locus of maximal 2DG uptake extended continuously across layers II-VI. Each locus of maximal 2DG uptake was bordered on one or more sides by irregularly shaped zones of below-background 2DG uptake that could extend without interruption from area 3b into area 3a, and/or from area 1 into area 2. In the absence of skin stimulation, little or no above-background 2DG uptake occurred at any locus within areas 3b and 1 of subjects in which the dorsal column pathway on the opposite side of the spinal cord was intact. In subjects with a complete transection of the spinal dorsal column the global 2DG pattern evoked by a repetitive skin stimulus in contralateral anterior parietal cortex was a near mirror image of the pattern evoked by the same stimulus in intact subjects. In the absence of the dorsal column path, neither 10-25 Hz vibrotactile nor brushing stimulation evoked above-background uptake at the topographically appropriate location(s) within contralateral area 3b and/or area 1. Instead, a prominent region of below-background 2DG uptake occupied the topographically appropriate location in area 3b and/or area 1, and the region of suppressed 2DG uptake was bounded by one or more regions of above-background 2DG uptake that extended from areas 3b or 1 into area 3a and/or into area 2. When a two-stage spinal tractotomy prevented stimulus-evoked short-latency input from reaching contralateral anterior parietal cortex via pathways other than the dorsal column, the 2DG activity patterns evoked in contralateral cortex by either brushing or vibrotactile stimuli were similar to the patterns obtained when the somatosensory pathways on the opposite side of the spinal cord were intact. A neural network model was developed to evaluate the hypothesis that the observed cortical effects of dorsal column transection might be attributable, at least in part, to inhibitory interactions among anterior parietal cortical regions that receive their principal input from different spinal cord pathways. The model incorporated known features of (1) the cortical projection of spinal somatosensory pathways, (2) anterior parietal intrinsic and long-distance horizontal connectivity, and (3) certain neurotransmitter/receptor systems characteristic of sensory neocortex. Simulations of the model network provided results consistent with the idea that repetitive skin stimuli evoke maladaptive, time-dependent corticocortical interactions within anterior parietal cortex contralateral to a dorsal column lesion. The observations indicate that corticocortical interactions account for the (1) near mirror-image pattern (relative to the normal Mexican hat-like pattern) of anterior parietal stimulus-evoked 2DG uptake observed in subjects with a dorsal column lesion, (2) unusual time-dependent response properties of individual area 3b and 1 neurons or neuron populations deprived of dorsal column input (Dreyer et al., 1974; Vierck et al., 1990a; Makous and Vierck, 1994), and (3) abnormal time-dependent characteristics of tactile perception in monkeys with dorsal colum

Minicolumnar activation patterns in cat and monkey SI cortex.

The distribution of stimulus-evoked 14C-2-deoxyglucose (2DG) labeling in primary somatosensory cortex (SI) of monkey (Macaca fascicularis) and cat was investigated. Reconstructions of the global pattern of labeling reveal that discrete skin stimuli evoke activity within an extensive region of SI, and that the activation pattern typically consists of multiple, elongated regions of above-background labeling ("modules," typically 0.5-1.0 mm wide, and 1-4 mm long). Evidence obtained using recently developed methods (Tommerdahl, 1989) for quantitative analysis of 2DG activity patterns is shown to be consistent with the idea (Whitsel et al., 1991) that SI modules typically are bounded by zones dominated by stimulus-evoked inhibition. The labeling pattern within individual 2DG modules in SI of both cats and monkeys is analyzed quantitatively (in the frequency domain). Within-module spatial activation patterns are demonstrated to be periodic, consisting of radially oriented profiles of above-background labeling separated from each other by less strongly labeled radial profiles. The spectral characteristics of within-module 2DG labeling change systematically with location along the module's long axis: spatial frequencies between 18 and 35 cycles/mm are prominent in the labeling that occupies both the middle and upper layers at central locations in the module, but are a less obvious component of the labeling in both the middle and upper layers at locations remote to the module center. Since the radially oriented periodic variation both (1) in 2DG labeling in regions of SI outside modules and (2) in optical density in images of Nissl-stained sections of SI consists predominantly of spatial frequencies in the range of 18-35 cycles/mm, it is concluded that the radial profiles of labeling within individual 2DG modules correspond to groupings of minicolumns distinguishable from their neighbors on the basis of labeling intensity. The findings raise the possibility that highly structured, within-module spatial patterns of SI minicolumnar activation encode information about the physical properties of tactile stimuli.