Escrever um comentário

Review of: Large sharp-wave ripples promote hippocampo-cortical memory reactivation and consolidation

By Donghoon Shin and Loren Frank

Hippocampal sharp-wave ripples (SWRs) during non-rapid eye movement (NREM) sleep are known to coordinate the reactivation of neurons encoding recent experience. These reactivations broadcast recent memory representations in hippocampus (HPC) to the neocortex and related areas, and this pattern of information flow is known to be crucial for memory consolidation. However, only a subset of SWRs are associated with memory reactivation, and we do not know which types of SWRs drive memory consolidation. This study addresses this critical question by identifying the specific SWR characteristics that promote hippocampo-cortical reactivation and contribute to successful memory performance. The authors of this paper use closed-loop optogenetic SWR boosting to show two results: (1) long duration and high amplitude (large) SWRs detected in hippocampus during NREM sleep promote consolidation in a memory dependent task; and (2) large amplitude SWRs promote reactivation in HPC and prefrontal cortex (PFC), and increase HPC-PFC coherence in the theta band (~8Hz), which is known to be associated with cross-area communication and successful memory performance during awake periods. Their first result is valuable in confirming the predictions from their previous work in their other paradigms (Oliva et al., 2020; Fernandez-ruiz et al., 2019), and the second result is important as it demonstrates that SWR boosting can enhance HPC-PFC reactivation during SWRs and HPC-PFC coherence during awake behavior using an optogenetic intervention. The experimental evidence presented strongly supports both claims. Their experiment is well controlled and appropriate for their questions and the paper is well written. We have two suggestions that would make the paper stronger: (1) evaluation of the SWR categorization strategy and boosting effects by providing more detailed quantification of SWR properties, and (2) an attempt to synthesize the two results so that they can test the main hypothesis that hippocampal-cortical memory reactivation and coherence driven by SWRs directly contribute to memory performance.

Strengths

1.    The manuscript is clearly written, and the figures are well designed and easy to interpret.

2.    The research questions are important and relevant to the field, and the background provides strong motivation.

3.    The scientific claims are appropriate, and their experimental approach is rigorous and well-controlled.

4.    The optogenetic SWR-boosting technique (Fernandez-ruiz et al., 2019; Oliva et al., 2020) is very powerful and clearly demonstrates the contribution of SWRs in memory performance.

5.    The analysis is comprehensive, including current-source-density measurements and coupling with delta waves during sleep, providing mechanistic insight into large SWR generation.

Major points

1.    Evaluation of SWR categorization and characterization of boosting effects. The manuscript categorizes SWRs into two groups based on jointly high duration and amplitude, which is consistent with Fig. 1 demonstrating that large SWRs are associated with stronger CA1 and PFC reactivation. To help readers fully evaluate the categorization of SWRs and the effect of SWR boosting, a more detailed quantification of SWR properties would be valuable. First, presenting the full distribution of SWR amplitudes and durations across conditions would allow readers to determine if the categorization into 'small' vs. 'large' SWRs is supported by any underlying bimodality. This is critical because the central question that the paper addresses is which types of SWR contribute to memory consolidation. The authors claim that large SWRs are important for memory consolidation, but it remains unclear whether these events are a distinct class or part of a continuum. Visualizing the distribution would let readers evaluate the two hypotheses: (1) that all SWRs contribute but larger ones have a stronger effect, or (2) that two separate distributions exist and only the large SWRs contribute to memory consolidation. Even if these distributions are not bimodal (as in Fernandez-Ruiz et al., 2019), this visualization is essential for readers to evaluate the classification strategy on their own. Second, using the same duration–amplitude criteria as in Fig. 1F–K, reporting the number or proportion of SWRs in each category during post-task sleep after 5-minute exploration, 15-minute exploration, and 5-minute exploration with boosting would allow direct comparison across behavioral conditions. Third, while boosting increases the fraction of large SWRs (Fig. 3E) and ripple-band power (Fig. 3D), providing distributions (e.g., histograms or CDFs) of SWR duration and amplitude for boosted versus control and baseline SWRs would clarify whether the manipulation preferentially extends SWR duration, increases amplitude, or both. Fourth, as the stimulation is delivered ~20 ms after SWR onset, visualizing ensemble reactivation specifically for boosted SWRs—separated into activity immediately before and after stimulus onset (blue window in Fig. 3C, see Fig. 3I)—would test whether boosting maintains or modifies the ongoing ensemble pattern. Finally, it may be useful to explore whether early SWR features predict which events become long-duration or high-amplitude SWRs under stimulation, which could inform future targeted boosting strategies.

2.    Linking SWR-driven reactivation to memory performance. The manuscript presents two key findings: (1) large SWRs are associated with stronger reactivation in HPC and PFC, and (2) large SWRs enhance memory performance. However, these results are currently presented independently, and the central interpretation—that increased HPC–PFC reactivation during large SWRs contributes to improved memory performance—is not directly tested. Fig. S7 demonstrates that individual variability in memory performance correlates with the rate of large SWRs, which is compelling. Extending this analysis to the neural coordination metrics presented in the paper would substantially strengthen the argument. Specifically:

a.    Compare individual variability in HPC and PFC reactivation strength with memory performance across animals.

b.    Perform a similar analysis relating hippocampo–prefrontal theta coherence to behavioral performance.

These correlations would help establish whether variability in neural reactivation and functional coupling accounts for variability in memory outcome, thereby more clearly integrating the physiology and behavior results.

Minor points

3.    In addition to the analysis of CA1-PFC theta coherence during recall (Fig 4E/F), considering a follow-up analysis on the actual firing-rate level population ensemble activation during awake recall would be insightful. This would mirror the assembly analysis performed during sleep and directly test if the enhanced coordinated firing patterns observed during sleep persist into the functional retrieval period.

Competing interests

The authors declare that they have no competing interests.

Use of Artificial Intelligence (AI)

The authors declare that they did not use generative AI to come up with new ideas for their review.