Transdiagnostic monoamine-based subtyping for attention- deficit/hyperactivity disorder and autism spectrum disorder via unsupervised machine learning

Neuroimaging and molecular studies have attempted to elucidate the etiology of attention-deficit/hyperactivity disorder (ADHD) and autism spectrum disorder (ASD). However, their findings have been inconsistent because of within-disorder heterogeneity and cross-disorder phenotypic overlap. We sought to identify monoamine-based subtypes across ADHD and ASD and to clarify their distinct brain structural characteristics. In 83 children with ADHD and/or ASD, we applied unsupervised machine learning (NbClust with K-means) to identify novel neurodevelopmental disorder (NDD) phenotypes using urinary monoamine metabolite profiles. Behavioral symptoms, cognitive performance, cortical surface area, and gray matter volume (GMV) were then evaluated for each NDD phenotype and for 83 children with typical development (TD). Clustering identified two NDD phenotypes: NDD-A (n = 18, characterized by high levels of 4-hydroxy-3-methoxyphenylglycol, 5-hydroxyindoleacetic acid, and homovanillic acid) and NDD-B (n = 65, characterized by low levels of these monoamine metabolites). Moreover, urinary 4-hydroxy-3-methoxyphenylglycol levels correlated positively with social communication difficulties in NDD-A. Behaviorally, the NDD-B group showed significantly lower levels of cognitive control, cognitive flexibility, and inhibitory control than did the TD group. Structurally, compared to the TD group, the NDD-A group showed significant surface area enlargement in the isthmus cingulate gyrus, whereas the NDD-B group exhibited significant GMV reductions in the frontal lobe, precuneus, posterior cingulate cortex, and superior occipital area. These findings suggest that monoaminergic hyper- or hypofunction in individuals with NDD-A and NDD-B is associated with distinct brain structural differences. Such phenotype specificity may provide a novel framework for understanding within-disorder heterogeneity and cross-disorder phenotypic overlaps.