Perception of human walkers’ direction is important for social interaction. Observers infer walker’s trajectory from his gaze direction (Nummenmaa, Hyona & Hietanen, Psychological Science, 2009). We reported that perception of walker’s direction was accurate with motion display only for 117 ms (Sato et al., ECVP 2008), however the gaze did not affect it. We aimed to investigate effects of head direction and body appearance on low-level perception of walker’s direction. We presented a walker stimulus, whose direction was randomly chosen in the rage of left and right 24 deg to observers, for 500 ms (two steps per second). The first walking phase was randomly chosen for every trial. The direction of head was either left 9 deg to, right 9 deg to, or same as the body. We varied body and head appearances, and combined them to make a stimulus. Body was either a gray-scale computer graphics of human, a robot made by replacing body parts with 15 boxes, or point lights at 18 joints as biological motion stimulus. Head was either gray-scale human face or boxes-made robot face. Ten naive observers were asked to adjust direction of a pin-shape probe to match perceived walking direction of the body irrespective of the head. We found a significant main effect of head direction, and a significant interaction of head direction and body appearance. Perceived direction of walker was modulated in the direction of head, and the modulation was larger with point-light body than human and robot. Response variance was larger for point-light walkers than human and robots. A subsequent experiment revealed that the same effects occurred for inverted walkers. These results suggest that perception of walker’s direction is implicitly and automatically affected by head direction, and the effect of head direction is strong when the perceptual reliability of body direction is low.
Meeting abstract presented at VSS 2012