Work

Run, sleep, repeat: Replay in hippocampus and visual cortex during sleep

Year: 2025
Type: dissertation
Abstract: Hippocampal place cells reactivate during sleep in sequences corresponding to past trajectories in a phenomenon known as replay. Replay is proposed to spread through cortical areas, constituting a hallmark of system-levels consolidation theories, whereby memory traces are transferred from hippocampus to the cortex for long term storage. These theories provide clear predictions about coordination of replay across regions, whereby hippocampal reactivations should be associated with simultaneous reactivation of cortical representation of the same experience. Remarkably, sequences of activity in the hippocampus have been reported to correspond to future trajectories through both familiar and novel spaces, in a phenomenon termed preplay, reflecting either planning or pre-existing patterns of sequential activity. Whether a similar phenomenon can be observed outside the hippocampus and how it is related to hippocampal activity is unknown. In this study, we recorded simultaneously from the hippocampus and primary visual cortex during navigation and sleep to answer the following questions: 1) Can replay be detected in V1 during sleep? 2) Can hippocampal and V1 activity simulate novel trajectories through familiar space during sleep? 3) Are trajectories reactivated in V1 and hippocampus coordinated? To rigorously assess replay, we systematically evaluated different shuffling methods for establishing statistical significance of replay. We found that all tested shuffles were associated with inflated false positive rates, which we attempted to mitigate using a combined shuffle and adjusted significance levels. While controlling false positive rates, these methods may be overly conservative, leading to low replay and preplay detection rates in our data, which did not exceed expected false positive rates. Consequently, we were unable to draw firm conclusions regarding offline reactivations in V1 and their coordination with the hippocampus. These findings underscore the ongoing challenge of reliable replay detection due to the absence of a gold standard methodology, as well as highlight the trade-offs inherent in different shuffling methods and suggest that a nuanced approach is necessary for reliable replay detection.
Language: English
Open Access status: green
OpenAlex ID: W7115100948
Cites:
Cited by:
Related to: 9