Abstract
Humans show better recognition of moving shapes presented through stationary multiple slits than of static shapes, as well as a remarkable ability to recognize biological motion from impoverished inputs. This study investigated whether spatial pattern processing in shape recognition of a human figure can be influenced by motion processing, or whether the two processes are independent. A walker stimulus consisted of a silhouette of a person walking on a treadmill. The stimulus was viewed through a set of thin slits (1-pixel width) separated by noise-corrupted masking fields. Observers were asked to identify the person's facing direction. Experiment 1 included three conditions: a single frame of the walker rigidly translated horizontally back and forth, a forward walker, and a backward walker. Vertical slits moved towards the left or right. No difference in judging the walker's facing direction was found between forward and backward walking conditions. However, we found poorer recognition of a biologically moving human figure compared to recognition of a rigidly translating human figure. This impairment diminished when the walker was inverted. In Experiment 2, we found similar impairment of biological motion perception for both static vertical slits and moving vertical slits. But when the slits were horizontal and static, recognition performance was more accurate in walking conditions than in the rigid translation condition. Facilitation was reduced when horizontal slits moved upward or downward. The discrepancy between vertical and horizontal slits can be explained by the need for interpolation between accelerating and decelerating fragments separated by space and time in the former but not the latter case. These results provide clear evidence that the visual system uses motion mechanisms to modulate shape interpolation along the trajectory of object movement. For biological motion, integration of spatial pattern information is affected by trajectory complexity due to deformation or articulation.