Convergence of top-down and bottom-up thalamic inputs in medial entorhinal cortex

Learning and memory is profoundly impaired in patients with Temporal Lobe Epilepsy (TLE) and Alzheimer’s Disease (AD), with devastating consequences on everyday actions such as the ability to safely navigate back home. Temporal lobe structures such as the medial entorhinal cortex (MEC) and hippocampus are critical for

the brain’s ability to create, update, and use internal representations of a dynamic external world. While decades of research have revealed how MEC neurons—whose firing rate and patterns encode an animal’s position, orientation, and speed in the environment—can build a static, reliable “map” of the physical environment, less is

known about how these maps can be flexibly updated to meet changing behavioral demands. Specifically, how are abstract cognitive features, such as having to avoid an accident-prone traffic intersection or pick up groceries on the way home, integrated into neural maps of the environment to guide behavior? Better understanding the

neural circuits and computations which govern flexible spatial processing in MEC is a crucial step towards identifying vulnerabilities in the entorhinal-hippocampal network which may be targeted to alleviate cognitive impairments. Recent advances in high-density electrophysiological recording techniques have generated critical

insights about the organization and diversity of information encoded by MEC neurons. In parallel, advances in 3D video recordings techniques and machine-learning algorithms have unlocked access to the rich dynamics of rodent behavior. In this proposal, we combine these state-of-the-art techniques to densely sample single-unit

neural activity and behavioral dynamics at high resolution in freely-moving mice performing tasks associated with different cognitive demands, while deploying optogenetic perturbations and an unbiased statistical model of neural encoding. This strategy will allow us to rigorously assess how diverse physical and abstract features of

the environment are encoded by hundreds of simultaneously recorded neurons across brain regions. Our overarching hypothesis is that flexible transformations of MEC maps are shaped by distinct neural circuits conveying physical versus abstract features of the environment. In Aim 1, we will test the hypothesis that changes

in cognitive demands transform neural maps by recruiting a diverse repertoire of spatial and behavioral variables to be flexibly encoded by MEC neurons. In Aim 2, we will examine if changes in physical features of the environment drive coordinated transformations of spatial representations across MEC and upstream regions

such as the anterior thalamic nucleus. In Aim 3, we will determine if cognitive signals conveyed by the thalamic nucleus reuniens enable transformations of MEC spatial representations. This proposal will identify circuit-level substrates that influence which, when, and how transformations of MEC representations occur. By revealing how

flexible MEC spatial processing is orchestrated by two thalamic regions which, alongside MEC and hippocampus exhibit pathological hallmarks in TLE and AD, this work will advance our understanding of entorhinal- hippocampal network function in health and disease.