Electrophysiological Indices of Biological Motion and Audio-Visual Integration in Infants at Risk for Autism

Background: Perceptual sensitivity to biological motion (BM) is evident in infants as young as two days old and is hypothesized to underpin development of more sophisticated social behaviors. Similarly, audio-visual (AV) integration, or the detection of a temporal contingency between auditory and visual events, is also evident in infancy and is a critical developmental ability. Behavioral studies suggest that toddlers with autism spectrum disorder (ASD) demonstrate atypical BM perception and display preferential attention to AVS (audio-visual synchrony) in comparison to unimodal stimuli. These early processing atypicalities may represent early predictors of autism.  

Objectives: This study applied event-related potentials (ERPs) to contrast neural responses to BM and AV integration in infants at high-risk (HR) and normal-risk (NR) for ASD. By investigating electrophysiological markers of BM and AV events, we aimed to (a) assess potential hyposensitivity to BM in HR infants, (b) identify intact/enhanced integration of AV events in HR relative to NR infants, and (c) compare developmental trajectories of typical and atypical neural responses to BM and AVS prior to the emergence of behavioral symptoms in ASD.

Methods: HR (n= 25) and NR (n= 40) infants were assessed at three-month intervals between 3 and 12 months of age. EEG was recorded with a 128-channel Hydrocel Geodesic Sensor net while infants viewed point-light displays depicting BM and scrambled motion (SM; Experiment 1) or unimodal/bimodal auditory (tone) and visual (circle) stimuli (Experiment 2). In Experiment 1, ERPs evoked by BM or SM (N200, a negative deflection over right occipitotemporal scalp between 200-300 ms; PSW, a late anterior positive slow wave between 900-1500 ms) and event-related oscillations (EROs) in the mu range (6-9 Hz), indicating activity of the action perception system, were computed. In Experiment 2, ERPs evoked by audio, visual, and AV presentations (N100, a negative deflection over fronto-central scalp between 80-130 ms) and EROs in the gamma range (20-100 Hz), indicating activity of integration mechanisms, were computed.  

Results: In Experiment 1, neither NR nor HR differentiated between BM and SM at the N200 at 6 months (ps>.05). At 12 months, only NR infants displayed emergent differentiation of the N200 elicited by BM and SM (p=.019). At 6 months, NR infants exhibited enhanced amplitude of the PSW to BM relative to HR infants (p<.05). Between 9-12 months, HR infants failed to display expected attenuation of EEG mu rhythm to BM (p=0.65) relative to NR infants (p=.038).  In Experiment 2, NR but not HR infants displayed expected sensitivity to AV integration at the N100; the N100 AV response in NR infants was more negative than the summed response of audio-only and visual-only.

Conclusions: Infants at risk for autism demonstrated atypical neural responses at electrophysiological markers of both social and basic sensory perception. These findings suggest that, rather than focal anomalies in social perceptual brain circuitry, distributed perceptual differences, including those affecting multisensory integration, may underlie autistic symptomatology. The study demonstrates that use of complementary measures of brain function may shed light on factors associated with risk status and eventual diagnosis of ASD.