Neuroimaging of Instrumental Learning Networks in Adolescent Cannabis Use Treatment

PROJECT SUMMARY/ABSTRACT The candidate’s long-term goal is to shift his research focus from examining neural correlates of individual differences in adolescent substance use disorders toward neurobiological predictors of individual treatment outcomes. Through the research and training in this K23 proposal, the PI will acquire training in multivariate statistics/machine learning methods and clinical research necessary to make this transition. Adolescent cannabis use is a serious public health problem. 6% of high school seniors report daily cannabis use and 3% of adolescents meet criteria for CUD. Although cannabis is increasingly becoming legal at the state level, cannabis is associated with increased risk of psychiatric disorders, cardiovascular disease, and respiratory disease. However, CUD treatment responses are inconsistent. Contingency management (CM) is a critical component of CUD treatment and relies on instrumental learning to reduce cannabis use by rewarding abstinence from cannabis use. However, little work has investigated whether integrity of brain networks underlying instrumental learning predict CUD symptom severity, and no work has investigated whether performance on an instrumental learning task predicts cannabis use outcomes following CM. Therefore, the proposed K23 project will address the candidate’s training goals while gathering preliminary and feasibility data supporting the next step in his research. The proposed research study aims to: (1) identify associations between CUD symptomatology and neural activity during instrumental learning and (2) examine whether neural activity during instrumental learning is associated with reduced cannabis use frequency during CM treatment. The PI proposes to accomplish these aims by collecting fMRI and report-based data from a sample of 76 youth ages 14-18 who use cannabis at least once per week. All youths will complete a screening visit, an fMRI scanning visit, 10 virtual CM sessions, and a follow-up visit. We anticipate that findings will show the extent to which disruptions in neural systems underlying instrumental learning are associated with CUD symptoms and the extent to which neural activity instrumental learning predicts contingency CM outcomes. This K23 application proposes training and research that is directly in line with NIDA objectives and represents a logical progression from the PI’s prior experience to address career development goals in three areas: (1) use of multivariate and machine learning techniques in analyzing neuroimaging data; (2) clinical research. The experience and findings from the proposed K23 will position the PI to pursue NIH funding and build this line of research through a large mechanistic clinical trial of CM treatment on adolescent cannabis use and the role of neuroimaging data in individualized treatment outcomes.