Abstract
Neuroimaging shows volumetric alterations of gray matter in attention-deficit hyperactivity disorder (ADHD); however, results are conflicting. This may be due to heterogeneous phenotypic sampling and limited sensitivity of volumetric analysis. Creating more homogenous cohorts and investigating gray matter microstructure may yield meaningful biomarkers for scientific and clinical applications. Children with sensory processing dysfunction (SPD) have differences in white matter microstructure, but not gray matter volumetric differences. Approximately 40% of SPD children meet research criteria for ADHD. We apply deep learning segmentation of MRI to measure gray matter volume (GMV) and density (GMD) in SPD children with (SPD+ADHD) and without co-morbid ADHD (SPD-ADHD). We hypothesize GMV and GMD are linked to ADHD but with sex-specific regional patterns. We find boys with SPD+ADHD have widespread reduction of GMD in neocortex, limbic cortex, and cerebellum versus boys with SPD-ADHD. The greatest differences are in sensory cortex with less involvement of prefrontal regions associated with attention networks and impulse control. In contrast, changes of ADHD in girls with SPD are in brainstem nuclei responsible for dopaminergic, noradrenergic, and serotonergic neurotransmission. Hence, neural correlates of ADHD in SPD are sexually dimorphic. In boys, ADHD may result from downstream effects of abnormal sensory cortical development.