Parvalbumin interneuron ErbB4 controls ongoing network oscillations and olfactory behaviours in mice

Reviewer #1 (Public review):

Summary:

The authors have addressed many of the concerns. In sum, this study examined how ErbB4 signalling in parvalbumin-positive (PV) interneurons contributes to olfactory bulb (OB) microcircuit function and olfactory behaviour in mice. Using RNAscope, single-cell RT-PCR, conditional Erbb4 deletion, awake in vivo recordings, slice electrophysiology and optogenetics, the authors report that ErbB4 was expressed in PV interneurons of the EPL and IPL, and that its loss impaired odour discrimination and dishabituation, reduced GABAergic transmission onto mitral cells, and elevated ongoing LFP power. The experiments were well-designed, but several central claims were not yet adequately supported by the data as presented.

Strengths:

Pretty comprehensive set of experiments.

Weaknesses:

(1) Figure 1. The investigation time of the AAV-Cre-GFP group varied considerably across isoamyl acetate concentrations. Were there any differences between the AAV-GFP and AAV-Cre-GFP groups at the 0 (mineral oil) condition? Since the threshold was defined by a comparison against mineral oil, a baseline difference would propagate directly into the threshold estimate and could reduce exploratory drive rather than an olfactory deficit.

(2) Previous papers have reported that ErbB4 is expressed in neuroblasts of the SVZ and RMS, and that disruption of NRG1-ErbB4 signalling may decrease GABAergic neuron number and alter distribution (Anton et al., Nat Neurosci 2004; Ghashghaei et al., PNAS 2006). The authors showed in Figure 2L that ErbB4 expression in the OB was already decreased at P7, suggesting reduced neuron number and disrupted neuronal distribution. The authors should therefore quantify the number and laminar distribution of the different neuronal types in the OB. Cresyl violet staining showed that gross lamination was preserved, but could detect changes in cell density or in the relative proportions of interneuron subtypes.

(3) Figure 4C. First of all, there were no indicators marking the sessions, so we had no idea what state the mice were in. Also, why did the odour-evoked MFRs of PV-Erbb4⁺/⁺ and PV-Erbb4⁻/⁻ mice show opposite trends during the 0-2 s window in the first session - were there any differences in the behaviour of the two groups during this session? As the mice were head-fixed on a floating ball, locomotion and sniffing should have been monitored, since both strongly influence M/TC firing.

(4) Figure 4K-L. The authors did not explain why spontaneous and odour-evoked MFR were so much different among the "no response" M/TCs. Did this mean that ErbB4 knockout in PV interneurons disrupts the function? This would imply a tonic rather than a stimulus-locked disinhibition, which is a different claim from the one made in the Discussion. In addition, the authors stated that panel 4L indicates "the distribution of types of odour-evoked responses was significantly altered in PV-Erbb4⁻/⁻ mice compared with PV-Erbb4⁺/⁺ mice, suggesting compromised odour detection" - what exactly does this mean? A shift in the proportions of response categories is also compatible with a changed baseline firing rate altering the classification itself.

(5) Figure 5. The power of all frequency bands in PV-Erbb4⁻/⁻ mice was higher than in PV-Erbb4⁺/⁺ mice, including the gamma band. However, Tan et al. (2022) reported that inhibition of ErbB4 kinase activity decreases LFP power and GABA release, which is inconsistent with the present data. This discrepancy should be addressed explicitly. A uniform elevation across all four bands was atypical for a PV-dependent mechanism and could reflect a global scaling difference; separating periodic from aperiodic components and reporting normalised power would help.

(6) Photocurrent amplitude and mCherry⁺ PV cell density was not compared across genotypes; Figure 9-figure supplement 1B-C showed photocurrents for only one group. Since the authors themselves reported in Figure 7-figure supplement 1 that PV interneurons in knockout mice reduced mEPSC frequency and amplitude and increased AP frequency, the presynaptic population was clearly not in an equivalent state. The reduced oIPSC could therefore result from differences in ChR2 expression level or in the number of labelled cells. Photocurrent curves and mCherry⁺ cell density must be provided for both groups.

Reviewer #2 (Public review):

The authors aimed to understand how the transcription factor Erbb4 contributes local circuit dynamics and affects performance. This is an interesting question because Erbb4 is largely localized to PV cells in the olfactory bulb (OB) and upregulated following exposure to an olfactory discrimination task.

Strengths:

They have now validated their local, PV-specific knockout strategy, which more directly tests their hypothesis.

The inclusion of optogenetically evoked presynaptic responses from PV cells onto M/T cells also strengthened the conclusion that this synapse was altered.

The authors also now included slice recordings from PV cells, which helps to form a more comprehensive picture of the circuit.

Weaknesses:

(1) The data used to support their spike sorting approach is now included but is difficult to evaluate. In Figure 4 - supp 1B, they show what appears to be mean + sem of concatenated waveforms from one tetrode, but this is not explained anywhere. Moreover, the timescale seems to be off by an order of magnitude. It also seems that the spike widths in panel E are bimodally distributed, with many of the half-widths being more consistent with fast-spiking cells. There is a high density of PV cells in the same layer where they record M/T cells, so is it not possible that some of the units were PV cells (note: this was suggested in the previous review as a potential exciting direction to further validate their findings)? It is still unclear how the authors confidently record from only M/T cells when there is no mention of how other units (e.g., fast-spiking cells) are excluded. The methods state that only M/T cells can be recorded from this layer, which is an oversimplification and appears at odds with the apparent diversity in waveforms and responses.

(2) Upon review, the authors acknowledge that the PV-Erbb4+/+ and PV-Erbb4-/- mice were recorded in separate batches where the control mice were exposed to unwanted artefacts related to a solenoid valve. While we don't know how much that change influenced the results, even with careful analysis, it does draw attention to a major weakness in study design where controls and experimental mice are not run simultaneously. The presence of an artefact from the solenoid valve (perhaps a noise?) immediately before odor presentation for controls, but not experimental mice, should be considered by the reader when it comes to evaluating the odor-evoked responses.

Overall, I feel the manuscript is much improved and more solidly demonstrates that Erbb4 expression in PV cells plays a major role in regulating local OB circuit dynamics and influences odor processing. The remaining weaknesses surrounding the in vivo single unit recordings from PV-Erbb4+/+ versus PV-Erbb4-/- mice reduce my confidence in the validity of the results, but the general agreement of the in vitro work is more re-assuring.

Author response:

The following is the authors’ response to the original reviews.

Public Reviews:

Reviewer #1 (Public review):

Summary:

The study by Hu et al. investigated the role of olfactory ErbB4 in regulating olfactory information processing. The authors demonstrated that ErbB4 deletion impairs odor discrimination, sensitivity, habituation, and dishabituation by using an impressive combination of techniques from morphological to electrophysiology (both slice and in vivo) and from viral injection to cell-type-specific mutation to behavioral analysis. The findings underscore the crucial role of ErbB4 in olfactory PV neurons in modulating mitral cell function and odor perception.

Strengths:

This study contains a pretty comprehensive set of experiments.

Major concerns:

(1) Line 151 page 7, "PV-Erbb4+/+ mice (generated by crossing PV-Cre mice (Wen et al., 2010) with loxP-flanked Erbb4 mice". Does this mean mice carrying PV-Cre and ErbB4 floxed allele? Or with the WT allele? This is confusing. Figures 2B and 2C, ErbB4 expression was evident in many cells that were not positive for PV. What are the identities of those cells? Are they important?

Thank you for your positive comments regarding our study. PV-Erbb4^+/+ mice carried the PV-Cre transgene with the WT allele. In the revised manuscript, we have simplified the description of the genotype for clarity and have provided detailed breeding information in the "Materials and Methods" section (Line 532).

As shown in Figures 2B and 2C, ErbB4 expression was evident in GL and GCL cells that were not positive for PV. The activation of ErbB4-expressing cells in the GL and GCL has been suggested to upregulate theta and gamma oscillations, respectively (PMID:35853717). Mammalian OB are the most abundant and varied interneurons among brain tissues, and their functional importance and diversity remain unclear (PMID:31358754 and PMID:32208366). ErbB4-expressing PV interneurons in the EPL and IPL were relatively less abundant but accounted for the majority of inhibitory inputs to M/TCs. Importantly, one PV interneuron connects with over 1000 M/TCs to mediate recurrent and lateral inhibition, linearly controlling the OB output (PMID:31358754 and PMID:39186804). This study provides evidence that ErbB4 signalling in PV interneurons is required for olfactory discrimination and dishabituation.

(2) In Figure 4, the authors performed tetrode recordings in awake head-fixed animals. Although individual neuron spikes could be obtained by spike-sorting, this is not a "single-unit" experiment due to the nature of this approach.

What is the odor used in Figure 4? How did the authors clean up the odor to limit the stimulation within 2 seconds? In what layer were the tetrodes placed? What is the putative cell type presented in Figure 4C? If Figure 4C is a representative neuron recorded, the odor-induced suppression of spike activity seems to be impaired in PV-ErbB4-/- animals. However, Figure 4D shows that suppressed neurons were similar between the two types of animals. Such comparisons among individual mice are difficult for in vivo electrophysiological experiments because the recorded cell type and placement of electrodes would be different. The authors should apply ErbB4 inhibitors to the same animals and compare the effects before and after. This would ensure the recording of the same population of neurons.

Thank you for your valuable comment and suggestion. We agree that tetrode recording combined with spike-sorting is not a strict "single-unit" approach, although this method is widely accepted and consistently referred to as "single-unit recording" even "single-neuron recording" (PMID:30833706). Based on your comments, we modified the description as follows: “To obtain individual neuron spikes, units were identified by principal-component analysis scan (PCA) clustering of the spikes” (Lines 210-212).

As shown in Figure 4, a standardised panel with eight tubes delivers eight odours: isoamyl acetate, 2-heptanone, phenyl acetate, benzaldehyde, dimethylbutyric acid, n-heptanoic acid, n-pentanol, and 2-pentanone. Odours were delivered via an odour delivery system (Thinkerbiotech), and charcoal-filtered air was supplied to the mouse’s nose at a constant rate of 1 L/min to eliminate airflow interference, which effectively limited the stimulation to 2s. We have added the above description to the revised "Materials and Methods" section (Lines 786-789).

Author response image 1.

The tetrodes were placed within the ventral mitral cell layer based on stereotaxic coordinates. The putative cell type is mitral/tufted (Figure 4C). Despite the presence of diverse cell types in the OB, only the spiking activity of M/TCs can be captured via tetrode recordings (Kay and Laurent, 1999; Doucette et al., 2011). We added the above description in the revised "Materials and Methods" section (Lines 764-767). As you have indicated, Figure 4C shows the impairment of odour-induced spike suppression in PV-Erbb4^-/- mice. Consistently, Figure 4D shows the decreased suppression of odour-induced spike activity in blue (quantified in Figure 4I ΔMFR, ~25%), but the increased number of suppressed neurones in the PV-Erbb4^-/- group (Figure 4L). Considering that applying ErbB4 inhibitors to the same animals has no selectivity for blocking ErbB4 signalling in PV interneurons, in vivo neural signals typically drift over time when performing pre- and post-inhibition comparisons. Therefore, we compared in vivo electrophysiological data at the population level according to previous studies (PMID: 35173173, PMID:28425486, and PMID:28817803).

(3) At a glance in the heatmap in Figure 4D, excited neurons were reduced in PV-ErbB4-/- mice, but not inhibited neurons. This was different from Figure 4L. The authors need to have a criteria or threshold to show how they categorized each population.

In Figure 4L, red denotes excited neurones, whose number decreased from 66 to 16 in the PV-Erbb4^-/- group, and blue denotes inhibitory neurones, whose number increased from 81 to 148 in the PV-Erbb4^-/- group. This trend is consistent with that observed in Figure 4D. The classification criteria for each population were defined as follows:

“To determine whether the odour evoked a significant response, we used the Wilcoxon signed-rank test to compare the spontaneous (baseline) firing rate and odour-evoked firing rate of each unit–odour pair across all trials; unit–odour pairs with P > 0.05 were defined as nonresponsive, those with P < 0.05 as responsive, and responsive pairs were further categorised as excitatory (odour-evoked firing rate > spontaneous firing rate) or inhibitory (odour-evoked firing rate < spontaneous firing rate). Details are provided in the "Off-line spike sorting and statistics of the unit data" section of Materials and Methods (Lines 810-816).

(4) Figure 4D, 4F and 4J seemed to be inconsistent. In Figure 4D before odor, there was no clear increase in the spontaneous activity in PV-ErbB4-/- mice; in Figure 4F-4G and 4J-4K, clearly, there was a high spontaneous activity in PV-ErbB4-/- mice.

Figure 4D, 4F and 4H shows odour-evoked changes in mean firing rate (ΔMFR), calculated by subtracting the baseline firing rate. By contrast, Figures 4G and 4I show raw spontaneous mean firing rate (MFR), which was elevated in PV-Erbb4^-/- mice. We have revised the description of Figure 4D in both the Results section (Line 222) and figure legends (Line 1217).

(5) What are the neurons recorded in Figure 6E-6F? If they were MCs, loss of ErbB4 in PV neurons should not alter their intrinsic electrical properties. Rather GABAergic inputs could be altered. Indeed, the authors presented a reduction of GABAergic inputs from PV neurons to MCs.

The neurones recorded in Figures 6E-6F are indeed mitral cells (MCs). The firing frequency of action potentials reflects not only the intrinsic electrical properties but also the balance between excitatory and inhibitory synaptic inputs. As demonstrated in this study, the loss of ErbB4 in PV neurones leads to reduced GABAergic synaptic inputs from PV interneurons to MCs, which is consistent with our reasoning and is a key factor contributing to the observed changes in the action potential firing of MCs

(6) Figure 8E-8H, a better experiment would be specifically expressing ErbB4 or PV neurons. In Figure 8F and Figure 8I, was it the excitability after the current injection? Why not perform the spontaneous activity recording?

According to your suggestion, we used optogenetics to activate PV interneurons and recorded inhibitory postsynaptic currents (IPSCs) in MCs. As shown in the revised Figure 9 and Supplementary Figure 1F and 1G, deleting ErbB4 of PV interneurons in the OB weakened the inhibitory modulation of MCs by PV interneurons.

Figures 8F and 8I show the excitability after current injection. We did not measure spontaneous activity, primarily because the spontaneous firing frequency varies significantly among different cells, resulting in poor stability. In contrast, current-injection-induced excitability is more stable and quantifiable, making it more suitable for assessing specific changes in MC responsiveness caused by altered PV-MC inhibitory inputs.

Reviewer #2 (Public review):

Summary:

Hu et al investigate the role of PV neurons and their expression of Erbb4 in olfactory performance through a series of behavioral tests, selective knockout experiments, and in vivo and in vitro electrophysiology. Knockout of Erbb4, either in PV cells or the whole OB, resulted in impairment of discriminating complex odors. The authors present data that inhibition is impaired in MCs, which is likely underlying the abnormal odor-evoked responses of MCs in vivo and the impaired behavioral responses.

Strengths:

Overall, a key strength of this manuscript is the breadth of experiments to test the role of PV Erbb4 expression on circuit dynamics and behavior. The behavioral experiments were clear and sufficiently powered.

Weaknesses:

The major drawback of this manuscript is the lack of depth and rigor in experiments. Some experiments are preliminary, underpowered, and not quantified. As a result, many conclusions of the manuscript are weakly supported in its current form and would require significant revisions to address these shortcomings. Major weaknesses that should be addressed are as follows:

(1) AAV-PV-Cre-GFP is not described or validated. Is this the S5E2 enhancer or something else? What is the specificity and efficacy of this approach in selectively knocking out Erbb4 in PV neurons? Reduced Erbb4 expression in the entire OB with PCR does not validate the selectivity of this approach. At a titer of 10^12, it is unlikely to be specific. Even a small amount of off-target Cre expression will knock out the gene in non-PV cells, so the authors should show whether the gene is knocked out at the single-cell level from PV and non-PC cells. Without validation of this approach, this experiment is no different than the AAV-Cre-GFP experiments.

Thank you for your comment and thoughtful suggestion. The AAV-PV-Cre vector used in this study was purchased from BrainVTA (catalogue number: PT-0275; vector name: rAAV-fPV-CRE-bGH polyA), which contains an fPV promoter to ensure the selectivity and a β-globin intron to enhance Cre expression efficiency. Importantly, the specificity and efficacy of this AAV-PV-Cre construct in selectively targeting PV-expressing neurones have been validated in multiple published studies (PMID:40196908 and PMID:31987494). According to your suggestions, we also performed immunofluorescence staining (co-staining for PV and ErbB4) and found that ErbB4 expression was specifically reduced in PV-positive interneurons in OB tissues injected with AAV-PV-Cre, while ErbB4 levels remained unchanged in PV-negative cells (the revised Figures 9B and 9C). This single-cell-level validation directly demonstrated that our AAV-PV-Cre-mediated knockout approach specifically targets ErbB4 in OB PV interneurons, thereby avoiding off-target effects in non-PV cells.

(2) Figure 1D - three mice per group is insufficient. There is no control group error (the same as Figure 9). Why is it a paired t-test when there is a control group? The authors should be comparing go/go vs. go/no-go. The methods for normalization are unclear and are likely to hide the fact that n=3 is insufficient to capture a difference without extra measures to normalize the data.

Thank you for the comment. According to your suggestions, we expanded the sample size to four mice per group and modified the statistical test from a paired t-test to an unpaired t-test in the revised Figure 1C and 1D. Since not all samples were exposed to the same blot for direct comparison, we performed normalisation of each sample relative to its internal control prior to statistical analysis to ensure the reliability of group comparisons. There was an error regarding the control group in the revised Figure 1D. Detailed information on the normalisation method and updated statistical results have been provided in the revised figure legends (Lines 1116-1118).

(3) The analysis of LFP is limited. During what period was this quantified? Are there any differences in task-related LFP changes? Also related to in vivo electrophysiology, the authors should show examples of isolated units, including their waveforms and how units were clustered and assigned to M/TCs.

Thank you for the comment. The spectral power of each frequency band was quantified using a 4-s sliding window during the pre-odour stimulation period (4–0 s before odour onset). Regarding task-related LFP changes, the initial recordings were affected by unavoidable interference, leading to unstable signals that precluded reliable analyses. Based on your suggestions, we have supplemented the revised manuscript with examples of isolated single units, including their representative waveforms, clustering results (e.g. principal component analysis plots), and unit assignment criteria (the revised Figure 4—figures supplement 1A and 1B). As previously noted, the tetrodes were placed in the mitral cell layer based on stereotaxic coordinates. Consistent with established literature (Kay and Laurent, 1999; Doucette et al., 2011), the putative cell type recorded using this approach was mitral/tufted cells (M/TCs). Detailed methodological parameters for spike sorting, unit isolation, and cell type identification are provided in the "Materials and Methods" section (Lines 764-767 and Lines 803-805).

(4) The authors use 80pA and 100pA to elicit equivalent AP spiking in MCs to determine if recurrent inhibition differs, but do not actually show that AP spiking is the same across groups. This should be quantified.

There seems to be a prominent increase in the firing of MCs in PV-Erbb4+/+ mice before odor presentation, but not in PV-Erbb4-/- mice. What is the significance of this?

Thank you for your valuable comment and suggestion. According to your suggestions, we quantified the AP firing frequency of MCs in both groups under the respective stimulation currents, and statistical analysis confirmed that there was no significant difference in the AP spiking levels between the two groups (the revised Figure 8—figure supplement 1A).

We speculate this prominent pre-odour firing increase in PV-Erbb4^+/+ mice (but not in PV-Erbb4^−/− mice) stems from interference caused by the opening of the solenoid valve during odour clearance prior to stimulation—this artefact was present in the initial recordings of control mice but was eliminated during the data acquisition for PV-Erbb4^−/− mice. This was not statistically significant. To avoid this interference, we strictly defined the analysis windows, and the spontaneous firing rate was calculated by averaging the spikes recorded during the 4–2 s period before odour stimulation, excluding the interference-prone interval immediately before odour onset. This rigorous window selection ensured the reliability of our comparisons of spontaneous and odour-evoked firing between the groups.

(5)There is a disconnect between the in vivo firing rates of MCs and ex vivo firing rates. In slice, the authors note that the spontaneous activity of MCs is elevated in the KO, but this is not observed in vivo, where conditions are physiological. Therefore, it is unclear whether the concept of signal-to-noise changes in slice (higher spontaneous, lower evoked), indeed translate to something in vivo. It would be important to know what the PV cells are doing in vivo. Perhaps they have low firing rates prior to odor onset, which may explain the lack of observed difference in baseline FRs in MCs. The authors should have this data in their tetrode recordings, which would offer insight into when inhibition is recruited.

Thank you for your valuable comments, which helped clarify the consistency between in vivo and ex vivo electrophysiological data. In ex vivo slice recordings, we measured both spontaneous and evoked excitatory firing rates of MCs, and found that the spontaneous activity of MCs was elevated in PV-Erbb4^−/− mice (Figures 6A and 6B). Importantly, this observation is consistent with our in vivo data: M/TCs in PV-Erbb4^−/− mice exhibited significantly higher spontaneous firing rates during excitatory responses to odour presentation (Figures 4F and 4G). However, for inhibitory responses during odour presentation, M/TCs in PV-Erbb4^−/− mice showed significantly lower spontaneous firing rates in vivo (Figures 4H and 4I). Thus, the overall in vivo spontaneous activity of MCs appeared unchanged owing to opposing changes in excitatory and inhibitory subpopulations, and there was no true discrepancy between the ex vivo and in vivo results. We agree that in vivo recording of PV interneuron activity would be invaluable for understanding their functional state under physiological conditions. Unfortunately, this is technically unfeasible with the current experimental setup, primarily because of the sparse distribution of PV interneurons in the OB. As a complementary alternative, we characterised the key electrophysiological properties of PV interneurons in ex vivo slices (e.g. action potential and mEPSC, the revised Figure 7—figure supplement 1), which will provide additional insights into the underlying cellular mechanisms.

(6) Since PV neurons are required for gamma oscillations, why is it that KOs have higher gamma oscillations? Is it indeed the case that PV cells have a hypofunctional phenotype in this model? Again, recording from PV cells in vivo would help make sense of this.

Thank you for your valuable comments, which were critical for clarifying the functional consequences of ErbB4 knockout in PV interneurons. The elevated gamma oscillations in the OB of PV-Erbb4^-/- mice appear to be primarily driven by the increased action potential firing frequency of PV interneurons that we observed in OB slice recordings (the revised Figure 7—figure supplement 1). A similar phenomenon has been reported in the hippocampus (PMID:24050403).

Indeed, the PV cells displayed a hypofunctional phenotype in this model. Our optogenetic experiments provided direct evidence for this (the revised Figure 9—figure supplement 1). Specifically, optogenetic activation of PV interneurons in AAV-PV-Cre-injected mice (selective ErbB4 knockout in OB PV interneurons) induced weaker inhibitory postsynaptic currents (IPSCs) in MCs than in controls (the revised Figure 9—figure supplement 1). This indicates that ErbB4 loss induces PV interneurons to be more excitable, but impairs their ability to inhibit neighbouring MCs.

We acknowledge that in vivo recordings of PV interneurons would further clarify their functional state under physiological conditions. As noted earlier, this remains technically challenging using our current setup; however, our ex vivo mechanistic data and optogenetic findings collectively support these conclusions.

(7) A clearer picture of how PV cell inhibition changes with Erbb4 KO would be achieved with optogenetically evoked IPSPs, rather than changes in mini frequency.

Thank you for the comment. Based on your suggestions, we used optogenetics to activate PV interneurons and recorded the optogenetically evoked inhibitory postsynaptic currents (oIPSCs) in MCs. As shown in the revised Figure 9 and Supplementary Figure 1F and 1G, deleting ErbB4 in PV interneurons in the OB weakened the inhibitory modulation of MCs by PV interneurons.

Reviewer #3 (Public review):

Summary:

The authors investigate the role of ErbB4 in parvalbumin (PV) interneurons within the olfactory bulb (OB) and its regulation of odor discrimination behavior in mice. They demonstrate that odor discrimination increases ErbB4 kinase activity and that the loss of ErbB4 in the OB impairs the dishabituation of odor response and discrimination of complex odors. The study also characterizes the expression of ErbB4 in the OB, showing it is enriched in PV neurons. Furthermore, the authors utilize a mouse model in which ErbB4 is knocked out in PV neurons and perform a variety of behavioral, electrophysiological, and local field potential (LFP) recording experiments to characterize alterations in olfactory bulb activity. They then use a model in which ErbB4 is specifically knocked out in PV neurons in the OB and show that this manipulation disrupts odor-related behaviors in mice.

Strengths:

The study's strengths lie in its use of a diverse range of techniques, including RNAscope, IHC, and Western blotting, to assess the presence of ErbB4 in PV neurons within the OB. Additionally, the authors employ various behavioral tests to evaluate the effects of ErbB4 manipulation in different mouse models, alongside comprehensive electrophysiological experiments and LFP recordings to examine the impact of these manipulations on OB physiology.

Weaknesses:

While the data presented in this paper are interesting, several major concerns reduce my enthusiasm for this study, as outlined below:

(1) In reviewing Figure 1C/D, there are several concerns regarding the clarity and interpretation of the data:

(a) While the Western blot for ErbB4 in other figures (Figure 1F, 2I) of the manuscript shows a clear single band, the blot presented in Figure 1C (for both p-ErbB4 and total ErbB4) shows multiple bands, which is unexpected. This discrepancy raises concerns about the consistency of the results.

(b) The data presented in Figure 1D uses only 3 mice per group, and the reported p-value of 0.0492, while technically significant, is very close to the threshold. This raises concerns about the robustness of the finding, especially given the small sample size. Additionally, the p-ErbB4 band intensity in the Go/No-Go condition in Figure 1C does not appear to show a clear increase over the Go/Go condition, which is not congruent with the bar graph in Figure 1D showing a 50% increase in p-ErbB4/ErbB4 levels.

(c) It is a standard practice in many journals to include full, uncropped Western blot images as supplementary material. This transparency helps ensure that no bands are selectively shown or omitted and increases confidence in the presented data.

Thank you for your comment and thoughtful suggestion.

(a) The multiple bands observed for both p-ErbB4 and total ErbB4 in the original Figure 1C are likely attributable to the high concentration of the primary antibody used initially (Santa Cruz Biotechnology), which has since been discontinued. In the revised manuscript, we have replaced the original antibody with an anti-ErbB4 antibody (Cat# 4795, Cell Signalling Technology), which showed a single ErbB4 band (the revised Figure 1C). To detect p-ErbB4, we first performed immunoprecipitation with a phosphorylated tyrosine antibody to enrich the total phosphorylated proteins, followed by immunoblotting for ErbB4.

(b) According to your comments, we have expanded the sample size to four mice per group and modified the statistical test to an unpaired t-test in the revised Figure 1D.

(c) We included all relevant uncropped blot images in the Supplementary Material (the revised Figure 1, 2—figure supplement 1).

(2) In Figure 2, the authors used the anti-ErbB4 antibody sc-283 from Santa Cruz to assess the expression of ErbB4 in PV neurons and the absence of its expression in PV-ErbB4 knock-out mice. However, this particular antibody has been shown to produce non-specific bands in Western blotting and also generate non-specific labeling in IHC. This non-specificity has been demonstrated in Vullhorst et al. (2009, J Neurosci), raising significant concerns about the reliability of the data generated using this antibody.

Thank you for the comment. As you have indicated, the non-specificity of the anti-ErbB4 antibody sc-283 has been reported in J Neurosci (Vullhorst et al, 2009). These antibodies have been widely used in numerous recent studies (PMID:39543371 and PMID:41297670), indicating their utility for assessing ErbB4 expression under appropriate experimental conditions (primary antibody dilution and incubation conditions). As shown in the revised Figure 2I and 2J, the absence of ErbB4 expression in PV-ErbB4 knock-out mice proved the specificity of the antibody for immunofluorescence staining. To address these potential concerns, we conducted a rigorous validation tailored to our model. In the revised manuscript, we complemented the protein-level detection with RT-PCR, which directly validated the specific knockout of Erbb4 mRNA in PV neurones (the revised Figure 2F-H). Additionally, we performed follow-up immunofluorescence staining experiments using a well-validated anti-ErbB4 antibody (19943-1-AP, Proteintech; see PMID: 37060105) to assess the expression of ErbB4 in PV neurones (the revised Figure 9B and 9C).

(3) In reviewing the statistical analysis for the series of odor discrimination tests, there could be a potential issue with the clarity of the significance testing. Although the figure legend reports the F and p values from the two-way ANOVA, it is unclear whether these values represent the main effects or the results of a post hoc test. Additionally, it is not clear whether the asterisk in the figures reflects significance from a post hoc test or from the overall ANOVA. The methods section does not explicitly state whether a post hoc test was performed to assess differences between the knockout and control groups. Given that the tests were conducted across multiple days or conditions, a post hoc test that can adjust for multiple comparisons would be necessary to accurately identify where specific differences between the groups exist.

Thank you for your comment and suggestion. For the odour discrimination tests, two-way ANOVA was followed by Tukey’s post hoc adjustment for multiple pairwise comparisons, which is explicitly stated in the revised Methods section (Lines 832-833). According to your suggestions, we have revised the figure asterisks to indicate significant post hoc differences between the groups.

(4) Throughout the manuscript, the authors use different mouse models, including ErbB4 knockout specifically in the OB (AAV-Cre-GFP), ErbB4 knockout in PV interneurons throughout the brain (PV-ErbB4-/-), and ErbB4 knockout in PV interneurons within the OB (AAV-PV-Cre-GFP). For Figures 4 and 5, the authors use the PV-ErbB4-/- model to examine odor-evoked activity and neural oscillations within the OB. Since the knockout affects PV interneurons across the entire brain, it is difficult to disentangle whether the observed changes in the OB are due to local effects or broader network alterations elsewhere in the brain.

Thank you for the comment. We first clarify the rationale for initially using the whole-brain PV-Erbb4^-/- model: this model is well-validated and widely adopted in in vivo electrophysiological studies targeting brain region-specific phenomena, such as regulating synaptic plasticity in the hippocampus (PMID:24050403) and modulating sensory coding in the cortex (PMID: 37069344).

Additionally, PV interneurons in the OB are local-projection neurones that exert stronger and more extensive inhibitory effects on M/TCs, a key feature that restricts their functional impact primarily to the local OB circuit rather than broader cross-brain network interactions, further supporting the relevance of the global model for OB-focused research.

Furthermore, local AAV-Cre injections have notable experimental limitations in our electrophysiological recordings. Stereotaxic infusion induces mechanical trauma and local inflammation in the OB, altering basal circuit activity independent of ErbB4 deletion. AAV also yields incomplete mosaic gene ablation among PV interneurons, introducing a variable knockout efficiency that complicates the interpretation of local electrophysiological readouts. In comparison, the global PV-Erbb4^-/- line mediates uniform, complete ErbB4 deletion across all PV interneurons, avoiding viral artefacts and rendering our OB electrophysiological measurements more reliable for assessing endogenous local circuit function.

(5) While the electrophysiological experiments shown in Figures 6-8 provide valuable insights into the reduced inhibition to MCs in PV-ErbB4 knockout mice, it appears that the authors did not record from PV interneurons themselves. Since PV interneurons are central to the proposed mechanism, directly recording them would provide critical information on how the ErbB4 knockout affects their intrinsic properties, synaptic inputs, and firing behavior. Without these direct recordings, the conclusions about the specific role of PV neurons in regulating MC activity remain somewhat indirect. Prior studies have established that knockout of ErbB4 in PV interneurons reduces mEPSC frequency in PV neurons (Del Pino et al., 2013).

Thank you for your valuable comment and suggestion. According to your suggestions, we have supplemented the electrophysiological experiments (the revised Figure 7—figure supplement 1) with direct recordings from PV interneurons in ex vivo slices. These recordings revealed that ErbB4 knockout alters some properties of PV interneurons (e.g. action potential frequency), reduces mEPSC frequency (consistent with Del Pino et al., 2013), and modulates their firing behaviour, directly confirming that ErbB4 deletion affects PV interneurons.

(6) In Figure 9, the authors knock out ErbB4 in PV neurons in the OB with AAV-PV-Cre-GFP and show with western blotting that ErbB4 expression is reduced in the mouse injected with AAV-PV-Cre-GFP. However, it is not clear whether ErbB4 was selectively knocked out in PV neurons without the quantification from IHC assays.

Thank you for the comment. To verify the selective knockout of ErbB4 in OB PV neurones, we performed an additional experiment: immunofluorescence co-staining for PV and ErbB4; quantitative analysis confirmed a significant reduction of ErbB4 in PV interneurons in AAV-PV-Cre–injected mice (the revised Figure 9B and 9C).

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Other concerns:

(1) No supplemental Figure 1 was shown.

We apologise for the oversight. Supplemental Figure have been included in the revised manuscript (the revised Figure 2—figure supplemental 2).

(2) In line 59, page 3, abbreviations of mitral/tufted cells (M/TCs) or "MCs" were confusing.

We have clarified the abbreviation by explicitly defining "mitral cells (MCs)", with "M/TCs" retained for mitral/tufted cells to avoid ambiguity (Line 56-59).

(3) The quality of blots in Figure 1C was less satisfactory, with lots of noisy bands. Maybe antibodies from Cell Signaling should be used.

We optimised the experiment as suggested. In the revised Figure 1C, we have used an anti-ErbB4 antibody (Cat# 4795, Cell Signalling Technology) combined with the immunoprecipitation-immunoblotting workflow.

(4) The authors should describe the serotype as well as the promoters of AAV-Cre-GFP used in the experiments.

As you suggested, we have supplemented the details of AAV-Cre-GFP in the Methods section: Viruses driven by the CAG promoter, including pAOV-CAG-EGFP-T2A-Cre (catalogue No. CN396, serotype AAV2/8) and pAOV-CAG-EGFP-2A (catalogue No. CN394, serotype AAV2/8) were purchased from Obio Technology, Shanghai, China (Lines 568-571).

(5) The authors should describe the binary mixtures of isoamyl acetate Vs. Heptanone. Simply 6/4 Vs. 4/6 is not easy to follow.

As you suggested, we have clarified the binary odour mixtures in the Results section: (60%/40% binary mixtures of these odours, "6/4" refers to 60% isoamyl acetate + 40% heptanone, and "4/6" refers to 40% isoamyl acetate + 60% heptanone, in volume) (Lines 111-113).

(6) Figure 1G, "0 limonene" AAV-Cre-GFP group did not show the error bar.

We apologise for the omission—error bars for the "0 limonene" AAV-Cre-GFP group have been added to the revised Figure 1.

(7) Figure 2A, the green signal for ErbB4 was not that strong, it was hard to visualize the expression of ErbB4 in five layers of OB.

Thank you for the comment. The revised Figure 2A shows an enhanced green signal for ErbB4, clearly indicating its distribution in all OB layers.

(8) The authors should clearly label or indicate the cell type of neurons recorded in Figure 6, Figure 7, and Figure 8A-8C.

As you suggested, we have clearly labelled the cell types in the revised Figures 6, 7, and 8A–8C: all recorded neurones are mitral cells (MCs).

(9) line 320-321, page 14, "Figures 8F and 8G, a burst of conditioning 320 stimuli (40 pulses at 100 Hz)", the authors should indicate the stimuli as a tick or scale on Figures 8F and 8I.

As per your suggestion, we have added scales and vertical ticks to the revised Figures 8F and 8I to indicate the timing of the conditioning stimuli (40 pulses at 100 Hz).

Reviewer #2 (Recommendations for the authors):

(1) Figure 1F is not quantified.

As suggested, we have supplemented the quantitative analyses of Figure 1F (see the revised Figure 1G) and the statistical details in the revised legend.

(2) Figures 2G and H are not quantified.

As suggested, we have supplemented the quantitative analyses of Figures 2G and 2H (see the revised Figures 2H and 2J), and statistical details are provided in the revised figure legends.

(3) Bar plots with standard error are not suitable for visualizing the distribution of data. Either show violin plots, histograms, or raw data points throughout.

As suggested, we included the raw data points in the revised manuscript.

(4) Figure 3 legend - (Wen et al., 2010) citations repeated.

Thank you for the comment. We have removed the repeated (Wen et al., 2010) citation in the revised manuscript.

(5) Does Ebbr4 deletion change the waveform shape of PV+ cells? This could influence spike sorting and cell classification.

Thank you for the comment. According to your recommendations, we performed additional ex vivo patch-clamp recordings of PV+ interneurons, which showed that ErbB4 knockout did not alter their action potential waveform shape (e.g. spike amplitude and half-width, the revised Figure 7—figures supplement 1D and 1E). Importantly, this knockout also did not directly affect the spike waveform shapes of M/TCs in vivo (e.g. peak-to-valley ratio and half-valley width, the revised Figure 4—figures supplement 1C-1E). Therefore, we speculate that spike sorting and cell classification were not affected.

(6) Line 327 - is this really a measure of inhibitory circuit plasticity?

Thank you for the comment. We have replaced "inhibitory circuit plasticity" with "inhibitory circuit activity" and comprehensively updated all relevant instances throughout the manuscript to ensure consistent use of the term.

(7) In the discussion: "Another interesting finding is that the changes in spontaneous firing rate differed for odor-evoked excitatory and inhibitory responses in vivo, suggesting that ErbB4 in PV interneurons dynamically regulates the spontaneous activity of M/TCs in response to different odors." The significance of this finding is not discussed and it is not clear what this result means. For example, does this relate to a "winner-take-all scenario"?

Thank you for the comment. According to your recommendations, we have discussed the significance of this finding in the revised Discussion section (Lines 438-449):

This result reveals a dual regulatory role of ErbB4 in PV interneurons in olfactory coding. For odour-evoked excitatory responses, PV interneurons (via ErbB4) constrain M/TC spontaneous firing at a low baseline, minimising background noise and enhancing the SNR. In contrast, the inhibitory responses maintain a higher spontaneous firing baseline to accentuate odour-evoked suppression. This regulation amplifies the "excitation of excited M/TCs and inhibition of inhibited M/TCs," a hallmark of the OB’s "winner-take-all" mechanism that sharpens odour identity coding, which is critical for odour discrimination. These data confirm that ErbB4 in PV interneurons does not exert a global modulation, but dynamically tailors M/TC spontaneous activity to odour-evoked responses, thereby optimising the ability of the OB to resolve olfactory cues.

(8) The discussion on linking AD with odor discrimination impairments, and the potential role of Erbb4 was interesting. However the discussion of "fluctuation status" is not defined. What does this refer to?

According to the comment, we have added the following wording to the Discussion section (Lines 478-497):

As referenced by Iemi et al. (2022), "fluctuation status" herein refers to the spontaneous, endogenous fluctuations in neural activity that occur in the brain prior to sensory stimulation (i.e. pre-stimulus) or in the absence of external sensory input. External cues do not drive this state but arise from intrinsic neural dynamics, and it is primarily characterised by changes in the power of neural oscillations across multiple frequency bands (e.g., alpha, beta, and their neighbouring frequencies, collectively termed "alpha+ band" in Iemi et al. (2022)). Critically, this fluctuation status is distinct from stimulus-evoked neural responses; it reflects a pre-existing "baseline" state of the neural system that modulates subsequent sensory processing and behaviour. Consistent with Iemi et al. (2022), previous work has established that such ongoing neural fluctuations strongly predict behavioural variability in both health (e.g. slower reaction times and reduced perceptual accuracy) and disease (e.g. age-related cognitive decline and schizophrenia) by regulating neuronal excitability. In our study, we extended this concept to olfactory function, focusing on "broad ongoing fluctuations" (encompassing multiple frequency bands) and linking their dynamics to OB output. Specifically, these fluctuations influence the ability of the OB to support rapid recovery from olfactory adaptation and to mediate complex odour discrimination, thereby uncovering a novel role for endogenous neural fluctuation status in olfactory circuit function.

Reviewer #3 (Recommendations for the authors):

(1) For Figure 1C, I strongly recommend that the authors include full, uncropped Western blot images for both p-ErbB4 and total ErbB4 as supplementary material to improve the transparency and credibility of the findings.

As you suggested, we have supplemented the full, uncropped western blot images for both phosphorylated ErbB4 (p-ErbB4) and total ErbB4, as the revised Figures 1 and 2 and Supplementary Figure 1.

(2) The authors should either validate the specificity of the sc-283 antibody in their experimental conditions or consider using a more specific antibody for ErbB4 to ensure the reliability of their data.

Thank you for your comments. As shown in the revised Figure 2I and 2J, the absence of ErbB4 expression in PV-ErbB4 knock-out mice proved the specificity of the antibody for immunofluorescence staining. To address these potential concerns, we conducted a rigorous validation tailored to our model. In the revised manuscript, we complemented the protein-level detection with RT-PCR, which directly validated the specific knockout of Erbb4 mRNA in PV neurones (the revised Figure 2F-H). Additionally, we performed follow-up immunofluorescence staining experiments using a well-validated anti-ErbB4 antibody (19943-1-AP, Proteintech; see PMID: 37060105) to assess the expression of ErbB4 in PV neurones (the revised Figure 9B and 9C).

(3) I recommend that the authors clarify in the methods section whether a post hoc test with multiple comparison adjustments (e.g., Tukey's HSD or Bonferroni) was conducted for the behavioral tests presented throughout the study. If no such test was performed, it would be beneficial to include one to ensure accurate identification of significant differences between groups across the different experimental days or conditions.

For the behavioural tests, we confirmed that two-way ANOVA was followed by Tukey’s post hoc adjustment for multiple pairwise comparisons, which has been clarified in the revised Methods section (Lines 832-833). Additionally, we have revised the asterisks in the relevant figures to clearly indicate significant post-hoc differences between the groups.

(4) The rationale for using a whole-brain PV-specific knockout model (PV-ErbB4-/-) to assess OB-specific phenomena in Figures 4 and 5 should be clarified. I also recommend the authors provide or conduct additional experiments with OB-specific knockouts to determine whether the effects observed in Figures 4 and 5 are directly attributable to the loss of ErbB4 in PV interneurons within the OB.

To clarify the rationale, the whole-brain PV-Erbb4^−/− mouse line was initially used to characterise the overall functional relevance of ErbB4 in PV interneurons for OB activity and odour-evoked responses. We clarified this rationale in the revised Results section (Lines 210-211 and Lines 250-251).

As elaborated above, AAV-mediated OB-specific knockouts suffer from injection-induced tissue injury, incomplete mosaic recombination, and uncertainties in local electrophysiological recordings, which would compromise data reliability. Moreover, OB PV interneurons act predominantly as local circuit neurones; therefore, the phenotypes observed in Figures 4 and 5 most plausibly arise from intrinsic ErbB4 deletion within bulbar PV cells rather than from long-range brain-wide network changes. For these reasons, we consider our current global PV-Erbb4^−/− model appropriate to interpret the local OB circuit effects here, and we outline OB-targeted knockout manipulations as a direction for future independent follow-up research.

(5) To strengthen their conclusions drawn from Figures 6-8, the authors should perform direct recordings from PV interneurons in the OB to assess how the ErbB4 knockout affects their intrinsic properties, synaptic inputs, and firing behavior. This would provide more direct evidence of PV interneuron dysfunction in the knockout mice.

According to your recommendations, we have performed direct recordings of OB PV interneurons using ex vivo patch-clamp electrophysiology (the revised Figure 7—figures supplementary 1). These recordings reveal that ErbB4 knockout affects the intrinsic properties of PV interneurons, synaptic inputs (reduced mEPSC frequency and amplitude, the revised Figure 7—figures supplementary 1A and 1B), and firing behaviour (increased action potential frequency, the revised Figure 7—figures supplementary 1C), directly confirming that ErbB4 deletion affects PV interneurons themselves. These results provide direct evidence of PV interneuron dysfunction in the knockout mice.

(6) I recommend that the authors assess molecular markers of PV interneuron inhibition, such as GAD67, in PVIs. This would help link the ErbB4 knockout to impaired PV interneuron function and its downstream effects on MC inhibition.

Thank you for the comment. We attempted to detect GAD67 in OB PVIs using multiple commercial antibodies but encountered consistent technical limitations (e.g. weak signal specificity and high background) that precluded reliable quantitative analysis.

To address this gap, we used optogenetics to activate PV interneurons and recorded optogenetically evoked inhibitory postsynaptic currents (oIPSCs) in MCs. As shown in the revised Figure 9 and Supplementary Figure 1F and 1G, deleting ErbB4 of PV interneurons in the OB weakened the inhibitory modulation of MCs by PV interneurons. This result provides direct electrophysiological evidence linking ErbB4 knockout to impaired PV interneuronal function and its downstream effects on MC inhibition.

(7) For Figure 9, I recommend that the authors perform IHC to show that ErbB4 is selectively knocked out in PV neurons but not in non-PV neurons within the OB, as the Western blot analysis alone does not provide this specificity.

According to your recommendations, we performed an additional key experiment: immunofluorescence co-staining for PV and ErbB4, with quantitative analysis confirming a significant reduction in ErbB4 in OB PV interneurons, but not in non-PV cells, in AAV-PV-Cre–injected mice (the revised Figure 9B and 9C).