Toward Scalable Biomarker-Based Prediction of ASD in High-Risk Infants

点击翻译按钮获取中文摘要

ABSTRACT Despite tremendous effort by parents, researchers, clinicians, and educators, autism spectrum disorder (ASD) continues to present a significant, lifelong challenge to most affected individuals and their families. Studies of early development in infants at risk for ASD (such as infants with older siblings with ASD: “HR infant siblings” – who have a ~20% chance of developing ASD) can identify early, presymptomatic predictors of ASD that can then improve early screening and promote presymptomatic intervention. Behavioral studies of these HR infant siblings have identified atypical behaviors in the second year of life in the social domain, with some evidence of motor delays and differences in social attention within the first year. However, in part because of the limited behavioral repertoire of infants, investigators have struggled to identify consistent first-year-of-life behaviors that predict later ASD in a clinically actionable manner. We propose that the earliest measures of atypical development should directly assay brain function. The Infant Brain Imaging Study (IBIS) Network has used MRI methods to reveal functional and structural brain changes in the first year of life in HR infant siblings. These brain changes are present prior to the emergence of behavioral features of ASD and accurately predict ASD at 24 months of age (positive predictive value >= 80%). While scientifically promising, MRI's cost and reduced availability limit its potential scalability for use in HR infants to use as a general population screener in clinical settings. Electroencephalography (EEG) and eye tracking (ET) represent two scalable methods that can measure brain function and can help to identify predictive biomarkers of ASD in early infancy. EEG and ET are developmentally sensitive and accessible in community, real-world settings. In spring 2019, the Infant Brain Imaging Study (IBIS) Network will launch a new study of 250 HR infants designed to replicate and extend its predictive MRI findings. Here, we propose to add EEG and ET measures of brain function to this study, testing HR infants from IBIS at 6 and 12 months of age, and assessing clinical outcomes at 24 months of age with clinical outcomes assessed at 24 months of age. We will examine brain network function at rest, during low level sensory processing, and during social information processing. Our hypothesis is that these scalable EEG/ET biomarkers will (Aim 1) accurately identify infants with a later diagnosis of ASD and will (Aim 2) relate to ASD-associated behaviors at 24 months of age. Capitalizing on this unprecedented opportunity to integrate EEG/ET with neuroimaging in the same cohort of infants, in Aim 3 we also propose to explore the predictive power of these combined measures, and the association between EEG/ET and MRI features. The overarching goal of this proposal is to lower the age of detection of autism to early infancy, making intervention before the emergence of ASD-specific behavioral features feasible and more effective. Positive findings in the proposed study will also facilitate the future extension of presymptomatic predictive testing from HR infants to infants in the general population.