Project Summary/Abstract Despite the annual $268 billion cost of ASD in the U.S. and the tens of millions spent annually on research, “precision” medicine does not exist in any meaningful way for ASD infants and toddlers. The heterogeneity of early neural and behavioral developmental trajectories in ASD has stymied the search for explanations, and the identification of clinically useful biomarkers of prognosis as well as the discovery of biotargets that could be used to develop maximally effective treatments. In our proposed studies, 175 ASD, typical, language delayed (LD) and global developmental delayed (GDD) toddlers will participate in a series of language- relevant (nursery rhymes vs music) and social emotion fMRI paradigms (own mother’s voice vs stranger’s) as well as resting state connectivity paradigms to begin to address this major gap in the field. Toddlers will be recruited using our novel general population based screening approach that provides unique and complementary data to those from baby sibling studies. In order to generate a rich clinical profile of each toddler, multiple language and social measures will be taken, including CELF-R, Mullen and Vineland. In order to examine change and leverage powerful longitudinal modeling approaches, toddlers will be clinically assessed and imaged at both 1-2 and 3-4 years. State-of-the-field MEMB and ME-ICA denoising approaches will be utilized that yield highly reliable high signal-to-noise functional imaging that outperforms previous fMRI approaches and enhances effect size estimates and statistical power; this greatly benefits robustness in our analyses, reliability, split sample feasibility, and exploratory prognostic biomarker modeling. Multiple analytic methods (e.g., PLS, seed-PLS, ICA, spectral DCM, PPI) will be applied to identify brain-language and brain- social emotion relationships; model neural and clinical trajectories from 1-2 to 3-4 years; reveal language- and social emotion-relevant fMRI activation and connectivity patterns at 1-2 years that are predictive of language and social outcomes; discover underlying neural-clinical subtypes; model continuous variation in still other neural measures that predict continuous language and social measures; identify fMRI-MRI relationships; define how language and social emotion neural deficits tap into shared neural network resources in early development; and examine similarities and differences in brain-behavioral relationships across multiple groups which then allows for sensitive tests of whether brain-behavioral patterns are common across diagnostic boundaries (e.g., LD, GDD and ASD poor language toddlers in an RDoC fashion) or specific to a subgroup of individuals. Our studies will identify clinically meaningful early-age neural biomarkers that predict which ASD children will go on to have good language outcomes and which poor ones, and others that predict social outcome. Compelling ASD language and social biotargets will be found th...