Preparing to reach: selecting an adaptive long-latency feedback controller

Abstract

In a voluntary movement, the nervous system specifies not only the motor commands but also the gains associated with reaction to sensory feedback. For example, suppose that, during reaching, a perturbation tends to push the hand to the left. With practice, the brain not only learns to produce commands that predictively compensate for the perturbation but also increases the long-latency reflex gain associated with leftward displacements of the arm. That is, the brain learns a feedback controller. Here, we wondered whether, during the preparatory period before the reach, the brain engaged this feedback controller in anticipation of the upcoming movement. If so, its signature might be present in how the motor system responds to perturbations in the preparatory period. Humans trained on a reach task in which they adapted to a force field. During the preparatory period before the reach, we measured how the arm responded to a pulse to the hand that was either in the direction of the upcoming field, or in the opposite direction. Reach adaptation produced an increase in the long-latency (45-100 ms delay) feedback gains with respect to baseline, but only for perturbations that were in the same direction as the force field that subjects expected to encounter during the reach. Therefore, as the brain prepares for a reach, it loads a feedback controller specific to the upcoming reach. With adaptation, this feedback controller undergoes a change, increasing the gains for the expected sensory feedback.

Figure 1.

Protocol for Experiment 1 and results from a representative subject. A, The subject held the handle of a manipulandum and reached to a target. A velocity-dependent curl force field was applied to the hand during the reach period only. On occasional trials, a force pulse was applied to the hand during the preparatory period before the reach to quantify the state of the feedback controller of the arm. We asked whether this controller changed as the subject adapted their reaching movements in the field. B, On each trial, the subject heard three beeps (100 ms in duration) and was instructed to start the reach in such a way so that during the third beep the hand crossed the boundary of an imaginary 5 mm radius circle centered on the start position. The arrows indicate onset time of the 50 ms duration force pulses that were occasionally given during the preparatory period. C, The gray line shows the state of the force field during the reaching movements. The dashed lines represent set breaks. The arrows represent the three periods in which we measured the feedback response by giving force pulses on occasional trials. D, Feedback response of a representative subject who trained in a CCW field. This plot shows hand displacement parallel to the direction of the pulse. The upward and downward displacements represent responses to CW and CCW pulses. The left figure shows the average response to a pulse at −750 ms. The right figure shows the average response to a pulse at −350 ms. The arrow indicates onset of the pulse. This subject showed an increased gain in the adapted state for CCW pulses but not CW pulses. Error bars are SEM.

Figure 2.

Performance during reaching movements. A, The plot shows a measure of error during reaching (maximum displacement perpendicular to the target direction) for subjects in the CW and CCW groups (error bars are SEM). The arrows mark the range of trials in which we probed the feedback gains in the preparatory period via force pulses. Bin size is 5, except for the first 10 trials of the adapt and washout blocks, in which bin size is 2 (to highlight the rapid changes). B, Reach start times with respect to the go cue (third beep) are plotted for the CW, CCW, and control groups. The baseline, adapt, and washout periods refer to the trials marked with arrows in A. Bin size is 10. Error bars are SEM. C, Displacement of the hand parallel to the target direction is plotted for the CW and CCW groups in the baseline, adapt, and washout periods. Error bars are SEM (they are very small). The dashed vertical line marks across-subject averaged reach start time, and the yellow bar indicates SEM. D, Muscle activations during reaching were normalized for each subject with respect to the activity of that muscle in a separate task in which a known force was gradually imposed on the hand (see Materials and Methods). The term RMS refers to the root-mean-squared value of EMG from that muscle during this control task. This plot shows across-subject averaged EMG activity from two muscles in the reach trials. After subjects adapted to the CCW force field, triceps EMG showed a marked increase near reach onset. After subjects adapted to the CW force field, biceps EMG showed a marked increase near reach onset. Error bars are SEM.

Figure 3.

Hand displacement in response to force pulses in the preparatory period before the reach. Displacements were measured parallel to the direction of the force pulse. A, Displacements for the CW group. Positive displacements are for CW force pulses, and negative displacements are for CCW force pulses. The arrows indicate onset time of the pulse (−750 and −350 ms) with respect to the go cue. The bar graphs show maximum displacement in the baseline, adapt, and washout conditions. Error bars are SEM. B, Same format as in A, but for the CCW group. C, Same format as in A, but for the control group. This group trained only in the null field. The adapt condition refers to the block of trials that correspond to those for the CW and CCW groups. The p values of t tests are indicated by asterisks as follows: *p < 0.05 and **p < 0.01.

Figure 4.

EMG responses to force pulses in the preparatory period. A, Across-subject averaged EMG responses in the CW and CCW groups in the baseline and adapt conditions. The response to a CW pulse in biceps and CCW pulse in triceps are plotted. The arrows indicate onset time of the force pulse. The left subfigure is for the −750 ms pulse, and the right subfigure is for the −350 ms pulse. B, Within-subject change in EMG in the adapt condition with respect to baseline. The traces are within-subject changes, comparing pulsed trials in the adapt condition with pulsed trials in the baseline condition. The dashed lines indicate periods 45–75 and 75–100 ms after the pulse. C, Same data as in B, but binned in time to reflect short-, medium-, and long-latency within-subject EMG changes in reflex responses. Error bars in all plots are SEM. The p values of t tests are indicated by an asterisk: *p < 0.05.

Figure 5.

Experiment 2: context-dependent changes in the feedback response. A, Experimental setup. Subjects reached to two targets. Reach to target 1 involved interaction with a field that flexed the elbow, whereas reach to target 2 involved interaction with a field that extended the elbow. B, Training protocol. We measured feedback response of the arm during the preparatory period in the period marked with red arrows. C, Reach errors (maximum displacement perpendicular to the target direction) for each target. Bin size is 5. The shaded region is SEM. E, Hand displacement in response to force pulses in the preparatory period before the reach. L-pulse refers to the force pulse directed toward 225°. R-pulse refers to force pulse directed toward 45°. Displacements were measured parallel to the direction of the force pulse. The shaded region is SEM. D, Hand trajectories from pulse trials. The dots represent across-subject average hand position, sampled at 10 ms intervals. The filled circles identify hand position at 50 ms intervals. Importantly, note that the movement toward the target does not begin until 200 ms or more after pulse onset. F, Maximum displacement in the baseline and adapt conditions in response to the force pulse. Error bars are SEM. G, Within-subject change in EMG response to force pulses in the adapt condition with respect to baseline. The data were binned in time to reflect short-, medium-, and long-latency within-subject EMG changes. Error bars are SEM. The p values of t tests are indicated by an asterisk: *p < 0.05.