Abstract
We measured the movements of the ball and of the hands of a thrower and a catcher in a natural one-handed catching task. We investigated how visual information is used by randomly occluding the catche's sight. We can capture the probability of successfully catching the ball by combining two kinds of information. The first is information that allows one to predict the ball's trajectory. This ability is larger when more of the initial part of the trajectory (including the motion of the thrower's hand before the ball is released) is visible, and is modelled as a cumulative Gaussian centred near the moment of release. The second is information that allows one to guide one's hand to the ball. This ability decreases as the ball approaches the hand, because the time left to adjust the movement becomes too short. It is modelled as an inverted cumulative Gaussian centred about 200 ms before the catch. Both Gaussians have widths of about 170 ms. The ability to make good use of either source of information shows how flexible human movements are despite the limitations imposed by neuronal resolution and delays.
Research supported by Grant SEJ2007-60488/PSIC of the Spanish Ministry of Education.