Structured and unstructured reactivations during REM sleep are modulated by novel experiences
Choi, J.; Carmichael, J. E.; Williams, S.; Etter, G.
Show abstract
Mammalian sleep is composed of two distinct phases characterized by unique neural activity patterns: rapid eye movement (REM) and non-REM (nREM) sleep. While both phases are essential for memory consolidation, neural replay of awake experience as a candidate mechanism for memory consolidation has only been shown for nREM sleep and it remains unclear whether awake experiences are reactivated during REM sleep. Here, we evaluated whether awake experiences are reactivated by hippocampal CA1 pyramidal neurons during REM sleep using a combination of approaches including Bayesian decoding, sequence factorization, and co-activation analysis. We confirmed that representations of awake exploration are recapitulated during subsequent REM episodes with varying levels of temporal structure. We also found that temporally organized sequences were replayed with a higher frequency during REM sleep following a novel experience. Finally, we find internalized structure in CA1 activity prior to novel task exposure that is expressed in subsequent behaviour. Altogether, these results suggest that while hippocampal ensembles appear as a result of structural priors, salient experiences refine temporal sequences during subsequent REM sleep, suggesting a unique role for REM sleep in salient-experience consolidation.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- REM sleep has minute-scale rhythms in mice and humans: A non-binary continuum between phasic and tonic microstates 96%
- Single neuron firing cascades underlie global spontaneous brain events 95%
- Coupling between slow-waves and sharp-wave ripples organizes distributed neural activity during sleep in humans 95%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.