Multiple motor learning strategies in visuomotor rotation

Abstract

Background: When exposed to a continuous directional discrepancy between movements of a visible hand cursor and the actual hand (visuomotor rotation), subjects adapt their reaching movements so that the cursor is brought to the target. Abrupt removal of the discrepancy after training induces reaching error in the direction opposite to the original discrepancy, which is called an aftereffect. Previous studies have shown that training with gradually increasing visuomotor rotation results in a larger aftereffect than with a suddenly increasing one. Although the aftereffect difference implies a difference in the learning process, it is still unclear whether the learned visuomotor transformations are qualitatively different between the training conditions.

Methodology/principal findings: We examined the qualitative changes in the visuomotor transformation after the learning of the sudden and gradual visuomotor rotations. The learning of the sudden rotation led to a significant increase of the reaction time for arm movement initiation and then the reaching error decreased, indicating that the learning is associated with an increase of computational load in motor preparation (planning). In contrast, the learning of the gradual rotation did not change the reaction time but resulted in an increase of the gain of feedback control, suggesting that the online adjustment of the reaching contributes to the learning of the gradual rotation. When the online cursor feedback was eliminated during the learning of the gradual rotation, the reaction time increased, indicating that additional computations are involved in the learning of the gradual rotation.

Conclusions/significance: The results suggest that the change in the motor planning and online feedback adjustment of the movement are involved in the learning of the visuomotor rotation. The contributions of those computations to the learning are flexibly modulated according to the visual environment. Such multiple learning strategies would be required for reaching adaptation within a short training period.

Figure 1. Experimental procedure and computational scheme for visually guided reaching.

A, Time sequence of one trial. There were two conditions: reaching with and without online cursor feedback. See Materials and Methods for details. B, Total trial procedure. The angles are the visuomotor rotation angles. Each experiment consisted of four phases: pre-training, training, post-training, and washout phases. The cursor online feedback was shown in experiment 1 but not in experiment 2. In experiment 1, subjects were divided into two groups and the visuomotor rotation was suddenly or gradually introduced in the training phase for each group. The 60° CCW visuomotor rotation was continued in the post-training phase and then suddenly changed to 0° in the washout phase. C, Computational scheme for visually guided reaching, where represents desired movement of the arm, indicates feedback information of current movement, and and indicate feedforward and feedback motor commands, respectively. This scheme assumes that target presentation triggers the motor planning and that the feedforward controller transforms the desired movement into motor commands. The manipulandum system converts the hand movement into cursor movement in the rotationally biased direction on the screen. The feedback controller adjusts the ongoing movement using the feedback information of the cursor movement.

Figure 2. Changes in reaching movement characteristics by the sudden and gradual visuomotor rotation learning.

A, Mean characteristics indices of reaching movement across subjects as a function of trial block in experiment 1. Thick blue and green lines in each panel indicate the sudden and gradual conditions, respectively. The visuomotor rotation angles (black and gray thin line for the sudden and gradual conditions, respectively) are superimposed on the panels of the E-DE and I-DE. Shaded areas represent the SE. Open red triangles indicate the trial blocks in which the indices were significantly different between the sudden and gradual conditions (t test, p<0.05). B, Changes in the averaged indices between pre- and post-training phases. In all panels, the asterisks (*, **, and ***) denote the significance of differences: p<0.05, p<0.01, and p<0.001, respectively. Error bars represent the SE.

Figure 3. Catch trial effects.

Each panel shows the mean I-DEs and E-DEs of cursor feedback trials (black bars) and catch trials (gray bars) across subjects. The asterisks (* and **) denote the significance of differences: p<0.05 and p<0.01, respectively. Error bars denote the SE. Left and right panels indicate the pre- and post-training phases, respectively. Top and bottom panels indicate the sudden and gradual conditions, respectively. Note that the cursor-feedback trials with the same visual target as the catch trials were selected to calculate the mean I-DE and E-DE.

Figure 4. Cursor perturbation effects.

A, Cursor perturbation. The cursor movement direction relative to the hand movement direction was suddenly shifted ±20° from 150 ms after the cursor was turned on. B, Averaged acceleration patterns in the y-direction for the negative (solid) and positive (dashed) direction perturbations. Time zero corresponds to the onset of the cursor perturbation. C, Mean amplitude of acceleration response across subjects. Error bars denote the SE. The response amplitude was calculated from the temporal average of the y-acceleration difference between the perturbation directions for the interval of 250–300 ms from the perturbation onset (shaded areas in B). The double asterisks denote the significance of differences (p<0.01). D, The relationship between the I-DE aftereffect and the response amplitude change in individual subjects in the gradual condition. The I-DE was averaged from 41^st to 46^th blocks for each subject. The negative I-DE indicates a large aftereffect. The positive value of response amplitude change indicates the increase in the response amplitude to the cursor perturbation after the training, which corresponds to the increase in the feedback gain due to the learning of the gradual visuomotor rotation.

Figure 5. Response amplitudes to cursor perturbation and RTs after learning with and without online cursor feedback.

Left and right panels show the mean amplitudes of acceleration responses to the cursor perturbation and the mean RTs for the reaching initiation across subjects, respectively. Black lines indicate the gradual condition with online cursor feedback (exp. 1) and gray lines indicate the gradual condition without online cursor feedback (exp. 2). Note that the data indicated by the black lines are the same as those represented by the gray line in Fig. 4C (response amplitude) and the green line in top-right panel of Fig. 2B (RT). The asterisks (*, **, and ***) denote the significance of differences: p<0.05, p<0.01, and p<0.001, respectively.

Figure 6. Reaching performances in the gradual visuomotor rotation learning without online cursor feedback.

A, Averaged hand trajectories of typical subjects in the pre-training (left panels), post-training (middle panels), and washout phases (right panels) of the “without online cursor feedback” condition (experiment 2). Upper and lower panels correspond to the trajectories of typical subjects in the small error group (SEG) and large error group (LEG), respectively. Small white disks in each panel indicate the reaching target locations. The hand was moved from the center outwards. The trajectories are averaged 1 sec from the reaching onset in each reaching direction. The angles placed aside of the endpoint of the trajectories indicate the directions of the visual targets. B, Averaged E-DE in pre- and post-training phases for all subjects. Error bars denote the SD. Dashed lines indicate the reaching success margin (±5.7°, see Data analysis).

Figure 7. Changes in reaching movement characteristics by the gradual visuomotor rotation learning without online cursor feedback.

Thick purple and orange lines in each panel indicate the SEG and LEG, respectively. Gray shaded blocks indicate the middle stage of the training phase (from the 18^th to 22^nd blocks, see Results). The notation is same as in Fig. 2A.