Abstract
Different heading-based strategies could be used to guide walking to a target. Observers could potentially align the visual direction of the target with their physical direction of movement, or their visual heading from optic flow, or some integrated estimate of heading. I investigated the relative influence of visual and physical heading on control of walking, and how this varies over the course of a movement. Previous studies have distinguished control strategies using conditions in which the visual heading specified by optic flow differs from an observer’s physical direction of motion. I tested similar conditions but used a novel analysis to estimate the controller’s goal state at different moments in time. Observers walked to a target pole in a virtual environment presented with a head-mounted display. On perturbed trials, visual heading differed from physical heading by ±10°. The target appeared after observers walked 1m in a pre-cued direction, and was randomly positioned in a range around the initial direction. I analyzed how path corrections across a brief time window depended on the current heading error relative to the target. Across a given time delay, change in heading was an approximately linear function of heading error. By fitting this function, one can estimate the heading error that would produce no steering adjustment across a time window. This goal state was estimated for 250 ms windows starting at target onset. Initial steering adjustments on perturbed trials were consistent with a goal state of aligning the physical heading direction with the target. Over the course of 2 s of walking, the estimated goal state gradually shifted to be approximately midway between the physical and visual heading. These results are consistent with control based on an integrated estimate of heading, which is initially determined by non-visual information but gradually incorporates heading information from optic flow.
Meeting abstract presented at VSS 2012