Early Sleep-Dependent Sensory Gating in the Olfactory System

Reviewer #1 (Public review):

Summary:

The authors of Serantes et al. produced a well-designed set of experiments to address the mechanisms of olfactory disconnection during sleep. In contrast to other sensory modalities, olfaction is not filtered or potentially gated by the thalamus, potentially opening the door to unimodal sensory stimulation during sleep. Recent work (Schreck, 2022) used optogenetically activated Olfactory Sensory Neurons to show that local field potential and activity across the olfactory pathway, not only remained open during sleep but were potentially even accentuated under these brain states. However, their optogenetic manipulation is an artificial perturbation to the system that could override naturalistic early-gating mechanisms. In a set of careful experiments, Serantes et al. show that coupling between airflow and brain activity at the Olfactory Bulb is diminished under sleep and anesthetic brain states. In contrast to a peripheral gating mechanism proposed by Schreck, this lack of respiration-locked activity, measured with EEG and LFP, persists even in the presence of intense respiration and even when nasal airflow is artificially induced and controlled. Their results point to nonthalamic early sensory gating of olfactory information during sleep, which is independent of nasal airflow but dependent on internal brain states. Their work elicits questions about potentially undiscovered mechanisms at the level of the early sensory pathway.

Strengths:

The strengths of this paper lie in the level of control afforded by the multiple preps and the wide array of physiological recordings. Specifically, both their control of airflow with a dual tracheotomy and their control of internal states using both sleep and urethane anaesthesia have a cumulative impact on the results.

The paper is simple, well-written, well executed, has clear questions, describes the literature comprehensively, and points out conflicting results with precision and transparency. The same transparency and judgment should be used on their own results.

Another strength of the paper is the clear, unambiguous results. The effect sizes presented in the paper are sizable and convincing.

Weaknesses:

The paper's shortcomings include open questions and a lack of a full mechanistic understanding of the suggested internal gating process. There are some open questions about the relative importance of airflow sensing vs. odorant sensing. Recent work by Mahajan et al., Sci.Adv 2025 points to OSN as sensing both odorants and airflow to produce anemotaxis. Potentially, other cells could contribute to anemosensation as well, so that gated or non-gated information might depend on the ratio of airflow to odorant information. Perhaps, optogenetic stimulation of OSN acts as an unnatural sensory stimulation that can alter both olfaction and anemosensation.

Detailed ablation, pharmacological, and optogenetic experiments may be needed to elucidate the suggested mechanisms and determine the correct answer to the question posed by the authors.

Reviewer #2 (Public review):

Summary:

In this study, Serantes and colleagues analysed how sleep and anesthesia impact the processing of olfactory inputs, focusing on early sensory processing (occurring at the first or second synaptic contacts). First, they show that the transition to sleep has a major impact on breathing-dependent gamma activity. Second, they show that this decrease originates at the first synaptic contact and is independent of respiration itself. Third, they show a decrease in connectivity associated with neocortical slow waves. These results are very interesting and supported by a robust methodology. However, I have two major concerns regarding this work.

First, the authors fail to adequately contextualize their work. For example, the impact of sleep on respiration-locked gamma activity was reported several years ago and is, in fact, used in some laboratories to score sleep using data from the olfactory bulb.

Second, the authors should exercise much more caution when comparing the urethane anesthesia model with NREM/REM sleep cycles. There are very significant differences between the two. Yet, the title and abstract of the article mention only sleep and anesthesia. More concerningly, the results obtained under urethane anesthesia are uncritically generalized to sleep.

In conclusion, the first finding was already shown in previous studies, and the second and third results were obtained not during sleep but during an anesthetic state that only resembles certain aspects of sleep.

Strengths:

The authors deploy an interventional approach that allows them to determine with compelling evidence the relationship of the gamma activity time-locked to breathing and different aspects of breathing, proving in particular that the disconnection is independent of respiratory dynamics. They leveraged invasive recordings that allow them to pinpoint at which level the disconnection occurs.

Weaknesses:

(1) My first comment concerns how this work fits within the state of the art. The introduction of the article leaves out very important and highly relevant work.

(1a) First, "disconnection" is not a defining feature of sleep; "unresponsiveness" is. It is often assumed that this unresponsiveness (which can be directly measured, contrary to disconnection) is due to a form of disconnection, but there has been substantial work over the past decade showing that disconnection is not as extensive as initially expected. It is therefore incorrect, in my view, to state that "most models attribute sensory gating to thalamocortical mechanisms". Most models attribute sensory gating to a combination of thalamocortical and cortical mechanisms.

(1b) The rationale of the article appears unclear ("the olfactory system-bypassing the thalamus-offers a unique window into earlier stages of sensory disconnection"). If the idea is to investigate gating mechanisms before the thalamus, then any sensory modality would suffice, since even modalities that later relay through the thalamus involve pre-thalamic processing stages. I assume that the authors instead mean that, because olfactory information does not relay through the thalamus, gating mechanisms in the olfactory stream could occur very early. However, this also implies that focusing on olfactory processing would say little about other sensory modalities.

(1c) Key previous results have been completely overlooked. First, the impact of sleep on respiration-locked gamma activity was reported several years ago (Bagur et al., Plos Biology 2018). Second, important articles investigating olfactory processing during sleep have been overlooked (e.g., Arzi et al., Nature Neuroscience 2012; Arzi et al., Journal of Neuroscience 2014). I am not providing an exhaustive list here, but these articles are not only extremely relevant to the present study; they have also become classics in the sleep literature.

(2) For most of their findings (Figures 2 to 5), the authors used urethane anesthesia. They show that this pharmacological manipulation results in alternation between periods of high-amplitude delta waves (SWSt) and a desynchronized state (ASt). However, the parallel with NREM and REM sleep, respectively, is rough and insufficiently justified. Differences can already be noted by contrasting the short examples provided in the figures. While NREM and REM sleep differ in terms of muscle tone (EMG), no such difference is discernible between SWSt and ASt. In SWSt, the slow waves appear to overlap with fast activity at the cortical level (M1, S1), which is not typically the case during NREM sleep. In addition, because the time scale is not the same in Figures 1 and 2 (1 s vs 2 s), yet the slow waves appear to have similar durations, it is also possible that the slow waves generated during SWSt and NREM differ. To better support the proposed parallel between NREM and SWSt on the one hand, and ASt and REM on the other, the authors should provide a thorough comparison of these states (spectral features, properties of the slow waves, duration and frequency of each state, etc.). Without this, inferences from results obtained under urethane anesthesia to sleep are not warranted.

The authors acknowledge this issue in the Discussion ("These findings suggest that there is no functional equivalence between urethane-activated states and REM sleep"), but this caveat should be integrated from the very beginning (title, abstract, and introduction).

(3) In some graphs, the power spectrum is normalized. Under anesthesia, this normalization was performed "within each animal to the SWSt maximum for that signal". However, I could not find equivalent information for sleep. This is key information needed to correctly interpret the results shown in Figure 1.

(4) The authors should also clarify their criteria for concluding on the absence or presence of a given effect. For example, in the legend of Figure 1c, they write: "Note the presence of coherence during wakefulness, demonstrating the internalization of the respiratory signal, and its drop during sleep". Unless coherence is exactly zero, some degree of coherence is always "present". Figure 1 instead shows that coherence is modulated across frequencies during wakefulness, with peaks in the delta and theta ranges.

In Figure 2, they write: "PAC between respiration and OB gamma amplitude was present during ASt but disappeared during SWSt". Again, the authors should clarify what is meant by "disappeared", as they only tested for differences between ASt and SWSt.

Given that the authors implemented a strategy to test for above-chance coherence using surrogate datasets, they should consistently provide statistical tests showing which conditions or frequency bands exhibit coherence above chance in order to justify claims about the presence or absence of an effect.

(5) Likewise, comparisons across states should always be supported by statistical tests, for example, in Figure 4. In addition, despite the apparent absence of coherence during SWSt in Figures 4f and 4g (which again should be formally tested), Figure 4h shows an increase in coherence around 2 Hz, which suggests some degree of coherence between nasal airflow and the olfactory bulb.

(6) Figures should more clearly distinguish results based on a single "representative" animal from population averages. For example, were Figures 4g and 2h computed at the population level?

Reviewer #3 (Public review):

Summary:

Sleep is typified by a behavioural attenuation of responsiveness to external stimuli (higher arousal thresholds). There are various mechanisms through which sensory perception could be dampened, and while thalamic and cortical gate points have been well studied, the focus here is on peripheral ones - at the level of the olfactory bulb (OB). While something conceptually similar has been shown in insects, this paper represents an important contribution to understanding attenuation of sensory perception during rodent sleep and anaesthesia.

This paper shows that respiration-locked potentials and gamma activity in the olfactory bulb, which are important for olfactory coding, are diminished during sleep and when under anaesthesia compared to wake. Further, this state-dependent activity in OB is likely to be locally generated. Using a tracheotomy procedure aimed to dissociate nasal airflow from natural inhalations, authors demonstrate that local field potentials (LFPs) in the OB phase lock with artificially generated air pulses (delivered into the nasal cavity) during the active phase of anaesthesia but not during a more passive state. LFPs did not synchronise with respiratory signals during either anaesthesia state. Lastly, the authors showed that as delta power increased (typical of slow-wave-sleep), the coherence between nasal inhalation rhythms and OB LFP coherence decreased, indicating that as rats experienced something akin to slow-wave-sleep (during anaesthesia), disconnection from the external environment could be augmented. Taken together, the authors argue that the change in activity observed in the olfactory bulb during sleep and anaesthesia provides a non-permissive state for sensory processing and manifests as sensory dissociation

Strengths:

The manuscript is well-written, and the experiments are thorough. Experiments examining coupling of nasal respiration with OB potentials and delta activity are particularly interesting as they point to augmented sensory disconnection during a sleep phase typically associated with higher arousal thresholds.

Weaknesses:

(1) An experiment addressing the following points, is missing:

Does odour stimulation that wakes up a subject restore gamma activity and respiration-locked potentials?

Is OB/respiration desynchrony maintained when presented with a non-rousing stimulus?

Is waking upon stimulus delivery less likely as delta activity increases and coherence between OB/respiratory rhythms weakens?

(2) Many of the experiments are performed under anaesthesia, which I understand is for practical reasons. While authors are forthcoming about limitations of using anaesthesia in lieu of natural sleep states, I would have preferred to see more experiments performed on sleeping animals.

Author response:

We thank the reviewing editor and the reviewers for their careful evaluation of our manuscript “Early sleep dependent sensory gating in the olfactory system”, and for their constructive feedback. We are encouraged by the overall positive assessment of the work.

In the revised version, we will address all the points raised by the reviewers. Below, we outlined the main aspects of the revision.

(1) Contextualization within prior literature.

We will expand the text to better situate our findings within the existing literature and clarify the specific contribution of our work, particularly with respect to state dependent changes in olfactory bulb activity.

(2) Distinction between sleep and urethane anaesthesia.

We will revise the text to more clearly distinguish findings obtained during natural sleep from those obtained under urethane anaesthesia. While avoiding direct equivalence between states, we will clarify that the comparison is intended to highlight shared features of slow wave brain dynamics associated with sensory gating.

(3) Clarification of analytical methods and statistical criteria.

We will provide additional details regarding normalisation procedures, surrogate based analysis, and statistical criteria used to assess the presence or absence of coherence and phase amplitude coupling, ensuring consistency across figures.

(4) Improvements in figures in terminology.

We will revise figure annotations to improve clarity (axis, colour scales, units and labelling) and ensure consistent terminology throughout the manuscript.

We believe these revisions will further strengthen the manuscript while preserving its central conclusions.