Event boundaries drive norepinephrine release and distinctive neural representations of space in the rodent hippocampus

Reviewer #1 (Public review):

Summary:

This study investigates the role of norepinephrine (NE) signaling in the hippocampus during event transitions, positing that NE release serves as a mechanism for marking event boundaries to facilitate episodic memory segmentation. The authors use a genetically encoded fluorescent indicator (GRABNE) to measure NE release with high temporal precision, correlating these signals with changes in hippocampal firing dynamics. By integrating photometry data, behavioral analyses, and analysis of neuronal activity from publicly available datasets, the work addresses fundamental questions about the relationship between neuromodulatory signals and memory encoding.

Strengths:

The authors present a compelling framework linking NE signaling to event boundaries, offering insight into how episodic memory segmentation may occur in the brain. The writing is clear and the data are well-described. It is easy to follow. The pharmacological validation of the GRABNE sensor enhances confidence in their NE measurements, an important methodological strength given the potential limitations of fluorescence-based neuromodulatory indicators. Moreover, the authors carefully disentangle NE signals from confounding behavioral variables, providing evidence that NE release is time-locked to event boundaries rather than movement or arousal-related behaviors. This level of analytical rigor strengthens their central claims. Additionally, the observation of NE signal dynamics that decay over hundreds of seconds is interesting, as it aligns with timescales relevant to hippocampal plasticity reported in prior literature.

Weaknesses:

While the authors establish correlations between NE signaling and hippocampal activity changes, causation is not demonstrated. Future studies using perturbative approaches (e.g., optogenetic or chemogenetic manipulation of NE release) would be necessary to establish a direct causal link. Furthermore, the persistence of NE signals over long timescales (hundreds of seconds) raises questions about its role in encoding rapid event boundaries, as it is unclear how this prolonged signaling might affect memory encoding for closely spaced events. The lack of a discussion about how NE dynamics would operate in such scenarios weakens the proposed framework. Finally, while the authors acknowledge the limitations of the GRABNE sensor, a more detailed exploration of how sensor sensitivity might influence their results would enhance the interpretation of their findings.

Reviewer #2 (Public review):

Summary:

The authors use a genetically encoded fluorescent sensor, GRABNE, to measure NE dynamics in the dorsal hippocampus of mice in response to multiple behavioral manipulations. A non-linear model and regression were used to quantitatively assess the contribution of multiple behavioral covariates to changes in NE signaling, with the result that NE signal dynamics were best predicted by time from event transitions, with the signal exponentially decaying over a period of seconds to minutes after transitions. Event transitions were implemented as a transfer from a home cage to a novel arena, a transfer to a familiar linear track, and the introduction of novel objects. Additional experiments showed that spatial context transitions dominate NE signaling over novel object presentations, and experience accelerates the decay of the NE signal after spatial context transitions. Correspondingly, the hippocampal CA1 spatial code takes minutes to stabilize after context transition in both novel and familiar spaces.

Strengths:

A strength of the study is the use of the NE sensor with sub-second resolution, non-linear modeling, and regression to identify the prominent variable of interest as time from event transition, and multiple behavioral controls. The use of multiple behavioral designs to investigate the effect of familiarity, experience, and interaction of spatial context transitions and novel object introduction is a strength. Relating the dynamics of NE signal decay to the rate of CA1 spatial code changes is also a strength.

Weaknesses:

A minor weakness is that the concept of an event boundary needs to be more broadly discussed. The manuscript uses event transitions such as spatial context changes and novel object introduction to implement an event boundary. However, especially in episodic memory studies in humans, event structure and boundaries have also been shown to occur through the automatic segmentation of experiences into discrete events (Baldassano et al., Neuron, 2017; Radvansky and Zacks, Curr. Opi. Behav. Sci, 2017). The rodent experiments in the current manuscript explicitly introduce event boundaries through changes in context or objects, which can potentially be conflated with novelty. A discussion of these differences, and whether NE can also have a role in event boundary transitions based on automatic segmentation of experiences, will add to the impact of the manuscript.

Reviewer #3 (Public review):

Summary

The manuscript investigates the role of norepinephrine (NE) release in the rodent hippocampus during event boundaries, such as transitions between spatial contexts and the introduction of novel objects. It also explores how NE release is altered by experience and how novelty drives the amplitude and decay times of extracellular NE. By utilizing the GRABNE sensor for sub-second resolution measurement of NE, the authors demonstrate that NE release is driven primarily by the time elapsed since an event boundary and is independent of behaviors like movement or reward. The study further explores how hippocampal neural representations are altered over time, showing that these representations stabilize shortly after event transitions, potentially linking NE release to episodic memory encoding.

Strengths

Overall, the work provides novel insights into the interplay between NE signaling and hippocampal activity and presents an intriguing hypothesis on how NE release may help push hippocampal activity into unique attractor states to encode novel experiences. The experiments are well-controlled, and the analysis is well-presented, with a detailed and engaging discussion that points towards several new and exciting research directions. The use of several behavioral paradigms to demonstrate the strongest predictor of NE release is a strength, as well as the regression analysis to disambiguate the contribution of other correlated variables. The suggestion that NE does not select ensembles for subsequent replay is also an interesting result.

Weaknesses

The authors have not convincingly established a link between hippocampal neural activity and NE release, showing qualitative rather than quantitative correlations. Therefore, at this stage, the role of NE on hippocampal function remains speculative.

Another general concern is that the smoothing/ kinetics of the sensor impacts the regression analyses. Most of the other variables, such as speed, acceleration, and even reward time points are highly dynamic and it is possible that the limitations of the sensor decorrelate the signal from (potentially) causal variables, therefore resulting in the time since the event start having the most explanatory power for most of the analyses.

More broadly, the figure legends should be expanded to better describe error bounds, mean vs median, sample sizes, and averaging choices for plots.

There are also some concerns regarding the nearest neighbor analysis and the reported differences in the rate of reactivations after familiar and novel environments, as outlined below.

(1) Lines 657-658. How far away in time can the top three nearest neighbor time points be? Must they lie in different trials, or can they also be within the same trial? Is there a systematic difference in the average time lags for the nearest neighbors over the course of the session?

The authors should only allow nearest neighbors to be in a different lap because systematic changes in behavior (running fast initially) might force earlier time bins in a certain location to match with a different trial, while the later time bins can be from within the same trial if the mice are moving slower and stay in the same spatial bin location longer. The authors should also provide information on how the averaging is performed because there are several axes of variability - spatial bin locations, sessions, different environments, and animals.

(2) Figure 8: These results are very interesting. However, I am confused by the differences between Figure 8B and D because the significant reactivations in A and C are very similar. The 1-minute and 10-minute windows seem somewhat arbitrary and prone to noise and variability. Perhaps the authors should fit a slope for the curves on A and C and compare whether the slope/ intercept are significantly different between the novel and familiar environments.