Abstract
Optic flow is one source of visual information relied on for various locomotion tasks, including steering, obstacle avoidance, and path integration. In the real world the information from optic flow is redundant with information from body senses (vestibular, proprioceptive, and efferent signals). People are relatively accurate in reproducing distances traveled based solely on optic flow (Bremmer & Lappe, 1999). However, when body senses are also available, they tend to dominate (Kearns, Warren, Duchon, & Tarr, in press; VSS 2001). In addition, participants fail to notice large (+/− 50%) changes in the gain of optic flow during walking, if their attention is not drawn to it. This leads to the question: how sensitive are observers to such gain changes? If the body senses are more useful in determining the distance traveled, one might expect low sensitivity to gain changes in flow speed. On the other hand, if optic flow is a robust source of information about self-motion, one might expect sensitivity may be high. Participants wore a head-mounted display (60 × 40 deg) while walking on a straight path (6–8m per trial) in a virtual environment; head position was recorded with a sonic/inertial tracker. Gain-change thresholds were measured using a 2AFC task in a 2-up, 1-down staircase procedure. The baseline gain of flow was manipulated (50%, 100% (normal), 200%), with incremental differences in gain presented on each trial. Results indicate high sensitivity (comparable to similar tasks) regardless of the baseline optic flow rate, indicating that flow sensitivity is relatively constant even in less familiar ranges. In a second experiment, we examined whether the actual walking speed affects sensitivity to flow gain by manipulating the speed of walking (normal, fast, and slow) with a baseline gain of 100%.
NSF LIS IRI-9720327, NIH KO2 MH01353