Early correlates of agency perception in interaction scenes: preliminary evidences from Williams syndrome

Our interaction and social skills are deeply rooted in the ability to quickly and properly detect other agents and grasp their intentions and emotions from their behaviour. According to direct access theories, those abilities are supported by innate “smart” perceptual processes, which extract relevant cues from complex scenes since the first stages of information-processing. Again, the atypical development of those abilities might be associated to clinical manifestations of Williams Syndrome (WS), a rare genetic disorder characterized by peculiar cognitive-affective profile and anomalous social behaviour. The present study aims at investigating early neural correlates of agency perception in interaction and electrophysiological specificities of WS during initial steps of social understanding. One participant with WS and 20 volunteers took part in the study. The genetic diagnosis of WS was established using Fluorescence in Situ Hybridization (FISH) probes for elastin gene. All participants were asked to carefully observe some dynamic visual stimuli (two static frames presented in close succession) showing realistic interactions, while EEG activity was continuously recorded. The interaction scenes included a human or artificial agent executing gestures. Event-Related Potentials (ERPs) were considered after the morphological analysis. Based on previous empirical evidences, we focused on two negative ERPs components – N2 and N3 – that proved to be sensitive to the manipulation of agent’s nature. In order to properly compare controls’ and patient’s data, all signals have been re-referenced to individual averages. Moreover, the morphology and relative amplitude of selected components have been compared taking into account the statistical dispersion of group data for each data point (resolution=1 ms). Differences in patient’s responses were defined as significant only if they were outside a ±1 SD range from the control group mean. Direct comparisons of patient’s and controls’ ERPs responses highlighted similar early sensory components, but an enhanced N2 peak regardless of agent’s nature – which has been associated to atypical processing of perceptual information for the detection of biological entities – and an increased N3 deflection – which has been associated to an increase in resource demand for pre-reflective detection of intentionality – for the WS participant. We suggest that our empirical evidences might mirror the peculiar uneven profile of WS social skills, with abnormally heightened sensitivity to social stimuli and remarkable difficulties in detecting and decoding others’ mental states even in perceptual tasks, despite their spared social perception skills