PROJECT SUMMARY Survival across species hinges on an animal’s ability to switch ongoing behaviors to optimize goals, such as obtaining rewards, in dynamic environments. This process of adapting behaviors based on changes in environmental contingencies, commonly referred to as behavioral flexibility, is impaired across many neuropsychiatric disorders including depression, schizophrenia, and obsessive-compulsive disorder. A key element of adaptive reward computations is reward prediction error (RPE) signals, which reflect the discrepancy between obtained and expected rewards. The lateral habenula (LHb) plays a well-established role in signaling RPE and has recently been more broadly implicated in adjusting choice behavior in response to changes in reward contingencies. This suggests the LHb as a potential target for therapeutic intervention for diseases characterized by inflexible behavior. However, the LHb is also involved in a diverse range of emotional and motivational behaviors including stress, fear, aversion, aggression, and social and parental behavior. Therefore, developing treatments that behavioral flexibility without risking a cascade of side effects first requires identifying specific LHb subpopulations underlying flexible behavior. In this proposal, we use cell-type-specific approaches to monitor and manipulate a genetically-defined neuronal subpopulation found to preferentially signal RPE to determine how LHb RPE signals facilitate flexible behavior. In our first aim, we will identify the neuronal coding mechanism used to encode reward prediction error in the LHb. In our second aim, we will determine how these signals contribute to flexible reward-guided decision making. Together these experiments will reveal the functional role of LHb RPE signaling in flexible behavior, which will provide more refined insight into potential targets for tailored treatment of disorders characterized by inflexible behavior.