Morphometric Latent Factors in Autism and Their Association with Receptor Profiles and Behavior.

The profound heterogeneity of Autism Spectrum Disorder (ASD) is a major barrier to developing targeted therapies. While dimensional subtyping using functional connectivity (FC) has advanced the field, the intrinsic instability of FC limits its power to identify stable trait-like biomarkers. This study provides a direct comparison of latent factors derived from both FC and a stable morphometric measure, Morphometric Inverse Divergence (MIND), within the same ASD cohort. We hypothesized that stable structural factors would provide a more behaviorally relevant and biologically grounded account of ASD heterogeneity. Our findings reveal an important dissociation between functional and morphometric features. Latent factors derived from morphometric similarity (MIND) significantly correlated with core ASD behavioral traits (SCQ, SRS), while functional factors showed no association. This dissociation was also found at the neural level. Specifically, we observed weaker structure-function correspondence in the ASD group compared to the healthy control group (HC). Moreover, the stable structural factors were linked to trait-like neurodevelopmental mechanisms: the association with the CB1 receptor (important for synaptic pruning) observed in the HC was notably missing in the ASD group. Conversely, the functional factors were associated with a state-like arousal system (the norepinephrine transporter, NET). Collectively, our results demonstrate that stable morphometric-based factors, rather than time-varying functional ones, are predictive of behavioral traits in ASD. This work validates MIND as a robust approach and suggests that the link between structural organization and its neurodevelopmental (CB1) underpinnings is a more powerful and stable target for developing biomarkers in autism.