ABSTRACT. There is a general consensus that substantial heterogeneity underlies the neurobiology of autism spectrum disorder (ASD). As such, it has become increasingly clear that a dissection of variation at the molecular, cellular, and brain-network domains is a prerequisite for identifying biomarkers. Neuroimaging has been widely used to characterize atypical brain patterns in ASD, though findings have varied across studies. This is, at least in part, due to a failure to account for neurobiological heterogeneity. Here, we summarize emerging data-driven efforts to delineate more homogenous ASD subgroups at the level of brain structure and function - i.e., neurosubtyping. We break this pursuit into key methodological steps: the selection of diagnostic samples, neuroimaging features, algorithm and validation approaches. Although preliminary and methodologically diverse, current studies generally agree that at least 2-4 distinct ASD neurosubtypes may exist. Their identification improved symptom prediction and diagnostic label accuracy above and beyond group-average comparisons. Yet, this nascent literature has shed light onto challenges and gaps. These include 1) the need for wider and more deeply transdiagnostic samples collected while minimizing artifacts (e.g., head motion), 2) quantitative and unbiased methods for feature selection and multimodal fusion, 3) greater emphasis on algorithms’ ability to capture hybrid dimensional and categorical models of ASD, and 4) systematic independent replications and validations that integrate different units of analyses across multiple scales. Solutions aimed to address these challenges and gaps are discussed for future avenues leading towards a comprehensive understanding of the mechanisms underlying ASD heterogeneity.
Jump to: Navigation