Early Neural Activation During Emotional Face Processing in Children with Autism

Background: Impaired social skills are a hallmark of Autism Spectrum Disorders (ASD). The ability to accurately perceive and interpret emotional faces is integral in successful social functioning.  Our group has shown significant differences in neural activity associated with emotional face processing in adolescents with and without ASD. Examining neural activity during emotional face processing in a younger cohort using the same paradigm will determine when differences in neural activity during emotional face processing emerge during development.

Objectives: We examined neural responses associated with emotional face processing in children with and without ASD through the use of magnetoencephalography (MEG). We hypothesized that a distinct spatiotemporal profile of neural activity during emotional face processing would be seen in children with ASD relative to typically developing controls.

Methods: We have preliminary data on eight children with high-functioning ASD, determined by ADOS and ADI, (7M, M=9.4+1.2 years) and nine typically developing controls (8M, M=9.2+1.1 years). Participants completed an implicit emotional face processing task while in the MEG scanner. Emotional (happy or angry) or neutral faces were shown concurrently with a scrambled pattern (target), on either side of a central fixation cross. Children indicated the location of the target. Individual structural MRIs were acquired for co-registration with MEG data. To examine the time course of neural activation during emotional face processing, global field power (GFP) plots were generated for emotional and neutral faces. The children also completed two subtests of the WASI and the Affect Recognition subtest of the NEPSY-II.

Results: To angry and happy faces, GFP plots showed similar trends in neural activity between groups, with peaks at approximately 140ms and 180ms.  Control children showed consistently greater overall amplitudes of neural activity relative to children with ASD. A different pattern was seen for neutral faces. GFP plots showed greater activity in controls peaking at approximately 140ms but greater neural activity in children with ASD at a second peak at approximately 180ms. Further analyses will be conducted, on a larger cohort of 15 children per group. MANOVA showed no significant effects of group or emotion on response latencies. Independent sample t-test showed significant differences in performance on the Affect Recognition subtest, with children with ASD (N=5) scoring lower (M=9.8, SD=2.0) than controls (N=8, M=12.0, SD=1.2), t(11)=2.5, p=.031.

Conclusions: Our preliminary results indicate that while children with ASD respond at similar latencies to emotional faces compared to controls, the underlying neural activity shows differences early on (100ms), similar to our previous findings in adolescents with and without ASD. This suggests that neural differences underlying emotional face processing are present in children with ASD. Of particular interest was the reversed pattern of neural activity to neutral faces, with children with ASD showing greater overall activity relative to control children.  This raises a caveat for studies that use neutral as an emotional baseline, especially in autism. That emotional face processing deficits exist in childhood is further supported by behavioural findings that children with ASD recognized facial affect more poorly than children without autism.