The Spectrum within the Spectrum: Neural Heterogeneity in Autism

Autism Spectrum Disorder (ASD) encompasses diverse behavioral, cognitive, and neural profiles that vary widely across individuals. Because diagnosis remains predominantly behavior-based, clinically meaningful within-spectrum differences are often missed. Recent advances in neuroimaging analysis have created opportunities for the identification of distinct neural profiles within ASD. These new avenues provide ways to examine intra-ASD heterogeneity directly, identifying person-specific brain-behavior patterns that group-level comparisons fail to capture. The present systematic review focused on neuroimaging methods (i.e., resting-state fMRI; rs-fMRI, cortical morphology) that emphasized heterogeneity-informed outcomes based on normative modeling, data-driven subtyping, clustering analysis and similar analytic choices. For rs-fMRI, findings showed that differences in connectivity within the default mode and salience networks were linked to social-communication and language impairments, with subtype-specific patterns emerging across samples. In cortical thickness and surface area, studies identified distinct developmental trajectories and individual deviation profiles, with prefrontal and temporal regions often tied to repetitive behaviors and social-communication difficulties. Voxel-based analyses revealed subgroup-specific gray matter alterations, including cerebellar and cortical differences, that aligned with variation in repetitive behaviors and social difficulties. Collectively, this emerging evidence supports the notion that neural heterogeneity in autism is a core feature rather than a limitation. Embracing individual variability can provide crucial insight into the inconsistencies and contradictions of earlier case-control and group-level findings. Moving forward, autism research must prioritize heterogeneity-based methods to establish reproducible subtypes and individualized brain-behavior links.