Abstract
Visual tasks that are sensitive to in-plane orientation (e.g. perception of biological motion; face recognition) depend not only on the orientation of stimuli with respect to the egocentric reference frame, but also with respect to external reference frames such as gravity and the visual environment. Here we use immersive virtual environments (VEs) to investigate how egocentric and external reference frames interact in a screen orientation preference task. Participants (206 UC Santa Cruz undergraduates) observed an 8-minute video of a soccer match while sitting upright or tilting sideways at 30°, 60°, or 90° and wearing an Oculus Rift CV1 head-mounted display. This paradigm allowed us to dissociate vestibular and visual orientation cues. Participants in the "fixed room" condition (n=54) were immersed in a stable VE that was always aligned with gravity; in the "changing room" condition (n=64) the VE changed across blocks and was aligned with participants' own body orientation; in the "dark room" condition (n=64), the VE was dark and only the display screen was visible. At each of the four body positions, the display screen randomly changed its orientation every 20 seconds, and participants were instructed to readjust its orientation until they felt it most comfortable for viewing. Average screen orientation preferences depended systematically on both body tilt and the orientation of the VE. In both the "dark room" and "fixed room" conditions, as participants tilted (counterclockwise) at 30°, 60°, and 90°, their preferred screen orientations were (clockwise) 10°, 13°, and 12°, respectively (relative to their egocentric frame). In contrast, in the "changing room" condition when the VE was aligned with participants' body, preferred screen orientations were close to 0°: (clockwise) 2°, 4°, and 5°, respectively. Our results indicate a significant interaction between body tilt and visual orientation cues in predicting screen orientation preferences.
Meeting abstract presented at VSS 2017