Abstract
The ability to rapidly assess visual motion information is critical for daily tasks that require rapid interceptions such as catching a falling object. When visual information is sparse, interception decisions have to be made under uncertainty. To investigate this in the laboratory, we can constrain the availability of visual information or the time over which information has to be extrapolated. We recorded human observers’ (n=10) eye and hand movements while they viewed the launch of a simulated flyball on a screen. The ball was occluded shortly after launch, and observers had to manually intercept it along its predicted trajectory within a hit zone. We measured interception accuracy in eye and hand movements for different trajectory shapes yielded by ball speed variations. In two sessions, we manipulated uncertainty by either varying ball presentation duration (100-500ms) at constant occlusion duration (500ms), or by varying occlusion duration (100-500ms) at constant presentation duration (500ms). Reducing uncertainty by increasing presentation duration rapidly improved eye and hand interception accuracy and reduced a bias toward the center of the trajectory space that was observed at the shortest presentation duration. These improvements occurred within the first 300ms of ball presentation. By contrast, long occlusion durations yielded a much weaker center-bias. Reducing uncertainty by decreasing occlusion duration from 500-100ms linearly improved interception accuracy until eye and hand interception errors were minimal. The availability of visual information and time available to extrapolate information both determine uncertainty in rapid interceptive control, but they do so at different rates. Effects of presentation duration plateaued early, indicating that 300ms are sufficient to accurately read out visual trajectory information. Occlusion duration improved performance at a later time, as interception switched from predictive to visually-guided control.