|
Abstract:
Reach control strategies in tasks with many viable targets
What is the brain’s control strategy for a reaching task that does not have a unique goal point, like reaching to a line? One model is that the brain selects a particular target point on the line before reach initiation by minimizing generic motor costs (e.g. hand-to-target distance, expected jerk, expected endpoint noise, total spent energy, etc.). During the reach the online control system guides the hand toward the pre-selected target location. Another model is that the brain does not select a particular target point, but rather maintains a representation of all viable target points (e.g. landing anywhere on the line). During reach execution, the online controller can guide the hand toward any of the viable points, based on the current cost-to-go. The two models behave differently when a random perturbing force-field is applied mid-reach: the first model corrects for the perturbations so that the pre-selected target point is reached, the second model only corrects force perturbations that threaten task success. To determine which model more accurately reflects human behavior, we conducted a psychophysical experiment to test their specific predictions. Participants used a PHANToM haptic stylus to virtually touch graphically rendered lines that appeared at various orientations and depths on a virtual tabletop. The line length filled the viewable tabletop. Trials were terminated when the stylus came to a stop on the line. During some trials, a perturbing force (in one of several directions chosen randomly) was applied during the reach. Participants who use the first model should exhibit target selection. This means they will exert control to counter the effect of the perturbing force and end their reaches at the pre-selected target. Participants who use the second model should “go with the flow”. This means they will allow the perturbing force to alter their reach trajectory as long as they still achieve their goal. Our results strongly suggest participants “go with the flow”. This means the brain maintains a representation of all viable target points and only corrects perturbations that prevent target points from being contacted. These results are strengthened by a second condition, in which we reduced the number of viable target points by shortening the lines. In this condition, participants applied more corrective control to ensure at least one viable target point was contacted. One obvious advantage of maintaining a representation of all viable target points is that noise that does not affect task success can be ignored. These results suggest that the control objective in redundant tasks like reaching to a line is task success rather than a particular target point. |