Abnormal Neural Correlates of Audiovisual Multisensory Integration in Autism Spectrum Disorders

Background: Individuals with autism spectrum disorders (ASD) exhibit abnormalities in multiple modalities of sensory functioning and in multisensory integration (Iarocci & McDonald, 2006; Rogers & Ozonoff, 2005). For example, utilizing a multisensory “flash-beep” illusion wherein presentation of a single visual flash along with two temporally proximal auditory beeps results in perception of an illusory second flash (Shams et al., 2000), research has revealed capacity for cross-modal integration in ASD (Van de Smagt et al, 2007), but an extended temporal window over which stimuli are integrated (Foss-Feig et al, 2010). Research using event-related potentials (ERP) with typical adults has demonstrated differences in neural P180 and N270 responses when the illusory second flash is perceived versus not perceived, reflecting the neural signatures of multisensory integration (Mishra et al., 2007). The substrates underlying atypical audiovisual integration processes in ASD have not yet been investigated. 

Objectives: The present study investigates disruption in neural mechanisms subserving multisensory integration in ASD using the flash-beep illusion. Specifically, this study examines ERP responses to the illusion in children with ASD and typically-developing (TD) controls.

Methods: Participants were 41 children (19 with ASD, 22 with TD) between 10 and 13 years of age (ASD = 11.98 years; TD = 12.05 years, on average).  ERPs were recorded with a 128-channel net. Participants were presented with several trial types including a condition with two beeps and one flash and a condition with two flashes and two beeps. Children responded to each trial with the number of flashes they perceived. This allowed for the separation of trials in which the subject did or did not perceive the illusion. Data analysis computed within-group grand-averaged ERPs corresponding to either perceiving or not perceiving the illusion. Results were compared between groups.

Results: Overall, group differences in brain response were found in the N2 and P2 components, reflecting abnormalities in general sensory and illusion-specific multisensory processes in ASD. Across groups, the ERP response over occipital cortex was characterized by a prominent P1 (80-140ms) component that was equivalent whether or not the illusory second flash was perceived. However, across conditions, children with ASD showed a larger than typical N2 (170-230ms) amplitude over visual cortex, suggesting a general difference in sensory processing in ASD. Analysis of ERP response over central and parietal electrodes revealed group differences between illusion and no-illusion conditions, providing neural correlates of multisensory integration. In children with TD, illusion perception was reflected in a double-peaked positive component beginning 180ms post-stimulus and peaking at 230 and 315ms. Children with ASD did not display this double peak; instead, illusion perception elicited heightened amplitude beginning at 180ms at both parietal and central electrodes. 

Conclusions: Results help to clarify the neural underpinnings of cross-modal integration in ASD. Specifically, it appears that though both groups are susceptible to a multisensory illusion, the brain mechanisms by which integration occurs differ between ASD and TD.  These results contribute to a broader understanding of sensory processing impairments in ASD, having potential to inform refinement of diagnostic assessments and therapeutic interventions.