Developing ERP Biomarkers of Social Cognition: The Journey from Task Design to Treatment Outcome Measures

Background: Successful development of effective treatments targeting social cognition in children with autism spectrum disorder (ASD) depends in part on the availability of objective outcome measures. Multidimensional approaches combining traditional behavioral assessments with psychophysiological measures yield more detailed information about treatment effects than standardized tests alone.

Objectives: The purpose of this series of studies was to develop and validate the nonverbal measure of social interest (incidental face memory) as a brain-based marker of social cognition suitable for use as a treatment outcome measure in children with autism.

Methods: Brain responses associated with spontaneous face perception and memory were evaluated using visual event-related potentials (ERPs) in a pre-post design in four groups of children with ASD (age 7-16 years, total n=90), including three treatment cohorts (peer-implemented social skills program). Additional data from typical controls (age 7-46, total n=37) and children with Prader-Willi syndrome (PWS; age 7-16 years, n=28) who demonstrate social deficits overlapping with ASD were used to further validate the proposed ERPs as a measure of social interest. All subjects viewed a series of color photographs depicting unfamiliar smiling young adults or house facades. One image of each type was randomly selected and repeated throughout the test session while the rest were shown only once. Social interest was quantified as the increased parietal ERP amplitude (300-600 ms) to repeated faces vs. those seen once. Behavioral measures of social cognition included NEPSY Memory for Faces test, caregiver reports (Social Communication Questionnaire, Social Responsiveness Scale, Child Behavior Checklist), and coded observations of playground behaviors with peers.

Results: All participants demonstrated perceptual discrimination of faces vs. houses as reflected in the larger N170 responses to the former. Increased parietal ERP responses to the repeated faces indicative of incidental memory traces were present in typical controls and in children with PWS with more typical social functioning, but not in children with ASD prior to treatment or in children with PWS with greater ASD-related symptomatology. Following the social skills intervention, children with ASD in the treatment group demonstrated increased ERP amplitudes to repeated faces compared to the baseline assessment, while the waitlist control group showed no change. This pattern of results was replicated in each of the three treatment cohorts. The ERP index of social interest did not correlate with IQ or ADOS severity scores, but showed associations with behavioral measures of social skills and behavior.

Conclusions: ERP-based measure of social cognition obtained in the incidental memory paradigm is a promising treatment outcome measure. It fits the biomarker definition as an objective and reproducible measure of neural activity differentiating typical and atypical social information processing. The replicated evidence of its sensitivity to treatment effects, good test-retest stability in the control groups, correlations with behavioral endpoints, and reduced susceptibility to the placebo effects due to the passive nature of the task further support its value as a treatment outcome measure.