EEG Coherence and Social Behavior in Children with Autism Spectrum Disorders

Background: Research investigating the neurological causes of autism indicates that social impairments in the disorder result from disturbance in a ventromedial “social brain” circuit and the frontal lobe. As social processing requires the integration of various brain areas, understanding how these regions interact is imperative. EEG coherence, or connectivity, is a marker of interaction between brain regions, providing insight into how they communicate during social interactions. Also, connections between neural measures and behavioral symptoms are important for developing interventions for autism. Objectives: This research explores EEG coherence between brain regions in children with autism during baseline, familiar person, and unfamiliar person video conditions. Coherence is related to parent report of behavioral symptoms of autism. Methods: Eighteen children with high-functioning autism (ASD), ages 8-12 (mean age 10.06; 17 males), viewed a baseline condition, a video of an unfamiliar person and a familiar person reading a story. Thirteen typically developing children (TYP) (mean age 9.92; 10 males) were tested for comparison. Brain activity in the frontal and parietal-temporal lobes was measured utilizing EEG. A 64-channel electrode cap collected data, which was analyzed using the Scan 4.3 program. Coherence was computed between electrodes in the frontal and right parietal-temporal lobes, frontal and left parietal-temporal lobes, and right and left frontal lobes using Compumedics-Neuroscan and Mathworks MATLAB programs. Caregivers completed the Social Skills Rating System (SSRS) and Social Responsiveness Scale (SRS). Coherence was entered into three repeated-measures ANOVAs. Diagnosis served as the between groups factor and video condition served as a three-level within groups factor. Coherence values were correlated with the SRS and SSRS. Results: A significant interaction between diagnosis and condition was found: the TYP group had higher coherence between the left and right frontal lobe than ASD during all conditions, Wilks Lambda=.59, F(2, 23)=8.14, p=.02, partial η2=.42. No significant interaction existed between diagnosis and video condition between the left parietal-temporal lobe and frontal lobe, Wilks Lambda=.91, F(2, 21)=1.08, p=.36, partial η2=.09), or between the right parietal-temporal lobe and frontal lobe, Wilks Lambda=.996, F(2, 23)=.04, p=.96, partial η2=.004. A main effect for video condition was found, Huynh-Feldt=.04, F(1.28, 23)=4.54, p=.03, partial η2=.16; ASD showed decreased coherence at baseline compared to TYP. No main effect for diagnosis was found, F(1, 23)=1.26, p=.27. Correlations for complete sample indicate higher intra-frontal coherence values during baseline were related to increased social skills (SSRS Social Skills scale; r=.44, p<.05), and social motivation (SRS Social Motivation scale; r=-.46, p<.05). Within the ASD sample, higher levels of social motivation (SRS Social Motivation Scale) were correlated with higher intra-frontal coherence values during baseline (r=-.52, p<.05), familiar condition (r=-.53, p<.05), and unfamiliar condition (r=-.64, p<.05). Higher coherence values between the frontal lobe and left parietal-temporal lobe during baseline were related to lower levels of social awareness (SRS Social Awareness Scale; r=.57, p<.05), social cognition (SRS Social Cognition Scale; r=.79, p<.01), and social communication (SRS Social Communication Scale; r=.56, p<.05). Conclusions: Conclusions will discuss the overconnectivity hypothesis in the frontal lobe in ASD, as well as possible lateralization of the brain in ASD.