A Snapshot seminar hosted by the School of Physiology, Pharmacology and Neuroscience
Abstract:
In volatile environments, humans and animals face different forms of uncertainty to which they must adapt to thrive. However, our understanding of the neural basis of this adaptation is incomplete. Here, we take advantage of the well-known spatial remapping of hippocampal place cells in the face of environmental change to interrogate these processes. We performed calcium imaging in CA1 place cells using a new Uncertain Reward virtual reality Task (URTask), where mice run along a linear track and lick for a water reward. The reward location could be stable, indicating low uncertainty, or substantially variable, a form of expected uncertainty. Expected uncertainty in reward location enables us to disentangle reward and position reference frames in the neural code. We identified cells that consistently encode position, and cells that follow the reward on a single trial basis, adopting a reference frame anchored to the reward location. We also found that the spatial selectivity of some of the reward-anchored neurons followed a warped spatial metric, normalizing the distance from the reward to the end of the track. Once the environments are familiar, the reward location translates to a narrow zone far away from predictive bounds without warning, thus triggering a form of unexpected uncertainty that may be more pronounced in mice accustomed to regularity than in those used to expected uncertainty. Surprisingly, we did not observe a significant difference in the ultimate format of the maps between groups of mice that had first experienced low uncertainty vs those that had experienced high expected uncertainty. When tracking the same cells across the switch, we even found that the same relative proportion of cells maintained their position-relative firing before and after the switch. However, we found that mice that had first experienced high expected uncertainty had a higher proportion of cells whose spatial encoding was tied to the reward location both before and after the switch. We also observed that the flexible warped sequence induced by experience in expected uncertainty flexibly translates to the new reward location. Our results suggest that uncertainty in reward location may cause an increased over-representation of the reward, which can adapt to new reward locations. Additionally, the reward location appears to act as a strong local reset for hippocampal activity, similar to salient visual inputs like teleportation points in virtual reality, thereby influencing the spatial metric of hippocampal networks. This indicates that the hippocampal code supports flexible behavior by integrating elements within varying and distorted metrics to achieve stability in stochastic environments.