Predicting ASD Outcomes Using Quantitative Movement Metrics in Infancy

About This Grant

Project Abstract Motor impairments are one of the first signs of atypical development in infants who go on to have autism spectrum disorder (ASD) and are more closely tied to abnormal neurobiological processes in ASD. Motor behavior has high potential to serve as a measurable early behavioral difference to advance early detection of ASD—a necessary step for access to earlier and more effective interventions. However, standardized assessments of infant motor function have not been able to capture motor behaviors that are specific to ASD due to their focus on categorical ratings of milestone-attainment that are present across many neurodevelopmental conditions. There is a critical need for methods that objectively capture deeper aspects of underlying infant movements that are more sensitive and specific to ASD outcomes. The current study aims to use wearable sensor technology and advanced computational techniques to develop and validate objective, specific, reliable, and scalable measures of infant motor function that serve as predictive biomarkers of ASD, with the ultimate goal of advancing earlier detection and intervention that can improve long-term outcomes in ASD. The proposed study will use wearable sensors to measure motor development in the first year of life in infants at increased likelihood for ASD (ILA, defined as having an older sibling with ASD). Our team’s preliminary data strongly supports a theoretical model that lower sensor-based quantitative measures of infant movement symmetry and variability are specifically associated with later ASD outcomes, and that lower infant movement variability in the first year of life is associated with later forming repetitive motor behaviors (RMB) seen in ASD. We will enroll 120 ILA infants and examine an external validation cohort of infants. Infants will be assessed at 3, 6, 9, and 12 months of age with wearable sensors worn on bilateral upper and lower extremities and with standardized motor and behavioral assessments. Behavioral measures of ASD symptoms and developmental level will occur at 12 and 24 months. The assessments from 3-12 months will occur in the infant’s home, capture ecologically valid movement data, and remove barriers for participation for rural and underserved populations that cannot easily access major academic areas. We will apply sophisticated signal processing and machine learning techniques on the multidimensional quantitative infant movement data collected to: (1) validate our existing quantitative measures of infant movement variability (complexity and curvature) and symmetry; (2) create and validate new quantitative measures that improve detection of atypical movement characteristics across different developmental stages; and (3) advance early detection of ASD by creating prediction models that include the quantitative measures of infant movements and measures of other atypical behaviors associated with ASD. We will further examine the performance of our prediction models in briefer time subsets that mirror pediatric well child visits. This study has great potential to advance our understanding of motor impairments in ASD and drive a paradigm shift in scalable ASD identification in the first year of life.