DYNAMIC ALTERATIONS IN DEFAULT MODE AND EXECUTIVE CONTROL NETWORK CONNECTIVITY DURING A DISTRESS TOLERANCE TASK PREDICT DISTRESS TOLERANCE AND POST TREATMENT SUBSTANCE USE

Low distress tolerance (DT) is associated with decreased resting-state functional connectivity within the executive control network (ECN) and increased connectivity between the salience (SN) and default mode (DMN) networks among individuals who use cocaine and healthy controls, as well as decreased connectivity between select regions of the ECN and SN during stress exposure among those who use cocaine. Coordinated activity within and between the SN, DMN, and ECN is important for task-oriented action in the presence of distress and neurobiological models of SUD emphasize inter-relations among these networks as central to withdrawal-precipitated relapse to substance use. Yet, dynamic interactions among these networks in response to stress have not been examined in relation to DT or post treatment risk and rate of substance use. Twenty-five treatment-seeking adults with SUD (Mage = 38.08 ± 11.35 years, 84% male, 72% Caucasian, Meducation = 12.36 ± 2.10 years) underwent functional magnetic resonance imaging (fMRI) at three distinct temporal phases: while at Rest and during both Task and Stress phases of a behavioral DT task. Weighted Least Squares linear modeling was used to derive phase-specific within and between network connectivity estimates of SN, DMN, and lateralized (left and right) ECN which were entered into a Lasso-regularized Cox proportional hazard regression as predictors of DT. Phase-specific functional connectivity estimates significantly predicting DT were then tested as predictors of abstinence duration and substance use frequency over a three-month follow-up period. Increased neural network connectivity of the LECN at Rest and during Stress as well as between the LECN and posterior cingulate cortex (PCC) of the DMN during Task predicted DT, while increased DMN connectivity from Task to Stress predicted higher substance use frequency over the three-month follow-up. Study findings support neurobiological models of SUD and highlight the importance of evaluating dynamic interactions among neural networks in relation to behavioral outcomes relevant to SUD.Doctor of Philosoph