Item

Abstract

Whilst moving through the environment humans use vision to discriminate different self-motion intensities and to control their action, e.g. maintaining balance or controlling a vehicle. Yet, the way different intensities of the visual sensory stimulus affect motion sensitivity is still an open question. In this study we investigate human sensitivity to visually induced circular self-motion perception (vection) around the vertical (yaw) axis. The experiment is conducted on a motion platform equipped with a projection screen (70 x 90 degrees FoV). Stimuli consist of a realistic virtual environment (360 degrees panoramic color picture of a forest) rotating at constant velocity around participants’ head. Visual rotations are terminated by participants only after vection arises. Vection is facilitated by the use of mechanical vibrations of the participant’s seat. In a two-interval forced choice task, participants discriminate a reference velocity from a comparison velocity (adjusted in amplitude after every presentation) by indicating which rotation felt stronger. Motion sensitivity is measured as the smallest perceivable change in stimulus velocity (differential threshold) for 8 participants at 5 rotation velocities (5, 15, 30, 45 and 60 deg/s). Differential thresholds for circular vection increase with stimulus intensity, following a trend best described by a power law with an exponent of 0.64. The time necessary for vection to arise is significantly longer for the first stimulus presentation (average 11.6 s) than for the second (9.1 s), and does not depend on stimulus velocity. Results suggest that lower sensitivity (i.e. higher differential thresholds) for increasing velocities reflects prior expectations of small rotations, more common than large rotations during everyday experience. A probabilistic model is proposed that combines sensory information with prior knowledge of the expected motion in a statistically optimal fashion. Results also suggest that vection rise is facilitated by a recent exposure.