Abstract
Human actions involve subparts of body movements nested within a hierarchical structure. However, numerous motion hierarchies are consistent with observed body movements. It remains unknown how the visual system identifies a particular hierarchic structure of body movements from visual input. To examine how multiple constraints interact to identify the hierarchy of body movements, and what mechanisms link different layers of the hierarchical representation, we used the behavioral oscillation paradigm, as stimulus-locked fluctuations in behavioral performance elicited by rhythmic brain activity. On each trial, moving limbs of a leftward or rightward walker, shown as a stick figure, were briefly presented for 100 ms. After an interstimulus interval (ISI) densely sampled from 0 to 1000 ms with 31 levels, participants viewed a point-light walker for 200 ms and judged its facing direction. Experiment 1 showed bipedal leg movements with opponent motions in the first stimulus; Experiment 2 showed arm-leg movements with opponent motions (as the arm and the leg were on the same side of the body), and Experiment 3 showed arm-leg movements with the same motion direction (as the arm and the leg were on different sides of the body). An analysis of response times (RT) showed that bipedal leg movements yielded the strongest priming effect, with the subparts of body movements for the congruent walking direction facilitating recognition of the subsequent point-light walker. Fourier analysis of RTs revealed theta-band (3-7 Hz) oscillations for all three experiments. However, only sub-body movements with opponent motions in Experiment 1 and 2 elicited alpha-band (10-12 Hz) oscillations. Our results suggest that motion opponency plays an important role in determining the hidden layer of the hierarchical representation for walking actions, and that neural oscillation may serve as a key mechanism to bind layers of hierarchical representation for complex visual stimuli.
Meeting abstract presented at VSS 2018