INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Tactile inputs from the different classes of skin receptors are conveyed to the cerebral cortex where they activate multiple processing areas in the contralateral hemisphere of most mammalian species. Two of the principal areas, in both primate and nonprimate mammals, are the primary and secondary somatosensory areas, known as SI and SII, respectively. As both SI and SII receive direct anatomical projections from the ventroposterior (VP) thalamus (for review see Jones 1985), it has been proposed that cortical somatosensory processing depends on parallel, distributed processing in SI and SII (Mountcastle 1978, 1986; Rowe et al. 1985).

There has been unequivocal experimental support for this hypothesis in a variety of nonprimate species, in particular, in cat (Burton and Robinson 1987; Manzoni et al. 1979; Turman et al. 1992), rabbit (Murray et al. 1992), possum (Coleman et al. 1999), and in the tree shrew and the prosimian galago (Garraghty et al. 1991). Furthermore, there is experimental support in the cat for the hypothesis that SI and SII occupy hierarchically equivalent positions in a parallel, distributed scheme for thalamocortical processing. This evidence was obtained in studies based on an examination of tactile responsiveness, first, in SII when SI underwent reversible inactivation by localized cortical cooling (Turman et al. 1992), and, second, in SI, when the procedure was reversed, that is, with SII reversibly inactivated by localized cooling (Turman et al. 1995). The results in each area were very similar; the majority of neurons were unaffected in their peripheral responsiveness, although a proportion showed some reduction, thought to be attributed to the removal of reciprocal facilitatory influences exerted via the intracortical connections between SI and SII, or via similar subcortical influences emerging from SI and SII (Ghosh et al. 1994; Turman et al. 1992).

In the case of simian primates, the organization of thalamocortical somatosensory systems appeared, from studies in the late 1980s, to be very different, with reports that SII responsiveness was abolished by surgical ablation of SI (Burton et al. 1990; Garraghty et al. 1990; Pons et al. 1987, 1992). This dependency of SII responsiveness on SI provided evidence for a serial processing scheme in which tactile information is conveyed from the thalamus to SI and then to SII via intracortical connections (Garraghty et al. 1990; Pons et al. 1987, 1992) and led to the hypothesis that there are fundamental differences between simian primates and other eutherian mammals in the organization of thalamocortical systems for tactile processing (Garraghty et al. 1991; Mackie et al. 1996; Murray et al. 1992; Turman et al. 1992). However, as the ablation method used for SI inactivation is clearly irreversible and does not permit examination of SII responsiveness (either in terms of evoked potential or individual neurons) before, during, and after SI inactivation, we have utilized the reversible, localized cortical cooling procedure to re-investigate the issue of SI-SII organization in the marmoset (Zhang et al. 1996). This study demonstrated direct thalamic input to SII indicating that parallel organization of SI and SII applies in this primate species as it does in nonprimate species.

The purpose of the present study was to test the hypothesis that the SI and SII areas occupy hierarchically equivalent positions in a parallel, distributed scheme for thalamocortical processing in the marmoset. The localized cortical cooling procedure has been used once again, in this case to selectively and reversibly inactivate the SII area of cortex to determine whether the incidence of altered responsiveness among SI neurons differs from that observed earlier among SII neurons when SI is inactivated (Zhang et al. 1996). Although there is general agreement that peripheral tactile inputs project directly to SI from the thalamus in the marmoset (Höhl-Abrahão and Creutzfeldt 1991; Krubitzer and Kaas 1992), it is only with the use of rigorously controlled tactile stimuli and with precise quantification of neuronal responsiveness that one can identify, first, whether any component of SI responsiveness to the stimulus-generated tactile input might be dependent on an indirect, serial path from the thalamus to SI via SII, and, second, whether SI responsiveness to these peripheral inputs is modulated by influences from SII. Preliminary accounts of the results have been reported in abstract form (Zhang et al. 1998).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal preparation

The experiments were carried out in 17 adult marmosets (Callithrix jacchus) of either sex (200-300 g body wt) and conformed with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes, the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and the Guiding Principles in the Care and Use of Animals approved by the Council of the American Physiological Society.

Anesthesia was induced with halothane followed by intramuscular injection of ketamine (25 mg/kg) and xylazine (2 mg/kg), and maintained with intravenous infusion of ketamine (~20 mg · kg¹ · h¹) mixed with either xylazine (~1 mg · kg¹ · h¹) or diazepam (~1 mg · kg¹ · h¹) in 0.18% sodium chloride containing 4% dextrose. Dextran 40 (10% in normal saline, iv) was administered if blood pressure fell below 80 mmHg to maintain satisfactory cerebral circulation. Rectal temperature was maintained at 38 ± 0.5°C.

The anesthetic infusion rate was adjusted depending on anesthetic depth as evaluated by autonomic indexes of blood pressure and heart rate, and by tests for the presence of the flexion reflex. To minimize the use of the infused anesthetic agents, a 4:1 ventilation mixture of N₂O and O₂ (at times supplemented with halothane 0.5-2% adjusted according to the depth of anesthesia) was given through tracheal intubation, and diazepam (~1 mg/kg) was administered intramuscularly at 1- to 2-h intervals. All incision sites except the scalp were smeared with lignocaine cream in an attempt to minimize nociceptive inputs and reduce the level of anesthetic agents needed. Atropine sulfate (0.08 mg/kg) was administered subcutaneously at the outset to minimize respiratory secretions, and dexamethasone phosphate (Decadron, 1.5 mg/kg im) was given to reduce the risk of brain edema. The femoral artery and vein were cannulated for the purpose of monitoring blood pressure and for administration of fluid and anesthetic agents. The animal was placed in a stereotaxic frame, and a craniotomy was performed to expose the frontoparietal cortex. A skin pool filled with warm (38°C) liquid paraffin was set up to prevent the cortex from drying. The distal forelimb and hindlimb on the side contralateral to the craniotomy were shaved and fixed in a perspex trough or plasticine mold to stabilize the limbs and allow accurate positioning of the mechanical stimulator. For some animals used in the present cortical study, anesthesia was maintained for a further 12-18 h, during which an investigation of response properties of tactile afferent fibers was conducted (Coleman et al. 2001). At the end of the experiment, an overdose of sodium pentobarbitone was administered.

Mapping of distal limb representations in SI and SII

The exposed cerebral cortex was photographed and mapped initially by means of evoked potential recording from the surface with a platinum ball electrode (1-mm tip diameter) to determine the locus of representation for the contralateral hand and foot in SI and SII (Fig. 1). These were identified as the areas where short-latency (usually <10 ms for forelimb and <15 ms for hindlimb), initially positive-going evoked potentials were elicited by tap stimuli (3 ms, 200-800 µm amplitude) applied to the central palm or sole by means of a feedback-controlled mechanical stimulator (Turman et al. 1992, 1995; Zhang et al. 1996).

View larger version (71K): [in this window] [in a new window]   Fig. 1. Experimental arrangement for reversible inactivation of SII. A: a parieto-temporal craniotomy was performed to expose the primary (SI) and secondary (SII) areas of cortex. B: hand and foot representations within SI and SII were determined by recording short-latency, positive-going evoked potentials from the cortical surface in response to a brief tap stimulus (3-ms, 400-µm amplitude) delivered either to the contralateral forelimb (FL, left traces) or hindlimb (HL, right traces) at the time indicated by the arrows. The recording locations of the evoked potentials correspond to the labels in A. C: a silver cooling block, 5 mm diam and equipped with a Peltier element, was positioned to cover both the forelimb and the hindlimb representations in SII.

Microelectrode mapping of the SII region was also carried out in detail (200- to 400-µm separation of electrode tracks) in 13 of 17 experiments to locate precisely the SII hand and foot areas. Together with evoked potential mapping, it revealed that these representations are located at the upper convexity of the lateral sulcus and are present largely in the exposed surface regions of cortex. The SII hand and foot representations revealed by evoked potential mapping were, if anything, slightly larger than those revealed by microelectrode mapping. However, for these experiments, we did not distinguish the so-called parietal ventral (PV) area (Krubitzer and Kaas 1990) from "SII proper" as we found no dichotomy in the focus of evoked potentials from a given peripheral source within this lateral somatosensory region of cortex. Furthermore, there is no cytoarchitectonic distinction between SII and PV, nor has any difference been identified in the source of thalamocortical inputs to these areas (Krubitzer and Kaas 1992). In the present experiments the 4- to 5-mm diameter cooling blocks used would have covered both of these putative divisions of the lateral somatosensory area that we identify as SII (Zhang et al. 1996).

Inactivation of SII by cooling

Localized cooling provides a well-established method for selective and reversible block of transmission within particular regions of the cerebral cortex or other areas of the CNS (see review by Brooks 1983 and Turman et al. 1992; Zhang et al. 1996). The procedure is known to achieve a block of synaptic transmission at temperatures of ~20°C, well above the temperature of ~5°C needed for axonal conduction block and the temperature of ~0°C, where permanent damage may be induced in the neural tissue (Brooks 1983). The localized cooling procedure has been used in many previous investigations of cortical interactions (e.g., Brinkman et al. 1985; Girard and Bullier 1989; Hupe et al. 1998; Sandell and Schiller 1982; Sherk 1978) and has been used previously in our laboratory for examining interactions between different somatosensory cortical areas, in particular, for examining SII responsiveness when SI was inactivated in the cat (Turman et al. 1992), rabbit (Murray et al. 1992), marmoset (Zhang et al. 1996), and possum (Coleman et al. 1999), and for examining SI responsiveness in the cat when SII was inactivated (Turman et al. 1995).

In the somatosensory areas of cortex in the marmoset, the hand areas of SI and SII are separated by ~4 mm, with an intervening facial representation (Zhang 1994; Zhang et al. 1996), a distance that allows the cooling method to be used in one area without significant spread of cooling to cause disruption to neural activity in the other area (Hupe et al. 1998; Zhang et al. 1996). When the SI surface temperature was lowered to 8°C by a 5-mm-diam block, cortical tissue at a distance of 4 mm had a small temperature drop from 38 to 35°C [see Fig. 2 in Zhang et al. (1996), in agreement with measurements by Lomber et al. (1996) of a temperature gradient of 10°/mm]. Temperature changes of this magnitude cause no reduction in responsiveness of cortical neurons according to previous studies (Sherk 1978; for review, see Brooks 1983; Turman et al. 1992; Vanduffel et al. 1997). Whenever lower surface temperatures (4-5°C) were required to ensure inactivation of SII in the present experiments (perhaps on account of the SII representation extending a little down the upper bank of the lateral sulcus), the SI recordings were made within the hindlimb representation area that had a separation of 7-8 mm from the SII cooling block (Fig. 1).

The inactivation of the hand and foot areas of SII was accomplished in the present experiments by localized cooling with a cylindrical silver block (Fig. 1), placed in contact with the cortical surface over the regions identified as the hand and foot areas of SII from the mapping study. The block was 5 mm diam (covering an area of ~20 mm²) in all experiments except one where the identified body representation was smaller than usual, permitting the use of a 4-mm block. A Peltier thermoelectric device was attached to the upper surface of the block, and the temperature at the face of the block was monitored by means of a thermistor embedded in the block.

Evoked potentials were recorded from the SII cortical surface beneath the block with a platinum ball electrode inserted through a hole in the block (Fig. 1). Heat generated at the upper surface of the Peltier element was removed by circulating cold alcohol (15°C), a procedure that conferred a faster response time on the cooling device. The temperature at the face of the block overlying SII was normally held at 38°C (control temperature) but could be changed within 1-3 min to 5-10°C, the temperature used for the inactivation of forelimb and hindlimb areas of SII. Rewarming to the control temperature had a similarly rapid time course.

The platinum electrode in the cooling block was used to record evoked potentials from SII, often from both hand and foot representations on account of their proximity (for example, position J in Fig. 1). A second surface electrode was placed on either the hand or foot representation in SI (Fig. 1A), and evoked potentials were recorded simultaneously from SI and SII by averaging 20-60 successive responses in a laboratory computer. Inactivation of the SII area was confirmed by the disappearance of the positive-going cortical evoked potential under the cooling block (Fig. 2).

View larger version (27K): [in this window] [in a new window]   Fig. 2. Effect of reversible SII cooling on SI responsiveness as assessed by evoked potential recording. Evoked potentials were recorded simultaneously from SII (left column) and SI (right column) in response to a tap stimulus (arrow, 3 ms, 600 µm) applied to contralateral hand (top traces) and foot (bottom traces). Each trace displayed was averaged from 40 successive sweeps. The amplitude of SI-evoked potentials appeared unaffected by SII inactivation (8°C for hand, and 5°C for foot), although there was some prolongation of the response. SII responses recovered fully on rewarming to 38°C.

Microelectrode recording at the margin of the SII cooling block was also carried out to confirm the effectiveness of the cooling procedure for inactivation of SII, in particular, for any part of the SII body representation located within the lateral sulcus. These recordings, at depths up to 4,270 µm, in oblique tracks next to the cooling block revealed an abolition of responsiveness in the individual SII neurons usually at a higher temperature than that required for abolition of the SII-evoked potential in the same experiment. For example, in the case illustrated below in Fig. 3, although cooling to 10-12°C abolished responsiveness of the SII neuron to tactile stimulation of the contralateral hand, 8°C was needed to abolish the SII evoked potentials. As the initial, positive-going component of the cortical-evoked potential reflects the excitatory postsynaptic potentials (EPSPs) generated by synchronous thalamocortical input, it might be expected that the evoked potentials would be more resistant to cooling than the neuronal spike activity that depends on the EPSPs exceeding threshold. Nevertheless, it was the temperature required in each experiment to abolish the SII-evoked potential that was always used to study the effect of SII inactivation on SI responsiveness.

View larger version (30K): [in this window] [in a new window]   Fig. 3. Abolition of responsiveness of an SII neuron located beneath the cooling block as the temperature of the cortical surface was twice cooled from 38°C to 10-12°C. This SII neuron, recorded at a depth of 2,650 µm in an oblique electrode penetration (~60° to the cooling block), had a receptive field that spanned digits 2-4 and responded to a 5-Hz train of rectangular pulses (100 µm, superimposed for 1 s on a 300-µm step indentation; A, bottom). On both occasions, responsiveness was abolished by cortical cooling. Response level (imp/s) is plotted in B () along with background activity (), and response traces are displayed in A (duration 2.5 s). Cortical spreading depression (CSD) was observed soon after the SII temperature was returned to 38°C following the 2nd cooling episode.

Analysis procedures

The effects of inactivation of SII on the responsiveness of SI neurons were also examined by evaluating stimulus-response (S-R) relations before, during, and after SII inactivation (see Fig. 6), by constructing peristimulus time histograms (PSTHs), and by examining the extent of phase locking in the responses of SI neurons to vibrotactile stimuli. Details of these analyses are given in Zhang et al. (1996). Quantitative measures, in particular of percentage entrainment, were derived from the cycle histograms to evaluate the effect of SII inactivation on temporal patterning in the SI responses to vibrotactile stimuli. The percentage entrainment represents a quantitative measure of the tightness of phase locking and is the highest percentage of impulse occurrences in any continuous half-cycle of the cycle histogram distribution (Ferrington and Rowe 1980; Ferrington et al. 1987a,b; Mountcastle et al. 1969). Percentage entrainment values range from a minimum of 50%, in a histogram with a rectangular distribution, indicating an absence of phase locking, to 100%, in a distribution in which all impulse counts are confined within a continuous half-cycle period.

The final measure obtained was the resultant R, which ranges from 0 to 1 and is a measure of phase coherence in a cyclic distribution such as a cycle histogram (Mardia 1972). This angular or directional statistic was developed for analyzing information about phase relations in cyclic events and has been applied in sensory physiology to measurements of the degree of synchronization, or vector strength, in the responses of auditory neurons to tonal stimuli (Goldberg and Brown 1969; Lavine 1971) and in our earlier studies of vibrotactile coding (Greenstein et al. 1987; Vickery et al. 1994).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effectiveness of localized cooling for SII inactivation

The effects of localized SII cooling on SI responsiveness were examined first by means of evoked potential recording. The inactivation of SII was usually achieved when the temperature at the face of the cooling block fell to 5-10°C as shown for both the hand and foot areas of SII in Fig. 2. On rewarming to the 38°C control temperature, the SII responses usually recovered within 1-2 min (Fig. 2), and the inactivation procedure could be repeated without evidence of impairment. In some cases, in particular when the cortical temperature changes occurred rapidly, the phenomenon of cortical spreading depression was observed (see Fig. 10 and DISCUSSION) necessitating a repetition of the inactivation procedure.

The effectiveness of cooling for reversible SII inactivation was also confirmed by single neuron recording from SII around the margin of the cooling block (Fig. 3). As the overlying cortical temperature fell, the spike responsiveness to skin stimulation was abolished by cooling to 12°C as revealed in the response traces in Fig. 3A and the plot of response level () in Fig. 3B. Furthermore, background activity plotted in the bottom graph was also abolished by the cooling. When the temperature of the cortex was restored, the responsiveness of the SII neuron recovered promptly (Fig. 3, A and B), although there was an episode of cortical spreading depression (CSD) in association with the second rewarming (Fig. 3B).

Effect of SII inactivation on SI-evoked potentials

Evoked potentials recorded in 16 experiments from the SI focus of representation for the contralateral hand or foot were never abolished by SII inactivation (Figs. 2 and 4). Although the evoked potentials vary somewhat even from one control record (at 38°C) to another (e.g., for hindlimb SI evoked potentials in Fig. 2), there appeared to be little or no effect on the initial positive-going component in the majority (11/16) of experiments (Figs. 2 and 4, A-G), while small reductions in amplitude were apparent in a minority (3/16) of instances (Fig. 4, H-J) and inconsistent effects in the remaining two cases. The records in Fig. 2 show forelimb and hindlimb evoked potentials recorded simultaneously from SI and SII in association with SII cooling and inactivation. In each set of averaged evoked potentials, it may be seen that the SI-evoked potentials from the hand (top set) and foot (bottom set) survive the disappearance of the SII-evoked potential response as the SII surface temperature fell to 5-10°C. In the top set, the failure initially of the SII-evoked potential to reappear on rewarming to 38°C was associated with an episode of spreading depression.

View larger version (24K): [in this window] [in a new window]   Fig. 4. Effect of SII inactivation on SI-evoked potentials for 9 different experiments. Control-evoked potentials, recorded when the temperature over SII was 38°C, are shown in the left column. Evoked potentials on the right were recorded when SII had been inactivated by cooling to the temperature indicated beside each trace. Traces were averaged from 20-60 successive sweeps, in response to a brief tap (3 ms, 200-800 µm) delivered to the hand (traces marked FL) or foot (remaining traces). Evoked potentials in B and C were recorded from the same experiment for the response to stimulation of the forelimb and hindlimb, respectively. SI-evoked potentials were unaffected by cooling in 7 cases (A-G) but showed some reduction in amplitude in the remaining 3 experiments (H-J).

Figure 4 shows, for 10 different experiments, representative pairs of SI-evoked potentials, in the control circumstance in the left column when SII was at 38°C, and in the right column, in the test circumstance, when SII had been inactivated by cooling to the indicated temperatures. There was little or no effect of the SII inactivation on the amplitude of the initial positive-going potential in A-G, but some reduction or slowing of the SI response in H-J.

Effect of SII inactivation on responsiveness of individual SI neurons

The failure of SII inactivation to abolish SI-evoked potentials demonstrated that the input responsible for the SI-evoked potential traverses a direct path from the thalamus to SI. However, there are two major limitations with observations based on the evoked potentials. First, the amplitude of the cortical-evoked potential is known to saturate at low stimulus intensities (McIntyre 1962) and therefore may not be a very sensitive index of changes in cortical responsiveness. Furthermore, the evoked potential is generated by a brief afferent volley, perhaps before there is time for any putative indirect serial path through SII to contribute to the response. Therefore the evoked potential observations do not permit firm conclusions about the relative contributions that might be made, by direct and indirect paths from the thalamus, to the responsiveness of individual neurons to tactile stimuli that are maintained over periods that allow signals to reach the neuron from both these sources. For these reasons it was essential to examine quantitatively the effects of SII inactivation on the responsiveness of individual SI neurons to maintained and controlled forms of tactile stimulation as was done in our earlier analysis of SII responsiveness in association with the inactivation of SI (Zhang et al. 1996).

Tactile responsiveness was examined for 31 SI neurons before, during, and after SII inactivation by cooling, the majority of them (27/31) located in the SI hindlimb area as the greater separation of the foot representations in SI and SII, compared with those for the hand, eliminated any risk of significant direct spread of cooling from SII to the neuron under study. The four remaining SI cells had receptive fields on the glabrous and/or hairy skin of the contralateral hand. The sampled neurons came from parietal cortical areas in which the SI-evoked potential was maximal (see METHODS) and should correspond to cytoarchitectonic area 3b, as this area is reported to contain the only cutaneous representation within the postcentral strip (Krubitzer and Kaas 1990). Among the 31 SI neurons studied, tactile responsiveness was unchanged by SII inactivation in two-thirds of the sample (20/31 neurons) and showed minor to more substantial reductions in the remainder. In none was responsiveness abolished by SII inactivation. The neurons examined were purely dynamically sensitive except for one, responsive to static skin displacement, that was designated a slowly adapting (SA) neuron.

The functional classes represented (Table 1) were identified in terms of the known primary afferent fiber classes supplying the marmoset hand (Coleman et al. 2001). The 30 dynamically sensitive neurons included a small Pacinian-corpuscle (PC)-related class (3 neurons), based on their high sensitivity to skin vibration at 100 Hz (Coleman et al. 2001; Ferrington and Rowe 1980; Hunt 1961; Talbot et al. 1968; Vallbo and Johansson 1984), and a class sensitive to low frequency vibration (<50 Hz) or mechanical pulse trains (23 neurons) whose input appeared to come from glabrous skin rapidly adapting (RA) afferents, probably associated with Meissner corpuscles in primates (Coleman et al. 2001; Talbot et al. 1968; Vallbo and Johansson 1984; Zhang et al. 1996), and a small sample (4 neurons) activated by hair follicle afferent (HFA) input (Table 1).

SI neurons unaffected by SII inactivation

All four classes of tactile neurons sampled in SI, the PC-, RA-, HFA-related, and the one SA neuron, were represented among the 20 neurons unaffected by SII inactivation (Table 1). This absence of effect on SI responsiveness is illustrated in Fig. 5 for one of the PC-related neurons whose response traces and peristimulus response histograms in A were obtained in response to skin vibration (200 Hz, 25 µm; lasting 1 s; superimposed on a background 1.5-s step indentation; lower waveforms in A) before, during, and after SII inactivation. The expanded, superimposed records of this larger spike in the insets in Fig. 5A, show no prolongation of the spike during SII cooling (12°C) and therefore no evidence for significant direct spread of cooling to the recording site (Zhang et al. 1996). The response level () to successive vibratory stimuli and background activity () are plotted in Fig. 5B and show considerable variability but no systematic change in association with SII inactivation (12°C). The mean response level (15.6 ± 4.7 imp/s, mean ± SD, n = 26) during SII inactivation was not significantly different (P > 0.05, t-test) from control levels before (13.3 ± 4.7, n = 35) and after (13.8 ± 3.6, n = 30) SII inactivation.

View larger version (37K): [in this window] [in a new window]   Fig. 5. Failure of SII inactivation to affect responsiveness of an SI neuron. This neuron responded best to high-frequency vibration delivered in the vicinity of the knee. A: impulse traces and peristimulus time histograms (PSTHs) illustrating the response of this Pacinian corpuscle (PC)-related neuron (larger amplitude spikes) to 200-Hz sinusoidal vibration (1 s, superimposed on a 1.2-mm step indentation) before (left), during (middle), and after (right) SII inactivation by cooling to 12°C. The PSTHs show accumulated impulse activity over a 2.5-s time segment and responses for 15 successive repetitions of the stimulus. B: plots of background activity () and response level () of the neuron as a function of time, as the temperature of the SII cooling block was lowered from 38 to 12°C, then returned to 38°C.

Whenever recording stability allowed, the responses of the SI neurons were studied at more than one stimulus intensity to determine whether responsiveness at different levels of peripheral drive might vary in its susceptibility to the effect of SII inactivation. In Fig. 6 this analysis, for an RA-related neuron, shows no effect in the impulse traces in A where responses are shown at four amplitudes of the 20-Hz vibration stimulus before, during, and after SII inactivation in the left, middle, and right hand columns, respectively. The stimulus-response relations in Fig. 6B plot the mean response level (±SD) as a function of vibration amplitude for these three circumstances and show no significant effect of the SII inactivation (P > 0.05, 1-way ANOVA). The cycle histograms plotted as an inset in Fig. 6B are discussed below.

View larger version (34K): [in this window] [in a new window]   Fig. 6. Stimulus-response relations for an SI neuron before, during, and after SII inactivation by cooling (to 8°C). This rapidly adapting SI neuron, recorded at a depth of 1,440 µm below the cortical surface, had a localized receptive field on the glabrous skin, of the sole of the foot just proximal to digit 1 and responded best to 20-Hz sinusoidal vibration. A: impulse traces showing the neuron's response at a range of amplitudes (25-150 µm) before (left), during (middle), and after (right) SII inactivation by cooling to 8°C. This neuron responded on almost all vibration cycles with a pair of closely spaced spikes (and on occasional cycles with just 1 or 3 spikes). However, the time scale of the impulse traces does not permit this detail of the spike responses to emerge. B: stimulus intensity-response curves derived from 5 responses (mean ± SD) of the neuron as a function of vibration amplitude before (), during (), and after () SII cooling. The inset cycle histograms show the phase locking of the neuron's response to vibration was unaffected by SII inactivation (R, resultant).

SI neurons affected by SII inactivation

The SI neurons whose responsiveness was reduced by SII inactivation ( of the sample) were made up of either RA-related or HFA-related neurons (Table 1). The most severely affected case, for an SI neuron whose impulse rate fell significantly during cooling (P < 0.001; see legend) and was restored on rewarming, is illustrated in the response traces (A) and plot of successive response levels (B) in Fig. 7. The expanded spike waveform traces for this neuron, in the insets in Fig. 7A, show no evidence of a prolongation, when SII was at 8°, that might be indicative of direct cooling spread to SI.

View larger version (35K): [in this window] [in a new window]   Fig. 7. Reduction in responsiveness of an SI neuron to tactile stimulation in association with SII inactivation. A: impulse records before, during, and after 2 episodes of SII inactivation (8°C). Expanded, superimposed spike traces (5 in each) show that there is no prolongation in the SI spike waveform in association with SII cooling. The bottom traces in A represent the mechanical stimulus, a 5-Hz rectangular pulse train superimposed on a background step indentation. B: response levels () and background activity () for the SI neuron during 2 SII cooling episodes. Receptive field: glabrous skin proximal to toe 2 of the contralateral foot. The mean response levels in the 3 control segments at 38°C in B were 7.7 ± 1.5, 5.6 ± 1.8, and 5.8 ± 1.2 imp/s, significantly different (P < 0.001) from those during the 2 cooling segments, when SII was at 8°C, of 1.6 ± 0.7 and 1.7 ± 0.7 (based, in each case, on the counts in the last 10 responses prior to each temperature change).

Effect of SII inactivation on temporal patterning of SI responses to vibration

If the SI responses were mediated by both a direct input path from the thalamus and an indirect path through SII, one might expect that the temporal differences arising between the inputs coming over the two paths might contribute to temporal dispersion in the responses of individual SI neurons to vibrotactile stimuli. We tested this hypothesis by blocking the putative indirect path through SII and examining whether there was a tightening of phase locking in the responses of nine SI neurons to skin vibration at frequencies of 30 Hz. Even where the response level was unaffected, it was possible that SII inactivation could have altered the phase locking in SI responses if the indirect path through SII involved a combined excitatory-inhibitory influence on the SI neuron (Zhang et al. 1996).

The PSTHs on the left in Fig. 8 show qualitatively for two SI neurons that the tight phase locking of impulse occurrences to a 50 Hz (A) and 30 Hz (B) vibration train was unchanged during SII inactivation (5° in A and 8° in B) compared with controls (38° in A and B). Cycle histograms on the right in Fig. 8 enabled the tightness of phase locking to be quantified (in terms of measures of percentage entrainment; see METHODS) and confirm the absence of any persuasive effect when SII was inactivated.

View larger version (48K): [in this window] [in a new window]   Fig. 8. Absence of effect of SII inactivation on phase locking for 2 SI neurons in response to 50-Hz (A) or 30-Hz (B) vibrotactile stimulation. PSTHs (left hand side) were constructed from 10 successive responses to a 1-s vibration before (top), during (middle), and after (bottom) SII inactivation by cooling (5°C in A and 8°C in B). Values for each PSTH (top right of histogram) represent the mean response level (mean ± SD). The intervals between individual peaks in the PSTHs correspond roughly to the cycle period of the vibration frequency used, demonstrating preferential grouping, or phase locking in impulse activity. The degree of phase locking was evaluated quantitatively by constructing cycle histograms (CHs; right hand graphs), with x-axis corresponding to the cycle period of the applied vibration. Each CH represents accumulated responses for 10 consecutive repetitions of 1-s vibration. Percentage entrainment values are shown for each CH.

Quantitative measures of phase locking, based on the percentage entrainment, or Resultant (R) measure obtained from circular statistics (see METHODS, and Fig. 6B, inset) were derived for nine SI neurons from cycle histogram data of the type shown in Fig. 8 and the Fig. 6B inset. The nine included three PC-related neurons that responded to high-frequency vibration (200 and 300 Hz) with rather poor phase locking (percentage entrainment values below 65%), which was unchanged when SII was inactivated (Fig. 9). The results for all nine neurons are plotted quantitatively in terms of percentage entrainment values, before, during, and after SII inactivation in Fig. 9 and show no improvement (or decline) in phase locking in association with the SII inactivation. Absolute values are plotted in Fig. 9A and fall into two groups, the three PC-related neurons (with low entrainment values in their responses to high-frequency vibration) and six RA-related neurons (with high entrainment values in their responses to lower frequencies of vibration). In Fig. 9B, the precooling control values were normalized to illustrate the change in phase locking values from the control circumstance to, first, the SII inactivation value and, second, the postinactivation control value. The absence of any systematic change with SII inactivation indicates that SI responses to vibrotactile inputs are unlikely to be mediated via both a direct path from the thalamus and an indirect path via SII.

View larger version (24K): [in this window] [in a new window]   Fig. 9. Quantitative evaluation of effect of SII inactivation on phase locking in SI responses to vibrotactile stimuli. A: percentage entrainment values for 9 SI neurons obtained from cycle histograms constructed before (left), during (middle), and after (right) SII inactivation by cooling. Note that rapidly adapting (RA) SI neurons had better phase locking (90%) than neurons deriving input predominantly from Pacinian corpuscles (PC; 75%). B: precooling control values have been normalized and data in A replotted to show change in phase locking during SII inactivation.

Cortical spreading depression triggered by localized cortical cooling

A not infrequent phenomenon observed in these experiments and in the earlier ones involving SI inactivation (Zhang et al. 1996) was an abrupt and rather unpredictable disappearance of unitary and evoked potential activity in association with the localized cortical cooling, or indeed, in some cases, in association with rapid rewarming of the cortex to the control temperature (see Fig. 2). Where this was observed for unitary activity, it was normally associated with a transient high-frequency burst of impulses followed by a silent unresponsive period lasting usually 1-3 min (Fig. 10). In the three episodes marked by the asterisks in Fig. 10A, the transient burst was very brief and fell between the 1 per 9-s sampling points. The burst was longer lasting (~20 s) for another SI neuron in Fig. 10B and detected in the sampled points. Clear silent periods in both the peripheral-evoked response () and background activity () follow the burst in four of the five examples in Fig. 10, A and B. In these and other instances, the recovery of responsiveness occurred within the period when SII remained inactivated by cooling, and therefore the disappearance of responsiveness was not caused by SII inactivation per se. Almost certainly this phenomenon reflects the occurrence of cortical spreading depression (Leao 1944) for which cortical cooling is known to be a predisposing trigger factor (Marshall 1959; Zacharová and Zachar 1961). As spreading depression could be a confounding factor in experiments where localized cortical inactivation is induced by cooling, the phenomenon needs to be documented and recognized in investigations of this kind.

View larger version (28K): [in this window] [in a new window]   Fig. 10. Effect of SII localized cooling on the activity of 2 SI neurons that were affected by cortical spreading depression. The response levels () of the neurons to 30-Hz (A) or 50-Hz (B) vibration applied to the foot are plotted along with background activity (). Despite frequent occurrences of cortical spreading depression (*) triggered by SII cooling, the response level of both neurons recovered to control levels while the SII temperature was held at 5°C.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effectiveness of localized cooling for SII inactivation

Cooling has been shown to provide an effective means for rapid reversible inactivation of localized regions of the cerebral cortex in earlier studies from our laboratory (Coleman et al. 1999; Turman et al. 1992; Zhang et al. 1996) and elsewhere (e.g., Girard and Bullier 1989; Hupe et al. 1998; for earlier review see Brooks 1983). As the SII area of cortex in the marmoset is located on the upper bank of the lateral sulcus and may extend into the wall of the sulcus, it was necessary in the present study to ensure the completeness of the inactivation of the fore- and hindlimb areas of SII. First, our mapping, in agreement with Krubitzer and Kaas (1990) confirmed that the hand and foot representations were almost entirely located on the exposed surface of the upper bank of the lateral sulcus. Second, we established that not only surface-evoked potentials from SII were abolished in association with cooling (Fig. 2) but that SII unitary responses below the surface of the lateral sulcus were also abolished (Fig. 3). Furthermore, as the cooling temperature needed to block underlying single neuron responses was not as low as that required to abolish the evoked potentials, and the routine criterion for inactivation that we applied was the disappearance of the evoked potential, we can be confident that inactivation of the fore- and hindlimb areas of SII was complete.

On average, the cooling temperature needed to block SII-evoked potentials (6.5°C) was about 3° lower than had been required in our earlier study (Zhang et al. 1996) for abolition of SI-evoked potentials, presumably because of the slight extension of the SII representation into the wall of the lateral sulcus, and possibly also because of the prominent vessels and blood supply along the lateral sulcus.

As indicated in METHODS, we did not attempt any distinction in this lateral somatosensory area between an SII proper and the PV representation. However, as the entire area responsive to tactile stimulation of the hand and foot was covered, both these putative components of the lateral somatosensory area, that we term SII (Höhl-Abrahão and Creutzfeldt 1991; Krubitzer and Kaas 1992; Zhang et al. 1996), would have been inactivated by the cooling.

Effect of SII inactivation on SI evoked potentials

As there was never an abolition of the SI-evoked potential in association with SII inactivation, and, in the substantial majority of cases (11/16) the SI-evoked potential appeared unchanged, the results demonstrate that the transient afferent volley responsible for generating the SI-evoked potential is generated by a direct thalamic input to SI as might have been expected on the evidence of the well-recognized direct anatomical projection from the ventral posterior thalamus to the SI area in the marmoset (Höhl-Abrahão and Creutzfeldt 1991; Krubitzer and Kaas 1992) as in other species.

Although there was a reduction in the SI-evoked potential in 3 of 16 experiments (with inconsistent effects in the two remaining cases), the overall outcome was similar to earlier experiments investigating the effect of SI inactivation on SII-evoked potential responses (Zhang et al. 1996), where in 8/20 experiments there was some (usually small) reduction in the magnitude of the SII-evoked potential in association with the SI inactivation. Furthermore, these evoked potential-based observations on the SI-SII interactions in the marmoset are qualitatively similar to earlier results in the cat (Rowe et al. 1996; Turman et al. 1992, 1995) in which inactivation of SI or SII never abolished the peripherally generated evoked potential from the other (SII or SI) cortical locus, but, in a proportion of cases (~50%), led to some reduction in its amplitude. In the marmoset, as in the cat, it appears [based on single neuron data (see Effect of SII inactivation on responsiveness of individual SI neurons) and Turman et al. (1992, 1995) and Zhang et al. (1996)] that where there was any reduction in the response it may be attributable to the removal of a background facilitatory influence mediated via the reciprocal intracortical connections that link SI and SII.

Why such an influence was apparent in only a proportion of cases is unclear, but, as we have suggested previously, it may be related to the intracortical modulatory influences operating in a rather sparse and context-dependent way (Turman et al. 1995). There has been no evidence in the present study or earlier ones that any observed reductions in evoked potentials were associated only with weaker levels of input (Zhang et al. 1996). Furthermore, the analysis of single neuron responsiveness in SI revealed that affected and unaffected neurons could be encountered in the same experiment irrespective of whether the SI-evoked potential was unaffected or was somewhat reduced by SII inactivation in that experiment. At a quantitative level, the somewhat lower incidence of any effect on the SI-evoked potentials with SII inactivation compared with the reverse circumstance (Zhang et al. 1996) may be related to the smaller size of the SII area compared with its SI counterpart in the marmoset. However, at a qualitative level it appears that the evoked potential observations on the SI-SII relations in the marmoset (based on present observations and Zhang et al. 1996) are entirely consistent with those in the cat (Turman et al. 1992, 1995).

Effect of SII inactivation on responsiveness of individual SI neurons

The analysis of single neuron responsiveness in SI in association with the SII inactivation was carried out with tactile inputs generated by maintained and controlled forms of stimuli (generally 1 s in duration), which would allow time for inputs that might traverse any putative indirect serial path via SII to contribute to the response in the SI neuron. However, the failure of SII inactivation to abolish responsiveness in any of the 31 SI neurons studied and the absence of effect in two-thirds of the sample extends the argument arising from the evoked potential analysis, that tactile inputs to SI, whether transient or maintained, are largely or exclusively conveyed over a direct path from the thalamus to the SI neuron.

Although a reduction in responsiveness was observed for some SI neurons, this was reminiscent of the effect of SI inactivation on SII single neuron responsiveness in both the cat (Turman et al. 1992) and the marmoset (Zhang et al. 1996) and is similar to the observation in the cat that a proportion of SI neurons show some reduction in response when the SII area is inactivated by cooling. However, as we observed in the cat (Turman et al. 1995), SI neurons that showed some reduction in tactile responsiveness also showed some reduction in background activity, suggesting that the SII inactivation removed a background facilitatory influence from SII, which in the present experimental circumstances operates, as it does in the cat (Turman et al. 1995), on a minority of SI neurons. A further observation consistent with this interpretation, and with earlier results in the cat (Turman et al. 1995), was that SII inactivation led to no enhancement in the tightness of phase locking in SI responses to vibrotactile stimuli. An improvement in this measure might have taken place if the peripheral input to individual SI neurons came from the thalamus over both a direct path and an indirect path via SII, as the SII inactivation may have eliminated temporal dispersion associated with the two putative input paths.

In none of the SI neurons whose responses were affected by SII inactivation could the effect be attributed to direct spread of cooling to the recording site as most (10/11 affected neurons) were recorded within the SI hindlimb area in the medial part of SI near the sagittal sulcus, at a distance of ~10 mm from the cooling block over SII. Temperature profiles measured previously around the cooling block have shown negligible spread of cooling at these distances (Lomber et al. 1996; Turman et al. 1992; Zhang et al. 1996). Furthermore, monitoring of spike waveforms from the recorded SI neurons even in the SI forelimb area showed no prolongation of the waveform indicative of spread of cooling to the recording site (Figs. 5A and 7A) (Sherk 1978; Turman et al. 1992, 1995; Zhang et al. 1996). In addition, affected and unaffected SI neurons were encountered at similar distances from the SII cooling site.

Extent of cortico-cortical modulatory interactions between SI and SII

In our earlier study of marmoset SI-SII interactions (Zhang et al. 1996) ~70% of SII neurons displayed some reduction in responsiveness in association with SI inactivation compared with the 35% of SI neurons in the present study that showed a reduction with SII inactivation. The different incidence of cortico-cortical influences may, in part, be related to anatomical asymmetries present (at least in certain primates) in the SI-SII linkages (for review see Burton 1986; Jones 1986). In particular, the projection from SI to SII is focused on granular and supragranular layers within SII, whereas the less dense, more diffuse reverse projection from SII is directed predominantly to layers I and VI of SI. The differential nature of the anatomical linkages has led to the hypothesis that the former represents a feed-forward connection for the flow of sensory information, whereas the projection from SII to SI is a feedback connection perhaps contributing to a more nonspecific maintenance of excitability in SI (Jones 1986).

Although this anatomical asymmetry may contribute to the higher incidence of modulatory influences operating in the SI to SII direction in the marmoset, another factor may be the differential size of the two areas. In this species the hand area of SI appears to occupy ~5-10 times the area of its counterpart within SII (Zhang 1994).

It should also be emphasized that in the cat, which has no anatomical asymmetry in the cortico-cortical connections between SI and SII (see Burton 1986 review), we have also found evidence for a differential incidence of cortico-cortical modulatory influences, but in the reverse direction from that seen in the marmoset. Thus ~20% of cat SII neurons displayed some reduction in responsiveness with SI inactivation (Turman et al. 1992), but ~40% of SI neurons showed a reduction when SII was inactivated (Turman et al. 1995). We argued at that time that these quantitative differences probably did not represent a fundamental differentiation of SI and SII function in the cat, in particular as the reductions in responsiveness for individual neurons were usually small (Turman et al. 1992, 1995). As this was also the case for the SI-SII interactions observed in the marmoset (Zhang et al. 1996 and present study), it remains somewhat uncertain whether the quantitative differences in the SI-SII modulatory influences reflect a fundamental differentiation of function for SI and SII.

Role of cortico-cortical modulatory interactions between SI and SII

The functional significance of the cortico-cortical linkages between SI and SII has been a matter of some speculation, in particular, based on the anatomical asymmetries described above for certain primates. Although these speculations have stimulated discussion and further research, there is little functional evidence at this stage in support of them. Regrettably, we can add little more that is concrete to this debate, in particular, from the context of experiments in anesthetized animals. However, in functional terms, the intracortical linkages from either area may provide a gain adjustment for responsiveness in the other area, or impose some spatial shaping on the activity profile in that area.

This hypothesis for a "gain" control influence mediated by cortico-cortical linkages is reminiscent of the interpretation suggested by Burton et al. (1990) to account for the virtual abolition of SII responses in the macaque monkey after SI ablation. Rather than interpreting their results as direct evidence for a simple serial scheme of organization for SI and SII in the macaque, they proposed that SI may "gate" thalamic inputs to SII, either at the level of the cortex or the thalamus, and thereby exert an "enabling" influence on SII responsiveness. Furthermore, as we have suggested before (Rowe et al. 1996; Turman et al. 1992, 1995; Zhang et al. 1996), these cortico-cortical influences may modulate the sensitivity or responsiveness of neurons in either area in the context of voluntary motor actions or in the switching of selective attention in the conscious behaving animal.

The question of how such modulatory influences might operate differentially from one area on the other, or be switched in direction, also remains a matter for conjecture. However, this might be determined, in part, by the nature of the sensory inputs being processed at the time and the quantitatively differential representation of different receptor sources within SI or SII. Thus SII appears to have a predominant role in processing PC inputs, whereas RA and SA inputs appear to have a more prominent representation in SI (present observations, and Bennett et al. 1980; Burton 1986; Ferrington and Rowe 1980; Fisher et al. 1983; Zhang et al. 2001). Therefore in circumstances in which tactile inputs are biased toward PC sources, for example, in processing high-frequency vibrotactile information, SII may assume processing priority and modulate SI activity. In contrast, when inputs are biased toward RA or SA sources, SI may assume processing priority and exert the dominant cortico-cortical modulatory role. However, the operation of these facilitatory or disfacilitatory influences may, of course, be determined by many factors, including the anatomical asymmetries of the linkage, together with the exigencies of consciousness, directed attention, and voluntary motor control.

Hierarchical equivalence of SI and SII in tactile processing

Despite the quantitative differences in the incidence of cortico-cortical modulatory interactions between SI and SII in the marmoset, it must be emphasized, as indicated above, that the effects of SII inactivation on SI responsiveness are qualitatively very similar to the effects of SI inactivation on SII responsiveness (Zhang et al. 1996). Furthermore, the marmoset results are qualitatively very similar to those observed for SI-SII interactions in the cat (Turman et al. 1992, 1995). The functional data from both the single neuron and the evoked potential analyses suggest that SI and SII in the marmoset, as in the cat, occupy a position of hierarchical equivalence in a thalamocortical processing network that is largely parallel in its organizational scheme. This conclusion is also reinforced by our findings, in the accompanying paper (Zhang et al. 2001), that SI- and SII-projecting neurons of a given class (e.g., PC-, RA-, or SA-related neurons) within the marmoset ventral posterior thalamic nucleus show no systematic differences in their functional capacities, and that both these groups of thalamocortical projection neurons convey high-acuity tactile information directly to SII as well as to SI.

In concluding that SI and SII occupy positions of hierarchical equivalence in the processing of tactile information, we should emphasize that our reference to equivalence is strictly confined to the issue of the sequence or temporal staging in which the two areas process incoming tactile information under the experimental conditions in which we have been able to examine these issues. We do not wish to imply an equivalence in processing capacities for SI and SII. Indeed, there is good evidence, much of it from our earlier studies, for both qualitative and quantitative differences between the two areas in their processing capacities. First, there appear to be differences in the proportional representation of particular classes of sensory fibers within the two areas (see above; and Bennett et al. 1980; Ferrington and Rowe 1980; Fisher et al. 1983; Rowe et al. 1985; Zhang et al. 2001). Second, we have provided evidence for quantitative differences in the cat between SI and SII neurons in their capacities for coding information about the frequency component of vibrotactile stimuli (Ferrington and Rowe 1980; Rowe et al. 1985). These studies revealed that SII neurons have a greater capacity for responding with phase-locked impulse patterns to high-frequency vibration (100 Hz) than do their counterparts in SI, a finding that points to a more important role for SII than for SI in this aspect of vibrotactile coding and processing.

We would therefore argue that, although SI and SII are clearly not equivalent in their processing capacities, they appear to operate at a level of hierarchical equivalence in terms of the temporal staging of their processing, at least under the experimental conditions of our analysis.

Parallel organization of SI and SII in different species

Consistent evidence for the parallel scheme of thalamocortical organization for SI and SII in the marmoset monkey has now been obtained from the accompanying functional analysis of SI- and SII-projecting thalamocortical neurons (Zhang et al. 2001) and from the investigations of SI-SII interactions based on reversible inactivation of either SI (Zhang et al. 1996) or SII (present study). This accumulated evidence from the marmoset monkey necessitates a revision of the hypothesis that tactile processing at thalamocortical levels in simian primates is based on a serial scheme in which signals are conveyed from the thalamus to SI, and thence to SII. Thus it is no longer appropriate to sustain the hypotheses that, for primates in general, 1) the major somatic drive to SII is from the postcentral, or SI cortex (Pons 1996), 2) the thalamic input to SII plays only a modulatory rather than an activating role (Garraghty et al. 1990), and 3) the relationship between SI and SII differs fundamentally from that in nonprimates (Garraghty et al. 1990).

These conclusions about the parallel organization of SI and SII in the marmoset are consistent with the consensus that these cortical processing areas are organized in parallel in a diverse range of nonprimate placental mammals (Burton and Robinson 1987; Garraghty et al. 1991; Manzoni et al. 1979; Murray et al. 1992; Turman et al. 1992, 1995) and in mammals of the marsupial order (Coleman et al. 1999), which diverged from the placental line in the Cretaceous period, ~140-70 million years ago (Rowe 1990). Our confirmation of a parallel organizational scheme for SI and SII in the marmoset monkey, in contrast to the earlier support (based on SI ablation studies) for a serial scheme of organization (Garraghty et al. 1990), demonstrates that there is no fundamental distinction between primates in general and other mammals in the organization of these thalamocortical systems for tactile processing.

Although there is reported to be serial organization of the SI and SII areas in the macaque monkey (Pons et al. 1987, 1992), this was also based on a study involving surgical ablation of the SI area and should, if possible, be re-investigated with a less disruptive reversible procedure, such as localized cooling, for SI inactivation to allow examination of SII responsiveness before, during, and after the SI inactivation procedure. This is necessary lest the ablation procedure itself induced effects that confounded the interpretation of results in the macaque, as we have suggested (Zhang et al. 1996) may have occurred in the earlier marmoset studies (Garraghty et al. 1990). We argued then that the surgical ablation of SI in the marmoset may have induced a depressant effect on SII responsiveness, for example, by setting up an injury discharge in corticocortical neurons that project from SI to the topographically related region of SII, leading to changes in extracellular ion concentrations, in particular, K⁺ ion accumulation, and accommodation block of neurons within this region of SII (Zhang et al. 1996). We did not attempt in the present study to verify the generality of this explanation by ablating SII to determine whether this abolished SI responsiveness. We believe that this would not be a satisfactory test because SII is very much smaller than SI in the marmoset, and thus the number and probable density of cortico-cortical projections from SII to SI would be much lower than in the reverse direction. The injury discharge set up in SII may therefore be insufficient to generate the putative accommodative block in SI.

A re-investigation of the parallel/serial issue for SI and SII in the macaque becomes all the more important with new data emerging first, from Nicolelis et al. (1998), emphasizing that multiple cortical areas in the owl monkey (including SII, and areas 3b and 2 within SI), operate almost simultaneously in processing tactile information about stimulus location, and second, from studies based on noninvasive recording of evoked neuromagnetic responses from the human SI and SII areas. The latter observations indicate that the SII area is co-activated with SI in the processing of somatic afferent inputs of median nerve origin (Karhu and Tesche 1999). Furthermore, other human studies based on the effects of parietal lobe lesions have shown that vibrotactile perception was unaffected in patients with lesions limited to the SI area, leading the authors to the conclusion that vibrotactile information is not dependent on the SI region but is processed in a nonserial, probably parallel cortical network (Knecht et al. 1996).

In this regard, a recent study in humans (Ffytche et al. 1995) has challenged the established view (Drasdo et al. 1993; Probst et al. 1993) that visual cortical area V5, the putative visual-motion processing site, is hierarchically beyond the striate area, V1, in a predominantly serially organized processing scheme. This re-evaluation, using visually evoked electroencephalography coupled to magnetoencephalography, indicates that there is parallel input of visual information to both V5 and V1 (Ffytche et al. 1995).

In conclusion, the accumulating evidence in studies of somatosensory cortex in particular, but also perhaps visual cortex, emphasizes the importance of parallel, distributed mechanisms as the dominant organizing principle, at least for those stages of sensory processing within the cerebral cortex that involve SI and SII. Furthermore, there are no longer grounds for the view that there has been an evolutionary shift in the organization of the somatosensory thalamocortical networks from parallel to serial mode with the emergence of the primate line in mammalian evolution.