Objectives: The present study measured electrophysiological brain responses to biological motion and AVS in infants at HR and normal-risk (NR) for ASD. By contrasting biological markers for human motion and temporal contingency of audio-visual stimuli, we aimed to detect (a) undersensitivity to biological motion and (b) oversensitivity to AVS in order to characterize indicators of atypical social development prior to the emergence of behavioral symptoms.
Methods: Participants included HR (n=20) and NR (n=40) infants assessed longitudinally at three-month time points between 3 and 18 months of age. EEG was recorded with a 128-channel Hydrocel Geodesic Sensor net while infants were presented with: point-light displays depicting biological motion and non-biological (scrambled) motion (Experiment 1); or unimodal and bimodal auditory (tone) and visual (circle) stimuli (Experiment 2). In Experiment 1, event-related potentials (ERPs) evoked by biological or scrambled motion and event-related oscillations (EROs) in the mu range (6-9 Hz), indexing activity in the action perception system, were extracted. In Experiment 2, ERPs evoked by unimodal and bimodal stimuli presentations and EROs in the gamma range (20-100 Hz), indexing binding and feature integration mechanisms, were extracted. All infants were administered the Mullen Scales of Early Learning and a comprehensive battery of social and communicative assessments at each time point.
Results: In Experiment 1, preliminary analyses from a subsample showed that NR infants exhibited a negative deflection in ERPs over right occipitotemporal scalp between 200-300 milliseconds that distinguished biological motion from scrambled motion at 12 months (p=0.019) but not at 6 months (p=0.132). Conversely, differentiation was not observed in HR infants at either 6 (p=0.206) or 12 months (p=0.870). Likewise, a significant difference in evoked mu suppression between scrambled and biological motion was detected in NR [t=-2.17, p=0.038] but not in HR infants [t=0.47, p=0.65]. In Experiment 2, preliminary analyses from a subsample indicated audio-visual integration between 78 and 198 ms over fronto-central scalp in both the HR group and the NR group; however, HR infants showed a shorter latency for bimodal components than NR infants (105.6 ms for HR versus 133.3 ms for NR).
Conclusions: Compared to NR infants, HR infants showed attenuated responses to biological motion on multiple, convergent electrophysiological markers. In contrast, HR infants showed intact multi-sensory integration, with more efficient detection of AVS compared to NR infants. The current study demonstrates, for the first time, a neural dissociation between perception of social and non-social information in infants at elevated risk for ASD.