Movement Function Follows Spatial Form: Coordinate System Implications for Online Visual Feedback Control of the Hand

Closed loop visual feedback control of the hand is essential for accurate reaching movements. Without visual signals of either the hand or target position, reaches are inaccurate and imprecise; therefore the brain uses a relative positional signal to control movements online. Previous studies suggest that movements are planned and represented in a polar coordinate system and that the dimensions, direction and extent are independently specified and processed. We were interested to find out whether there was behavioral evidence for the independent control of hand direction and extent as a movement unfolded. We asked subjects to make a reaching movement in a virtual reality environment in which we singularly removed the visual presentation of direction and extent information of the moving hand. Results from this experiment suggest that people control the direction and extent of hand position independently during the course of a movement. With that in mind, we reasoned that if these two dimensions were controlled independently then human responses to actual visuomotor perturbations would reflect that processing. Therefore we asked subjects to perform point to point reaching movements with a visuomotor displacement and recorded their hand positions. The trajectories were then compared to model predictions of visuomotor control to determine what type of coordinate system and control architecture most closely mimicked human behavior. We chose 2 model systems, the Next State Planner: NSP) which could be implemented in different coordinate systems, and the Stiff-Viscous: SV) model which implements control by generating corrective feedback torques. We used systems engineering metrics to evaluate human and model responses. Results from this study suggest that feedback responses of humans to visuomotor perturbation are most closely modeled by a controller than compares spatial locations of the hand and target in a polar coordinate system. Our final investigation into the numerical underpinnings of this type of control system demonstrates that using a polar coordinate system to represent movement space naturally optimizes feedback control of point to point reaches. Taken together our works suggests that the brain has evolved to represent this particular movement space in polar-like coordinates in part to efficiently enact closed loop control of the hand