Abstract
We evaluated the influence of moving visual scenes and knowledge of spatial and physical context on visually induced self-motion perception in an immersive virtual environment. A sinusoidal, vertically oscillating visual stimulus induced perceptions of self-motion that matched changes in visual acceleration. Subjects reported peaks of perceived self-motion in synchrony with peaks of visual acceleration and opposite in direction to visual scene motion. Spatial context was manipulated by testing subjects in the environment that matched the room in the visual scene or by testing them in a separate chamber. Physical context was manipulated by testing the subject while seated in a stable, earth-fixed desk chair or in an apparatus capable of large linear motions, however, in both conditions no actual motion occurred. The compellingness of perceived self-motion was increased significantly when the spatial context matched the visual input and actual body displacement was possible, however, the latency and amplitude of perceived self-motion were unaffected by the spatial or physical context. We propose that two dissociable processes are involved in self-motion perception: one process, primarily driven by visual input, affects vection latency and path integration, the other process, receiving cognitive input, drives the compellingness of perceived self-motion.