Brain imaging research strongly suggests that autism spectrum disorder (ASD) involves altered function in multiple neural networks. Neuroanatomical and neurofunctional findings in ASD support changes in brain development and brain connectivity. Neural oscillations reflect organized cycling of excitability within neural ensembles, and communication across distributed cortical regions can be measured in their synchronized neuro-oscillatory activity. Mounting evidence for disordered neuro-oscillatory activity suggests a sensitive and direct assay of altered information processing and communication in ASD within and across neural networks that has the potential to be applicable to any age and functioning level, to be informative with regard to underlying mechanisms and translatable to animal models of ASD.

Using high-density electroencephalographic (EEG) recordings of human brain activity, this program of research aims to investigate local and network neuro-oscillatory function in children with ASD. Although previous findings clearly endorse the potential of such measures, a comprehensive comparative approach is needed to realize their promise as biomarkers of altered neural processing.

Specifically, Sophie Molholm and colleagues plan to record EEG data under a range of conditions (e.g., during motor activity, sensory stimulation, cognitive task performance, social cognition and rest) and address specificity of dysfunctional local and network neuro-oscillatory processes. They will assess the relative strength of potential biomarkers with machine learning and graph theory-based network analysis to determine the most informative combination of test conditions and EEG features to predict diagnosis and dimensional scores on ASD-relevant traits/behaviors as measured by standardized scales.

Children with ASD will be compared to age- and sex- matched typically developing children. An unaffected ASD-sibling cohort will also be included as part of the study design to address whether observed differences in oscillatory measures of neural information processing are likely related to heritable mechanisms underlying the development of ASD, or if, alternatively, they are more likely the consequence of frank disease expression. Although this work will be performed in higher-functioning children, Molholm and colleagues will emphasize approaches that are translatable to infants, more severe (e.g., syndromic) forms of ASD and animal models.

The project aims to: (1) test the hypothesis that neuro-oscillatory function is affected in children with ASD compared to typically developing children; (2) determine the test conditions and neural measures most informative with regard to the ASD phenotype, using machine learning and graph theory based network analysis; and (3) test the hypothesis that neuro-oscillatory function during social and motor processing, which have been identified as atypical in infants with heightened risk for ASD because they have a sibling with ASD, are endophenotypic.