Abstract
With simulated optic flow (OF), heading thresholds are little affected by the retinal position of stimulation (Warren & Kurtz, 1992; Crowell & Kurtz, 1993). The only retinal advantage is with radial flow in the central retina. This study examines whether these passive-judgment findings generalize to the use of OF in walking. With an immersive virtual environment, 5 subjects performed a walking task with full fields (48 diam) and with simulated visual field loss linked to eye position. Retinal stimulation was restricted to either the central 10, a 10 window located 10 left of fixation (parafoveal), or the full field with a central 10 mask (peripheral). The target was presented in an empty space that lacked OF cues, a richly textured forest with trees, and a “cloud of dots” that lacked reliable landmark cues. Subjects were instructed to walk to a pole 8m away. Head position, measured with an optical tracker, determined the point of view and was used to record the walking path for analysis. To identify the influence of OF, the OF pattern was offset 10 to the right or left of the direction of walking. A 10 heading error indicates no OF influence, whereas a 0 error indicates maximum influence. The results showed that with full fields, OF was used only modestly to guide walking and the amount was reduced when stimulation was in the central retina. The peripheral stimulation had the lowest heading error in the forest scene (6.9 ), comparable to that with the full field (7.0 ), and significantly lower than that with central or parafoveal stimulation (9.0, 10.4 ). Similar findings were obtained in the cloud scene. No difference in heading error was observed in the empty-space scene (8.9 –9.8 ). These findings show that retinal position matters in the use of OF for goal-directed walking, but not as predicted by the threshold results. The use of simulated field loss in mobile observers takes us one step closer in bridging the gap between lab and the real world.