# Computational Modeling-Informed Reward Subgroups in Adolescent ADHD > **NIH NIH R01** · HARTFORD HOSPITAL · 2022 · $680,965 ## Abstract Although there is notable evidence for ADHD abnormalities in the brain's reward system, that evidence is limited due to both its narrow focus on a limited number of reward constructs, as well as its inconsistencies across studies. We believe the inconsistencies are the result of unrecognized ADHD neurobiological heterogeneity. This project was designed to test the prediction that ADHD is a disorder where several, wholly different types of neurobiological dysfunction are capable of producing the same ADHD diagnostic phenotype. Clear, decisive evidence for such neurocognitively and neurobiologically distinct ADHD subgroups is needed to support an emerging paradigm shift in ADHD neuroscience away from the assumption that every ADHD patient has similar pathophysiology, to a multi-etiology model that ultimately should prove to have greater translational usefulness. Here, we will focus on better understanding reward dysfunction in ADHD, which is relatively under-studied compared to the other neurocognitive abnormalities often found in the disorder. We propose to examine a large (n=200) sample of ADHD-diagnosed and non-ADHD (n=150) adolescents with an ambitious reward-based phenotyping approach using a diverse reward test battery to assess numerous domains of reward behavior (e.g., different types of reinforcement learning, valuation, cost processing, effort expenditure, etc.) – an approach that proved highly successful in our preliminary ADHD study. This battery also will leverage the advances in reward phenotyping made over the past decade by sophisticated computational modeling of reward choice behavior and reaction time data. We then will map aspects of individual differences in these abilities to frontostriatal network connectivity during reinforcement learning `prediction errors.' Our ultimate goal is to use this information to classify ADHD patients into different biotypes. Although our preliminary data suggest there likely are at least two different reward-impaired ADHD subgroups with disparate profiles of reward dysfunction, our approach will be more rigorous than a simple replication attempt. We will test the fit of broad conceptual models of ADHD neurocognitive abnormality for the first time in the reward domain, then use those findings to refine our approach to biotyping individual cases with rigorous classification methodology. The resulting biotypes will be validated using other fMRI reward tasks. This project will create the most detailed and extensive database describing reward dysfunction in ADHD to date, which we will make freely available to the scientific community to accelerate the pace of discovery. A renewed, concentrated focus on reward dysfunction in ADHD is not just timely, but also likely to set the stage for important advances in future etiological and translational research. Different types of ADHD reward dysfunction could represent untapped new targets for novel intervention development, where treatments are matched to the ... ## Key facts - **NIH application ID:** 10322181 - **Project number:** 5R01MH119815-03 - **Recipient organization:** HARTFORD HOSPITAL - **Principal Investigator:** Michael C Stevens - **Activity code:** R01 (R01, R21, SBIR, etc.) - **Funding institute:** NIH - **Fiscal year:** 2022 - **Award amount:** $680,965 - **Award type:** 5 - **Project period:** 2020-03-15 → 2024-12-31 ## Primary source NIH RePORTER: https://reporter.nih.gov/project-details/10322181 ## Citation > US National Institutes of Health, RePORTER application 10322181, Computational Modeling-Informed Reward Subgroups in Adolescent ADHD (5R01MH119815-03). Retrieved via AI Analytics 2026-05-22 from https://api.ai-analytics.org/grant/nih/10322181. Licensed CC0. --- *[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*