Early Neural Network Activation During Emotional Face Processing in Adolescents with Autism

Background: Recognizing and allocating appropriate attention to emotional faces is integral to children’s ability to acquire social skills. In autism spectrum disorders (ASD), impaired social skills and interactions are one of the diagnostic indices of the disorder. In children and adolescents with ASD, neuroimaging studies have suggested that deficits in emotional processing reflect abnormalities in underlying neural networks, particularly the amygdalae (Baron-Cohen et al., 2000). ERP findings of delayed and smaller neural responses to emotional faces in children with ASD highlight the importance of utilizing neuroimaging techniques with high temporal resolution, such as MEG, to investigate the temporal and spatial properties of the neural network underlying emotional processing in ASD (Batty et al., 2011).

Objectives: To assess neural network activation during emotional face processing in adolescents with ASD through the use of magnetoencephalography (MEG). We hypothesized that adolescents with ASD would show distinct emotional processing patterns, relative to controls, and atypical activation of specific brain regions, including the amygdalae and anterior cingulate cortex.

Methods: Six adolescents with ASD (6M, M=14.16+1.28 years) and 9 typically developing controls (6M, M=14.28+1.09 years) performed an attentional paradigm (Hung et al., 2010) involving implicit emotional face processing with a neutral, happy or angry face presented concurrently with scrambled patterns, while in a MEG scanner.  Participants responded as quickly as possible to the scrambled pattern.  Event-related beamforming (ERB) analyses on the happy-neutral and angry-neutral comparisons were used to localize task-dependent brain activation.  MEG data were co-registered to the participant’s anatomical T1-weighted MRI scan.

Results: ERB analyses show early (time window of 80-120ms) bilateral middle frontal activation as well as left amygdala activation in adolescents with ASD in response to happy faces. In contrast, left anterior cingulate cortex and right middle frontal activation to happy faces were found in controls. In response to angry faces, adolescents with ASDs showed early left middle frontal and right inferior frontal activation while controls showed left medial frontal and right amygdala activation. Adolescents with ASD did not differ significantly from controls in response latency for either emotion.

Conclusions: Our preliminary results indicate differences in early neural network activation in adolescents with ASD compared to controls.  Of particular interest are our findings of left amygdala activation in response to happy faces in adolescents with ASD, which are not only distinct from that found in controls, but resembles activation found in younger children, suggesting the involvement of an immature neural network (Hung et al, under review). Data collection is ongoing and will allow us to further elucidate the neural mechanisms underlying emotional processing in ASD.