| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
The Journal of Neurophysiology Vol. 79 No. 2 February 1998, pp. 947-963
Copyright ©1998 by the American Physiological Society
1 Centre de Recherche en Sciences Neurologiques, 2 Département de Physiologie, and 3 École de Réadaptation, Faculté de Médecine, Université de Montréal, Montreal, Quebec H3C 3J7, Canada
 |
ABSTRACT |
|---|
 Abstract
 Introduction
Methods
Results
Discussion
References
|
|---|
Williams, Stephan R., Jafar Shenasa, and C. Elaine Chapman. Time course and magnitude of movement-related gating of tactile detection in humans. I. Importance of stimulus location. J. Neurophysiol. 79: 947-963, 1998. The time course and spatial extent of movement-related suppression of the detection of weak electrical stimuli (intensity, 90% detected at rest) was determined in 118 experiments carried out in 47 human subjects. Subjects were trained to perform a rapid abduction of the right index finger (D2) in response to a visual cue. Stimulus timing was calculated relative to the onset of movement and the onset of electromyographic (EMG) activity. Electrical stimulation was delivered to 10 different sites on the body, including sites on the limb performing the movement (D2, D5, hand, forearm and arm) as well as several distant sites (contralateral arm, ipsilateral leg). Detection of stimuli applied to the moving digit diminished significantly and in a time-dependent manner, with the first significant decrease occurring 120 ms before movement onset and 70 ms before the onset of EMG activity. Movement-related and time-dependent effects were obtained at all stimulation sites on the homolateral arm as well as the adjacent trunk. A pronounced spatiotemporal gradient was observed: the magnitude of the movement-related decrease in detectability was greatest and earliest at sites closest to the moving finger and progressively weaker and later at more proximal sites. When stimuli were applied to the distant sites, only a small (~10%), non-time-dependent decrease was observed during movement trials. A simple model of perceptual performance adequately described the results, providing insight into the distribution of movement-related inhibitory controls within the CNS.
The CNS has a variety of mechanisms at its disposal that modulate the quantity and quality of the sensory information that it processes. One such mechanism, gain control, enhances or diminishes sensory feedback. Several forms of gain control are found in the somatosensory system, including direct controls over receptor sensitivity (fusimotor control of muscle spindle sensitivity) and controls over the transmission of somatosensory signals within the CNS. An example of the latter is the suppression or "gating" of the transmission of cutaneous signals seen in association with voluntary movement. Single-unit and evoked-potential studies have demonstrated that the transmission of cutaneous signals through the dorsal column-medial lemniscal pathway to primary somatosensory cortex is decreased during movement (e.g., Chapman et al. 1988 Subjects
A total of 47 naive, paid volunteers (25 males and 22 females, ages 15-30 yr) participated in this study. Four subjects were left handed for writing and 43 were right handed. The experimental protocol was approved by the institutional ethics committee, and all subjects or their legal guardian gave their informed consent before participating in the study. Data from each subject were gathered in one or two sessions lasting 1-3 h each. At the beginning of each session, subjects received verbal instructions about the motor and perceptual tasks that they were to perform. This was followed by a small block of practice trials. In brief, the motor task consisted of an active abduction of the right index finger (D2), whereas the perceptual task consisted of determining whether or not a weak near-threshold electrical stimulus had been delivered to the skin. One stimulation site was investigated in each experiment, and two to three different sites were tested in each experimental session. Altogether, 118 experiments were performed, using 10 different stimulation sites mainly located on the right arm.
Motor task
A simple reaction time (RT) task was employed, whereby subjects made a rapid abduction of the right D2 as soon as possible after the appearance of a visual "GO" cue (illumination of a 3 × 3 array of LEDs placed at eye level, 1 m in front of the subject). As shown in Fig. 1A, subjects were seated in a chair beside a small table on which the right arm rested. The index finger rested on a small pivoting plate (1.8 × 10 cm) the base of which was aligned with the axis of rotation of the D2 metacarpophalangeal joint (Fig. 1B). The subjects were asked to relax with D2 in a neutral position. An oscilloscope was used to display D2 position (Fig. 1A) along with two reference lines corresponding to the starting (neutral) position and to the minimal required amplitude of movement (15°). No maximal movement amplitude was specified, but subjects were instructed to avoid hitting the stop situated 45° from the start position. Subjects also were requested to limit muscle contraction to the relevant agonist muscle (1st dorsal interosseous, 1st DI) during the initial phase of the movement to the extent that this was possible.
Perceptual task
Subjects were asked to report whether or not they detected the occurrence of weak electrical stimuli delivered to a site on the skin (see further) under two experimental conditions: at rest and while performing the motor task. No information regarding the proportion of trials with and without a stimulus was given to the subjects. No feedback was given with regard to the accuracy of the subject's perceptual judgements. The stimulus consisted of a single 2-ms square wave pulse generated by a Grass S88 stimulator and delivered via a SIU7 constant current photoelectric stimulation isolation unit. The stimulus was applied to the skin via surface electrodes (7 mm diam) spaced 30 mm apart, and in all cases only a localized sensation was produced. Before applying the electrodes, skin resistance was minimized and electrode adhesion maximized by vigorous cleansing of the area immediately beneath the electrodes with 70% alcohol. For experiments involving stimulation of a digit, the electrodes were not in contact with the supporting surface but were recessed in a specially designed cavity in the foam padding. At the beginning of each experiment (subject at rest), a modified sequential tracking procedure using six observations per stimulus intensity tested (as described by Wetherhill and Levitt 1965 Stimulation sites
Ten stimulation sites were employed (Fig. 2). Six sites were situated on the arm ipsilateral to the moving digit: the glabrous skin of the middle and distal phalanges of digits 2 (iD2) and 5 (iD5), the dorsum of the hand (iHA), the mid-forearm (iFA), the lower arm (iA), and the shoulder (iSH). The remaining sites were one on the ipsilateral pectoral girdle just below the sternocleidomastoid joint (iPG), two sites on the contralateral arm (cSH and cD2), and one on the ipsilateral thigh (iTH). Three subjects were tested at 8 stimulation sites, 6 at 7 sites, 1 at 6 sites, 5 at 3 sites, and 31 at 1 site (iD2). Those subjects tested at three or fewer sites participated in other experiments, not reported in this paper, which will form the basis of other publications. The order of testing for the different stimulation sites in each subject was randomly determined. The exact number of subjects tested at each site is given in Fig. 2.
Data collection
The data acquisition and task were under the control of a PDP 11/73 minicomputer. Each trial was initiated by the experimenter and lasted for 2 s. Three types of trials were presented. In the majority of trials (70%), subjects were invited to move and a stimulus was presented (movement + stimulation); these trials were used to gather data on the effects of movement on the perception of the test stimulus. In 20% of the trials, subjects were asked not to move during the trial, and the stimulus was applied at rest (rest + stimulation); these trials monitored the stability of perceptual performance at rest. Finally, catch trials (no stimulus) represented 10% of the total trials and were divided between the movement and no-movement trials; these trials evaluated each subject's rate of false positive responses. Subjects were not aware of the proportion of trials in which a stimulus was presented. Subjects were instructed verbally before the onset of each trial whether to move or not; no information was given as to whether or not there would be a stimulus, and the experimenter could not be seen by the subjects. Figure 1C schematizes the events occurring during a movement + stimulation trial. Data collection was initiated at time 0. The light signaling the subject to move was illuminated 500 ms after the beginning of the trial, allowing a period during which any spontaneous movement or EMG activity could be observed. The stimulus was given at one of five or nine delays relative to the onset of the GO cue (see further). At the end of each trial, subjects indicated verbally whether or not they had perceived a stimulus, and their response was entered into the computer by the experimenter and stored with the trial. The intertrial interval varied from 1 to 10 s. Subjects were observed carefully by the experimenter during each trial. Any trial during which the subject produced movements unrelated to the motor task, as well as any trial after which the subjects reported inattention, discomfort, or lack of readiness was rejected and repeated later in the session.
Data analysis
For each movement trial, the onset of EMG was determined visually. Several timing values were calculated off-line automatically, including RT (determined by an algorithm that found the first of 5 consecutive position samples that changed in the same direction), the lead time between 1st DI EMG onset and movement onset, and movement duration. Kinematic parameters also were calculated, including movement amplitude, peak velocity (computed by 3-point numerical differentiation and appropriate digital filtering of the displacement trace), and peak acceleration. Approximately 1% of the trials were discarded at this stage, usually because EMG activity was observed in the 500-ms interval before the GO signal. The presence of a causal relationship between EMG and movement was assessed by performing linear regression analyses between EMG onset and movement onset. The rationale for this test is that EMG activity responsible for the movement should not only precede the onset of movement but also show a high degree of correlation with RT with a slope approaching 1 (see e.g., Chapman et al. 1986 Bias
Our experimental approach was designed to minimize the duration of each experiment, so minimizing drift in performance and fatigue (recognizing that there was a motor component to the task), and to maximize the collection of data by presenting a relatively high proportion of trials with a stimulus. Positive bias was monitored continuously throughout each experiment using catch trials (Green and Swets 1988
Performance of the movement task
A total of 118 experiments were performed in 47 subjects (Fig. 2). Figure 4 shows an example of the movement traces and EMG records from one subject. First dorsal interosseus EMG activity, the principal agonist, consistently preceded movement onset by an average of 27 ms in this example (Fig. 4, A and B) and showed a strong linear correlation with movement onset (Fig. 4C). Cocontraction was evident in abductor digiti minimi (Fig. 4, A, B, and D), but the activity was weaker (note the change in scale), more irregular, and always followed the onset of 1st DI EMG activity (average delay, 10 ms). At the forearm extensor site, EMG activity was weak, frequently absent, and followed movement onset. In 118 experiments, the mean correlation coefficient between 1st DI and movement onset was 0.96 ± 0.02 (mean ± SE) and the mean slope of the linear regression line was 0.95 ± 0.04. First DI EMG activity led movement onset by 50 ± 14 ms (Table 1).
Performance of the perceptual task
Table 2 summarizes the global performance of all subjects in the three trial types. Overall, pooled data showed that subjects perceived 94% of the stimuli presented at rest. A one-way ANOVA failed to discern any significant difference in pooled perceptual performance at rest across the experimental test sites, thus providing a constant baseline against which performance during movement could be evaluated. Of 1,479 catch trials (pooled data), only 12 false positive responses were noted (0.8%), indicating that subjects used a very conservative response strategy throughout the series of experiments. As detailed in the legend for Table 2, no stimulation site was associated with significantly more false positives than any other. Pooled data showed that subjects perceived 58% of the stimuli presented during movement + stimulation trials, and a one-way ANOVA (Table 2) showed that there was significant intersite variability in pooled perceptual performance for these trials. Inspection of Table 2 reveals that sites closer to the body part in motion, iD2, showed greater reductions in the proportion of stimuli perceived during movement than did those that were more distant from the movement site. Comparisons between performance in the rest + stimulation trials and the movement + stimulation trials showed that significantly fewer stimuli were perceived during movement at all 10 stimulation sites (P < 0.001, Fischer exact tests for a 2 × 2 contingency table).
Variation over time of perceptual performance at rest
For each stimulation site, perceptual performance in the first 20% of immobile trials was compared with perceptual performance in the last 20% of immobile trials. No significant differences were observed at any of the sites (P > 0.01, Fischer exact probability tests). For data from all sites pooled together, the proportion of stimuli perceived in the first 20% of immobile trials was 0.95, whereas performance in the last 20% was 0.93.
Time-dependent change in the detection of stimuli applied to the moving digit (iD2)
Figure 5 shows that the effects of iD2 movement on the ability of 41 subjects to detect stimuli applied to the moving digit were not uniform over time. Although modest but significant decreases in the proportion of stimuli perceived were observed
Effects of changing the site of stimulation on the movement-related decrease in tactile detection
Figure 6 summarizes the results obtained at the six other stimulation sites located on, or adjacent to, the ipsilateral arm, in increasing order of distance from the moving digit (Fig. 6, A-F). All six sites showed time-dependent modifications in perceptual performance that were best described by logistic functions. Two major observations were made. First, the depth of modulation of tactile detection, reflected in logistic function minima, lessened as the distance between the stimulation site and the site of movement increased. Second, the timing of the peak decrease in perceptual performance shifted as distance increased, from just around EMG onset for the sites immediately adjacent to iD2 (Fig. 6, A-C) to well after EMG onset for the sites located on the proximal arm and adjacent trunk (Fig. 6, D-F).
Relationship between movement parameters and perceptual performance
It has been reported that movement-related gating is a function of the kinematics of the movement, with faster movements producing larger gating effects (Angel and Malenka 1982
The results of the present study have shown that the movement-related decrease in detection of near-threshold cutaneous stimuli can precede movement onset and the onset of EMG activity. In addition, both the time course and amplitude of the modulation were dependent on the distance between the stimulation site and the body part in motion.
Methodological considerations
Electrical stimulation was used in this study for the following reasons. It had been shown previously that detection thresholds for electrical stimuli the parameters of which were identical to those used in this study are increased during movement (Chapman et al. 1987 Distribution of non-time-dependent decreases in tactile detection
The three distant sites showed no time-dependent decrease in perceptual performance, but rather a constant, ~10%, decrease. Chapman et al. (1987) Spatiotemporal characteristics of decreases in tactile detection
All sites on the moving limb (including the adjacent trunk, iPG) showed time-dependent reductions in the proportion of stimuli perceived. At the four sites closest to the moving digit, these reductions began before EMG onset (Fig. 9). The timing of these reductions is in agreement with the earliest modulation reported in previous psychophysical investigations (Coquery et al. 1971 Sources and mechanisms of the movement-related decrease in tactile detection
The spatiotemporal characteristics of the movement-related decrease in detectability, discussed earlier, help to define the necessary attributes of the gating mechanism(s). Both centrally mediated inhibition, originating from central and/or peripheral sources, and also physical factors in the periphery may have contributed to generating the observed spatiotemporal gradient.
CENTRAL SOURCES.
One finding in favor of a central source for the gating signals is the observation that at the sites closest to the moving digit, detection began to decline 120 ms before movement onset and 70 ms before the onset of EMG, i.e., well before any peripheral feedback could have been generated. The earliest time-dependent decrease may be related to the preparation and execution of the movement, as has been suggested by others (Chapman et al. 1988 PERIPHERAL SOURCES.
Other spatial and temporal characteristics of the gating argue in favor of a role for peripheral sources in mediating the movement-related gating. The timing of the peak decrease at the stimulation sites closest to the moving digit coincided with the onset of EMG activity. It is reasonable to assume that peripheral feedback begins at this moment, e.g., muscle spindle discharge elicited by alpha-gamma coactivation, and that this feedback contributed to the sharp decrease in detection performance at this time. In support of this, passive movements can be as effective as active movements in generating movement-related decreases in tactile inputs (e.g., Chapman et al. 1987 CENTRAL MECHANISMS.
Whatever the source of the gating signal(s), any explanation for the results needs to explain the widespread spatial distribution of the time-dependent decrease in detectability and the pronounced temporal gradient across the homolateral limb. Clearly a single inhibitory signal distributed simultaneously across a given somatosensory relay would not produce the complex spatiotemporal gradient observed here. Several, not necessarily exclusive, explanations can be advanced to explain the large changes in time and magnitude at the more proximal sites on the arm. The spatial distribution may be the result of lateral inhibition generated by the gating signals. The timing of the modulation would reflect the time necessary for spread of the lateral inhibition, whereas the amplitude of the observed decrease in detectability would reflect the strength of the inhibitory signal. The distant sites that showed no obvious time-dependent modulation would either be outside the zone of effective inhibition or the inhibition might have occurred at delays >100 ms after movement onset (the longest delay tested here). Given that synaptic delay is only 0.5 ms and that conduction distances within any candidate relay (dorsal column nuclei (DCN), ventrobasal thalamus, primary somatosensory cortex) are short, then this suggestion requires an extremely slow rate of spread through multiple synapses to explain delays of the order of 60 ms. This makes it more likely that multiple loops within the CNS may be involved in generating this spatiotemporal gradient, possibly via descending cortical projections to somatosensory relays, including cortico-thalamic, cortico-DCN, cortico-reticulo-DCN, and corticospinal projections. If such is the case, then it may well be that a combination of signals interact to produce the observed spatiotemporal gradient, with only a subset of these mechanisms being effective at the more distant sites. Such a suggestion is supported by our observation that only the sites closest to the moving digit showed evidence of a relationship between perceptual performance and movement kinematics.
CONTRIBUTION OF PERIPHERAL FACTORS.
Although central mechanisms can explain the results, the potential contribution of physical factors (distance, nerve conduction velocity) to the spatiotemporal gradient also should be considered. The shift in timing of the peak decrease over distance might be explained by the decrease in distance (and therefore travel time) between the more proximal stimulation sites and the CNS. Assuming that conduction velocity was similar at all stimulation sites, that low-intensity electrical stimuli preferentially activated large diameter (A beta) cutaneous fibers with a conduction velocity of Functional significance
We recently suggested that simple reductions in the signal-to-noise ratio, produced by an increase in background noise, could not explain all of the changes in perception that accompany movement (Chapman et al. 1996
INTRODUCTION
Abstract
Introduction
Methods
Results
Discussion
References
; Ghez and Lenzi 1971; Rushton et al. 1981; see Chapman 1994 for a recent review). Movement-related gating also exerts powerful influences on perception. Thus the detection of near-threshold cutaneous stimuli is decreased during movement (Coquery et al. 1971; Dyhre-Poulsen 1978), and detection threshold is elevated correspondingly (Chapman et al. 1987; Post et al. 1994). Magnitude estimates of clearly supra threshold innocuous stimuli, including vibrotactile stimuli, also are diminished during movement, although relative differences are preserved and so discrimination thresholds are unchanged (Chapman et al. 1987; Milne et al. 1988; Post et al. 1994).
; Coulter 1974; Dyhre-Poulsen 1978; Ghez and Lenzi 1971; Jiang et al. 1990b). Peripheral feedback, generated during movement execution, also plays an important role in the modulation that follows movement onset (e.g., Chapman et al. 1988; Huttunen and Hömberg 1991; Jones et al. 1988). Estimates of the time of onset of gating influences from cortical somatosensory evoked-potential (SEP) studies and psychophysical studies vary widely. Some studies observed modulation 60-100 ms before the onset of EMG activity and so well before movement onset, whereas other studies reported modulation only after the onset of EMG and movement (Chapman et al. 1988; Cohen and Starr 1987; Coquery et al. 1971; Dyhre-Poulsen 1978; Jiang et al. 1990b). These different results may arise from a number of factors, including differences in the intensity of the stimulus, differences in the spatial relation between the stimulus and the movement, and differences in the motor tasks, including the movement studied (digit vs. elbow) as well as the associated movement kinematics (Chapman et al. 1988; Ghez and Pisa 1972; Rauch et al. 1985).
; Tapia et al. 1987). In monkeys, on the other hand, the gating effects have been described as widespread and nonspecific (Jiang et al. 1990b). The latter contrasts with the topographically organized gating influences elicited by intracortical microstimulation applied to motor cortex (area 4), a potential source of central gating signals: Jiang et al. (1990a) reported that these influences show a proximal to distal gradient whereby the modulation is directed toward cutaneous inputs from the skin either overlying or distal to the motor output. Several psychophysical studies have examined the spatial extent of the gating influences on perception (Coquery et al. 1971; Post et al. 1994; Schmidt et al. 1990), and all reported that proximity between the site of stimulation and the movement was an important factor with sites closer to the movement showing a greater modulation. In contrast to the results of Jiang et al. (1990a), however, Coquery et al. (1971) reported that magnitude estimates were decreased at sites located both distal and also proximal to the site of movement. The latter results were confounded potentially by differences in the associated movements because subjects were required to make a variety of movements (digits, wrist, elbow, and shoulder). Furthermore, potential differences in the time course of the gating influences at the different sites may have contributed to the results.
).
METHODS
Abstract
Introduction
Methods
Results
Discussion
References
View larger version (40K):
[in a new window]
FIG. 1.
A: experimental set-up showing the subject's position facing the visual cue and the oscilloscope that displayed movement-related information. B: close-up of hand showing the pivoting plate on which the right index finger rested (here in initial rest position). Shown also are the positions of the stimulating electrodes applied to the index finger and the electromyographic (EMG) recording electrodes over the 1st dorsal interosseous muscle. C: time course of events in movement + stimulation trials. Stimuli were applied at one of 5 (solid lines) or 9 delays, encompassing the time of EMG onset and the reaction time (RT).
) was used to find the stimulus intensity level at which ~90% of the stimuli were detected. This intensity (range 0.18-1.32 mA) then was used throughout the experiment.
View larger version (27K):
[in a new window]
FIG. 2.
Stimulation sites tested relative to the moving digit (iD2). Distance between each stimulation site and iD2 as a proportion of subject height is shown in parentheses. Number of subjects tested at each site also is indicated. A, arm; c, contralateral; D2, digit 2; D5, digit 5; HA, hand; i, ipsilateral; PG, pectoral girdle; SH, shoulder; TH, thigh.
; NASA 1978). Distances were estimated from the metacarpophalangeal joint of iD2; iD2 itself was assigned a distance of 0, being the body part in motion.
120 ms, RT 80 ms, RT 40 ms, RT, and RT + 40 ms. Four additional delays (RT 160 ms, RT + 80 ms, RT + 120 ms, and RT + 160 ms) were used in 10 experiments at the iD2 site. Natural trial-to-trial variations in RT, combined with the use of multiple stimulus presentation delays, resulted in the collection of an adequate sample of perceptual ability during a 200-ms time interval, which included EMG onset and movement onset. Because of variation in RT, trial-by-trial delays between the onset of peripheral movement-related activity and stimulus delivery were not known to either the subject or the experimenter at the time of the experiment. Data were collected in blocks of 
22 trials, one stimulus delay being tested in each block, with the order of delay testing being varied between experiments. Within a block, 15 movement + stimulation trials, 5 rest + stimulation trials, and 2 catch trials were performed. The different types of trials were intermixed randomly, the order of presentation varying from block to block. Thus at each stimulation site, a minimum of 110 trials were recorded.
). The possibility of significant intersite differences in the temporal and kinematic parameters was evaluated by calculating individual subject averages for each parameter and then comparing across the stimulation sites using one-way analyses of variance (ANOVA, level of significance P < 0.05).
2 distribution [df = number of points that defined perceptual performance at the site (number of parameters in the fitting equation + 1)]. The best fitting model (linear or logistic) was retained if it provided an adequate description of the data, i.e., if the probability of obtaining the observed amount of total squared error was >0.05. If a linear model was retained, the presence of a slope significantly different from zero was evaluated using a t-test. If a logistic descriptor was retained, four parameters then were determined: the maximum predicted perceptual performance, the minimum predicted perceptual performance, the peak slope (measure of the peak rate of decrease in perceptual performance), and the timing of the peak slope (the time at which perceptual performance decreased most rapidly). The effects of distance on these parameters were evaluated using adjusted correlation coefficients and F tests.
). The existence of systematic and significant positive or negative biases induced by the method or the experimenter was further evaluated in experiments that compared detection performance obtained with the main experimental method to detection performance obtained using a bias-free two alternative forced-choice (2AFC) procedure (3 subjects: 2 females, 1 male; 2 of 3 right handed). Stimuli were delivered to the iD2 site. Both experimental methods were tested in a single session, ensuring that stimulating conditions were similar. The main experimental method was as described above. For the 2AFC procedure, in each "trial," the GO cue was presented twice, i.e., there were two observation periods per trial, separated by a delay of ~1 s. Observation period duration, stimulus presentation delays, etc., remained the same. The stimulus was assigned randomly to one of the GO cue presentations in each trial: 50% of the trials contained the stimulus in the first interval and 50% in the second interval. At the end of each trial, the subject was asked to report in which of two intervals a stimulus was presented. Immobile trials were again interspersed between the movement trials to provide for a continuous monitoring of performance at rest.
(i) = [p(i) + 1]/2, where p(i) = the proportion of stimuli perceived in a given time interval, and a logistic function was fitted to the resulting data. No significant differences were observed when this logistic function was compared with that describing the 2AFC data (P > 0.05, Kolmogorov-Smirnov statistic), making it unlikely that experimenter- or procedure-induced biases played a significant role in the results obtained with the main experimental procedure.
View larger version (30K):
[in a new window]
FIG. 3.
A-D: effects of iD2 abduction on the detection of stimuli applied to the moving digit in 3 subjects using a 2 alternative forced-choice (2AFC) version of the experiment (A and B) and using the main experimental procedure (C and D). Perceptual performance over time is plotted relative to movement onset (A and C) and the onset of 1st DI EMG (B and D). Error bars represent the 95% confidence intervals. The shaded area shows the 95% confidence interval for perceptual performance at rest. For all bins (20 ms), 
, perceptual performance during movement + stimulation trials was significantly lower than that observed at rest (P < 0.01); 
, no change.
RESULTS
Abstract
Introduction
Methods
Results
Discussion
References
View larger version (44K):
[in a new window]
FIG. 4.
A: sample movement and EMG traces for 1st dorsal interosseous (1st DI), abductor digiti minimi, and the forearm extensors aligned on the onset of the GO cue. B: averaged movement and EMG traces for the same muscles, aligned on the onset of movement (same scales as A). C and D: scatter plots of EMG onset as a function of RT for 1st DI (SE of the residual, 1.0 ms; regression constants: m = 0.99, b = 21 ms) (C) and abductor digiti minimi (SE of the residual, 3.5 ms; regression constants: m = 1.10, b = 51 ms; D). Linear regression lines are shown along with the slope (m) and the correlation coefficient (r).
View this table:
TABLE 1.
Temporal and kinematic parameters describing the performance of the motor task, iD2 abduction and the results of ANOVAs comparing values across 10 stimulation sites
View this table:
TABLE 2.
Proportion of stimuli detected in the three trial types as a function of the site of stimulation
200 ms before the onset of movement (Fig. 5A), performance declined precipitously beginning 60-80 ms before movement onset so that practically no stimuli were perceived after the onset of movement. A logistic function was found to best fit the data. A number of parameters were calculated from the logistic function (see METHODS), and these are detailed in Fig. 5A. In particular, the time of the peak decrease in perceptual performance preceded movement onset by 50 ms.
View larger version (23K):
[in a new window]
FIG. 5.
A and B: effects of iD2 abduction on the detection of stimuli applied to the moving digit in 41 subjects. Perceptual performance over time is plotted relative to movement onset (A) and the onset of 1st DI EMG (B). Plotted as in Fig. 3.
View larger version (39K):
[in a new window]
View larger version (16K):
[in a new window]
FIG. 6.
A-F: effects of iD2 abduction on perceptual performance over time for the 6 ipsilateral limb stimulation sites. Data plotted relative to EMG onset as in Fig. 5B. At each stimulation site, there was a significant, time-dependent reduction in the proportion of stimuli perceived during movement trials ( : P < 0.01) as compared with at rest. As distance increased, peak reductions occurred later and the minimum proportion of stimuli perceived increased.
View larger version (23K):
[in a new window]
FIG. 7.
A-D: logistic equation parameters plotted as a function of distance (proportion of body height) from the site of movement, iD2, for the 7 closest stimulation sites, all of which showed a time-dependent decease in perceptual performance. Both the timing of peak decreases in perceptual performance (A, ±10 ms) and the minimum predicted perceptual performance (B) showed strong positive correlations with distance. Maximum predicted perceptual performance varied very little with distance (D) and peak slope varied irregularly (C). R2A, coefficient of determination adjusted for the number of parameters in the fitted equation.
View larger version (27K):
[in a new window]
View larger version (16K):
[in a new window]
FIG. 8.
A-C: effects of iD2 abduction on perceptual performance over time for the 3 most distant stimulation sites. At each site, performance was best described by a linear function with a slope not significantly different from zero (P > 0.05). D: data were pooled to generate a single "Distant" data set. Plotted as in Fig. 5B.
70 ms were included in the analyses. The effect of site on performance only became significant (P < 0.05) when data from bins later than 50 ms were included in the analyses. The data from the three earliest bins at each site were pooled, and significant differences with perceptual performance at rest were seen (P < 0.05) at all sites but iSH (P = 0.09). Performance observed over the initial three bins (60 ms) for the seven locations closest to iD2 was not significantly different from perceptual performance during the movement + stimulation trials at the Distant sites (P > 0.05, Fischer exact probability test). All of these results suggest that site- and timing-independent modulation was present at all sites, along with superimposed time- and location-dependent modulation at the seven closest sites. To provide a precise estimate of the time of onset of the time- and location-dependent effect at each of the seven locations closest to iD2, performance in each time bin at each of these sites was compared with perceptual performance in the same time bin for the Distant data set (P < 0.05). The results are summarized in Fig. 9. The earliest sustained decrease (>3 consecutive bins) was at iD2 where performance was significantly decreased beginning in the bin centered 70 ms before EMG onset. The latest decrease was seen at iPG, 50 ms after EMG onset. These data were best described by a linear function with a positive slope significantly different from zero (P < 0.05), indicating that the time of onset of perceptual suppression increased significantly as the distance between the stimulation site and the moving digit (iD2) increased.
View larger version (15K):
[in a new window]
FIG. 9.
Time of onset of the 1st significant time-dependent decrease in perceptual performance during movement trials (20-ms bins), measured relative to 1st DI EMG onset, as a function of the distance (proportion of body height) from the site of movement, iD2. See text for further description.
View larger version (57K):
[in a new window]
FIG. 10.
A-C: surfaces showing time and location dependence of perceptual performance. A: pooled performance data during movement trials (z axis) plotted as a function of time (x axis), and distance (as a proportion of body height) between the moving digit and the site of stimulation (y axis). B: best-fit model surface for the data shown in A. Model surface was not significantly different from the actual performance data (P > 0.05). Model parameters are described in the text. C: contour map of the squared error between the surfaces plotted in A and B.
; Chapman et al. 1988, 1996; Rauch et al. 1985). With this in mind, for each of the seven stimulation sites that showed a time-dependent decrease in the proportion of stimuli perceived, movement + stimulation trials were examined to see if modifications in perceptual performance could be related to variations in kinematic parameters (peak amplitude, peak velocity, peak acceleration). Two different analyses were performed. First, the entire data set was divided into two groups: trials in which stimulation was applied before the onset of EMG and trials in which stimulation was applied after the onset of EMG. For each group, the kinematic parameters were compared across trials in which the stimulus was, or was not, perceived (t-tests). No significant differences were found. Second, the analysis was repeated, this time using only trials in which a stimuli was delivered within 30 ms of the time of the peak decrease in perceptual performance. In effect, this served to minimize the confounding effect of the time-dependent variation in detection previously described by concentrating on the time period where changes in detection were occurring. As shown in Fig. 11, significant differences were now obtained (filled symbols), especially at the stimulation sites closest to the moving digit, iD2. At these sites, the kinematic parameters were consistently, and usually significantly, smaller when the stimulus was perceived as compared with when the stimulus was not perceived. Sites further away showed more variability in their kinematic relationships.
View larger version (19K):
[in a new window]
FIG. 11.
A-C: comparison of kinematic parameters during movement + stimulation trials at the 7 sites showing a time-dependent decrease in detection, as a function of whether the stimulus was perceived (circles) or not perceived (diamonds). Analysis was restricted to those trials in which the stimulus was delivered within 30 ms of the time of peak decrease in perceptual performance. Significant t-test results (P < 0.05) are shown as filled symbols (nonsignificant, open symbols). Kinematic parameters were often significantly smaller in those trials where the stimulus was perceived than in those trials where the stimulus was not perceived. Peak velocity (B) was more frequently a significant factor (5 of 7 sites) than peak amplitude (A) or peak acceleration (C).
DISCUSSION
Abstract
Introduction
Methods
Results
Discussion
References
). In addition, electrical stimulation represented a reliable means of providing stimuli of identical duration and detectability to different body parts. The stimulating electrodes were placed at sites where stimulation affected mainly cutaneous afferents, although the electrodes on the finger sites also may have stimulated articular afferents traveling in the digital nerves. Subjects reported that the electrical stimuli did not feel particularly unnatural, just "extremely weak" at all stimulation sites. By calibrating stimulus intensity relative to a given detection level (~90% detected), baseline detection performance at rest was identical from site to site and subject to subject. This considerably facilitated intersite comparisons and permitted the pooling of data from multiple subjects. Finally, the inclusion of two contralateral sites identical to the ipsilateral sites made it possible to ensure that the observed distance-related effects could be distinguished from possible differences in the psychometric function at each stimulation site.
), was 0.965 and at most 0.995. Maximum negative bias at rest was therefore between 0.025 and 0.055.
had seen small elevations in detection threshold for sites contralateral to the movement, but previous experiments using suprathreshold stimuli showed no significant change in either perception (Chapman et al. 1987; Milne et al. 1988) or cortical SEPs (Cohen and Starr 1987; Rushton et al. 1981) when the stimuli were delivered to a limb other than the one producing the movement. Analyses indicated that the non-time-dependentdecrease was present at all stimulation sites and that the time-dependent decreases were superimposed on this. As explained above, an increase in negative bias could have produced this sort of result; the low false positive rate at rest makes it impossible to exclude this hypothesis. On the other hand, similar reductions were observed in test experiments using a bias-free experimental procedure. Previous studies have shown that spatial and cross-modal shifts in attention can modify tactile detection (Butter et al. 1989; Meyer et al. 1963; Post and Chapman 1991). Given that the subjects had to divide their attention between the motor and perceptual tasks, it is suggested that attentional influences might have been responsible for the non-time-dependent decrease in performance.
; Dyhre-Poulsen 1978), as well as the earliest modulation reported in evoked-potential and single-unit recording studies (Chapman et al. 1988; Cohen and Starr 1987; Jiang et al. 1991). On the other hand, the timing of the peak decrease in perceptual performance at the four distal sites coincided with the onset of agonist EMG (Fig. 7B). The possible significance of this with regard to the role of peripheral feedback in the reduction of detection during movement is discussed in the text following.
; Duysens et al. 1995; Post et al. 1994), with the largest change occurring at sites closest to the moving segment.
; Coulter 1974; Ghez and Lenzi 1971; Jiang et al. 1990a,b, 1991). Although the timing is consistent with the precentral motor areas playing a role, the spatial distribution observed here is not consistent with the pattern of modulation of cutaneous transmission elicited by intracortical microstimulation of primary motor cortex in nonhuman primates (Jiang et al. 1989, 1990a). In the latter studies, only sites on the same segment or distal to the activated muscle showed evidence of sensory gating, and microstimulation did not modulate transmission from the glabrous skin of the hand (although inputs from the hairy dorsum of the digit and hand were gated). In contrast, the present findings showed that sensory modulation extends to sites located proximal to the moving segment and to the glabrous skin of the digits. Providing the mechanisms are similar in humans and monkeys, this suggests that area 4 is unlikely to be the only source for the widely distributed gating effects reported here. Other areas (e.g., premotor cortical regions) also must be involved if the present findings are to be explained by a centrally originating signal.
, 1988; Huttunen and Hömberg 1990), although in a study of cortical SEPs elicited by suprathreshold cutaneous stimuli, Chapman et al. (1988) found that the time course for modulation was delayed considerably for passive as compared with active movements, i.e., the modulation followed movement onset instead of preceding it. A variety of mechanoreceptors are activated by passive movements, and evidence suggests that muscle spindle feedback contributes to the decrease in cortical SEPs during active and passive movement (Brooke et al. 1997). Cutaneous feedback also diminishes the amplitude of cortical SEPs (Jones et al. 1988).
; Scherrick 1964; Schmid 1961; Weisenberger 1994). Although the spatial pattern is similar to that seen here, no temporal shift over distance has been reported in masking studies, leaving the importance of these observations unclear with regard to the present findings.
50 m/s (Diabetes Control and Complications Trial Research Group 1995) and that the average distance between the proximal and distal sites was ~70 cm, then at most a 10-ms shift in the timing of the modulation would be explained by differences in travel times. This suggests that differences in path length did not contribute significantly to the results.
). In summary, physical factors likely made a small contribution to the results and it appears more likely that central mechanisms need to be invoked.
; Post et al. 1994). Although such a mechanism explains well the decreases in tactile detection with unchanged discrimination thresholds, it also predicts reduced magnitude estimates at low, but not high, intensities of stimulation. Using spatially distributed vibrotactile stimuli, however, we found that magnitude estimates were decreased at high, but not low, intensities (Post et al. 1994). Consideration of other models, linear and nonlinear inhibitory surrounds, led us to suggest that the latter provided the closest approximation to the experimental data (Chapman et al. 1996). The model developed here, based largely on logistic functions, thus provides some support for our suggestion that the underlying inhibitory processes may have a nonlinear distribution. Relative distances between body parts were used in the model. This permits the generation of simple, testable predictions for the amount of gating that will be observed at a given distance on the body surface from the site of motion. However, because the gating effects are exerted within the CNS where the representation reflects peripheral innervation density and not absolute size, it would be interesting to incorporate the relative distance between body parts at each of the somatosensory relays within the CNS into the model. Unfortunately, such data for humans are not currently available. In summary, the model represents an initial step toward developing a complete description of the effects of inhibitory mechanisms associated with movement-related gating on detection performance and provides a framework for future physiological studies aimed at characterizing the exact sources and sites of action of these mechanisms.
, movement-related gating controls may suppress redundant inputs that can be predicted from the motor command so that the detection of other unexpected or novel stimuli is enhanced. The spatial distribution and large temporal shift in the gating actions (hand vs. shoulder) may reflect the spread of nonspecific lateral inhibition originating from the body part in motion. Alternatively, the spatiotemporal gradient could be the result of "hard wired" gating, which reflects common patterns of use for the upper limb, thus providing an automatic reduction in afferent input during natural goal-directed movements.
| |
ACKNOWLEDGEMENTS |
|---|
We gratefully acknowledge the late R. Bouchoux for the construction of the apparatus, the excellent technical assistance of R. Albert, M. Bourdeau, D. Cyr, G. Filosi, C. Gauthier, and C. Valiquette, and the mathematical assistance of C. Bourget and J. F. Anger. We also thank Drs. T. Drew and W. Jiang for helpful comments on the manuscript.
This work was supported by the Medical Research Council of Canada. S. R. Williams was supported by bursaries from the Groupe de Recherche sur le Système Nerveux Central (GRSNC) and Fonds pour la Formation de Chercheurs et l'Aide à la Recherche (FCAR), Fonds de la Recherche en Santé du Québec (FRSQ), and Unimédia. J. Shenasa was supported by a bursary from the GRSNC. C. E. Chapman was a chercheur-boursier of the FRSQ.
| |
APPENDIX A |
|---|
Logistic equation used to fit perceptual data for each stimulation site, where P represents the proportion of stimuli perceived, t represents time (ms), mpk represents the peak slope, tpk represents the time of the peak slope, max represents the maximum proportion of stimuli perceived, and min represents the minimum proportion of stimuli perceived.
![<IT>P</IT>= [(max − min)(1/1 + <IT>e</IT><SUP>−4<IT>mpk</IT>(<IT>t−tpk</IT>)</SUP>)] + min](/content/vol79/issue2/fulltext/947/img001.gif)
|
(A1) |
| |
APPENDIX B |
|---|
The equation for the model surface was constructed using the following reasoning. Because logistic functions provided a good description of the effects of time on perceptual performance (Figs. 5 and 6) and distance on function minima (Fig. 7B), this type of function again was used to model these effects. A linear function was used to model distance-dependent modifications in the timing of peak decreases in performance based on the finding that a linear model provided an adequate description of the effect of distance (Fig. 7A). Because maximum performance was invariant between sites, it was modeled with a constant. The resultant model can be described as follows, where Fdist is a function representing the proportion of stimuli detected, i represents time (ms), j represents the distance between the stimulation site and iD2 (proportion of subject height), and max represents the maximum proportion of stimuli detected.
Distance-dependent component (affects mostly minimum predicted perceptual performance)
![Distlogistic<SUB>(<IT>j</IT>)</SUB>= 1/[1 + <IT>e</IT><SUP>(−96<IT>j</IT>+41)</SUP>]](/content/vol79/issue2/fulltext/947/img002.gif)
|
(B1) |
|
(B2) |
![Timelogistic<SUB>(<IT>i</IT>)</SUB>= 1/[1 + <IT>e</IT><SUP>(0.046<IT>i</IT>+(0.36+Shift<SUB>(<IT>j</IT>)</SUB>))</SUP>]](/content/vol79/issue2/fulltext/947/img004.gif)
|
(B3) |
|
(B4) |
![<IT>F</IT>dist<SUB>(<IT>i,j</IT>)</SUB>= max [Distlogist + Timelogist(1 − Distlogist)]](/content/vol79/issue2/fulltext/947/img006.gif)
|
(B5) |
|
(B6) |
| |
FOOTNOTES |
|---|
Address for reprint requests: C. E. Chapman, Centre de Recherche en Sciences Neurologiques, Faculté de Médecine, Université de Montréal, PO Box 6128, Station Centre Ville, Montréal, Québec H3C 3J7, Canada.
Received 27 March 1997; accepted in final form 15 October 1997.
| |
REFERENCES |
|---|
|
Abstract
Introduction
Methods
Results
Discussion
References
|
|---|
This article has been cited by other articles:
|
|
 |
|
  C. E. Chapman and E. Beauchamp Differential Controls Over Tactile Detection in Humans by Motor Commands and Peripheral Reafference J Neurophysiol, September 1, 2006; 96(3): 1664 - 1675. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. E. Chapman and E.-M. Meftah Independent Controls of Attentional Influences in Primary and Secondary Somatosensory Cortex J Neurophysiol, December 1, 2005; 94(6): 4094 - 4107. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E.-M. Meftah, J. Shenasa, and C. E. Chapman Effects of a Cross-Modal Manipulation of Attention on Somatosensory Cortical Neuronal Responses to Tactile Stimuli in the Monkey J Neurophysiol, December 1, 2002; 88(6): 3133 - 3149. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. R. Williams and C. E. Chapman Time Course and Magnitude of Movement-Related Gating of Tactile Detection in Humans. III. Effect of Motor Tasks J Neurophysiol, October 1, 2002; 88(4): 1968 - 1979. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. R. Williams and C. E. Chapman Time Course and Magnitude of Movement-Related Gating of Tactile Detection in Humans. II. Effects of Stimulus Intensity J Neurophysiol, August 1, 2000; 84(2): 863 - 875. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M-C Perreault, S J Shefchyk, I Jimenez, and D A McCrea Depression of muscle and cutaneous afferent-evoked monosynaptic field potentials during fictive locomotion in the cat J. Physiol., December 15, 1999; 521(3): 691 - 703. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Salimi, T. Brochier, and A. M. Smith Neuronal Activity in Somatosensory Cortex of Monkeys Using a Precision Grip. I. Receptive Fields and Discharge Patterns J Neurophysiol, February 1, 1999; 81(2): 825 - 834. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
