To validate a residual-based cognitive reserve (CR) index optimized for a pediatric sample with attention-deficit/hyperactivity disorder (ADHD). Participants were N = 115 children aged 9.5-13 years at baseline (Mage = 10.48 years, SDage = 0.61), and n = 43 (37.4%) met criteria for ADHD. Elastic-net regularized linear regression was used to generate baseline and longitudinal CR indices by maximally residualizing variance in fluid intelligence for demographics and brain structure. Academic and diagnostic outcomes were regressed onto CR indices, and interactions with brain integrity were assessed. Baseline CR predicted baseline math computation (estimate = 0.10, SE = 0.02, p < .001), while change in CR predicted change in word reading ability (estimate = 0.08, SE = 0.02, p < .001). Further, when grey matter volume tended to be lower, higher CR was associated with higher word reading score (estimate = -0.05, SE = 0.02, p = .019) and lower ADHD symptom severity (estimate = 0.04, SE = 0.02, p = .047) compared to lower CR, at baseline. Similarly, when longitudinal change in white matter hypointensity volume tended to be greater, higher change in CR resulted in more favorable word reading trajectory (estimate = 0.03, SE = 0.02, p = .048). A data-driven residual approach to operationalizing pediatric CR shows better evidence of construct validity over previous methods, with our index showing a novel ability to moderate the deleterious effects of lower grey matter on outcomes in ADHD. This approach may benefit future research aiming to study the early development of CR. (PsycInfo Database Record (c) 2024 APA, all rights reserved).