Differences in context and feedback result in different trajectories and adaptation strategies in reaching

Abstract

Computational models of motor control have often explained the straightness of horizontal planar reaching movements as a consequence of optimal control. Departure from rectilinearity is thus regarded as sub-optimal. Here we examine if subjects may instead select to make curved trajectories following adaptation to force fields and visuomotor rotations. Separate subjects adapted to force fields with or without visual feedback of their hand trajectory and were retested after 24 hours. Following adaptation, comparable accuracies were achieved in two ways: with visual feedback, adapted trajectories in force fields were straight whereas without it, they remained curved. The results suggest that trajectory shape is not always straight, but is also influenced by the calibration of available feedback signals for the state estimation required by the task. In a follow-up experiment, where additional subjects learned a visuomotor rotation immediately after force field, the trajectories learned in force fields (straight or curved) were transferred when directions of the perturbations were similar but not when directions were opposing. This demonstrates a strong bias by prior experience to keep using a recently acquired control policy that continues to produce successful performance inspite of differences in tasks and feedback conditions. On relearning of force fields on the second day, facilitation by intervening visuomotor rotations occurred only when required motor adjustments and calibration of feedback signals were similar in both tasks. These results suggest that both the available feedback signals and prior history of learning influence the choice and maintenance of control policy during adaptations.

Figure 1. Experimental set-up and paradigm.

Subject, with the head placed on a chinrest and the arm in a natural posture, reaches to a target projected onto a mirror using a robotic arm. Vision of the subject's hand and robotic arm is occluded. Trial flow (horizontal) - the sequence of events in a trial that could be one of 4 types (vertical): standard (S), force field without visual feedback (FFnv), force field with visual feedback (FFv), or visuomotor rotation (R−). Perturbations are only introduced in blocks with a single target and never in standard trials. In FFnv, the subject does not see the cursor during the reach but does in FFv (yellow circle and green arrow). In R−, the hand to cursor mapping is rotated 45° counterclockwise such that the subject has to move the hand to 45° in order to bring the cursor to a 90°-target.

Figure 2. Adaptation to force fields with and without visual feedback: Single subjects.

A–B, Day1 and day2 hand paths of two single subjects during the first 10 trials and late trials (ranging from trial 131–220) of force field without visual feedback (FFnv, left) and with (FFv, right). Aborted trials (see methods) are not shown. Hand paths, plotted from detected movement onset to movement end, show displacement from origin to a target at 90° (gray circle). C, Velocity profiles of the hand paths in the late trials shown in B. Representative single-trial hand paths and their corresponding velocity profiles are also shown. In FFnv, the smooth early phase of the velocity profiles showed that in most cases, the path curvature in the late trials was not due to online corrections. However, in some cases trajectory corrections were observed as reflected in the presence of inflections after peak velocity (gray arrows, trial 201). D–E, Endpoint variability. Shown are 95% confidence ellipses (per subject) for early and late trials of both days. Gray circle shows the size of the target for comparison. F, Aftereffects. Hand paths of subjects in FFnv (n = 6) and FFv (n = 6) corresponding to the first trial in the learned direction (90°) in the post-learning standard block. Starting points of hand paths were aligned at (0,0) for easy comparison of directional deviations. Hand paths were deviated in the direction opposite to that of force field. Since this first trial could occur after several trials in this block, the aftereffect could be smaller. For this reason, aftereffects are used here for illustration only.

Figure 3. Success rates.

Success rates in early and late adaptations to force fields with (FFv) and without visual feedback (FFnv) on day1 and day2. Success rate was calculated in 10 bins of 22 trials each. Each bar depicts the mean of the first or last 3 bins across all subjects in the group. The mean success rate for standard trials for the same direction (n = 22) is also shown for FFnv. Vertical line is ±1 standard deviation (* = p<0.05).

Figure 4. Adaptation to force fields with and without visual feedback.

Group data, A–C, Day1 and day2 time courses, showing trial-by-trial means and ±1 SEM of spatial errors, directional deviation, and path curvature respectively, for force field without visual feedback (left) and with visual feedback (right). Shaded areas correspond to early and late trials used for comparisons. The directional deviation of the first trial in FFnv was the mean across 3 subjects only (the other 3 were aborted trials). D, Endpoint variability ellipses for early and late trials for each group. Each subject's endpoint position for each trial was subtracted from his mean endpoint position. Gray circle shows the size of the target for comparison. E, As in D but using endpoints taken at near zero velocity.

Figure 5. Double perturbations of matched directions: Visuomotor rotation adaptations after force field adaptations.

A, Cursor and hand paths of representative subjects (one per group) during the first trial of the first exposure to rotation (left, Control R−), and following adaptations to force field without visual feedback (center, R− after FFnv) and with (right, R− after FFv). To reach a target at 90° in visuomotor rotation, subjects should direct their movements 45° clockwise from the target. Subjects see a rotated cursor feedback of their hand movement and final hand position such that they see the cursor reaching 90°-target while their hands end at 45°. Center and right, the hand path of the first rotation trial (solid gray line) are also shown to illustrate aftereffects of prior force field adaptations. B, As in A, cursor paths of the late rotation trials. The hand paths of the subject who had prior adaptation to FFnv were curved (center) as opposed to the straight paths of the subjects in control R− and in R− after FFv. C, Velocity profiles of the late trials shown in B. The profiles of the subject in R− after FFv show that the curved paths were not due to online trajectory corrections, as seen previously in single force fields without VFB.

Figure 6. Effects of prior force field adaptations on matched visuomotor rotations.

Group data. A–C, Time courses for the different movement parameters during adaptations to single rotation only (control R⁻) and to rotations after learning force field with visual feedback (R− after FFv) and without (R− after FFnv). D, Endpoint variability ellipses of the first 40 (dotted lines) and last 40 (solid lines) rotation trials for all subjects in each group.

Figure 7. Effects of visuomotor rotation on force field retention.

A–C, Improvement indices (IMPs) show increments in adaptations to force field from day1 to day2. Shown are mean IMPs of the first 40 trials for all groups. Vertical lines are 95% confidence interval of between-group differences in means. Numbered horizontal bars with asterisk indicate significant differences between paired groups (p<0.05). D, Endpoint variability ellipses of all subjects in the matched double perturbation groups.