Brain Research, 538 (1991) 127-135

127

Elsevier BRES 16217

Second somatosensory cortical area in macaque monkeys: 2. Neuronal responses to punctate vibrotactile stimulation of glabrous skin on the hand Harold Burton and Robert J. Sinclair Department of Anatomy and Neurobiology, Washington University School of Medicine, St. Louis, MO 63110 (U.S.A.) (Accepted 7 August 1990)

Key words: Tactile stimulus; Primate; Parietal cortex

Single neuron responses from the second somatosensory cortical area (SII) of macaque monkeys were studied using computer-controlled vibratory stimuli ranging in frequency from 10 to 300 Hz. Results were obtained using chronic recording techniques in awake or lightly tranquilized animals. Most neurons were unable to follow the temporal order of vibrations in excess of 10 Hz. A smaller sample of cells provided faithful reproduction of frequencies up to 50-75 Hz and another responded to low amplitude, high frequency stimulation in excess of 100 Hz. Cells that displayed temporally cohesive responses to lower frequencies demonstrated predictable, time-locked discharges to successive stimulus cycles. Cells activated by higher frequencies showed a lower probability of following successive stimulus cycles. These findings are discussed in reference to the hypothesis that SII may provide a parallel channel for processing high frequency vibrotactile inputs from Pacinian receptors. INTRODUCTION Functions of the second somatosensory cortex (SII) are not known. Ferrington and colleagues 2,1° suggested that SII in cats may be especially concerned with processing the temporal dynamics of high frequency vibrotactile inputs from pacinian corpuscles. There has b e e n n o convincing evidence of rhythmic firing patterns to high frequency stimulation in primate SI although a rate code has been noted 25. It appeared possible that comparable sensitivity to pacinian-like inputs would therefore exist in SII of primates, supporting the hypothesis of parallel processing of somatosensory inputs from different receptor classes. A problem with this notion is the idea that SII may depend on or even receive its sensory inputs through SI in primates 7,14,18,22,26,27. According to this hypothesis, SII is part of a hierarchical link that funnels processed signals from SI into higher cognitive and m e m o r y centers in the brain. Consequently, with n o temporally coded activity in SI, it seems unlikely that such responses would occur in SII. This paper will show that some n e u r o n s in SII reslSond to vibrotactile oscillations of higher frequencies but that the majority of SII n e u r o n s activated by vibration stimuli respond best to low frequency stimuli.

MATERIALS AND METHODS Most of the techniques used have been describeda,6. Single neuron recordings were obtained from SII in normal adult Macaca fascicularis (9 cases) and M. mulatta (2 cases) monkeys. Additional observations were made in two rhesus monkeys (1 infant and 1 adult) that had sustained lesions to the ipsilateral SI approximately 1 year prior to the recording experiments in SII. The lesions in these cases (Mml0 and Mml7) have already been illustrated 7. Stainless steel chambers were centered over SII as noted previously6. All surgery procedures were done with strict adherence to sterile techniques. Animals used in these studies had been acclimated to the experimental environment by positive conditioning with food and fruit-flavored water. Consequently, searching for and isolation of most cells was accomplished while the monkeys were awake. During most recordings, tranquilizing doses of xylazine (1 mg/kg) plus ketamine (2-3 mg/kg) were injected intramuscularly or a mixture of 70% nitrous, 30% oxygen and 0.5-1.0% halothane was administered through an inhalation mask. Data was obtained from -15 cells without any drugs from three very tame animals. There were no systematic differences between the behavior of cells recorded in awake and lightly tranquilized animals.

Cell identification Neurons studied with vibrotactile stimulation were a subset of those identified previously6 as having low threshold, cutaneous receptive fields on glabrous skin of the contralateral hand and digits. Receptive fields were initially determined using hand-held probes. Cells that appeared to be particularly sensitive to vibratory 3timuli were sought by first applying a vibrating tuning fork of different frequencies (64, 128, 256 and 512 Hz) to or near the receptive field. Responsive cells were further evaluated using controlled stimulation. A cell's action potentials (spikes) were isolated and separated into

Correspondence.. H. Burton, Department of Anatomy and Neurobiology, Washington University School of Medicine, 660 South Euclid Avenue, St. Louis, MO 63110, U.S.A. 0006-8993/91/$03.50 © 1991 Elsevier Science Publishers B.V. (Biomedical Division)

128 stochastic events using an amplitude-time window (BAK Electronics, DIS-I); the time intervals between spikes were stored with 100 Ms resolution, together with the times of zero crossing of vibratory stimulus cycles, on a PDP 11/34 computer (Digital Equipment) for subsequent analyses 1.

discharge trains was sufficient to predict sequential events. Cells with highly entrained discharges showed a high probability (ordinate axis of autocorreiogram) of peaks in their autocorrelograms at time intervals equal to the period of the simulating frequency for 10 or more successive cycles. Cells with these characteristics also displayed high percent entrainments and unimodal peaks during one-half of the time interval for cycle histograms. Average firing rates were compared during stimulation at different vibration amplitudes and frequencies to determine whether a simple rate code provided identification of vibration frequency. Degree of entrainment was numerically evaluated by calculating percent entrainment, which indicated what proportion of a cell's discharges occurred during one half of the cycle period 2. The number of impulses per cycle also provided a measure of entrainment.

Stimulation parameters All vibrotactile stimulation consisted of sinusoidal fluctuations in amplitude around an initial skin indentation of 400 #m. Peak vibration amplitude was dynamically controlled, cycle-by-cycle through digital feedback to a peak detector whose output was compared to expected displacement for a specified amplitude. Consequently, most vibration amplitudes were delivered at specified levels within a few cycles of stimulus initiation. The skin was contacted orthogonal to its surface by a 2-ram diameter flat or conical cylinder (with either a 30, 45 or 60 degree taper) made of Delrin. Frequency sensitivity was first tested using a series of vibrotactile stimuli ranging from 10 to 300 Hz generally at 50 #m peak amplitudes. Subsequent stimulation consisted of testing with different frequencies presented in ascending amplitudes of vibration, with each amplitude step applied for 2-4 s (10-50 #m in 10 Mm increments or 10-110 Mm at 25 Mm increments). Depending on responsiveness, records were obtained from 2-5 repetitions of each amplitude-frequency combination.

Histology The location of recording sites was illustrated previously6. All of the cells described here had receptive fields on the hand and digits and were judged, based on previous criteria 5 and histological localization, to be from the hand and digit region of SII proper. RESULTS R e s p o n s e s f r o m m o r e t h a n 500 n e u r o n s w e r e t e s t e d for sensitivity to v i b r o t a c t i l e stimuli. M o s t cells w e r e rela-

Data analysis

tively insensitive to any v i b r a t o r y f r e q u e n c y . A s a m p l e of

Analyses focused on whether responses provided temporal characteristics of the stimulus as defined previously32. This was evaluated using cycle, interval and autocorrelation histograms. These showed, respectively, whether a cell preferentially fired during one phase of the stimulus cycle, whether this firing followed integer multiples of the stimulus period, and whether rhythmicity in

A

20

- 1 2 0 cells s h o w e d s o m e sensitivity to v i b r o t a c t i l e stimulation and m o s t of t h e s e r e s p o n d e d o n l y to v e r y low f r e q u e n c i e s ( 5 - 1 0 H z ) p r e s e n t e d at a m p l i t u d e s in excess of 50 Mm. A s these cells w e r e r e l a t i v e l y insensitive to

20

50~m

•

/x \

i / / u

x

IOHz I u ---------I 50Hz

15 A ~ A

. ____.

A--

~ , / ~ - ~

- A / A

"1000

Hz I

10

-

05

&&

05

LI~

300Hz

AM45N5

00 0

i SO

i

100 Vibtolion

B

150

i

i

200

250

00 300

25 50 75 V i b r o l i o n Ampl i l u d e

Frequency (flz)

100 (~m)

30 IIOHz

10Hz

25 3

~.

20

20Hz

/ i _ _ i / I " 2 / ~ 5 0 H z --

I

75, 100Hz

O5 O0 Vibrolion

50 Amplitude (~m)

100

50 100 V i b r o t i o n Amplitude (/~m)

Fig. 1. Responses tuning by stimulus cycles. A: impulses/cycle are plotted from three low frequency cells to 50 Mm vibrations at various frequencies. Some responses followed cycle by cycle, but only for the lowest vibration frequencies. B-D: impulses/cycle for three additional cells showing that entrainment to the lowest frequencies was usually evident at minimal vibration amplitudes. Increased stimulus amplitudes recruited additional impulses/cycle, but again only for low frequencies. D: responses from this exceptional cell were able to follow higher frequence stimuli (75 and 100 Hz) at higher amplitudes. All records obtained from lightly tranquilized M. fascicularis.

129 10=m 5Hz

30~,m

C

t[LLII,t,L,[,L,I A

B

1 1 1 1 1

=

15Hz

~"

400

200

75Hz

150

35,m

20HZ

1000

20Hz

1000

30Hz

500

30Hz

500

50Hz

330

50Hz

20(]

100

50~m 10Hz

•

2000

25Hz

10Hz

100

330

D

2OHz

50Hz

100Hz

200

500

200

100

55

50

Time (ms) Fig. 2. Autocorrelation histograms for 4 different low frequency cells. Histogram display times have been manipulated to encompass 10 successive stimulus periods for different stimulus frequencies. A: cell M29T05 followed nearly every stimulus cycle and frequently responded during the other half-cycle at 5 Hz. It showed some progressive decline in rhythmicity up to 75 Hz and very little vibratory sensitivity at 95 Hz. Stimulus amplitude for all frequencies was 10 pm. B: cell M45N5 responded asynchronously due to multiple discharges during both phases of each stimulus cycle at 10 Hz. It followed, with declining probability, up to 50 Hz and showed no entrained activity to 75 Hz. Stimulus amplitude for all frequencies was 30/~m. C: cell M45G6 discharged during on and off phases of a 10-Hz vibration, followed 20 and 30 Hz, skipped some successivecycles at 50 Hz, and failed to follow 70 Hz. Stimulus amplitude for all frequencies was 50pro. D: cell M29U01 appeared to follow 50 and 100 Hz more rhythmically than 20 Hz due to multiple discharges to each cycle at the lower frequency, and still maintained some following to 200 Hz. Stimulus amplitude for all frequencies was 35 pro. All records obtained from lightly tranquilized M. fascicularis.

vibratory stimuli they were not studied further. A second group of 79 cells followed a broader range of lower frequencies (10-75 Hz) and a third group of 23 were selectively activated by higher frequencies (100-300 Hz). Detailed analyses were made for 15 of 79 and 13 of 23 cells in the latter two groups for responses obtained to low stimulus amplitudes. All studied cells had maiialy poor to moderately defined receptive fields on the glabrous hand or digits that were typical of those described previously in SII of primates 5. No distinctions were seen between the receptive fields of cells with different sensitivities to vibratory stimuli.

Low frequency cells These cells usually showed progressively declining sensitivity to higher frequency stimuli. As shown by 3 examples in Fig. 1A, only the lowest frequencies engaged responses for every stimulus cycle even at higher stimulus amplitudes (Fig. 1B,C). A rare exception to this inability to respond cycle by cycle for frequencies over 50 Hz is shown in Fig. 1D. For most cells, a steep fall in sensitivity occurred to frequencies higher than 30 Hz. Autocorrelation histograms (ACC) of these responses, however, indicate that despite a general inability to follow frequencies cycle by cycle, discharge sequences

preserved the rhythmicity of most lower stimulus frequencies at fairly low stimulus amplitudes (Fig. 2). Some cells were so responsive to the lowest frequencies that they discharged during both phases (half cycles) of vibration (Fig. 2A, 5 Hz; 2B, 10 Hz; 2C, 10 Hz). Where multiple spikes occurred to a low frequency stimulus, timing could be disrupted since the duration of the discharge carried into the next half cycle (Fig. 2B). Most low frequency SII cells (e.g. Fig. 2A-C) displayed very low probabilities of rhythmic following that replicated the period of the stimuli at or above 75 Hz. An exception, shown in Fig. 2D, demonstrated a high degree of entrainment even to 200 Hz, although this cell's best frequency was