Abstract
Purpose: Movement planning takes into account both movement uncertainty and rewards/losses associated with the possible outcomes of the movement. Previous research has focused on uncertainty in movement endpoint. Here, we manipulate reward/loss and examine the covariance structure of uncertainty along the movement trajectory. To do so we required subjects to pass to the left of an invisible virtual obstacle in reaching to targets to earn monetary rewards.
Methods: Targets were located at one of three positions on the screen (either on the subject's midline, or to the right of the midline) spaced 3.8 cm apart. A line on the screen marked the edge of an invisible vertical half-plane, parallel to the screen, and situated 66% of the way between the start point and the screen. Its edge was always .7 cm to the right of the target. Each of 6 naïve subjects completed trials across three different penalty conditions and for three locations of the target/obstacle. Subjects practiced touching targets, and were allowed to freely explore the location of the virtual half-plane with feedback before the main experiment. Trajectories were monitored with an Optotrak 3020 and feedback was immediate. Subjects were instructed to earn the greatest reward possible.
Results: Subjects avoided the obstacle more as the penalty Increased and as the edge was displaced to the right. We computed the covariance between the fingertip position in passing through the obstacle plane and in hitting the target plane. Surprisingly, the correlation of horizontal displacements in the two planes changed sign with position of the obstacle. For the leftmost position of the obstacle it was positive and then became smaller and finally negative for the middle and right target locations, respectively. We discuss these results and covariance along the entire trajectory In terms of current models of visual feedback control of movement.
NIH EY08266, NIH NS047178-01A1