Establishing Specificity and Scalability of Motor Imitation Assessments for Autism Spectrum Disorder

About This Grant

PROJECT SUMMARY AND ABSTRACT Autism spectrum disorder (ASD), a lifelong, neurodevelopmental condition defined by core behavioral symptoms encompassing social functioning, communication, restricted interests and repetitive behaviors, affects an estimated 1/36 children in the United States at a cost of $37B annually. Wide variations in ASD presentation and comorbidities affecting more than 50% of autistic individuals substantially impact diagnosis, and the planning and delivery of clinical care. Autistic children with more complex presentations due to medical and neurodevelopmental comorbidities (e.g., attention deficit hyperactivity disorder; ADHD, and/or developmental language disorder; DLD) can experience greater diagnostic and treatment delays, increasing the risk for poor developmental outcomes and quality of life. We propose to establish specificity and scalability of a promising biomarker for autism spectrum disorder (ASD) by combining innovative wearable high-density diffuse optical tomography (HD-DOT) functional neuroimaging with a Computerized Assessment of Motor Imitation (CAMI) developed by our team. Many fundamental social and communicative skills are learned through imitation and impaired imitation likely contributes to core difficulties in ASD. While movement difficulties are associated with several developmental conditions that commonly co-occur with ASD, including ADHD and DLD, deficits in motor imitation may help distinguish these overlapping conditions and capture variation relevant to underlying biology of ASD. Our team has pioneered an automated CAMI, using a brief, engaging task that can be readily scaled to clinic and home settings, with published studies showing CAMI has robust diagnostic discrimination for identifying ASD as compared with neurotypical children (NTC). However, the specificity of imitation deficits and corresponding brain mechanisms for ASD relative to other neurodevelopmental conditions has had limited study. Addressing these gaps, we propose to to establish specificity of impaired motor imitation to ASD through simultaneous assessment of motion imitation fidelity via CAMI and brain function/connectivity via HD-DOT in school-age children (6-10 years) with ASD, ADHD, DLD, and NTC. Crucially, as part of Aim 1, we will collect CAMI data using both 3D depth cameras and more readily available 2D cameras and use machine learning to build multimodal representations of movements and develop metric learning techniques to quantify motor imitation from 2D camera data alone. Additionally, we will identify brain-behavior relationships between cortical activity and connectivity with motor imitation fidelity and relate them to both diagnostic and transdiagnostic traits (Aim 2). Finally, we will extend the motor imitation brain-behavior assessments to preschool age children (3-5 years) with and without ASD (Aim 3). Across all Aims, we will assess specificity by examining associations of CAMI and HD-DOT measures with diagnoses as well as established dimensional measures of ASD, ADHD, and DLD. Our proposed study has substantial potential to profoundly improve predictive diagnostic utility over current subjective clinical assessments and thereby aid public health efforts to identify and support affected children.