Abstract
Purpose. When observers extrapolate occluded contours of constant curvature the extrapolated portions of the contours decrease in curvature with distance from the point of extrapolation (Singh & Fulvio, 2005, 2007). If an observer attempted to touch the occluded curve endpoint using his perceptual estimate, he would have failed. Consideration of everyday motor tasks within the environment, however, suggests that observers reach for partially-occluded objects accurately. We studied performance in a dynamic motor version of the perceptual task to test the hypothesis that different representations are used for extrapolation in the motor system.
Methods. Subjects attempted to touch a moving dot with their fingertip on a computer screen as it emerged (unseen) from behind an occluder. Dot trajectories could have one of two curvatures and two speeds. The occluder was a half-disk (two possible radii). At the beginning of each trial, a blinking dot signaled where the dot would begin its trajectory. On reach initiation, the dot traveled rightward toward the straight edge of the occluder and vanished. It was not shown emerging from behind the occluder. Subjects were rewarded for accuracy in the timing of their attempts to touch the dot where it would emerge from the far edge of the occluder. Reach trajectories, recorded with an Optotrak 3020, and the distribution of movement endpoints for each condition were analyzed.
Results: Because each start position could be continued as one of several dot trajectories, we analyzed the spatiotemporal reach trajectories to determine when subjects acquired information about substantial non-zero curvature in the dot trajectory. Despite similarities to perceptual performance for the zero-curvature trajectories, patterned deviations occurred for the non-zero curvature trajectories. For example, despite the usual speed-accuracy tradeoff for increased spatial variance with increased speed, we did not find strong evidence for the same increased variance with increasing occluder radius.
NSF Graduate Research Fellowship (JMF) NIH EY08266 (TEH, LTM).