INTRODUCTION: The illusion of self-motion induced by moving visual stimuli has typically been attributed to bottom-up perceptual processes. Here, we investigated whether a cognitive factor such as spatial presence can contribute to the illusion. Spatial presence was indirectly manipulated by presenting either a photorealistic image of a natural scene or modified versions of the same stimulus. Those were created by either scrambling image parts in a mosaic-like manner or by slicing the original image horizontally and randomly reassembling it. We expected scene modifications to decrease spatial presence and thus impair vection. METHODS: Twelve observers viewed stimuli projected onto a curved projection screen (FOV: 54 x40.5 ). Dependent measures included vection onset time, vection intensity, and convincingness of the illusion (0-100% ratings). Spatial presence was assessed with presence questionnaires. RESULTS: Scene modification led to both reduced presence scores and impaired vection: Modified stimuli yielded significantly longer vection onset times, lower perceived intensity, and lower convincingness ratings than the intact market scene. No clear difference was found between the sliced and scrambled stimuli or among the number of slices or mosaics (2, 8, or 32). Results suggest that high level information (consistent reference frame for the intact market scene) dominated over the low-level information (more contrast edges in the scrambled stimulus, which are known to facilitate vection). CONCLUSIONS: Results suggest a direct relation between spatial presence and self-motion perception. We posit that stimuli depicting naturalistic scenes provide observers with a convincing reference frame for the simulated environment which enables them to feel "spatially present". This, in turn, facilitates the self-motion illusion. This work has important implications for both self-motion perception and motion simulator design and applications.

EU grant POEMS-IST-2001-39223 (see www.poems-project.info) and Max Planck Society