Predicting ASD and other developmental outcomes in the first year of life using EEG in a diverse community-based sample

点击翻译按钮获取中文摘要

Project Summary Children with autism who receive early intervention services have better outcomes than those who do not. It is therefore imperative to lower the age of diagnosis. There is strong evidence that there are reliable behavioral signs/symptoms of the disorder that emerge in the second year of life. However, there is mounting evidence from our laboratory that there are patterns in the EEG that emerge as early as 3 months that are reliably associated with autism outcomes at 2-3 years. In this proposal we seek to extend our previous work in two important ways. First, we will deploy our high-dimensional EEG data collection in a large pediatric primary care clinic, thus demonstrating the potential scalability of EEG as a biomarker of autism risk. Second, we will focus our efforts on a population of infants who have historically been underserved and understudied: primarily Black and Hispanic infants growing up in low-income homes. We will enroll 720 infants over 3 years (240/year), and based on previous work, anticipate a retention rate of 85%. We will collect resting EEG data at 4, 9 and 12 months in conjunction with their well-baby visits at the clinic. At 24 months diagnostic outcomes will be evaluated using the ADOS, developmental measures, and expert clinical judgement. In addition to the EEG assessment in the first year of life, a general developmental screener will be included (Ages and Stages Questionnaire-3) and indices associated with a number of non-genetic variables associated with increased autism risk (e.g., infant sex, parental age, prenatal maternal health, etc.) will be obtained from a demographic questionnaire and medical records. The specific aims of the project are: Aim 1: Using a prospective study design in a racially, ethnically and socioeconomically diverse primary care population, we will identify EEG features measured <1 year of life that are associated with ASD at 2-years of age. Aim 2: To develop predictive models with EEG biomarkers and other risk factors that reliably predict later diagnosis of ASD. Aim 3: To determine the specificity of predictive features for ASD versus other neurodevelopmental outcomes such as language or cognitive delays. Our ultimate goal is to create a scalable, practical, neurobiologically-based tool that can be readily integrated into a pediatric primary care setting, and in so doing, greatly improve our ability to identify autism in the first year. We believe our approach will allow us to demonstrate scalability of EEG in the primary care setting, develop usable models for children at greatest risk of delayed diagnosis, and improve our understanding of the underlying neural mechanisms of idiopathic autism.