Do people only adjust ongoing movements vigorously when it is advantageous to do so?

Abstract

We previously found that arm movements towards a jittering target are constantly guided by the latest target position: the responses to target displacements became more vigorous as the movement proceeded, as required for the movement to reach the latest position smoothly within the remaining time. Here we examine whether this behaviour was a consequence of how that experiment was designed. We compared the vigour of adjustments in blocks of trials in which targets followed a random walk, as in our previous studies, with the vigour of adjustments in blocks of trials in which the target position varied at random with respect to a fixed position. For the random walk, the latest position is the best estimate of the final position, so neglecting earlier information can be useful. For random variability around a fixed position, the target's position at any instant is equally informative about the final position, so making vigorous adjustments in response to the latest information is pointless. In that case, the best estimate of the final position is the average of all the encountered positions. Some participants responded less vigorously in the latter case, but most did not. We discuss why tuning the adjustments to be complete within the remaining time may be a good strategy, even when the target does not follow a random walk.

Keywords: Interception; Learning; Motor control; Noise; Online adjustments; Optimal feedback control.

Fig. 1

Variations in the lateral position of the target in the random walk (blue) and stable jitter (red) blocks. A Example trials. Dots show the target’s positions at different times during a single trial for each block. Target positions during the time that was used to determine the response are indicated by the dots being black rather than white. In random walk blocks, the target stepped 1.67 mm on each image frame, randomly either to the left or to the right. In stable jitter blocks, each new position was picked from a normal distribution around some value. This baseline value (white line) was the horizontal midline of the screen most of the time, but was briefly shifted by 1.67 mm to the left (as shown) or to the right on some trials. B Average target positions in each block after the trials were split into two sets (first 66 frames; 550 ms). The random walk trials were split by whether the target stepped to the left or to the right at the selected moment (yellow arrow in A). The stable jitter trials were split by whether the baseline value briefly stepped to the left or to the right. C Histogram of the displacements between consecutive target positions in the two blocks. D Histogram of the target positions at the moment participants tapped the screen in the two blocks. For clarity, the frequency scale differs between the two blocks in C and D

Fig. 2

Determining the response and the vigour of the response. One participant’s data for the block with stable jitter. A The lateral position of the fingertip as a function of the time from when the baseline stepped to the left or to the right. Thin lines: individual trials. Thick lines: averages. The fainter curves are for leftward steps. Positive values are to the right. B The fingertip’s lateral velocity on the same trials. The difference between the average velocity after steps to the right and to the left (shaded area) is the response. C The response. Positive values are in the direction of the step. The average response between 100 and 200 ms after the step (grey bar) is the response vigour (point in inset)

Fig. 3

Overall performance in both blocks. Each participant is represented by two symbols, joined by thin lines. The blue and red bars indicate the mean values across participants. The three participants who responded less vigorously in the stable jitter block are indicated by squares rather than circles. A Participants hit fewer targets in the random walk block. B They did not systematically take more time to tap the screen in one of the blocks

Fig. 4

Time-course of responses to a 1.67 mm target step (with 95% confidence intervals based on the variability across participants). For the random walk, the 1.67 mm step was present on each trial (either to the left or to the right at the selected moment). For the stable jitter, the 1.67 mm is the step in the baseline; the step in individual trials varied around this value (see Fig. 1). The response is the difference between the lateral velocity of the hand after leftward and rightward steps (see Fig. 2). A positive response means that the fingertip moved in the direction of the step

Fig. 5

Individual participants’ responses. Each line or point represents one participant. Three participants who clearly responded less vigorously in the stable jitter block are represented by thicker lines in A and B, and by squares rather than circles in C and D. A Time course of the responses to the selected target step of the random walk. B Time course of the responses to the shift in the baseline of the stable jitter. C The relation between the response vigour in the two blocks. The grey bars in A and B indicate the time across which the response was averaged to obtain the vigour. D The relation between response vigour and the remaining time. Since the remaining time was matched across blocks for each participant, each participant’s two points are aligned vertically. The participants who hit fewest targets in each kind of block are represented by open symbols for the respective block (one participant has open symbols for both blocks; see Fig. 3A)