Stuber, Garret DHjort, Madelyn M2024-10-162024-10-162024-10-162024Hjort_washington_0250E_27422.pdfhttps://hdl.handle.net/1773/52421Thesis (Ph.D.)--University of Washington, 2024Although tightly associated with prefrontal cortex (PFC), concrete cognitive flexibility signals have historically been ill defined. One common test of cognitive flexibility involves reversal learning, where the contingencies of discrete learned cues are enhanced or degraded, and an individual subsequently must flexibly remap their behavior. This work presents meta-RPE (mRPE), a cognitive flexibility signal that peaks in the middle of reversal behavior and represents the average of repeated, concentrated errors over many trials. Allowing mRPE to modulate canonical single-trial reward prediction errors (RPEs) expedites reversal learning and fits observed animal behavior better than models with static learning rates. Using novel statistical and imaging methods, this work identifies a subpopulation of neurons in prelimbic cortex (PL), a PFC subregion, that selectively encode a contingency degradation-related mRPE signal and can directly modify RPE via preferential representation in projections to VTA. Otherwise stable PL dynamics across reversal suggest the mRPE signal is unlikely attributable to representational drift. Dopaminergic innervation to PL does not predict the mRPE signal, instead representing a contingency elevation mRPE signal from elsewhere in the brain. Deriving mRPE and identifying its neural correlate in the PL-VTA circuit represents a quantitative advance in the field’s understanding of cognitive flexibility signaling within the prefrontal cortex.application/pdfen-USCC BY-NCNeurosciencesBehavioral neuroscienceThesis