PROJECT SUMMARY/ABSTRACT The proposed K01 project will use multimodal magnetic resonance imaging (MRI) and machine learning (ML) to elucidate the neurocognitive processes underlying treatment failure in young adults with opioid use disorder (OUD). Young adults are at particularly high risk of OUD and fatal opioid overdose. The monthly injectable extended-release opioid antagonist naltrexone (XR-NTX) is a highly effective OUD treatment and is particularly well suited for young adults. However, XR-NTX adherence and relapse show considerable individual variability, and the behavioral and clinical factors associated with such variability remain inconclusive. Previous research has demonstrated the potential for multimodal MRI and ML techniques to elucidate the neurocognitive factors that contribute to treatment response beyond behavioral and clinical measures. This project will take advantage of the cutting-edge MRI and ML methods to model brain structures and functions that predict XR-NTX treatment outcomes in young adults with OUD. The study will evaluate 18–34 year-old OUD patients before and during the first three months of XR-NTX treatment, a period associated with the highest rate of dropout from treatment. The primary outcome will be opioid relapse confirmed by weekly urine toxicology and self-report. The secondary outcome will be non-adherence defined as failure to complete the first three injections. The study will focus on five baseline measures of brain structures and functions that are potentially predictive of treatment response: 1) grey matter volume; 2) functional connectivity with the ventral striatum; 3) reactivity to opioid cues; 4) inhibitory control; and 5) self-evaluation. ML techniques will be used to reveal the patterns of brain structures/functions that are associated with each outcome variable. Based on literature and preliminary findings, we anticipate that combining MRI with behavioral and clinical assessments will better account for individual variability in XR-NTX treatment outcomes in young adults with OUD, than using the behavioral and clinical variables alone. The data will unveil novel brain mechanisms that contribute to the risk of treatment failure in this critical population. The project will also serve as a training vehicle for Dr. Zhenhao Shi to improve his clinical and computational skills and facilitate his independent career development. Specifically, it will enable Dr. Shi to achieve five training goals: 1) to advance his knowledge in the methodology of clinical research; 2) to gain hands-on experience in leading clinical projects; 3) to master ML and multivariate methodologies; 4) to apply multimodal MRI techniques to translational and clinical research; and 5) to advance his general independent research skills including leadership, networking, collaboration, scientific writing and grantsmanship. Through a combination of didactic and hands-on activities, the project will fulfill Dr. Shi's training needs and ...