Spatiotemporal brain-state dynamics delineate executive function subtypes in school-aged autism: evidence from co-activation patterns and a four-year follow-up.
TL;DR
Findings provide evidence for neurodynamic heterogeneity in ASD, suggesting that similar clinical outcomes may be associated with divergent brain-state profiles and supports the move toward neuro-subtype-informed precision stratification and targeted intervention strategies.
Abstract
Background
Autism spectrum disorder (ASD) is characterized by profound clinical and biological heterogeneity. The neurodynamic profiles associated with divergent developmental trajectories of executive function (EF) during the critical transition from late childhood to early adolescence remain poorly understood. This study aimed to determine whether heterogeneity in EF development is associated with distinct neurodynamic profiles.
Methods
In a longitudinal study, 68 children with ASD (aged 6-9 years) and 50 age-matched typically developing (TD) controls underwent baseline resting-state fMRI. The ASD group was followed for approximately 4 years. EF was assessed using the Behavior Rating Inventory of Executive Function (BRIEF), alongside follow-up depression, anxiety, and sleep outcomes. Longitudinal EF trajectory clusters were identified to characterize developmental heterogeneity. In parallel, baseline neurofunctional subtypes were derived from ALFF using a normative-deviation framework, with fALFF used for sensitivity analysis. CAP analysis was then applied to examine brain-state dynamics across these complementary stratification approaches.
Results
EF declined longitudinally in the ASD group, particularly in behavioral regulation domains, and these changes were associated with depressive symptoms and sleep problems at follow-up. Three longitudinal EF trajectory clusters were identified, but baseline CAP dynamics showed minimal differences across these groups. In contrast, ALFF-derived neurofunctional subtypes exhibited distinct CAP dynamic profiles, with Subtype 1 showing greater engagement of visual-related states and Subtype 2 exhibiting enhanced transitions among DMN/FPN-related control states. fALFF-based analyses yielded similar subtype assignments, supporting the robustness of the neurofunctional stratification. Critically, similar EF deterioration was associated with distinct neurodynamic profiles, as reflected by subtype-specific state-transition patterns that showed opposite associations with EF changes.
Limitations
First, the TD group was not followed longitudinally, limiting precise quantification of deviation from normative developmental pathways. Second, the sample size and attrition during follow-up may affect the stability of subtype assignment. Finally, as inferences were based on resting-state fMRI, task-based or ecologically valid measures were not available to validate functional interpretations.
Conclusion
These findings provide evidence for neurodynamic heterogeneity in ASD, suggesting that similar clinical outcomes may be associated with divergent brain-state profiles. This work supports the move toward neuro-subtype-informed precision stratification and targeted intervention strategies.