Project Abstract Learning is one of the essential building blocks of cognition. Individuals with mental health conditions, such as bipolar disorders and schizophrenia, often show symptoms related to learning. A great deal of evidence exists indicating that the orbitofrontal cortex (OFC) and anterior cingulate cortex (ACC) are both heavily involved in flexible learning under uncertainty, with the basolateral amygdala (BLA) sending robust bidirectional connections to both cortical structures. I hypothesize that BLA provides OFC and ACC with dissociable signals that delineate a difference in the expected reward and what was received, vital information for adaptive learning. In Aim 1, rats will be tested on a novel, dynamic, restless bandit task for flexible learning under different probabilistic schedules where the reward contingencies switch within the session. While rats perform this task in some sessions, we will chemogenetically inhibit BLA. A subset of these rats will also express GCaMP6f in OFC and GRIN lenses directly in this region so that we can use miniscopes to record calcium traces while the freely-moving rat learns. In parallel Aim 2, rats learning on the same task will instead have BLA on- or off-line while I record calcium traces in ACC. These approaches will allow us to study how the uncertainty representations in OFC and ACC differentially depend on BLA input. My central hypothesis is that OFC and ACC use BLA-supplied information differently: OFC receives information from BLA that is needed to detect first reversals and setting future adjustments, whereas ACC uses information from BLA to approximate changes in probabilities over time.