On-line vs. off-line utilization of peripheral visual afferent information to ensure spatial accuracy of goal-directed movements.

Manual aiming movements are often initiated when one gazes at the target, while the hand is seen in peripheral vision. The objective of the present study was to determine whether vision of one's hand in peripheral vision and/or central vision as it progresses towards the target can be used to modulate the direction and the extent components of the initial movement impulse. Participants performed video aiming movements while vision of the cursor they were moving was permitted for its whole trajectory, 40 degrees to 15 degrees of visual angle, 15 degrees to 0 degrees of visual angle, or not visible at all. Movements were to be completed within prescribed movement times varying between 300 ms and 900 ms. The results did not reveal endpoint accuracy or variability differences between the 40 degrees -15 degrees and the 15 degrees -0 degrees visual feedback conditions. Both conditions yielded lower endpoint bias and variability than the no-vision condition from early on after movement initiation. This indicates that the visual afferent information available in the 40 degrees -15 degrees and the 15 degrees -0 degrees visual feedback conditions could be used to better plan upcoming movements than the no vision condition. From these data, it appears very unlikely that different portions of the retina are specialized for processing different movement attributes as has been suggested in the past (Paillard 1980; Paillard and Amblard 1985). Both the peripheral and central retina are apt at detecting on-line extent and direction errors in one's movement. In addition, the data cast serious doubts on the widely accepted proposition that the movement initial impulse is essentially ballistic.