Abstract
Hitting a baseball requires a two-stage decision: whether or not to swing, and when and where to hit. These decisions have to be made ultra-fast, in less than 400 ms, and before the entire trajectory of the ball can be viewed. Here we investigate the role of eye movements in sensorimotor decision-making and interception under uncertainty. We developed a go-no go manual interception task in which observers (n=26 varsity baseball players) tracked and predicted linear target trajectories shown briefly on a screen. In each trial, the trajectory either went through a designated strike box (hit) or past it (miss). Observers were instructed to intercept the target with their index finger in the strike box in hit trials, and to not move their hand in miss trials. Only the initial launch (100-300 ms) of the ball was shown, and balls moved at 36 or 40°/s. Eye and hand movements were recorded with a video-based eye tracker and magnetic hand tracker. Linear regression and random-forest models were used to relate movements of eye, hand, and decision performance. The decision whether or not to intercept was best predicted by smooth pursuit velocity during the earliest (open-loop) phase of the movement, possibly due to more reliable motion trajectory information as a consequence of accurate pursuit initiation. Hitting accuracy was best predicted by pursuit position error and velocity gain during the later (steady-state) phase. These findings indicate that different stages of task performance could be predicted by different pursuit measures. Interestingly, performance was significantly better for the fast speed (shorter decision time) as compared to slow speed, where players frequently intercepted too early. On-field baseball experience with fast-moving balls might affect performance, a conclusion supported by the finding that more experienced, senior players showed a stronger performance benefit at high speed than junior players.
Meeting abstract presented at VSS 2017