Multi level Analysis of Positive Valence Systems Across Mood Disorders

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): The guiding principle of the RDoC initiative is that DSM-based conceptualizations of psychiatric illness may fail to capture biological dimensionality, and the goal of RFA-MH-12-100 is to use translational neuroscience approaches to test dimensional domains consistent with underlying biology. Reward-processing domains within the Positive Valence Systems (PVS) matrix represent an excellent area of focus for clinical research, given the vast basic science literature on which to draw upon. PVS abnormalities, especially Reward Learning, are particularly salient in mood disorders, where both depressive and manic states are associated with aberrant reward behavior, albeit in opposite directions. While past studies have sought to use behavioral and neuroimaging measures of reward processing to discriminate between unipolar and bipolar depression, a central limitation of prior work is the reliance on DSM diagnoses that fail to capture the full dimensional range of the reward learning construct. To address this challenge, the PI developed a widely-used objective measure of reward learning, the Probabilistic Reward Task (PRT), which allow the assessment of participants' propensity to modulate behavior as a function of reinforcements. The PRT has been used in over 900 individuals across the world and provides the opportunity to estimate population norms for normal reward learning. In the proposed research, we plan to recruit 160 individuals seeking treatment for mood disorders at three mood disorder clinics who will be screened with the PRT. Patient performance will be classified relative to normed control data, and 50% (n=80) of the screening sample will return for further testing. Importantly, participants in this sub-sample will be selected so that each quintile of the PRT normative-reference distribution is equally represented. We will then investigate biological mechanisms of reward learning across four units of analysis: molecules, circuitry, physiology, and behavior. Data on 32 healthy controls will also be collected. Measures from each unit will be integrated into a Reward Learning Network composite score (RLN Composite). In Aim 1, we hypothesize that the RLN Composite score will show superior ability in predicting reward-processing symptoms (e.g., measures of anhedonia, impulsivity, mania) compared to DSM diagnoses. In Aim 2, we will test how well the RLN Composite score predicts symptom profiles at 3- and 6-month follow-up time-points. Specifically, we hypothesize that, relative to DSM diagnoses, the RLN composite score will have greater positive and negative predictive power for anhedonic, manic, impulsive and suicide-related symptoms as well as overall functioning assessed at follow-up. In sum, this proposal would study the full dimensional range of reward learning across multiple levels of analysis in individuals exhibiting a wide range of symptoms and impairments (from severe depression to hypomania/mania). This constitutes the first step in a research program that has the promise to reshape how mental illness is conceptualized and treated.