Abstract
Introduction: Visually guided reaches take ≈1s. Given unstable feedback control with neural transmission delay, stable reaching is assumed to require internal models that generate simulated feedback without delay that combines with actual feedback for stability. We investigated whether stable visually guided reaching requires internal models to handle such delay or whether perspective information might be used to overcome the problem of internal delay. Methods: Two experiments entailed reaches to match 9 target depth planes (20cm to 60cm) with a hand avatar in a virtual environment despite different mappings between speeds of hand and hand avatar. Four conditions in Experiment 1 were: (1) visually guided reaches and constant mapping (OG CM); (2) feedforward reaches with constant mapping and hand avatar only visible at reach start and end (FF CM); (3) visually guided reaches with mappings of 0.62, 0.71, 0.83, 1.2, 1.4, 1.6 of constant mapping (OG VM); (4) feedforward reaches with variable mapping (FF VM). In Experiment 2, conditions 1, 3, 1 from Experiment 1 controlled for practice. Results: In both experiments, reaches were accurate in OG and FF constant mapping conditions and in variable mapping conditions with online guidance. Implicitly, an internal model was calibrated in condition 1 (OG CM) and deployed in condition 2 (FF CM) in Exp 1. Errors in the feedforward variable mapping condition were large showing that visual guidance in the OG VM condition was successful despite an ineffective internal model. Performance in respect to accuracy and MTs was strictly comparable across CM and VM conditions with online guidance in Experiment 2. We reproduced reach accuracy and MTs using a proportional rate model with disparity Tau information delayed by 0msec, 50msec, and 100msec. Simulations remained stable with delayed feedback. Conclusions: Stable visually guided reaching did not require an internal feedforward model.