Abstract
The perception and control of body orientation is based on the ability to align body axes with gravity directions on surrogates axes such as visual ground surface. We hypothesized a 'horizontality assumption': similarly to vertical surfaces moving in depth (Lee & Aronson, 1974), a slight inclination of the ground will induce postural adjustments in the direction of the surface norm, in order to regain the visually assumed upright orientation. We tested this hypothesis in an experiment that took place in a virtual reality system free of extraneous visual gravity cues. The stimulus consisted of a 20x20 m ground-like surface, textured with an irregular mesh, presented with binocular disparity. In the static condition, the surface was presented 3 times with 4 slants from 2.5 to 20 deg deviation from the horizontal, and 5 tilts, the direction of maximum inclination, from -45 to +45 deg from straight ahead. In the dynamic condition, the surface was initially presented horizontally, and gradually inclined to same orientations. We measured center of pressure (COP) using a Wii Balance Board. We fitted COP trajectories by 95% confidence ellipses, and performed a linear regression between the angle of major axis and surface tilt, in order to quantify how postural reactions are aligned with the direction of surface inclination. In the static condition, mean slope was 0.21 (p=0.04, signed-rank test), with the correlation significantly strongest for slant of 5 deg. In the dynamic condition, mean slope was 0.55 (p=0.03), with a maximum at slant 10 deg. Time evolution of the regression slope indicates a delay followed by a plateau in the static condition, and an immediate and continuous rise in the dynamic condition. These results support the horizontality assumption in both dynamic and static conditions, but with postural reactions about two times smaller in the static condition.
Meeting abstract presented at VSS 2017