Heterogeneous decision-making in ADHD: computational subtypes and their distinct prefrontal connectivity patterns

Children with ADHD exhibit decision-making deficits characterized by lower efficiency in evidence accumulation and a preference for smaller, sooner rewards in intertemporal choices. This study aimed to preliminarily explore computational subtypes of ADHD in children and to investigate differences in prefrontal functional connectivity between ADHD subtypes. A total of 94 children with ADHD and 63 typically developing children matched for sex, age, and IQ were included. All participants completed the Two Choice Impulsivity Paradigm (TCIP) and the ADHD Symptom Rating Scale, and 43 children with ADHD underwent resting-state fMRI. A hierarchical drift diffusion model was applied to analyze TCIP task performance, followed by two-step cluster analysis to identify decision-making subtypes. Differences in prefrontal functional connectivity among ADHD decision-making subtypes were then compared. The ADHD group showed fewer square selections in the TCIP (p < 0.001), lower drift rates (pBayes < 0.001), and shorter non-decision times (pBayes = 0.002). Two decision-making subtypes were identified: a decision-making deficit subtype and a decision-making normal subtype, characterized by lower and higher drift rates, respectively. The ADHD decision-making deficit subtype exhibited stronger functional connectivity between the left medial superior frontal gyrus and both the right middle frontal gyrus and the right postcentral gyrus, as well as between the right medial superior frontal gyrus and the right postcentral gyrus (all p < 0.05). The results reveal decision-making deficits and internal heterogeneity in children with ADHD during intertemporal choice, with this heterogeneity possibly associated with abnormal functional connectivity between the medial superior frontal gyrus and the middle frontal gyrus/postcentral gyrus.