Abstract
When optic flow patterns are measured during natural outdoor locomotion, the head-centered focus of expansion (FOE) of the resulting optic flow patterns are highly unstable as a consequence of head oscillations in all three axes during locomotion (Matthis et al, 2018). This suggests that the head-centered FOE is unsuitable for controlling heading during locomotion, and visual direction could serve this purpose instead (Rushton et al, 1999). However, Warren et al (2001) pitted visual direction against the FOE in a virtual environment and found comparable influences from flow and visual direction. To further explore exactly how optic flow is used in locomotion we devised a similar experiment to that of Warren et al in a virtual environment using an HTC Vive Pro running on Unity. We simulated a prism rotation to separate visual direction and the FOE. We found similar effects to those reported by Warren et al, where optic flow combines with visual direction cues to result in straighter walking paths in more structured environments. However, the way that flow exerts this influence is unclear in light of the instability of the FOE in the natural optic flow stimulus. It is possible that the curl signal introduced in retinal flow from the ground plane drives walking trajectories rather than the FOE. We are exploring this possibility by independently manipulating flow from the ground plane vs the surrounding scene structure.
Acknowledgement: NIH grant EY05729 and K99EY028229