Item

Abstract

Self-motion perception refers to the ability to perceive the speed and direction of movement through space. Past measures of self-motion perception have failed to directly assess the dynamic, instantaneous properties of perceived self-motion in real-world coordinates. Here we developed a novel continuous pointing method to measure perceived self-motion during translational movements. This experiment took place in a large, fully tracked, free-walking space. Participants viewed a target and then, with eyes closed, attempted to point continuously towards the target as they moved past it along a straight, forward trajectory. Pointing behaviour was tracked using a high-precision optical tracking system which monitored a hand-held pointing device. By using arm angle, we continuously measured participants' perceived location and, hence, perceived self-velocity during the entire trajectory. We compared the natural characteristics of continuous pointing in a control condition (sighted walking) with that during conditions in which particular sensory/motor cues were reduced, including: blind walking, passive transport, and imagined walking in the complete absence of physical movement. Results demonstrate that under all reduced cue conditions involving actual movement, perceived self-velocity and displacement were relatively accurate. Specifically, the pattern of pointing in the blind walking condition did not differ from that of the passive transport condition. This indicates that, for simple, linear trajectories with a raised-cosine velocity profile, inertial cues alone can be used to perceive self-motion. Perhaps most interestingly, the “signature” pattern of pointing observed during true self-motion (notably an increase in arm azimuth velocity upon target approach) was absent during imagined pointing. Consequently, continuous pointing reveals a characteristic arm trajectory that is unique to actual self-motion. This appears to be an automatic, obligatory process that is not reproduced during a purely cognitive representation of self-motion in the absence of movement. This method has direct implications for several research areas, including spatial cognition and navigation.