Abstract
Slow smooth-pursuit eye movements help maintain the image of an object on the fovea during visible object motion for increased visual acuity. When a visually tracked object disappears, the visual system no longer receives sensory information regarding the current position or velocity of the target and becomes dependent upon previous knowledge related to the target velocity that is held in a short-term velocity memory store. Previous research indicates that the velocity and occlusion duration of a target influence the accuracy in extrapolating an object’s motion. The present study measured both explicit location predictions and eye positions to investigate how the smooth pursuit system supports motion prediction. Participants observed a target moving along a circular path at one of three speeds and eventually disappearing. After a delay of one of three durations, participants indicated whether the target would have passed a probe line. Slowly moving targets were significantly more difficult to follow than faster targets, with decreased response accuracy and increased eye position error and eye velocity. Responses were less accurate in the long occlusion duration compared to the short occlusion duration. However, eye position was more accurate in the long occlusion duration, suggesting that the ability of the smooth pursuit system to accurately track an occluded object increases as occlusion duration increases. This divergent finding suggests that response accuracy, while the most common dependent variable in target concealment studies, does not provide a complete understanding of the smooth pursuit system. Accessing the short-term velocity memory store is not instantaneous and requires time to integrate that knowledge with the oculomotor system. While initial decay is evident, top-down information continues to provide the feedback needed throughout extended occlusion durations to correct for errors in eye position.