Sex Differences and Executive Function Profiles in Childhood Attention-Deficit/Hyperactivity Disorder: Evidence from...

To test sex differences and executive-function (EF) heterogeneity in childhood attention-deficit/hyperactivity disorder (ADHD) using a standardized digital battery.

We studied 208 clinically referred children aged 7-12 years with ADHD diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5). CNS vital signs yielded 11 domains overall; all prespecified inferential analyses used 10 domains (Composite Memory excluded), with age as covariate and false-discovery-rate control (q = 0.05). Frequentist analysis of covariance were complemented by Bayesian information criterion (BIC)-based Bayes factors and two one-sided tests equivalence (|d| < 0.30). Heterogeneity was examined via k-means clustering on z-scored profiles (complete-case n = 170) with validation by silhouette and Adjusted Rand Index from hierarchical clustering.

The multivariate sex effect was not significant (Pillai's trace = 0.04, p = .74). After false discovery rate (FDR) control, only Psychomotor Speed and Motor Speed differed, favoring boys (Psychomotor Speed: F = 11.43, q = 0.005, η2p = 0.053, d = 0.48; Motor Speed: F = 21.82, q < 0.001, η2p = 0.097, d = 0.66). For Complex Attention, Cognitive Flexibility, Executive Function, Simple Attention, Processing Speed, Reaction Time, Verbal Memory, and Visual Memory, between-sex differences were not reliable and Bayesian evidence typically supported similarity (BF₀₁ ranged from 3 to 13). Clustering produced three severity-graded EF profiles (n = 48/39/83; average silhouette = 0.17) with moderate cross-method agreement (Adjusted Rand Index = 0.55) and no association with sex or DSM-5 presentation. Sensitivity analyses (additional covariates, exclusions, construct-level indices, and 11-domain clustering) were convergent.

Boys and girls with ADHD show largely similar EF architectures, with small speed-specific differences favoring boys. EF heterogeneity is captured by three severity-graded profiles that transcend sex and DSM-5 presentations, supporting dimensional, performance-based assessment.