Learning-based cognitive control in ADHD: a multicentric study

Background: Learning-based cognitive control (CC), the ability to adapt control strategies based on contextual regularities, has been studied in typically developing (TD) children but remains underexplored in children with attention deficit hyperactivity disorder (ADHD). This study investigates whether children with ADHD show intact learning-based CC and how it is affected by increased cognitive demands. Methods: In a multicentric study, 7–14 years old children (145 ADHD [20F, mean = 10.4 ± 2.0], 97 TD [51F, mean = 10.4 ± 2.0]) completed two experimental tasks: a modified Flanker task and a cued go-noGo task. Both included a List-Wide Proportion Congruency (LWPC) manipulation to create different contextual predictability (mostly congruent/valid vs. half congruent/valid blocks). Reaction times (RTs) and accuracy were analyzed across blocks, and we also examined associations between task adaptation and neuropsychological (NPS) as well as questionnaire-based measures of CC, including inhibition, cognitive flexibility, working memory, and behavioral difficulties. Results: Children with ADHD showed similar LWPC effects to TD peers in the Flanker task, suggesting preserved learning-based CC in low-demand contexts. In the cued go-noGo task, which involved greater attentional and inhibitory demands, children with ADHD displayed less efficient modulation of RTs and reduced speed-accuracy trade-off adaptation compared to TD peers. A composite NPS score predicted task adaptation. No associations were found with parental questionnaires. Conclusions: Learning-based CC appears preserved in children with ADHD in simple contexts but is hindered under complex, multi-demand conditions. Neuropsychological performance may be a key mechanism underlying these group differences. These findings emphasize the need for ecologically valid paradigms and tailored interventions targeting complex CC challenges in ADHD.