Abstract
Vision is used to program the initial, feedforward component of a reach as well as to control accuracy during the feedback component of a reach. Previous research has shown that binocular information is important for both components. Vergence can be used to program the feedforward component of a reach by providing distance information. It's been proposed that disparity matching could be used in the feedback component, however, this would not specify how to control the hand's approach velocity. One well known control strategy for visually controlling approach is the classic monocular tau-dot strategy: move so as to preserve tau-dot at −.5. This yields constant decleration to stop at a target. Monocular tau cannot be used to guide the hand to a target. Instead, we propose a disparity tau-dot strategy. Disparity tau is defined as the ratio of the relative disparity between hand and target to the temporal rate of change of that disparity. What types of trajectories would be produced by moving so as to hold this disparity tau-dot constant? We generated simulated reaching profiles based on our disparity tau-dot model. Both position and velocity profiles were output by integrating the model over time given a constant tau-dot value, initial hand and target distances from the eye, and initial hand velocity. Using representative values produced characteristic reach trajectories. The simulated reaches exhibited acceleration to a peak velocity followed by deceleration to soft contact at the target. Constant temporal change of disparity tau does not correspond to constant change of time-to-contact of the hand relative to a target! This constant tau-dot model is different from monocular tau-dot because it generates both acceleration and deceleration. This result is important because the trajectories look exactly like those of natural reaches meaning that it may well indeed be used to guide such reaches.