Abstract
The egocentric direction strategy and the optic flow strategy both play a role in the online control of steering during locomotion. Their relative contribution depends on the structure of the visual environment (i.e. the amount of optic flow) and the flow rate. Moreover, optic flow serves to calibrate the walking direction in the egocentric strategy. However it remains unclear precisely how these two strategies are integrated. Our previous results suggest that they may be linearly combined, with weights of about 0.8 (flow) and 0.2 (egocentric) in visually structured environments. However, this was observed when the heading specified by optic flow was displaced at a single ±10° offset (right or left) from the locomotor axis. Here, we test this weighted combination by investigating a broader range of offsets. Participants actively walk in a 40 x 40 ft. virtual environment while wearing a head-mounted display (63° H x 53° V), and head position is recorded with a sonic/inertial tracker (70 ms latency). The virtual environment contains a textured ground, ceiling, frontal wall, and randomly positioned vertical posts. Participants repeatedly walk to a target (a doorway in the wall) while the heading direction specified by optic flow is displaced randomly to the right or left of the actual walking direction, blocking adaptation. The displacement increases from ±5° to ±25° in 5° steps, in blocks of 8 trials. If the two strategies are linearly combined, we expect that the weights will be constant over some range, although there may be an upper limit due to biomechanical and proprioceptive constraints. Results for human heading direction are compared with model predictions based on a control law for steering in which the turning rate is a weighted linear sum of egocentric direction and optic flow.
NIH R01 EY10923, Brown Center for Vision Research.