Clinical and Neurocomputational Effects of Behavioral Activation in Veterans With Impaired Social Functioning

Overview

Poor psychosocial functioning, including social disconnection, is common and disabling among Veterans, and often fails to improve following the best available evidence-based treatments, particularly if more severe pathology is present. Behavioral activation (BA) is a promising, low-burden intervention, which could help improve social functioning impairments in Veterans, as this treatment targets social reward sensitivity, an important driving mechanism of social functioning. Thus, the proposed research will test whether BA can specifically improve social functioning in Veterans. Leveraging a neurocomputational framework of social reward seeking behavior, this research will help to improve the treatment and assessment of social function and the prediction of psychosocial treatment needs in Veterans, while providing neurobehavioral targets to develop new interventions that can restore social functioning.

Perceived social disconnection and social isolation are disabling and common among Veterans. Yet, these psychosocial impairments can fail to improve following the best available evidence-based treatments. Behavioral activation (BA), an evidence-based treatment for depression has shown promise in improving social functioning. However, most of this research was conducted in primarily non-Veteran populations and has typically relied on patients' subjective symptoms and self-report, in contrast to objective behavioral probes, to assess social functioning. Moreover, the cognitive mechanisms promoting social behavior remain poorly understood, particularly in Veterans. For instance, social anhedonia, i.e., a reduced sensitivity to social rewards, such as positive human interactions, has been implicated as an important driving mechanism of poor social functioning. Yet, the degree to which BA successfully target social reward sensitivity remains unclear and has not been evaluated in Veterans. This lack of understanding limits the ability to develop more targeted and effective interventions for psychosocial recovery. To address these gaps, this study aims to test the clinical and neurobehavioral effectiveness of BA in targeting social dysfunction in Veterans. Computational models of reward learning (RL), particularly Bayesian RL learning models, can offer an important bridge to link in-vivo behavior, such as social reward sensitivity, and recovery outcomes like psychosocial functioning, by providing a rich quantitative representation of reward learning and its dynamic updating during reward exploration. This proposal aims to capitalize on this approach, combining a translational behavioral paradigm, Bayesian reinforcement learning modeling, and neuroscientific evaluation of change processes, in order to assay social anhedonia and improve prediction of social functioning as well as understanding and effective dissemination of BA treatment. The specific aims are to: 1) test the effectiveness of BA in improving social functioning outcomes in Veterans (Aim 1); 2) examine the impact of BA on neurocomputational markers of social reward sensitivity and their relationship to social functioning improvement in Veterans (Aim 2); and 3) assess the degree to which neurocomputational markers of social reward sensitivity predict social functioning recovery in Veterans undergoing BA (Aim 3). This project will also explore whether the relationships between neurocomputational markers of social function and clinical outcomes are moderated by primary diagnosis (exploratory Aim 4). Veterans (N=136) with impaired social functioning and anhedonia will be recruited and randomized to either behavioral activation (BA) or supportive care therapy (SCT; active control) over 12 weeks. At baseline, immediately post-treatment, and 12-weeks post-treatment (24-weeks follow-up), participants will complete a computerized task aimed at probing social reward-seeking behavior, in which they must choose between unknown partners with the goal of maximizing the number of compliments they receive from these partners. Concurrent brain activity will be collected with functional magnetic resonance imaging (fMRI) in a subset of participants (N=68; at baseline and post-treatment). Therapy sessions and follow-up (24-weeks) assessments will be administered through VA-approved telehealth methods. Computational modeling will be used to derive individual parameters of social reward (i.e., compliments) learning and maximization, and to extract associated neural activity, e.g., neural responses to prediction error signals (discrepancies between expected and actually obtained reward/compliment). Randomization of participants to treatment arms will be be stratified based on fMRI eligibility status, sex, and primary psychiatric diagnosis. This project is expected to 1) determine if BA offers an effective, low-burden option to improve social functioning in Veterans; 2) identify potential neurocognitive targets, which could be used in future research to develop and test novel treatments to improve psychosocial functioning in Veterans, such as computerized neurocomputational training protocols to boost social reward sensitivity; and 3) help identify precise markers of BA responsiveness for functional recovery.