TESTING A VISUAL MOTION HYPOTHESIS OF PATH INTEGRATION: HUMANS CAN ESTIMATE DISTANCE AND DIRECTION USING ONLY OPTIC FLOW

T.M. Ellmore*; E.R. Lindstedt; B.L. McNaughton

NSMA, Univ Arizona, Tucson, AZ, USA

Navigation by path integration requires that self-motion cues be used to keep track of the translatory and rotatory components of a journey so that one s position with respect to a starting location may be updated accurately. We hypothesized that ideothetic information from the optic flow field alone is sufficient for navigation. Sixty-one humans actively explored with a joystick a computer generated virtual environment. Then, after removal of all local and distal visual landmarks, participants estimated their position relative to previously visited locations during either passive linear displacement or angular rotation. Participants were able to estimate their location with respect to targets in a heads-down journey during which the only continuous self-motion cue available was the speed of motion of the ground. Estimation errors varied significantly as a function of distance to target, the speed of movement and the number of landmarks available at the beginning of the journey. Estimation errors and the variance of the estimation errors increased significantly as a function of distance to the target. The results suggest that navigation may be guided by visual motion alone, and that the accuracy of visual motion navigation may be influenced by the distribution of stable references in the environment. The results are also consistent with a model of path integration, a recurrent process in which errors in the current measure of position changes are cumulative.

Support Contributed By: NS020331, JST CREST & MH01565

path integration, navigation, motion, virtual reality