Dynamic endocrine factors shape hippocampal spatial representations

The computational study of neural circuits canonically assumes that the underlying connectivity of the network is static in the absence of learning. However, it has long been understood that endocrine factors can influence network connectivity in a cyclical manner, as demonstrated by a reliable 30% fluctuation in synaptic density observed across the estrous cycle in vitro. Despite this, the mechanisms underlying endocrine modulation of the hippocampal network remain poorly understood. Here, we used chronic two-photon (2P) imaging to demonstrate that both dendritic spine turnover and place cell remapping are modulated by cyclic fluctuations in sex steroid hormones. This phenomenon was observed using a custom chronically implanted glass microperiscope in a cohort of awake, behaving mice. To concomitantly track endocrine state, we developed a supervised deep neural network for unbiased classification of estrous stage at expert levels. Using these techniques, we first measured dendritic spine turnover across the course of several weeks. We found approximately 15.0% fluctuation in turnover, with spine density peaking during periods of high estradiol. To understand how these structural changes affected the functional properties of the circuit, we tracked place cell responses across environments defined by differential visual cues. The mice were first introduced to environment A, then transferred to environment B, then finally reintroduced to environment A (A > B > A’). The degree of hippocampal remapping between environments A and B was greatest during periods of elevated estradiol and synaptic density, while the stability of place fields between environments A and A’ was not significantly different across days. Shifts in dendritic spine turnover along with place cell remapping during periods of high estradiol suggest that endocrine state modulates both the structural and functional plasticity of the hippocampal network.