Abstract
The attribution of meaning in biological motion displays is highly dependent on the correct spatio-temporal information. Here we assess the importance of the temporal coherence, between and within actors, for the interpretation of non-verbal cues conveyed in video sequences depicting two actors engaged in interlocution. The actors read 6 excerpts from plays for 15 seconds (depicting the communication of fear, aggression, happiness, love and neutral emotions) which were videoed and converted to point-light biological-motion displays with one point on the head, two on the shoulders, elbows, wrists, hips, knees and feet. In exp.1 subjects were required to discriminate a pair of actors who were videoed together from 3 dyads who were videoed with a different partner. In exp.2 intra-subject temporal coherence was manipulated by disrupting the spatio-temporal order of the video sequence. In exp.3 inter-subject temporal coherence was disrupted by offsetting the two actors by n frames until the discrimination of the correct natural dyad fell to chance. In exp. 4 we investigated whether subjects could classify the emotions depicted by the actors. Performance in exp.1 was close to perfect for all subjects which suggests that the correct inter-personal emotions, displayed by the non-verbal behaviours, were easily discriminable from spatio-temporal information contained in point-light displays. This ability was invariant to intra-personal manipulations where performance remained at above chance levels with disruptions up to 320 msecs. The disruption of inter-personal coherence had an effect on performance, which fell to chance with perturbations of only 3 frames (120 msecs). Exp. 4 showed that subjects could decode and categorise the emotions displayed. These results suggest that the temporal coherence of inter-personal behaviour, which controls turn taking and the communication of emotion, can readily be discerned in the spatio-temporal pattern of point-light displays.