Estimation of traveled distance in a virtual environment

We investigated the role of optic flow information in the estimation of distance travelled and its relative importance in the presence of other sensory and non-sensory information (e.g., proprioceptive cues and elapsed time). Subjects wore a head-mounted display and rode a stationary bicycle along a straight path in a seemingly infinite hallway with random surface texture and without any distinctive features. The distance that subjects were required to travel varied randomly (equivalent to 8 to 64 meters) with the end of the path signalled auditorily. A percentage estimation of relative distance was required following a reference distance subjects initially travelled. The rate of pedalling motion was converted to optic flow velocity in the display. Such a proportional factor could be either held constant or varied across trials (much like changing the gears of a bicycle, but without changing required pedaling force). When a constant proportional factor was maintained over trials, subjects were able to make reliable distance estimations. The proportion of variance in distance estimations accounted for by distance was on average 86. When the proportional factor was varied randomly between trials, the distance information based on visual input was often incongruent with that based on pedalling movement and/or elapsed time. Subjects' distance estimations in this condition were found to rely more heavily upon the visual source as opposed to relying on the information provided by pedalling motion. On average, the proportion of variance accounted for by distance information derived from visual input was 59, while the amount of variance accounted for by distance information predicted using pedalling motion was 10. By introducing a sensory conflict paradigm, the current study convincingly supported the notion that travelled distance estimation depends primarily on optic flow even in the presence of additional sensory and cognitive cues