Mapping ADHD Heterogeneity and Biotypes by Topological Deviations in Morphometric Similarity Networks

Area of Science:

• Neuroscience
• Psychiatry
• Developmental Psychology

Background:

• Attention-deficit/hyperactivity disorder (ADHD) presents significant clinical heterogeneity, challenging current classification systems.
• Existing frameworks limit the development of neurobiologically informed subtyping approaches for ADHD.

Purpose of the Study:

• To investigate if normative modeling of brain morphometry similarity networks can identify robust stratification markers for ADHD in children.
• To explore novel subtyping approaches for ADHD based on neurobiological characteristics.

Main Methods:

• Utilized multisite, cross-sectional neuroimaging data with longitudinal cognitive assessments.
• Constructed morphometric similarity networks and applied normative modeling to topological metrics (degree centrality, nodal efficiency, participation coefficient).
• Employed semisupervised clustering to delineate ADHD biotypes and examined their clinical and neural profiles, including neurochemical and functional correlates, with external validation.

Main Results:

• ADHD cases showed atypical brain network organization compared to controls, particularly in the orbitofrontal cortex.
• Identified three distinct ADHD biotypes: severe-combined with emotional dysregulation, predominantly hyperactive/impulsive, and predominantly inattentive.
• Each biotype exhibited unique clinical features, longitudinal trajectories, and distinct neurobiological underpinnings, with findings replicated in an independent cohort.

Conclusions:

• Normative modeling combined with semisupervised clustering offers valuable insights into ADHD heterogeneity.
• The identified ADHD biotypes possess distinct clinical-neural profiles, advancing the understanding of ADHD's neurobiological complexity.
• These findings provide a foundation for developing personalized management strategies for ADHD.