Abstract
The planning-control model proposed by Glover (2004), suggests that an unfolding visual representation gradually assumes control of goal-directed action. Can the gradual build up of control during movement time be quantified by looking at performance as a function of time from movement onset to which no further grasp corrections are made? Participants (N=7) were asked to grasp a bar of light presented briefly (40ms) on a computer display and specify it with two separate features, namely angle (30°-70°) and length (6cm-9cm). Hand and finger movements where acquired using a motion capture system that ensured sampling of the full grasp cycle. Statistical analysis of the continuous angle and aperture between index finger and thumb were conducted from movement onset to terminal position. A novel measure of accumulated variance was employed to each trial, quantifying them in five discrete stages going from very late to very early time of specification. Results show that successful performance for both features are highly dependent on the time to which no further grasp corrections are made (chi2 test of independence; angle: p<0.001 length: p<0.001), sooner the better. A simple speed-accuracy trade-off would have predicted the inverse relationship. Trial-by-trial accuracy can be predicted from the specification time. Motor specification of features in an open-loop condition updates during movement time in a feed-forward manner (no visual feedback). Thus, the results support the conjectures of the planning-control model.
Meeting abstract presented at VSS 2012