We have reported that searching an approaching walker among 2–5 deviating walkers was more efficient than searching a deviating walker among approaching walkers for small deviation angles (6–12 deg), but the search asymmetry got opposite/reversed for large deviations (30–60 deg) (Ono & Kitazaki, VSS 2010). In the present study we aimed to investigate inversion effect of walkers' search asymmetry and to apply point-light display (biological motion) to the search asymmetry. In Experiment 1, we presented 3, 4, or 6 human walkers (front view, smooth shaded 3-dimensional computer graphics), one of which was approaching to or deviating from the observer, while the other walkers were deviating from or approaching to the observer, respectively (deviation angle: 6, 12, 30, or 60 deg). The orientation of walkers was upright or inverted. Ten observers were asked to identify one approaching or deviating walker as accurately and quickly as possible. Identification of an approaching walker among deviating walkers was quicker than the opposite identification with small deviation (6 deg), but that of a deviating walker among approaching walkers was quicker with 30 and 60-deg deviations for both upright and inverted walkers. Visual search of inverted walkers was less efficient than upright walkers especially with small deviations. In Experiment 2, we conducted the same experiment using biological motion stimuli (18 point-lights on joints). We found that the reaction time was much longer than computer-graphics walkers with small deviations (twice for 6-deg and 1.5 times for 12-deg) and that the search asymmetry and its reversal were very similar to those of upright computer-graphics walkers. These results suggest that searching walkers with small deviations requires walker-specific social-cognitive processing in which approaching is more important than deviating, while the search with large deviations is related with ordinary object perception in which deviation properties are salient.
Supported by Grant-in-Aid for Scientific Research on Innovative Areas (22118504), and The Global COE program ‘Frontiers of Intelligent Sensing’.