Prospective Coding and Memory Retrieval

点击翻译按钮获取中文摘要

Memory evolved to predict outcomes, but the neural mechanisms that link outcomes and remembered episodes are unclear. Episodic memory requires the hippocampus (HPC), working flexibly with memory requires the prefrontal cortex (PFC), and both structures are needed for tracking outcomes that change over space or time. Spatial reversal learning entails learning to stop approaching a previously rewarded location and instead approaching a previously unrewarded one. The PFC and HPC are needed for spatial reversal learning, and each structure supports different functions. HPC inactivation impairs all spatial learning, whereas PFC inactivation spares discrimination but impair reversal learning. Different PFC circuits guide reversals depending on outcome history: medial prefrontal cortex (mPFC) supports switching between rapidly changing goals, orbitofrontal cortex (OFC) supports switching from one well-established goal to another. Neural representations in each structure predict choices that correspond with the inactivation effects: OFC predictions develop relatively slowly across episodes, while mPFC and HPC predictions develop quickly. The proposed experiments combine local circuit manipulations with high density unit recording in rats performing spatial reversals to test the extent to which coordinated HPC and PFC activity associates episodes and outcomes. Aim 1 will test how outcome history affects reversal strategies and functional interactions between PFC and CA1 by disrupting local circuits. Groups of rats will be trained in a HPC-dependent spatial task followed by reversals designed to require either OFC or mPFC while local OFC or mPFC circuits are temporarily inactivated. To investigate the required circuitry, mPFC and OFC will be infected with halorhodopsin (NpHR3.0) or channelrhodopsin (ChR2) containing viruses and PFC axon terminals will be modulated by light delivered to the n. reuniens, a thalamic relay between PFC and CA1. The results will determine if mPFC and OFC contribute independently or interactively to spatial outcome predictions, and test the extent to which these interactions require the n. reuniens. Aim 2 will investigate how outcome history affects PFC and CA1 coding interactions by recording simultaneously in the three regions or inactivating one while recording the other two as rats perform the tasks described in Aim 1. The results will identify neuronal signals and CA1 interactions with mPFC or OFC that predict outcome guided memory representations. Aim 3 will test PFC-HPC communication mechanisms by modulating the temporal coordination of PFC and HPC activity. Electrically stimulating the fimbria fornix synchronizes local field potentials in widespread cortical and subcortical networks and improves spatial learning and memory. “Pacemaker” stimulation of the medial septal area (MSA) and nucleus basalis modulate HPC and PFC synchrony independently. Enhancing or disrupting learning-related activity patterns recorded in Aim 2 will test if network synchrony is needed to link outcomes with memory.