Abstract
Sensory predictions made by the motor system during movement are likely to facilitate adaptive interaction with the sensory world. Recently, it has been suggested that agents may use the same sensorimotor models to anticipate the consequences of their own movements and to predict matching behaviours in others (i.e., imitative reactions). Under this hypothesis, underlying sensory models should exhibit considerable generality across time and viewpoint, as the actions of others are typically observed at delays in the order of seconds and from different perspectives. Here we examine this hypothesis by investigating how action influences the processing of visual outcomes presented at different viewing perspectives and with varied action-effect delays. We measured the perceived brightness of sensory outcomes of movement, given recent findings that events congruent with prediction are perceived more brightly than those unpredicted. In a series of psychophysical experiments, participants performed finger abductions (index or middle) while observing congruent or incongruent movements performed by an avatar hand. The observed lifted finger 'flashed' at one of seven intensities (10-70 % increase in local brightness) 50 ms after observed movement onset before being replaced by a comparison stimulus. Participants judged which event appeared brighter, and the point of subjective equivalence (PSE) was derived from the modelled psychometric function. Experiment 1 revealed that when hands are viewed from a first-person perspective, congruent action outcomes are perceived to be brighter, irrespective of the action-effect delay (0 -3600 ms). In contrast, Experiment 2 revealed no effects of action congruency when hands are viewed from a third-person perspective. These results suggest that predictive processes during action may operate with greater generality than previously thought. However, despite generalising across time, the underlying mechanisms may still be ill-suited to anticipating the behaviour of others given their lack of generalization across viewpoint.
Meeting abstract presented at VSS 2016