| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1Centre for Neuroscience Studies, 2Department of Anatomy and Cell Biology, and 3Department of Medicine, Queen's University, Kingston, Ontario, Canada
Submitted 31 January 2008; accepted in final form 28 April 2008
 |
ABSTRACT |
|---|
|
|
|
INTRODUCTION |
|---|
|
Many researchers have examined the flexibility of such rapid responses, traditionally called reflexes, in the upper limb by noting how they are altered by voluntary goals (Capaday et al. 1994
; Colebatch et al. 1979; Crago et al. 1976; Evarts and Granit 1976; Hagbarth 1967; Hammond 1956; Jaeger et al. 1982a; Lee and Tatton 1982; Lewis et al. 2006; Rothwell et al. 1980). These studies typically issued a verbal instruction such as "resist/let go" (Colebatch et al. 1979; Hammond 1956; Rothwell et al. 1980), "flex/extend" (Hagbarth 1967), or "compensate/do not intervene" (Crago et al. 1976) that indicated how a subject was to respond to a mechanical perturbation. Indeed, many of these studies have shown that long-latency reflexes are substantially modulated by voluntary goals. Although these pioneering studies revealed that long-latency reflexes are modifiable by subject intent, the use of verbal instructions severely limited the breadth of conditions that could be explored to the subset of behaviors where verbal cues have a reliable interpretation.
The limited range of reflex sophistication that can be explored via verbal instructions is an impediment to testing recent theories of motor control. Specifically, optimal feedback control posits that the extensive repertoire of behavior exhibited by the volitional motor system is accomplished via intelligent manipulation of sensory feedback (Todorov 2004; Todorov and Jordan 2002). We have previously suggested that one neurophysiological implication of such manipulation is that long-latency reflexes, like volitional movement, should be capable of a rich assortment of behaviors because reflexive and volitional control are intimately linked (Kurtzer et al. 2008; Scott 2004); effectively, they are part of the same ongoing control process. Such a close link between reflexive and voluntary responses makes many experimental predictions but belies the traditional distinctions made between reflexive and voluntary responses (see DISCUSSION). Therefore we have discarded the term reflex in favor of a simple empirical separation between slow (voluntary) and rapid motor responses that fits better within our framework (see METHODS).
Our general hypothesis is that these rapid motor responses are capable of all the sophistication attributed to voluntary control within the constraints of their limited processing time. To test for such extensive sophistication, we created a paradigm that can examine these responses under a wide range of behaviors akin to the prevalent methodology used to study the volitional motor system (Shadmehr and Wise 2005). Briefly, subjects are shown a target while they maintain posture at the start position and are trained to respond to an unpredictable perturbation by quickly placing their hand inside the target, thus allowing us to quantify how rapid responses change when targets are placed at various locations. Unlike the ambiguity of verbal instructions, target metrics such as position, size, and shape, along with timing constraints, explicitly define the goal of the task and can be monitored and controlled. In the present study, we tested whether rapid responses of muscles spanning the shoulder and elbow were sensitive to the direction and distance of spatial targets in both one and two dimensions. These factors are known to modify voluntary control and we hypothesized the same metrics would modify rapid responses.
In our first experiment we placed targets such that the same perturbation pushed the arm toward one of the targets and away from the other, thus allowing us to establish whether rapid responses are categorically sensitive to target direction, analogous to previous studies using verbal instructions. In the second experiment we placed targets at five positions along one axis to determine whether rapid responses scale continuously with target distance. Last, we placed targets radially around the hand to establish how rapid responses are spatially tuned to target direction. Our results indicate that there is a clear distinction between short-latency and long-latency rapid responses for all the muscles we examined, including shoulder flexors, shoulder extensors, elbow flexors, and elbow extensors. Short-latency rapid responses were never modulated by target position even in conditions of relatively high stimulus predictability. In contrast, long-latency rapid responses were robustly modified in all three experiments even in conditions of relatively low stimulus predictability and the changes they expressed progressively approached the voluntary response. This work was previously presented in abstract form (Pruszynski et al. 2007).
|
|
METHODS |
|---|
|
A total of 18 subjects (11 males and 7 females, aged 21–35 yr, 17 right-handed) participated in 37 experimental sessions. All subjects were neurologically unimpaired, had normal or corrected-to-normal vision, and gave informed consent according to a protocol approved by the Queen's University Research Ethics Board.
Apparatus and experimental paradigm
Subjects performed the experiments with a robotic exoskeleton (KINARM, BKIN Technologies, Kingston, Ontario, Canada) that permits combined flexion and extension movements of the shoulder and elbow to move the hand in the horizontal plane (Scott 1999). Furthermore, KINARM can independently apply mechanical loads to the shoulder and/or elbow and record kinematic variables of the joints and hand. Target lights and simulated hand feedback were presented to the subject in the horizontal plane via a heads-up display composed of an overhead projector and semitransparent mirror. Throughout the experiment, direct vision of the entire arm and hand was occluded and hand feedback was removed prior to perturbation onset so that responses were guided only by proprioception.
EXPERIMENT 1: CATEGORICAL MODULATION BY TARGET DIRECTION. Subjects (n = 11) maintained their hand in a small central area (radius = 0.3 cm) against a background load (±2 Nm) that activated either the elbow flexors or extensors. The central area was positioned such the shoulder and elbow angles were 45 and 90°, respectively. Subjects were then presented with a peripheral target (PT, radius = 20 cm) located on either the left or right side of their hand and thus requiring predominantly elbow flexion or extension movements (Fig. 1A). After a random hold (1–4 s) at the central area, a rapid torque perturbation (±2 Nm at elbow) displaced the hand either toward (IN condition) or away from (OUT condition) the center of the displayed PT (Fig. 1B). Subjects were instructed not to anticipate the perturbation and to place their hand in the PT as quickly as possible after perturbation onset. On completing the trial, subjects were given feedback to indicate success (PT, filled green) or failure (PT, filled red and an auditory tone) based on preset speed and accuracy criteria (hand within PT for 1,000 of the 1,300 ms after perturbation onset). A detailed timeline of each trial is presented in Fig. 1C. Fifteen repeats of each condition (2 muscle groups and 2 targets) were performed in a blocked sequence, randomized across subjects, for a total of 60 trials.
|
EXPERIMENT 2: GRADED MODULATION BY TARGET DISTANCE. The experimental protocol is similar to experiment 1 with the exception of PT placement and size. Subjects (n = 9) were presented with one of five potential PTs (radius = 5 cm) located at 0, 3, and 6 cm to the left and right of the hand and thus requiring predominantly elbow flexion or extension (Fig. 1D). Twenty-five repeats of each condition (2 muscle groups and 5 targets) were performed in a blocked sequence, randomized across subjects, for a total of 250 trials.
EXPERIMENT 3: SPATIAL TUNING OF RAPID RESPONSES. The experimental protocol is similar to experiment 1. Subjects (n = 6) were shown one of four potential PTs (radius = 5 cm) that were equally distributed around the hand's starting position (CT) at a constant distance of 6 cm and placed at locations requiring predominantly elbow flexion, elbow extension, shoulder flexion, or shoulder extension (Fig. 1E). Twenty repeats were collected in each condition (4 muscle groups and 4 targets) in a blocked sequence, randomized across subjects, for a total of 320 trials.
CONTROL EXPERIMENT: DETERMINING VOLUNTARY ONSET.
A common problem in previous experiments that investigate reflexes is deciding what components of muscular activity are reflexive. In fact, the philosophical and semantic legacy of the term has made it unclear what responses (if any) qualify (see DISCUSSION). Further, recent theoretical advances suggest that the distinction between reflex and voluntary is obsolete because the underlying processes are inherently linked (Scott 2004). We sought to explicitly avoid these historical considerations so we chose a simple operational definition based on a control experiment that empirically determined voluntary muscle onset; any muscle activity occurring prior to voluntary muscle onset is then termed a rapid motor response.
Voluntary muscle onset was determined by applying perturbations that did not engage significant early phasic responses (Evarts and Vaughn 1978; Hammond 1956; Jaeger et al. 1982b). In the absence of these early phasic responses, we could clearly identify the large and prolonged burst activity typically associated with voluntary muscle activity (compare traces in Fig. 2A). Subjects (n = 9) performed the same protocol as in experiment 1 with perturbations reduced to ±0.5 Nm and delivered at the shoulder joint. Importantly, although the perturbation was applied at the shoulder, background loads were applied at the elbow (±2 Nm) as in experiment 1 to tonically activate the muscle of interest. We applied the perturbation at the shoulder since it allowed us to generate a perceptually robust perturbation without evoking substantial responses at the elbow. Unlike our other experiments, no formal speed and accuracy criteria were used. Rather, subjects were encouraged to move to the PT as quickly as possible after perturbation onset. Twenty-five repeats were completed in each condition for a total of 50 trials.
|
; Hammond 1956; Jaeger et al. 1982b; Kimura et al. 2006; Lee and Tatton 1975; Matthews 1986; Rothwell et al. 1980; Tatton and Lee 1975). Muscle activity
Surface EMG recordings were obtained from six upper-limb muscles involved with flexion or extension at the elbow and/or shoulder: brachioradialis (Br, monoarticular elbow flexor), biceps long (Bi, biarticular flexor), triceps lateral (TLat, monoarticular elbow extensor), triceps long (TLo, biarticular extensor), deltoid posterior (DP, monoarticular shoulder extensor), and pectoralis major (PM, monoarticular shoulder flexor). Prior to electrode placement, the skin was cleaned and abrased with rubbing alcohol and the electrode contacts were covered with conductive gel. Electrodes (DE-2.1, Delsys, Boston, MA) were placed on the belly of the muscle oriented along the muscle fiber and the reference electrode (Dermatrode, American Imex, Irvine, CA) was attached to the ankle. To assess the quality of each EMG signal, we performed a set of maneuvers found to elicit high levels of activation for each muscle in the plane of the task. EMG signals were amplified (gain = 103 or 104) and band-pass filtered (20–450 Hz) by a commercially available system (Bagnoli, Delsys) then digitally sampled at 1,000 Hz.
Rapid response epochs
We were primarily interested in comparing different epochs of muscle activity occurring before our operationally defined voluntary muscle onset (see CONTROL EXPERIMENT); such responses are referred to as rapid motor responses. Taking into account previous reports (Bonnet 1983; Crago et al. 1976; Lee and Tatton 1975; Mortimer et al. 1981; Nakazawa et al. 1997; Rothwell et al. 1980) and our own observations of reliably timed phasic events across subjects and muscles, we defined three distinct epochs of activity in temporal order: Response 1 (R1, 20–45 ms), classically referred to as the short-latency or spinal stretch reflex (Pierrot-Deseilligny and Burke 2005); Response 2 (R2, 45–75 ms), often referred to as the long-latency reflex (Hammond 1954; Matthews 1986); and Response 3 (R3, 75–105 ms), sometimes referred to as long-latency reflex or triggered response (Crago et al. 1976; Rothwell et al. 1980). Note that the timing of our epochs most closely mirror M1, M2, and M3 as proposed by Lee and Tatton (1975) and our choice of R1, R2, and R3 is largely to avoid confusion with the common abbreviation for primary motor cortex (M1). For some analyses, we included time epochs between –100–0 and 120–180 ms, which were defined as baseline (PRE) and voluntary (VOL), respectively. Note that our results were not qualitatively changed by modestly altering epoch onset or duration.
Data analysis
FILTERING AND NORMALIZATION. The combined compliance of the robot and the subject's arm introduces a shift between the commanded perturbation by the motors and the corresponding start of limb motion. We determined this delay by tightly coupling an accelerometer (EGA Series, Entran, Montreal, Canada) to the arm of several subjects while they performed a perturbation task in the KINARM robotic system. The time between issuing the perturbation command and the onset of arm acceleration was about 10 ms and was similar across several subjects and load conditions. Thus all data were aligned on perturbation onset as determined by the KINARM and then shifted by a fixed delay of 10 ms. Joint and hand position were obtained directly from the KINARM and then low-pass filtered (20 Hz, two-pass, sixth-order Butterworth). EMG signals were band-pass filtered (25–250 Hz, two-pass, sixth-order Butterworth), full-wave rectified, and normalized by their mean activity for a 2-Nm load.
BEHAVIOR AND KINEMATICS.
We were most interested in comparing kinematic and muscle responses for the same imposed perturbation but different target positions. To quantify the static changes in behavior as a function of target position, we analyzed the hand's final position (350 ms after perturbation onset). We used principal component analysis to generate 95% confidence ellipses of final hand position. If these ellipses did not overlap, we deemed the positions to be significantly different. Furthermore, we used the receiver-operator characteristic (ROC) technique to determine when these differences in position were significant (Green and Swets 1966). For each time step (1 ms) we generated an ROC curve representing the probability that an ideal observer could discriminate the PT based on the joint position for the same perturbation. Values of 0 and 1 indicate perfect discrimination, whereas a value of 0.5 indicates performance at chance. Discrimination was deemed significant when the ROC area remained <0.25 or >0.75 for five consecutive samples (Corneil et al. 2004). Note that changing threshold criteria did not qualitatively alter our results.
MUSCLE ACTIVITY. When comparing responses across predefined epochs of muscle activity (PRE, R1, R2, R3, VOL), we calculated the mean level of activity for each muscle in that epoch on a trial-by-trial basis. For each muscle, we used the Wilcoxon rank-sum test (also known as a Mann–Whitney U test), a nonparametric comparison of medians, to evaluate the hypothesis that the median muscle activity differed with target position. Similar results were found using a t-test.
To identify when responses differed across target conditions without a priori assumptions about epochs, we generated an ROC curve for every 1-ms sample to calculate the probability that an ideal observer could discriminate the target position based on the EMG response for the same perturbation but different PTs (ROC <0.25 or >0.75). We also calculated the point when the ROC curve began to deviate from chance (Thompson et al. 1996), termed the "knee," by regressing the ROC values located 15 ms around the discrimination point then calculating the time when this line intersected the preperturbation ROC results. Note that calculating the ROC knee is a formalized method of determining when two signals initially diverge and replaces previous attempts to identify this point by visual inspection (Crago et al. 1976; Evarts and Vaughn 1978; Marsden et al. 1976).
In experiment 3, we determined how responses were tuned to two-dimensional target positions by fitting the data with a plane (Kurtzer et al. 2005, 2006). The resulting plane coefficients (X, Y) describe how the EMG activity is related to spatial target position and can be used to calculate the preferred target direction (PD), where PD = atan2 (Y, X) and 0, 90, 180, and 270° represent right, ahead, left, and behind in external space. A Rayleigh test was applied across the population of PDs (for a particular muscle and epoch) to determine whether the net tuning was unimodally distributed (Batschelet 1981).
AKAIKE'S INFORMATION CRITERIA.
For experiment 2, we used Akaike's Information Criterion (AIC) to judge the merits of potential models relating muscle activity within our predefined epochs to target position. AIC is a principled technique for choosing a parsimonious model from a set of candidates by providing a metric of model quality that balances fit and complexity (Burnham and Anderson 2002). In AIC, model quality is proportional to the likelihood (£) of a candidate model (
'), given the experimental data (x), and complexity is accounted for by K, the number of free parameters in the candidate model, expressed as
 | (1) |
AIC = AIC – min (AIC)]. If this difference was less than an arbitrarily chosen threshold (AIC <2), the model was deemed acceptable (Burnham and Anderson 2002).
We compared four possible models relating EMG activity to target position (for each muscle sample and epoch). 1) A constant function that had one free variable, f(p) = b, where p is the target position (defined as 1 to 5) and b is a positive constant. 2) A step function defined as two piecewise continuous constant functions with three free variables, two constant levels, and the switch point, f(p) = b, if P 
x and f(p) = c, otherwise, and constrained such that c >a and that x was within the range of tested positions. 3) A linear function; f(p) = ap + b, where a is a positive constant and b is unconstrained. 4) A sigmoid that had four free variables, f(p) = a/{1 + exp[–(p – x)/c]} + d and constrained such that a >0, x was within the range of positions. Note that as c approaches 0, a sigmoid approaches a step function; to avoid this we constrained the sigmoid to rise over a minimum of three consecutive target positions (c >0.34). In some analyses the constant function is referred to as "no modulation," the step function is referred to as "discrete modulation," and the grouping of linear and sigmoidal functions is termed "graded modulation." Although our candidates for constant and discrete representation are fundamentally limited to a constant and step function, in principle, any smooth and monotonically increasing function could have been included to describe graded modulation. We included both linear and sigmoidal functions to permit the possibility of no saturation, or saturation at either or both target extremes.
|
|
RESULTS |
|---|
|
FEATURES OF BEHAVIOR. This experiment examined whether subjects categorically modify their rapid motor responses as a function of spatial target direction when subjects are presented target and load conditions in blocked order. Subjects learned the task quickly and had little difficulty reaching the goal target within the imposed speed and accuracy constraints, performing with a mean success rate of 99% (SD 1). Figure 3A shows the mean hand trajectories and endpoint variability for the IN (gray) and OUT (black) conditions from a representative subject. Note that the IN and OUT correspond to those conditions where the perturbation displaces the hand into or out of the presented target, respectively. The initial segments of the IN and OUT conditions were overlapping, as expected, because they were generated by the same perturbation. However, hand paths rapidly deviated toward the PT (Fig. 3A), which resulted in significant differences in final hand position. We used ROC analysis (see METHODS) to determine when elbow and shoulder kinematics (Fig. 3, B and C) were modified by target position. For this subject, target-dependent differences in elbow and shoulder kinematics were significant 130 and 121 ms after perturbation onset, respectively (see where traces cross 0.75 or 0.25 in Fig. 3, D and E). These results were typical across the population with target-dependent effects becoming significantly different at 139 ms (SD 24) for the elbow and 136 ms (SD 22) for the shoulder.
|
The mean response exhibits increasing sensitivity to target position, that is, the difference between IN and OUT conditions increased with time after perturbation onset. The earliest rapid response (R1, 20–45 ms) was indistinguishable for the two target positions; medium-latency rapid responses (R2, 45–75 ms) were excited by the perturbation and additionally scaled by target position and late rapid responses (R3, 75–105 ms) were mostly evoked or suppressed by target position. For the exemplar muscle, binned analysis reveals that whereas all epochs are significantly elevated from baseline (Wilcoxon rank-sum, P < 0.01) there are no target-dependent differences in either the baseline [rank-sum statistic, Rs(15,15) = 240, P = 0.77] or R1 epochs [Rs(15,15) = 226, P = 0.81]. In contrast, target-dependent modulation of muscle activity was significant in the R2 [Rs(15,15) = 173, P = 0.01], R3 [Rs(15,15) = 120, P < 10–6], and VOL epochs [Rs(15,15) = 120, P < 10–6, Fig. 3H]. As in the exemplar muscle, target-dependent modulation across the population of muscle samples was rare for early periods and typical to universal for later periods (Wilcoxon rank-sum, P < 0.01): PRE (2 of 44 muscle samples, 5%), R1 (5%), R2 (32%), R3 (86%), and VOL (100%). Across the population of collected muscles, only the R2, R3, and VOL epochs show a significant change in activity as a function of target position (paired t-test, P < 0.01). Similar results were found in an analysis of the individual muscles (Table 1).
|
|
As in the blocked conditions, upper-limb muscle responses exhibited an increasing sensitivity to target position. Figure 5, A and B presents the same exemplar muscle as in Fig. 3, F and G; again, there is little difference in R1 and an increasing difference in the R2, R3, and VOL epochs. Across the population of muscle samples only the R2, R3, and VOL epochs show a significant change in activity as a function of target position (paired t-test, P < 0.01). ROC analysis reveals that the knee of the curve for random conditions, which occurs at 64 ms (SD 18), is not significantly different from the blocked condition (t-test, t59 = –1.7, P = 0.1). In contrast, there was a significant delay in the time of significance, which was pushed back to 95 ms (t59 = –3.6, P < 10–3).
|
|
Experiment 2: graded modulation by target distance
FEATURES OF BEHAVIOR. Here, we assessed whether rapid motor responses are modulated in a graded fashion as a function of target distance or whether they are limited to categorical changes according to target direction (as shown in experiment 1). Subjects had no difficulty learning the task and were extremely successful with a success rate of 96% (SD 3). Population mean hand trajectories for each target position in both flexor and extensor conditions are shown in Fig. 7, A and B, respectively. All subjects showed systematic differences in final hand position for the five target conditions and temporal differences in elbow kinematics between the extreme targets were significant 135 ms (SD 30) after perturbation onset (Fig. 7C). Similar temporal trends are found at the shoulder (Fig. 7D). Furthermore, these trajectories reveal a visible rank-order trend according to target position as early as any separation.
|
Since the qualitative observation of graded modulation across muscles and subjects may simply reflect the temporal smearing of discrete responses, we used Akaike's Information Criterion (AIC; see METHODS) to determine the best candidate model for each collected muscle sample (Fig. 8A). In the R1 epoch, almost all muscles (29 of 34, 85%) were best described by a constant function, indicating no modulation as a function of target position. The R2 epoch was generally best described by a constant function (59%) with some muscles exhibiting discrete (18%) or graded (linear or sigmoidal) modulation by target position (24%). R3 was most commonly described by a linear function (53%) with a total of 21 muscles that exhibited graded modulation by target position. Last, the VOL epoch was best described by graded modulation (74%) with most muscles exhibiting a sigmoidal (50%) or linear (24%) relationship. A similar progression from constant to graded modulation was found when inspecting individual muscles (Table 2).
|
|
AIC) to assess the adequacy of the candidate models (Fig. 8B). Based on this analysis, R1 activity was adequately explained by the constant function for 31 of 34 (91%) muscles. Furthermore, activity in the R2, R3, and VOL epochs was adequately explained by a linear or sigmoidal function (graded modulation) for 59, 82, and 88% of muscle samples, respectively. Taken together, these results suggest a progression of target selectivity, with R1 exhibiting constant or no modulation by target position, whereas R2, R3, and VOL epochs were modulated by target position in a graded, rather than discrete, fashion. Experiment 3: spatial tuning of rapid responses
FEATURES OF BEHAVIOR. We distributed four targets radially around the central area to determine whether rapid motor responses could be spatially modulated as a function of two-dimensional target position. As in previous experiments, subjects had little difficulty performing the task, with an average success rate of 93% (SD 8), and no substantive differences in performance across the targets or load conditions. Figure 9, A–D shows hand kinematics for an exemplar subject in all four load conditions. Note that the target locations are identical for all the load conditions and that the only difference was the background and perturbation loads chosen for each muscle group (elbow flexor, elbow extensor, shoulder flexor, shoulder extensor). As in previous experiments, spatial analysis of hand position reveals that every subject achieved significantly different hand endpoints for all targets under all load conditions. Temporal analysis of shoulder and elbow angles revealed rapid target-dependent changes (Fig. 9, E–H). ROC analysis between IN and OUT targets reveals significant differences between elbow and shoulder kinematics at 143 ms (SD 21) and 129 ms (SD 19) after perturbation onset, respectively.
|
|
|
|
|
DISCUSSION |
|---|
|
In the present study, we examined how rapid motor responses of the human upper limb are modulated by the metrics of a visuospatial task. In all of these experiments we noted a progression of motor sophistication whereby earliest responses showed little or no sensitivity to the spatial goal, whereas the later responses progressively approached the voluntary response. In experiment 1 we showed that long-latency rapid responses (R2: 45–75, R3: 75–105 ms postperturbation) were categorically scaled up and down by the direction to the spatial target. This pattern of scaling was robust across both subjects and muscles with 86% of muscle samples showing significant target-dependent modulation. In contrast, only 5% of samples were modulated within the short-latency rapid response epoch (R1: 20–45 ms). Comparable results were found when load and targets were presented in random order, although the effect size was reduced, suggesting that R2 and R3 epochs incorporate both prior information and current sensory evidence in generating their response. In experiment 2, we established that modulation of long-latency rapid responses, particularly R3, is not merely a categorical switch that is turned "on" for distant targets and "off" for close ones. Rather, the response is graded based on the distance to the spatial target. Last, in experiment 3, we showed that the R2/R3 responses are spatially tuned to two-dimensional target position, indicating that they incorporate the multijoint requirements of the task in their response.
Although many studies have investigated whether rapid motor responses can be modulated by subject intent at the elbow (Colebatch et al. 1979; Crago et al. 1976; Evarts and Granit 1976; Hagbarth 1967; Hammond 1956; Rothwell et al. 1980), wrist (Calancie and Bawa 1985; Jaeger et al. 1982a; Lee and Tatton 1982), finger (Capaday et al. 1994; Marsden et al. 1981; Rothwell et al. 1980), ankle (Gottlieb and Agarwal 1979, 1980; Ludvig et al. 2007), and jaw (Pearce et al. 2003), most of these have focused on individual muscles at that joint. An important feature of the present study is that our experiments were designed to test a large subset of muscles acting to move the arm in the plane of the task including monoarticulars, biarticulars, flexors, and extensors of the shoulder and elbow. Accordingly, our results demonstrate that although muscles exhibit some differences in detail, they all showed the same general pattern of modulation as a function of spatial goal, suggesting that modulation of rapid motor responses is a common feature for upper-limb muscles.
Advantages of our visuospatial perturbation paradigm
Our paradigm builds on a prominent methodology that uses verbal instructions to instruct a subject how to respond to an imposed perturbation (Capaday et al. 1994; Colebatch et al. 1979; Crago et al. 1976; Hagbarth 1967; Hammond 1956; Lewis et al. 2006; Rothwell et al. 1980). Although verbal instructions have varied widely across researchers and experiments, including "resist/let go" (Colebatch et al. 1979; Hammond 1956; Rothwell et al. 1980), "flex/extend" (Hagbarth 1967), and "compensate/do not intervene" (Crago et al. 1976), they are inherently categorical and restricted to the ambiguities of natural language. They are also not amenable to animal models that could explore the neural substrates of sophisticated rapid motor responses. In the present study we minimized the dependence on verbal instructions by specifying and enforcing an explicit spatial goal. In our first experiment we replaced the above-cited categorical verbal instructions with two potential targets placed such that the perturbation displaced the hand directly into one target and directly out of the other. Such an approach is advantageous because the task is explicitly defined and performance can be directly measured, allowing us to explore the flexibility of rapid motor responses under conditions that would be difficult with verbal cues. For example, we established that the modulation is graded by simply distributing visual targets at various distances from the start position. Furthermore, we showed that rapid responses could be adjusted to varying multijoint requirements at the shoulder and elbow by distributing targets in two dimensions. The analogous verbal cues for these experiments, perhaps "increase/decrease resistance by 25, 50, and 75%" for the gradation experiment, and "respond to the elbow perturbation by flexing your shoulder and elbow" for the multijoint experiment, would be difficult for subjects to adhere to and likely impossible to monitor or enforce.
It is important to note that a few researchers have used spatial targets to study the modulation of rapid motor responses (Brown and Cooke 1986; Day et al. 1983; Koshland and Hasan 2000). However, these studies differ from the present experiments because they provided precise information about when to generate a specific movement and thus tested how rapid motor responses change during the initiation of a reach. For example, Koshland and Hasan (2000) reported that R2 and R3 responses depend on the impending direction of arm movement. In that study, subjects were instructed to reach a target in response to either a "small" or "large" perturbation in the same direction, one of which always followed 1 s after an auditory warning. Notably, both perturbations used were small and transient resulting in similar displacements (both <1°) relative to the required movement (
20°). As a result, subjects could reasonably prepare and initiate the same motor commands prior to the mechanical perturbation. However, our recent work illustrates a profound change in neural processing in primary motor cortex during the transition from posture to movement (Kurtzer et al. 2005). This abrupt change in the state of neural control makes it difficult to ascertain whether their applied perturbations were probing a postural, movement, or a transition control state. Further, rapid motor responses are known to be modulated by motor-neuronal excitability (Matthews 1986), which is likely changing during the initiation of movement. Our study tried to control both of these factors by introducing a long and random hold period and/or randomizing the perturbation direction. This manipulation made it impossible for subjects to consistently anticipate perturbation onset and initiate a movement before perturbation onset. Rather, they had to remain in posture until perturbation onset. Thus our study establishes that changes in rapid motor responses as a function of target position do not require the overt initiation of a reaching movement; appropriate long-latency rapid responses can be chosen and implemented after perturbation onset.
What is a reflex?
Reflexes are broadly defined as a stereotyped transformation from a sensory stimulus to a motor event (Kandel et al. 2000). However, past research in physiology, psychology, and philosophy has yielded little agreement about what specifically constitutes a reflex response (Prochazka et al. 2000). One class of definitions offered by Prochazka and colleagues (2000) is that reflexes are those responses that are automatic, that is, they are transformations that cannot be consciously modified or suppressed. However, given that even the monosynaptic stretch reflex is modifiable under certain conditions (Stein and Capaday 1988), it is unclear that any muscular responses can be characterized as a reflex by this definition. A second class of definitions is based on the anatomical or physiological nature of the mapping between the input and output. When the mapping is simple and well defined, the response is a reflex and when the pathway is complex, the response is voluntary. However, the demarcation of complexity is inherently subjective and changes with time as pathways become better characterized.
In the present experiment we wanted to avoid the philosophical and semantic debate surrounding the term reflex because it is unclear that such a debate is fruitful in furthering our understanding about the underlying neurophysiological processes. In fact, our own experience suggests that the term is a substantial obstacle. Therefore we turned to an operational definition whereby we empirically defined a set of rapid muscular responses because they appear prior to the earliest muscle activity in a control experiment designed to determine kinesthetic reaction time (Evarts and Vaughn 1978; Hammond 1956; Jaeger et al. 1982b). Our reaction time manipulation eliminated the earliest phasic muscle responses by reducing perturbation magnitude, leaving only a large and maintained pattern of muscle activity commonly attributed to a voluntary response. The earliest of these responses occurred about 100 ms after perturbation onset, a result that is consistent with many other studies (Calancie and Bawa 1985; Hammond 1956; Jaeger et al. 1982b; Kimura et al. 2006; Lee and Tatton 1975; Matthews 1986; Rothwell et al. 1980; Tatton and Lee 1975) and corresponds to a similarly timed volley of activity in our main experimental tasks (Figs. 3, 5, 7, and 10). It should be noted that although we labeled this maintained volley of activity as voluntary, we could just as well have chosen R4. In summary, we specifically avoided the term reflex and replaced it with rapid motor response, which refers only to a temporal window of events occurring prior to kinesthetic muscle onset as determined empirically.
Short-latency (R1) rapid responses are not modulated by target position
In agreement with many previous studies, our results demonstrate that R1 responses of the upper limb during postural control are not quickly modified by subject intent (Crago et al. 1976; Hammond 1956; Marsden et al. 1972b; Rothwell et al. 1980). Such inflexibility is surprising given that spinal interneurons receive descending input from cortical areas and exhibit task-specific preparatory activity (Prut and Fetz 1999). However, in that study the muscles were silent in the preparatory period that may have allowed subthreshold changes in muscle activity that would not be visible in the recorded EMG (Capaday and Stein 1987). Because such changes are known to modify R1 responses (Capaday et al. 1994; Marsden et al. 1972a; Matthews 1986) it would be informative to establish whether changes in preparatory activity are present without changes in EMG when the muscles are above threshold.
The lack of R1 modulation within our postural task is contrasted by changes observed across behaviors such as the profound modulation of all rapid motor responses in the upper limb between posture and movement (Mortimer et al. 1981) or in the lower limb between stance and walking (Komiyama et al. 2000) or running (Duysens et al. 1993). In fact, rapid responses of both upper and lower limbs have been shown to change over the course of a cyclical movement such as gait (Akazawa et al. 1982; Capaday and Stein 1986; Forssberg et al. 1975; Zehr et al. 2003), sinusoidal tracking (Dufresne et al. 1980; Johnson et al. 1993), or hand cycling (Zehr and Chua 2000), illustrating that spinal circuitry can generate sophisticated and flexible responses under certain conditions.
In addition to changes across tasks, several studies have shown systematic changes in R1 responses over long training periods in a single task (Christakos et al. 1983), particularly when R1 magnitude is directly reinforced (Wolf and Segal 1996; Wolpaw 1983; Wolpaw et al. 1983). Although our experiments could have reasonably elicited modulation of the R1 response we did not find such modulation. In fact, two of the authors (AP and IK) have performed the task over several thousands of trials and neither shows modulation of R1 responses. This does not imply that R1 is completely immutable; rather, it is consistent with past observations that changes require extensive practice within a consistent training regime. We suspect that we could have observed R1 target dependence by providing subjects with a very long set of trials under the same perturbation and target conditions. In contrast, naive subjects were able to modify their R2 and R3 responses within seconds of the instruction and with minimal practice (Colebatch et al. 1979; Soechting et al. 1981).
Effect of stimulus predictability on long-latency (R2 and R3) rapid responses
Our three main experiments demonstrate that R2 and R3 responses are modulated in a sophisticated fashion as a function of spatial target position even when perturbation onset is unpredictable. However, in these experiments the perturbation direction was known in advance and it is possible that our observed modulation was due to the release of preplanned motor output in response to the perturbation such as that observed under startle conditions (Carlsen et al. 2004; Valls-Sole et al. 1999). We tested this possibility by having subjects perform experiment 1 in random order such that both the perturbation direction and onset were unpredictable. If the responses merely reflected a preplanned output then we would observe no mean target dependence and highly variable kinematic behavior because on half the trials subjects would release the inappropriate plan for the current perturbation. In fact, subjects showed only a small and inconsistent increase in variability when perturbation direction was unpredictable and never responded with the wrong movement. That said, there were subtle differences in the R2 and R3 responses that exhibited similarly timed but systematically smaller target dependence. These results are consistent with previous findings demonstrating that task dependence of R2/R3 responses is decreased but not eliminated by reducing the predictability of perturbation onset (Rothwell et al. 1980) or direction (Crago et al. 1976; Evarts and Vaughn 1978; Gottlieb and Agarwal 1980; Oriain et al. 1979). However, unlike the reports of some previous authors (Crago et al. 1976; Evarts and Vaughn 1978) we did not find a significant delay in the onset of target dependence (ROC Knee), indicating that sophisticated rapid responses are present at a fixed time after perturbation onset regardless of predictability (Gottlieb and Agarwal 1980; Oriain et al. 1979; Rothwell et al. 1980). Taken together, our present results suggest that R2 and R3 responses are not the result of a simple triggered response where one motor output is preprogrammed and released based on the occurrence of a sensory stimulus. Rather R2 and R3 responses combine both current sensory information and prior perturbation history (in addition to the spatial goal) to form an appropriate response; additional work is needed to establish the rules of how these factors are combined.
How are R2/R3 responses related to the volitional motor system?
The volitional motor system possesses an incredible capacity to control direction, distance, speed, accuracy, load, and many other parameters of movement. Our results demonstrate that R2 and R3 responses of multiple upper-limb muscles share some of the key functional attributes of volitional control. For example, we have recently demonstrated that R2/R3 responses possess an internal model of limb dynamics (Kurtzer et al. 2008) and here we show that these responses are appropriately tuned to the metrics of a spatial task. We suspect that such similarities are not accidental and can be readily understood if one appreciates that R2/R3 responses and volitional control share similar neural circuitry, including somatosensory cortex, motor cortex, cerebellum, and other subcortical structures (Cheney and Fetz 1984; Evarts 1973; Lewis et al. 2004; Matthews 1991; Strick 1983). Accordingly, the sophistication of R2/R3 responses would reflect the earliest volley of activity through the same neural circuit that is later engaged by voluntary control (Scott 2004). This concept of an evolving sensorimotor approximation through the same neural structures fits well within recent theories of volitional motor control that emphasize the importance of intelligently using feedback (Scott 2004; Todorov 2004; Todorov and Jordan 2002) and is related to previous work that attributed the task dependence of R2/R3 responses to a hastened voluntary response (Crago et al. 1976; Hasan 2005; Rothwell et al. 1980). It may also explain why a precise definition of reflex has proven difficult to establish.
|
|
DISCLOSURE |
|---|
|
|
|
GRANTS |
|---|
|
|
|
ACKNOWLEDGMENTS |
|---|
|
|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: S. H. Scott, Centre for Neuroscience Studies, Botterell Hall, Room 232, Queen's University, Kingston, Ontario, Canada K7L 3N6 (E-mail: steve{at}biomed.queensu.ca)
|
|
REFERENCES |
|---|
|
Batschelet E. Circular Statistics in Biology. New York: Academic Press, 1981.
Bonnet M. Anticipatory changes of long-latency stretch responses during preparation for directional hand movements. Brain Res 280: 51–62, 1983.[CrossRef][Web of Science][Medline]
Brown SH, Cooke JD. Initial agonist burst is modified by perturbations preceding movement. Brain Res 377: 311–322, 1986.[CrossRef][Web of Science][Medline]
Burnham KP, Anderson DR. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. New York: Springer, 2002.
Calancie B, Bawa P. Voluntary and reflexive recruitment of flexor carpi radialis motor units in humans. J Neurophysiol 53: 1194–1200, 1985.
Capaday C, Forget R, Milner T. A reexamination of the effects of instruction on the long-latency stretch reflex response of the flexor pollicis longus muscle. Exp Brain Res 100: 515–521, 1994.[CrossRef][Web of Science][Medline]
Capaday C, Stein RB. Amplitude-modulation of the soleus H-reflex in the human during walking and standing. J Neurosci 6: 1308–1313, 1986.[Abstract]
Capaday C, Stein RB. A method for simulating the reflex output of a motoneuron pool. J Neurosci Methods 21: 91–104, 1987.[CrossRef][Web of Science][Medline]
Carlsen AN, Chua R, Inglis JT, Sanderson DJ, Franks IM. Can prepared responses be stored subcortically? Exp Brain Res 159: 301–309, 2004.[CrossRef][Web of Science][Medline]
Cheney PD, Fetz EE. Corticomotoneuronal cells contribute to long-latency stretch reflexes in the rhesus-monkey. J Physiol 349: 249–272, 1984.
Christakos CN, Wolf H, Meyerlohmann J. The M2 electro-myographic response to random perturbations of arm movements is missing in long-trained monkeys. Neurosci Lett 41: 295–300, 1983.[CrossRef][Web of Science][Medline]
Colebatch JG, Gandevia SC, McCloskey DI, Potter EK. Subject instruction and long latency reflex responses to muscle stretch. J Physiol 292: 527–534, 1979.
Corneil BD, Olivier E, Munoz DP. Visual responses on neck muscles reveal selective gating that prevents express saccades. Neuron 42: 831–841, 2004.[CrossRef][Web of Science][Medline]
Crago PE, Houk JC, Hasan Z. Regulatory actions of human stretch reflex. J Neurophysiol 39: 925–935, 1976.
Day BL, Rothwell JC, Marsden CD. Interaction between the long-latency stretch reflex and voluntary electro-myographic activity prior to a rapid voluntary motor reaction. Brain Res 270: 55–62, 1983.[CrossRef][Web of Science][Medline]
Dufresne JR, Soechting JF, Terzuolo CA. Modulation of the myotatic reflex gain in man during intentional movements. Brain Res 193: 67–84, 1980.[CrossRef][Web of Science][Medline]
Duysens J, Tax AA, Trippel M, Dietz V. Increased amplitude of cutaneous reflexes during human running as compared to standing. Brain Res 613: 230–238, 1993.[CrossRef][Web of Science][Medline]
Evarts EV. Motor cortex reflexes associated with learned movement. Science 179: 501–503, 1973.
Evarts EV, Granit R. Relations of reflexes and intended movements. Prog Brain Res 44: 1–14, 1976.[CrossRef][Medline]
Evarts EV, Vaughn WJ. Intended arm movements in response to externally produced arm displacements in man. In: Cerebral Motor Control in Man: Long Loop Mechanisms (Progress in Clinical Neurophysiology, vol. 4), edited by Desmedt JE. Brussels: Karger, 1978, p. 178–192.
Forssberg H, Grillner S, Rossignol S. Phase dependent reflex reversal during walking in chronic spinal cats. Brain Res 85: 103–107, 1975.[CrossRef][Web of Science][Medline]
Gottlieb GL, Agarwal GC. Response to sudden torques about ankle in man: myotatic reflex. J Neurophysiol 42: 91–106, 1979.
Gottlieb GL, Agarwal GC. Response to sudden torques about ankle in man. II. Postmyotatic reactions. J Neurophysiol 43: 86–101, 1980.
Green DM, Swets JA. Signal Detection Theory and Psychophysics. New York: Wiley, 1966.
Hagbarth KE. EMG studies of stretch reflexes in man. Electroencephalogr Clin Neurophysiol Suppl. 25: 74–79, 1967.
Hammond PH. Involuntary activity in biceps following the sudden application of velocity to the abducted forearm. J Physiol 127: 23P–25P, 1954.
Hammond PH. The influence of prior instruction to the subject on an apparently involuntary neuro-muscular response. J Physiol 132: 17P–18P, 1956.
Hasan Z. The human motor control system's response to mechanical perturbation: should it, can it, and does it ensure stability? J Mot Behav 37: 484–493, 2005.[Web of Science][Medline]
Jaeger RJ, Gottlieb GL, Agarwal GC. Myoelectric responses at flexors and extensors of human wrist to step torque perturbations. J Neurophysiol 48: 388–402, 1982a.
Jaeger RJ, Gottlieb GL, Agarwal GC, Tahmoush AJ. Afferent contributions to stretch-evoked myoelectric responses. J Neurophysiol 48: 403–418, 1982b.
Johnson MT, Kipnis AN, Lee MC, Ebner TJ. Independent control of reflex and volitional EMG modulation during sinusoidal pursuit tracking in humans. Exp Brain Res 96: 347–362, 1993.[Web of Science][Medline]
Kandel ER, Schwartz JH, Jessel TM. Principles of Neural Science. New York: McGraw-Hill, 2000.
Kimura T, Haggard P, Gomi H. Transcranial magnetic stimulation over sensorimotor cortex disrupts anticipatory reflex gain modulation for skilled action. J Neurosci 26: 9272–9281, 2006.
Komiyama T, Zehr EP, Stein RB. Absence of nerve specificity in human cutaneous reflexes during standing. Exp Brain Res 133: 267–272, 2000.[CrossRef][Web of Science][Medline]
Koshland GF, Hasan Z. Electromyographic responses to a mechanical perturbation applied during impending arm movements in different directions: one-joint and two-joint conditions. Exp Brain Res 132: 485–499, 2000.[CrossRef][Web of Science][Medline]
Kurtzer I, Herter TM, Scott SH. Random change in cortical load representation suggests distinct control of posture and movement. Nat Neurosci 8: 498–504, 2005.[Web of Science][Medline]
Kurtzer I, Pruszynski JA, Herter TM, Scott SH. Primate upper limb muscles exhibit activity patterns that differ from their anatomical action during a postural task. J Neurophysiol 95: 493–504, 2006.
Kurtzer IL, Pruszynski JA, Scott SH. Long-latency reflexes of the human arm reflect an internal model of limb dynamics. Curr Biol 18: 449–453, 2008.[CrossRef][Web of Science][Medline]
Lee RG, Tatton WG. Motor responses to sudden limb displacements in primates with specific CNS lesions and in human patients with motor system disorders. Can J Neurol Sci 2: 285–293, 1975.[Medline]
Lee RG, Tatton WG. Long latency reflexes to imposed displacements of the human wrist: dependence on duration of movement. Exp Brain Res 45: 207–216, 1982.[Web of Science][Medline]
Lewis GN, MacKinnon CD, Perreault EJ. The effect of task instruction on the excitability of spinal and supraspinal reflex pathways projecting to the biceps muscle. Exp Brain Res 174: 413–425, 2006.[CrossRef][Web of Science][Medline]
Lewis GN, Polych MA, Byblow WD. Proposed cortical and sub-cortical contributions to the long-latency stretch reflex in the forearm. Exp Brain Res 156: 72–79, 2004.[CrossRef][Web of Science][Medline]
Ludvig D, Cathers I, Kearney RE. Voluntary modulation of human stretch reflexes. Exp Brain Res 183: 201–213, 2007.[CrossRef][Web of Science][Medline]
Marsden CD, Merton PA, Morton HB. Changes in loop gain with force in human muscle servo. J Physiol 222: 32P–34P, 1972a.[Medline]
Marsden CD, Merton PA, Morton HB. Servo action in human voluntary movement. Nature 238: 140–143, 1972b.[CrossRef][Medline]
Marsden CD, Merton PA, Morton HB. Servo action in human thumb. J Physiol 257: 1–44, 1976.
Marsden CD, Merton PA, Morton HB. Human postural responses. Brain 104: 513–534, 1981.
Matthews PBC. Observations on the automatic compensation of reflex gain on varying the preexisting level of motor discharge in man. J Physiol 374: 73–90, 1986.
Matthews PBC. The human stretch reflex and the motor cortex. Trends Neurosci 14: 87–91, 1991.[CrossRef][Web of Science][Medline]
Mortimer JA, Webster DD, Dukich TG. Changes in short and long latency stretch responses during the transition from posture to movement. Brain Res 229: 337–351, 1981.[CrossRef][Web of Science][Medline]
Nakazawa K, Yamamoto S, Yano H. Short- and long-latency reflex responses during different motor tasks in elbow flexor muscles. Exp Brain Res 116: 20–28, 1997.[CrossRef][Web of Science][Medline]
Oriain MD, Blair RDG, Murphy JT. Modifiability of the stretch reflex by prior information. Electromyogr Clin Neurophysiol 19: 57–63, 1979.[Web of Science][Medline]
Pearce SL, Miles TS, Thompson PD, Nordstrom MA. Is the long-latency stretch reflex in human masseter transcortical? Exp Brain Res 150: 465–472, 2003.[Web of Science][Medline]
Pierrot-Deseilligny E, Burke D. The Circuitry of the Spinal Cord: Its Role in Motor Control and Movement Disorders. Cambridge, UK: Cambridge Univ. Press, 2005.
Prochazka A, Clarac F, Loeb GE, Rothwell JC, Wolpaw JR. What do reflex and voluntary mean? Modern views on an ancient debate. Exp Brain Res 130: 417–432, 2000.[CrossRef][Web of Science][Medline]
Pruszynski JA, Kurtzer I, Scott SH. Examining the flexibility of upper-limb reflexes via a novel visuo-spatial task. Proc 17th Annu Meeting Soc Neural Control of Movement. Seville, Spain, 2007.
Prut Y, Fetz EE. Primate spinal interneurons show pre-movement instructed delay activity. Nature 401: 590–594, 1999.[CrossRef][Medline]
Rothwell JC, Traub MM, Marsden CD. Influence of voluntary intent on the human long-latency stretch reflex. Nature 286: 496–498, 1980.[CrossRef][Medline]
Scott SH. Apparatus for measuring and perturbing shoulder and elbow joint positions and torques during reaching. J Neurosci Methods 89: 119–127, 1999.[CrossRef][Web of Science][Medline]
Scott SH. Optimal feedback control and the neural basis of volitional motor control. Nat Rev Neurosci 5: 532–546, 2004.[Medline]
Shadmehr R, Wise SP. The Computational Neurobiology of Reaching and Pointing. Cambridge, MA: MIT Press, 2005.
Soechting JF, Dufresne JR, Lacquaniti F. Time-varying properties of myotatic response in man during some simple motor tasks. J Neurophysiol 46: 1226–1243, 1981.
Stein RB, Capaday C. The modulation of human reflexes during functional motor tasks. Trends Neurosci 11: 328–332, 1988.[CrossRef][Web of Science][Medline]
Strick PL. The influence of motor preparation on the response of cerebellar neurons to limb displacements. J Neurosci 3: 2007–2020, 1983.[Abstract]
Tatton WG, Lee RG. Evidence for abnormal long-loop reflexes in rigid parkinsonian patients. Brain Res 100: 671–676, 1975.[CrossRef][Web of Science][Medline]
Thompson KG, Hanes DP, Bichot NP, Schall JD. Perceptual and motor processing stages identified in the activity of macaque frontal eye field neurons during visual search. J Neurophysiol 76: 4040–4055, 1996.
Todorov E. Optimality principles in sensorimotor control. Nat Neurosci 7: 907–915, 2004.[CrossRef][Web of Science][Medline]
Todorov E, Jordan MI. Optimal feedback control as a theory of motor coordination. Nat Neurosci 5: 1226–1235, 2002.[CrossRef][Web of Science][Medline]
Valls-Sole J, Rothwell JC, Goulart F, Cossu G, Munoz E. Patterned ballistic movements triggered by a startle in healthy humans. J Physiol 516: 931–938, 1999.
Wolf SL, Segal RL. Reducing human biceps brachii spinal stretch reflex magnitude. J Neurophysiol 75: 1637–1646, 1996.
Wolpaw JR. Adaptive plasticity in the primate spinal stretch reflex: reversal and re-development. Brain Res 278: 299–304, 1983.[CrossRef][Web of Science][Medline]
Wolpaw JR, Braitman DJ, Seegal RF. Adaptive plasticity in primate spinal stretch reflex: initial development. J Neurophysiol 50: 1296–1311, 1983.
Zehr EP, Chua R. Modulation of human cutaneous reflexes during rhythmic cyclical arm movement. Exp Brain Res 135: 241–250, 2000.[CrossRef][Web of Science][Medline]
Zehr EP, Collins DF, Frigon A, Hoogenboom N. Neural control of rhythmic human arm movement: phase dependence and task modulation of Hoffmann reflexes in forearm muscles. J Neurophysiol 89: 12–21, 2003.
This article has been cited by other articles:
|
|
 |
|
  I. Kurtzer, J. A. Pruszynski, and S. H. Scott Long-Latency Responses During Reaching Account for the Mechanical Interaction Between the Shoulder and Elbow Joints J Neurophysiol, November 1, 2009; 102(5): 3004 - 3015. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Shemmell, J. H. An, and E. J. Perreault The Differential Role of Motor Cortex in Stretch Reflex Modulation Induced by Changes in Environmental Mechanics and Verbal Instruction J. Neurosci., October 21, 2009; 29(42): 13255 - 13263. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. P. J. Selen, D. W. Franklin, and D. M. Wolpert Impedance Control Reduces Instability That Arises from Motor Noise J. Neurosci., October 7, 2009; 29(40): 12606 - 12616. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Kimura and H. Gomi Temporal Development of Anticipatory Reflex Modulation to Dynamical Interactions During Arm Movement J Neurophysiol, October 1, 2009; 102(4): 2220 - 2231. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. R. Makin, N. P. Holmes, C. Brozzoli, Y. Rossetti, and A. Farne Coding of Visual Space during Motor Preparation: Approaching Objects Rapidly Modulate Corticospinal Excitability in Hand-Centered Coordinates J. Neurosci., September 23, 2009; 29(38): 11841 - 11851. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Pruszynski, I. Kurtzer, T. P. Lillicrap, and S. H. Scott Temporal Evolution of "Automatic Gain-Scaling" J Neurophysiol, August 1, 2009; 102(2): 992 - 1003. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. M. Herter, T. Korbel, and S. H. Scott Comparison of Neural Responses in Primary Motor Cortex to Transient and Continuous Loads During Posture J Neurophysiol, January 1, 2009; 101(1): 150 - 163. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. W. Franklin and D. M. Wolpert Specificity of Reflex Adaptation for Task-Relevant Variability J. Neurosci., December 24, 2008; 28(52): 14165 - 14175. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. White, N. Dowling, R. M. Bracewell, and J. Diedrichsen Hand Interactions in Rapid Grip Force Adjustments Are Independent of Object Dynamics J Neurophysiol, November 1, 2008; 100(5): 2738 - 2745. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
TOP
ABSTRACT
INTRODUCTION
P < 0.01) and differences as a function of target position within an epoch (*P < 0.01).
