Abstract
We investigated spatial and temporal characteristics of eye-hand coordination. Using a rapid reaching task we measured correlations of eye and hand movement latencies, taking into account the location (peripheral vs. foveal) of visual stimuli during preparation and execution of movements. Subjects repeatedly reached toward a peripheral target on a touch screen under time pressure (<550 ms) with constrained (fixated at the screen center) or unconstrained eye movements. Rewards depended on the proximity of hand endpoints to the target. The target either appeared at the same location within a block of trials ("predictable" condition) or randomly in one of two locations ("unpredictable" condition). We found that the shapes of eye and hand endpoint distributions were highly variable across subjects and conditions. Hand endpoints, but not eye endpoints, were concentrated along the line connecting the starting point and the target. The distributions of both eye and hand movement directions were symmetric (not skewed), but eye movements were systematically short of reaching the target by approximately 2.5 deg. Target predictability affected movement latencies and their relations. In the predictable condition, eye initiation latencies were longer and arrival times shorter than in the unpredictable condition. Relative latencies of eye and hand also depended on target predictability. Hand movements preceded eye movements in the predictable condition; the pattern was reversed in the unpredictable condition. Highest hand precision was attained when hand movements were initiated within 60 ms of the onset of eye movement. We conclude that target predictability determines the regime of eye-hand coordination. At high uncertainty about the upcoming visual target, early eye movements help to avoid corrective hand movements and expedite the visual feedback needed for good hand precision. At low uncertainty, subjects can afford initiating hand movement earlier and thus gain time for increasing hand precision.
Meeting abstract presented at VSS 2012