Autism is an early developing condition diagnosed on the basis of difficulties in the domains of social-communication and restricted repetitive behaviors. However, the single label of ‘autism’ masks a large degree of heterogeneity at multiple scales, from genome to phenome, and may not signal differential needs and services required for some types of profoundly affected individuals. In order to better understand autism and develop more impactful ways to facilitate positive outcomes at an individualized-level, we need better models that help to explain how autism may be differentiated in meaningful ways that help explain variability in developmental paths, responses to treatment, and differential underlying neurobiological mechanisms. This research program develops and tests the AUTISMS-3D (A3D) model – disability versus difference over development - as a first-level major distinction for explaining heterogeneity within the autism population. AUTISMS-3D postulates that the major first-level split within the autism population is between subtypes described by profound disability versus difference in developmental outcomes. Stratification models using standardized clinical measures of intellectual functioning, language, and adaptive behavior will be used to predict ‘disability’ versus ‘difference’ subtypes. These stratification models are based on highly reproducible data-driven clustering analyses from very large phenotypic datasets. We then test how the subtypes may develop differently with analyses of big clinical longitudinal datasets and also within the context of early intervention. Finally, we test the subtypes for neurobiological differences via neuroimaging (MRI, fMRI, DTI) and electrophysiological (EEG) and gaze phenotypes (e.g. eye tracking) under static and naturalistic viewing conditions. The research will help refine our understanding and concepts of autism and will be an important next step on the path to precision medicine for the autisms.
