Role of Anxiety in the Neural Response to Faces in Children with ASD: Results from the ABC-CT Feasibility Study

Background: Biomarkers of social-communication in autism spectrum disorder (ASD) are needed to obtain objective measures of social functioning. Furthermore, examining relationships among potential biomarkers and other measures of functioning within and across diagnostic groups is vital. There is high comorbidity of anxiety disorders in ASD; specifically, 40-84% of individuals with ASD have at least one co-morbid anxiety disorder. Exploring the role of anxiety on neural response to social stimuli measured with event related potentials (ERP) may provide insight into the complex relationship between anxiety and ASD and provide pertinent information regarding potential ERP biomarkers.

Objectives: Examine relationships between anxiety and neural response to faces in children with and without ASD.

Methods: Across five sites, ERPs were collected from 20 children with ASD [16 males; M=7.99±2.26 years] and 26 typically developing (TD) children [17 males; M=6.59±1.98 years] using a 128 electrode Geodesic Net. Children were presented with standardized photographs in two experiments: 1) 50 neutral and 50 fearful faces; 2) 72 upright faces, inverted faces, and upright houses. P100 and N170 ERP latencies were extracted from selected electrodes over the occipitotemporal cortex. Anxiety was measured with the Childhood and Adolescent Symptom Inventory, and analyses included T-scores from the General and Social Anxiety subscales. Group differences were examined using t-tests, and relationships between anxiety and neural responses were examined with correlational analyses.

Results: Participants were matched on sex, [X²(1,46)=1.19, p=.34], but not IQ [t(44)=-4.59, p<.01], or age [t(44)=2.24, p=.03]. ASD children had higher symptoms of general anxiety [M=66.62±13.67] than TD children [M=42.95±3.95], [t(31)=7.34, p<.01] and higher symptoms of social anxiety [M=57.92±13.43] compared to TD children [M=45.05±4.40], [t(31)=3.40, p<.01]. Significant negative relationships were found between general anxiety and P100 and N170 latency to stimuli across hemispheres and conditions in Experiment 1 [all r’s(20)≥-.59, p’s<.01] and Experiment 2 [all r’s(20)≥-.47, p’s<.04] in TD children only. There were no significant relationships between general anxiety and neural response in ASD children in either experiment. For ASD children, social anxiety was negatively correlated with N170 latency to fearful faces in the right hemisphere [r(13)=-.59, p=.03], and upright faces in the left hemisphere [r(17)=-.50, p=.04]. There were no significant relationships between social anxiety and neural responses in TD children in either experiment.

Conclusions: Results indicate that general anxiety, but not social anxiety, affect TD neural response to stimuli regardless of content or emotional valence. Faster activation in individuals with heightened anxiety symptomology may reflect a neural index of hypervigilance. A different pattern was observed for ASD children; only social anxiety impacted neural response to condition-specific stimuli. These findings suggest that increased anxiety specific to core deficits in ASD (i.e. processing social information) relates to faster neural processing when presented with 1) fearful compared to neutral faces and 2) upright faces compared to inverted faces and upright houses. Results indicate that potential ERP biomarkers were sensitive to the effect of anxiety on neural processing, and specific in distinguishing differences between diagnostic groups.