The efficiency of online visuomotor processes was investigated by manipulating vision based on real-time upper limb velocity. Participants completed rapid reaches under two control (full vision, no vision) and three experimental visual window conditions. The experimental visual windows were early: 0.8-1.4 m/s, middle: above 1.4 m/s, and late: 1.4 to 0.8 m/s. The results indicated that endpoint consistency comparable to that of full-vision trials was observed when using vision from the early (43 ms) and middle (89 ms) windows, but vision from the middle window entailed a longer deceleration phase (i.e., a temporal cost). The late window was not useful to implement online trajectory amendments. This study provides further support for the idea of early visuomotor control, which may involve multiple online control processes during voluntary movement.
Keywords: online control; optimal; reaching; real-time; vision.