Age-Adjusted Functional Connectivity Alteration in ADHD from Resting-State fMRI: A Covariate-Modulated Graphical Model Approach

Brain functional connectivity, measured through resting-state fMRI, can reveal how neurological conditions alter neural communication patterns. However, most current approaches assume connectivity remains constant across individuals and contexts, ignoring critical factors as age and clinical diagnosis. We apply a novel covariate-modulated functional graphical modeling framework that explicitly models how multiple covariates simultaneously influence brain connectivity to ADHD-200 Pre-processed data (NYU site; AAL-116), jointly modeling ADHD diagnosis and age. The resulting ADHD modulation graph shows predominantly attenuated connections, with prominent differences involving cerebellar and fronto-striatal circuits. This approach provides more accurate differ-ential connectivity estimates by jointly modeling factors of interest and confounders, yielding magnitude-enriched co variate-specific modulations that offer new insights into how clinical conditions reshape neural organization.