Rhythmic coordination and ensemble dynamics in the hippocampal-prefrontal network during odor-place associative memory and decision making

Article activity feed

1. 

   Version published to 10.7554/elife.79545 on eLife

   Dec 8, 2022
2. 

   eLife

   Nov 25, 2022

   Author Response

   Reviewer #1 (Public Review):

   The authors performed simultaneous extracellular recordings in brain regions (CA1, prefrontal cortex (PFC), olfactory bulb (OB)) that are key to odor-guided decision making to delineate the oscillatory and cell population dynamics that guide decision making based on learned associations. They used complementary analyses to assess the coordination between CA1 and medial PFC (mPFC), using coherence and phase-locking analysis as well as generalized linear models and Bayesian decoding methods.

   One of the strengths of this work is the comparison of beta and respiratory (RR) LFP coherence in several behavioral states to rule out confounds due to sniffing or preparatory motor behavior (e.g., coherence was assessed during decision making with and without an odor present, during reward …

   Author Response

   Reviewer #1 (Public Review):

   The authors performed simultaneous extracellular recordings in brain regions (CA1, prefrontal cortex (PFC), olfactory bulb (OB)) that are key to odor-guided decision making to delineate the oscillatory and cell population dynamics that guide decision making based on learned associations. They used complementary analyses to assess the coordination between CA1 and medial PFC (mPFC), using coherence and phase-locking analysis as well as generalized linear models and Bayesian decoding methods.

   One of the strengths of this work is the comparison of beta and respiratory (RR) LFP coherence in several behavioral states to rule out confounds due to sniffing or preparatory motor behavior (e.g., coherence was assessed during decision making with and without an odor present, during reward consumption). These controls allowed the authors to identify a specific enhancement of beta compared to RR coherence during decision making.

   The analyses of task-responsive putative interneuron and pyramidal cells suggest that accurate decision-making is associated with a stronger modulation of beta phase-locking in interneurons. Additional cross-correlation analyses between cell types across regions showed that cells, particularly interneurons, are temporally coordinated in the beta range. Their analyses did not identify a mechanism for this coordination, but the temporal lags between PFC and CA1 cells raise the possibility of top-down interactions mediated by a third brain region.

   The authors used the cellular activity to determine that the animal's upcoming behavior could be predicted from the ensemble activity during decision-making a few hundred milliseconds before the behavioral choice, but decoding accuracy diminished soon after the decision-making period. Interestingly, decoding accuracy increased after decision-making when using the spatially active cell ensembles. As indicated by the authors, these results suggest that different cell ensembles are engaged during decision-making and during the execution of the decision. It is possible that this change in ensemble dynamics before and after decision-making relates to the familiarity of the animals with the task, which makes it likely to involve procedural components (e.g., Jog et al., 1999). As pointed out by the authors in the discussion, several results have implications for the formation of associative memories and provide clues for future experiments. Thus, future work looking at the ensemble dynamics and at the occurrence of CA1 ripples in the early stages of task learning compared to when the animals are very familiar with the task (as in the current study), will provide a better understanding of the shifts that develop during the formation and consolidation of the association.

   One of the considerations in interpreting the results is that the odor sampling and decisionmaking periods overlap, making it difficult to disentangle the neural dynamics that are driven by the recall of the association (cued retrieval) and those that relate to the upcoming turning behavior after odor port disengagement. However, the author's analyses of odor and choice selectivity in correct and incorrect trials demonstrate a preferential association between spike activity and choice selection in this task.

   Overall, the results advance our understanding of odor-guided decision-making mechanisms in CA1 and PFC at the LFP and cell population level. This work will be of significance to further research on the cellular basis of memory-guided decision-making, and to future work characterizing the interactions between CA1 and PFC during learning.

   We thank the Reviewer for their detailed evaluation summarizing and highlighting the strengths of the study. In addition to beta and respiratory rhythm (RR) modulation of CA1-PFC activity and the relationship between spiking activity and choice selection, the Reviewer also highlighted the temporal coordination of CA1 interneurons and change in ensemble dynamics during the decision-making period at the odor-port vs. during the execution of the decision on the maze, which is further emphasized as a novel result in the revised manuscript.

   Reviewer #2 (Public Review):

   Symanski et al. investigated the communication between the medial prefrontal cortex (mPFC), the hippocampal CA1 region, and the olfactory bulb (OB) while rats underwent an odor-cued decision-making task. By recording local field potentials and spiking activity in the three regions, they found that all regions became synchronized at the beta band and respiratory rhythms during cue sampling/decision-making. Although the strength of inter-region synchrony was not predictive of correct choices, both CA1 and mPFC neurons showed stronger phase-locked firings to beta oscillations for correct than incorrect choices. Moreover, a subset of putative pyramidal and interneurons in both regions were selective for task variables, and as ensembles, they formed activity patterns differentiating choices. Also, their firings were temporally coordinated in a direction that the mPFC interneurons led CA1 interneurons and pyramidal neurons. Based on these findings, the authors propose that cue-evoked beta oscillations modulate the activity of interneurons to coordinate ensemble activity in CA1-mPFC networks supporting decision-making.

   Strength:

   The findings uncovered a new style of mPFC-Hippocampal communication through odorevoked beta oscillations, which contrasts with theta oscillations and sharp-wave/ripples reported during memory-guided spatial navigation tasks. The overall quality of the work is outstanding. The data collection and analysis were meticulously conducted with appropriate controls and statistical tests.

   Weakness:

   The initial analysis of LFP activity (Figure 2d) revealed strong coherence in the beta band in all region pairs; however, the subsequent analysis focuses on mPFC-CA1 interaction. To justify this approach, it is essential to establish that the mPFC-CA1 beta synchrony reflects their direct communication rather than a by-product of common inputs from the OB.

   The authors used cross-correlograms to reveal the directionality of mPFC-CA1 interaction. To strengthen the author's view that beta oscillations help coordinate neural activity, it is worth investigating if the same temporal relationship is also detectable within each cycle of beta oscillations. Specifically, mPFC interneurons may fire at earlier phases, followed by firings of CA1 interneurons and pyramidal neurons at later phases.

   We thank the Reviewer for their positive evaluation and constructive comments. We have addressed the weaknesses noted in the revised manuscript. In particular, we have added analyses and text that emphasize the change in beta synchrony in the OB-CA1PFC network during the task, and added analyses that examine phase locking of pyramidal cells and interneurons to beta rhythms in the mPFC, CA1 and OB.

   Reviewer #3 (Public Review):

   Symanski et al. describe a set of interesting results derived from analyzing electrophysiological recordings performed in rats well trained on an associative memory task on a spatial maze (a T maze), in which animals learned to associate a given odor delivered in an initial maze region (upon a nose poke) with a subsequent spatial choice (a left or a right turn) to receive a reward. The authors have obtained LFPs from the OB, PFC, and CA1 from 8 animals subjected to this task, along with single-unit activity from the PFC and CA1. The authors describe that, during odor sampling, there is prominent LFP activity in the beta range (20-30 Hz) as well as prominent activity of the respiration-entrained LFP rhythm (RR, 7-8 Hz). The authors convincingly show that beta activity - but not RR - is specific to odor sampling (RR also shows up during other immobility periods within the task and when animals breathed clean air). They further show that not only beta power but also inter-regional beta coherence significantly enhances during the odor sampling period. In addition, the authors find a higher beta phase modulation of spiking in a subset of neurons associated with subsequent correct decisions. Since the authors also prove - based on behavioral analysis - that the odor-sampling period corresponds to the decisionmaking period in this task, they propose a role for beta coordination of hippocampal-prefrontal networks in sensory-cued decision making. The paper also brings along a set of complementary findings looking at the single unit and ensemble activity in both regions (CA1 and PFC), which are capable of predicting future spatial choices.

   I consider the investigated topic relevant to modern neuroscience and likely to interest a broad audience. Nevertheless, while there is much to like about this paper (e.g., carefully done experiments, advanced computational data analyses, well-written text, and well-crafted figures), I caught some issues that called my attention upon a careful reading, which I list below:

   A) The paper is written in a way clearly centered on rhymical brain activity (c.f. title, abstract, introduction, and discussion). Yet, out of 7 main figures, only 2 of them show data related to oscillations (while 1 figure shows behavioral data and 4 figures show spiking analysis not related to brain rhythms). Therefore, the presentation of the results seems unbalanced and disconnected from the main story.

   B) Somewhat related to the point above, in a strict sense, the title is not well justified ("Rhythmic coordination of hippocampal-prefrontal ensembles (...)") since there is no analysis relating assembly activity with either beta or RR (their results show beta or RR modulating a subset of single units), nor there is a combined ensemble analysis of PFC and hippocampal units (i.e., interregional cell assemblies). Why not try to relate ensemble activity to the observed oscillations?

   C) The main result of increased interregional beta coherence specifically during odor sampling is very interesting and seems quite solid. Though I hate being the one raising questions about the level of advancement, I cannot avoid noticing that similar increases in beta coherence in odor-sampling-based tasks have been observed before (e.g., increased OB-HPC beta coherence during odor sampling has been shown in Martin et al 2007 and between LEC and HPC in Igarashi et al 2014), which is to say that there is overlap between this core finding and previous research. But that said, in times where the reproducibility of our scientific endeavor has been put into question, this particular reviewer favors the publication of similar findings by independent labs, especially given this neatly collected dataset. It is recommended to highlight which results constitute novel insights here and which results provide support for previously published results.

   D) It called to my attention that many of the spiking results were obtained for a small percentage of neurons. For instance, how can the authors be confident that the choice-selective neurons are actually coding for the choice as opposed to being randomly detected by statistical chance? As a case in point, the authors mention that 1309 units were recorded in CA1, and from these 42 cells were choice selective. If the authors have employed a typical alpha of 5% for detecting such neurons, chance alone would predict ~60 neurons being false positives. I apologize if I am missing something, but could the authors clarify? On a related note, even though most findings hold true for a small percentage of neurons, the writing also tends to generalize the findings to the whole population (e.g., "Beta phase modulation of CA1 and PFC neuronal activity during this period was linked to accurate decisions, suggesting that this temporal modulation influences sensory-cued decision making.").

   We thank the Reviewer for their detailed comments and feedback. We have addressed the issues raised by the Reviewer, which has significantly strengthened the manuscript.

   A) We have added several new analyses for rhythmic modulation of spiking activity, and elevated some of the Supplementary Figures related to oscillations to the main figures (Figures 2, 5). In addition, since several of our analyses provide novel results for spiking and ensemble dynamics before and after the decision making period, as noted by Reviewer 1, and we have emphasized these results as a novel advance in the revised manuscript , including the title and abstract.

   B) We agree that our analysis focuses on rhythmic coordination by beta and RR oscillations, phase modulation of single cell spiking activity in CA1 and PFC for accurate odor-cued decision making, and ensemble dynamics during decision making and execution of decisions. While relating ensemble activity to the observed oscillations is a long-term goal, we are limited by the size of simultaneously recorded ensembles within single sessions, since measures of ensemble dynamics per trial are required for such analyses. This is now noted in the Discussion section. We therefore focus our analyses separately on single cell modulation by rhythms and dynamics of ensemble activity during decision making.

   We have also retitled the manuscript to indicate this: “Rhythmic coordination and ensemble dynamics in the hippocampal-prefrontal network for odor-place associative memory and decision making”, to more accurately reflect our results.

   C) We appreciate the Reviewer’s favorable view on independent confirmation of previous results on beta coherence using our strong dataset. We have referenced previous results on OB-HPC, LEC-HPC and striatal beta coherence in the manuscript (e.g., Kay and Beshel 2010; Igarashi et al. 2014; Rangel et al. 2016; Leventhal et al., 2012).

   In addition, we also highlight the novelty of our results in the manuscript, as noted by Reviewers 1 and 2. Our findings in these specific circuits, namely the PFC-CA1 network, during odor-cued decision making are novel. Our results show that beta phase modulation of a sub-population of phase-coherent CA1 and PFC neurons is linked to accurate decision making, elucidate selectivity and ensemble dynamics in these regions during decision making, and show that independent ensembles are recruited during odor-sampling vs. the execution of decisions on the spatial maze. These results are emphasized in the revised manuscript.

   D) We apologize for the misunderstanding regarding the number of neurons. We had initially reported total number of neurons recorded across run and sleep sessions, including those with very few spikes during the task. In determining task-responsive and task-unresponsive neurons (Figure 3), the task-unresponsive set also includes a very large fraction of neurons that did not have sufficient spikes during the odor-sampling or decision making period (e.g. using a criterion of number of spikes equal to number of trials; similar numbers are seen with other criterion such as an absolute minimum number of spikes). These neurons should be more accurately denoted as “Odor Period Inactive”. Therefore a more accurate estimate of task-responsive neurons in CA1 and PFC indicating their task engagement is now shown in Figure 3, starting with neurons that had sufficient spikes for this analysis. Using this metric, a large fractions of neurons are task responsive and selective, similar to previously reported fractions in other studies (Igarashi, et al., 2014). We have added this description and numbers in the text (page 11 lines 230-241) and Methods (page 37 lines 795-797).

   We have also toned down the interpretation by avoiding generalizing to the whole population, and note that beta phase modulation of phase-locked neurons is related to behavior accuracy. Here, in particular, our results suggest a key role of CA1 interneurons in beta-mediated interactions.

   Read the original source
3. 

   eLife

   Jun 23, 2022

   Evaluation Summary:

   The authors report on the coordination mechanisms between oscillations recorded in the CA1 subfield of the hippocampus, prefrontal cortex, and olfactory bulb and cell ensemble activity in CA1 and prefrontal cortex that are associated with odor-cued decision making. The findings support the hypothesis that the beta rhythm plays a role in coordinating CA1-prefrontal cortex ensembles associated with an animal's accurate decisions. Sensory-guided decision-making is of broad significance to many readers who are studying executive functions and cognitive behaviors, and the observations reported in this manuscript provide insights into mechanisms that may support these functions and behaviors.

   (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private …

   Evaluation Summary:

   The authors report on the coordination mechanisms between oscillations recorded in the CA1 subfield of the hippocampus, prefrontal cortex, and olfactory bulb and cell ensemble activity in CA1 and prefrontal cortex that are associated with odor-cued decision making. The findings support the hypothesis that the beta rhythm plays a role in coordinating CA1-prefrontal cortex ensembles associated with an animal's accurate decisions. Sensory-guided decision-making is of broad significance to many readers who are studying executive functions and cognitive behaviors, and the observations reported in this manuscript provide insights into mechanisms that may support these functions and behaviors.

   (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)

   Read the original source
4. 

   eLife

   Jun 23, 2022

   Reviewer #1 (Public Review):

   The authors performed simultaneous extracellular recordings in brain regions (CA1, prefrontal cortex (PFC), olfactory bulb (OB)) that are key to odor-guided decision making to delineate the oscillatory and cell population dynamics that guide decision making based on learned associations. They used complementary analyses to assess the coordination between CA1 and medial PFC (mPFC), using coherence and phase-locking analysis as well as generalized linear models and Bayesian decoding methods.

   One of the strengths of this work is the comparison of beta and respiratory (RR) LFP coherence in several behavioral states to rule out confounds due to sniffing or preparatory motor behavior (e.g., coherence was assessed during decision making with and without an odor present, during reward consumption). These controls …

   Reviewer #1 (Public Review):

   The authors performed simultaneous extracellular recordings in brain regions (CA1, prefrontal cortex (PFC), olfactory bulb (OB)) that are key to odor-guided decision making to delineate the oscillatory and cell population dynamics that guide decision making based on learned associations. They used complementary analyses to assess the coordination between CA1 and medial PFC (mPFC), using coherence and phase-locking analysis as well as generalized linear models and Bayesian decoding methods.

   One of the strengths of this work is the comparison of beta and respiratory (RR) LFP coherence in several behavioral states to rule out confounds due to sniffing or preparatory motor behavior (e.g., coherence was assessed during decision making with and without an odor present, during reward consumption). These controls allowed the authors to identify a specific enhancement of beta compared to RR coherence during decision making.

   The analyses of task-responsive putative interneuron and pyramidal cells suggest that accurate decision-making is associated with a stronger modulation of beta phase-locking in interneurons. Additional cross-correlation analyses between cell types across regions showed that cells, particularly interneurons, are temporally coordinated in the beta range. Their analyses did not identify a mechanism for this coordination, but the temporal lags between PFC and CA1 cells raise the possibility of top-down interactions mediated by a third brain region.

   The authors used the cellular activity to determine that the animal's upcoming behavior could be predicted from the ensemble activity during decision-making a few hundred milliseconds before the behavioral choice, but decoding accuracy diminished soon after the decision-making period. Interestingly, decoding accuracy increased after decision-making when using the spatially active cell ensembles. As indicated by the authors, these results suggest that different cell ensembles are engaged during decision-making and during the execution of the decision. It is possible that this change in ensemble dynamics before and after decision-making relates to the familiarity of the animals with the task, which makes it likely to involve procedural components (e.g., Jog et al., 1999). As pointed out by the authors in the discussion, several results have implications for the formation of associative memories and provide clues for future experiments. Thus, future work looking at the ensemble dynamics and at the occurrence of CA1 ripples in the early stages of task learning compared to when the animals are very familiar with the task (as in the current study), will provide a better understanding of the shifts that develop during the formation and consolidation of the association.

   One of the considerations in interpreting the results is that the odor sampling and decision-making periods overlap, making it difficult to disentangle the neural dynamics that are driven by the recall of the association (cued retrieval) and those that relate to the upcoming turning behavior after odor port disengagement. However, the author's analyses of odor and choice selectivity in correct and incorrect trials demonstrate a preferential association between spike activity and choice selection in this task.

   Overall, the results advance our understanding of odor-guided decision-making mechanisms in CA1 and PFC at the LFP and cell population level. This work will be of significance to further research on the cellular basis of memory-guided decision-making, and to future work characterizing the interactions between CA1 and PFC during learning.

   Read the original source
5. 

   eLife

   Jun 23, 2022

   Reviewer #2 (Public Review):

   Symanski et al. investigated the communication between the medial prefrontal cortex (mPFC), the hippocampal CA1 region, and the olfactory bulb (OB) while rats underwent an odor-cued decision-making task. By recording local field potentials and spiking activity in the three regions, they found that all regions became synchronized at the beta band and respiratory rhythms during cue sampling/decision-making. Although the strength of inter-region synchrony was not predictive of correct choices, both CA1 and mPFC neurons showed stronger phase-locked firings to beta oscillations for correct than incorrect choices. Moreover, a subset of putative pyramidal and interneurons in both regions were selective for task variables, and as ensembles, they formed activity patterns differentiating choices. Also, their firings …

   Reviewer #2 (Public Review):

   Symanski et al. investigated the communication between the medial prefrontal cortex (mPFC), the hippocampal CA1 region, and the olfactory bulb (OB) while rats underwent an odor-cued decision-making task. By recording local field potentials and spiking activity in the three regions, they found that all regions became synchronized at the beta band and respiratory rhythms during cue sampling/decision-making. Although the strength of inter-region synchrony was not predictive of correct choices, both CA1 and mPFC neurons showed stronger phase-locked firings to beta oscillations for correct than incorrect choices. Moreover, a subset of putative pyramidal and interneurons in both regions were selective for task variables, and as ensembles, they formed activity patterns differentiating choices. Also, their firings were temporally coordinated in a direction that the mPFC interneurons led CA1 interneurons and pyramidal neurons. Based on these findings, the authors propose that cue-evoked beta oscillations modulate the activity of interneurons to coordinate ensemble activity in CA1-mPFC networks supporting decision-making.

   Strength:

   The findings uncovered a new style of mPFC-Hippocampal communication through odor-evoked beta oscillations, which contrasts with theta oscillations and sharp-wave/ripples reported during memory-guided spatial navigation tasks. The overall quality of the work is outstanding. The data collection and analysis were meticulously conducted with appropriate controls and statistical tests.

   Weakness:

   The initial analysis of LFP activity (Figure 2d) revealed strong coherence in the beta band in all region pairs; however, the subsequent analysis focuses on mPFC-CA1 interaction. To justify this approach, it is essential to establish that the mPFC-CA1 beta synchrony reflects their direct communication rather than a by-product of common inputs from the OB.

   The authors used cross-correlograms to reveal the directionality of mPFC-CA1 interaction. To strengthen the author's view that beta oscillations help coordinate neural activity, it is worth investigating if the same temporal relationship is also detectable within each cycle of beta oscillations. Specifically, mPFC interneurons may fire at earlier phases, followed by firings of CA1 interneurons and pyramidal neurons at later phases.

   Read the original source
6. 

   eLife

   Jun 23, 2022

   Reviewer #3 (Public Review):

   Symanski et al. describe a set of interesting results derived from analyzing electrophysiological recordings performed in rats well trained on an associative memory task on a spatial maze (a T maze), in which animals learned to associate a given odor delivered in an initial maze region (upon a nose poke) with a subsequent spatial choice (a left or a right turn) to receive a reward. The authors have obtained LFPs from the OB, PFC, and CA1 from 8 animals subjected to this task, along with single-unit activity from the PFC and CA1. The authors describe that, during odor sampling, there is prominent LFP activity in the beta range (20-30 Hz) as well as prominent activity of the respiration-entrained LFP rhythm (RR, 7-8 Hz). The authors convincingly show that beta activity - but not RR - is specific to odor …

   Reviewer #3 (Public Review):

   Symanski et al. describe a set of interesting results derived from analyzing electrophysiological recordings performed in rats well trained on an associative memory task on a spatial maze (a T maze), in which animals learned to associate a given odor delivered in an initial maze region (upon a nose poke) with a subsequent spatial choice (a left or a right turn) to receive a reward. The authors have obtained LFPs from the OB, PFC, and CA1 from 8 animals subjected to this task, along with single-unit activity from the PFC and CA1. The authors describe that, during odor sampling, there is prominent LFP activity in the beta range (20-30 Hz) as well as prominent activity of the respiration-entrained LFP rhythm (RR, 7-8 Hz). The authors convincingly show that beta activity - but not RR - is specific to odor sampling (RR also shows up during other immobility periods within the task and when animals breathed clean air). They further show that not only beta power but also inter-regional beta coherence significantly enhances during the odor sampling period. In addition, the authors find a higher beta phase modulation of spiking in a subset of neurons associated with subsequent correct decisions. Since the authors also prove - based on behavioral analysis - that the odor-sampling period corresponds to the decision-making period in this task, they propose a role for beta coordination of hippocampal-prefrontal networks in sensory-cued decision making. The paper also brings along a set of complementary findings looking at the single unit and ensemble activity in both regions (CA1 and PFC), which are capable of predicting future spatial choices.

   I consider the investigated topic relevant to modern neuroscience and likely to interest a broad audience. Nevertheless, while there is much to like about this paper (e.g., carefully done experiments, advanced computational data analyses, well-written text, and well-crafted figures), I caught some issues that called my attention upon a careful reading, which I list below:

   A) The paper is written in a way clearly centered on rhymical brain activity (c.f. title, abstract, introduction, and discussion). Yet, out of 7 main figures, only 2 of them show data related to oscillations (while 1 figure shows behavioral data and 4 figures show spiking analysis not related to brain rhythms). Therefore, the presentation of the results seems unbalanced and disconnected from the main story.

   B) Somewhat related to the point above, in a strict sense, the title is not well justified ("Rhythmic coordination of hippocampal-prefrontal ensembles (...)") since there is no analysis relating assembly activity with either beta or RR (their results show beta or RR modulating a subset of single units), nor there is a combined ensemble analysis of PFC and hippocampal units (i.e., interregional cell assemblies). Why not try to relate ensemble activity to the observed oscillations?

   C) The main result of increased interregional beta coherence specifically during odor sampling is very interesting and seems quite solid. Though I hate being the one raising questions about the level of advancement, I cannot avoid noticing that similar increases in beta coherence in odor-sampling-based tasks have been observed before (e.g., increased OB-HPC beta coherence during odor sampling has been shown in Martin et al 2007 and between LEC and HPC in Igarashi et al 2014), which is to say that there is overlap between this core finding and previous research. But that said, in times where the reproducibility of our scientific endeavor has been put into question, this particular reviewer favors the publication of similar findings by independent labs, especially given this neatly collected dataset. It is recommended to highlight which results constitute novel insights here and which results provide support for previously published results.

   D) It called to my attention that many of the spiking results were obtained for a small percentage of neurons. For instance, how can the authors be confident that the choice-selective neurons are actually coding for the choice as opposed to being randomly detected by statistical chance? As a case in point, the authors mention that 1309 units were recorded in CA1, and from these 42 cells were choice selective. If the authors have employed a typical alpha of 5% for detecting such neurons, chance alone would predict ~60 neurons being false positives. I apologize if I am missing something, but could the authors clarify? On a related note, even though most findings hold true for a small percentage of neurons, the writing also tends to generalize the findings to the whole population (e.g., "Beta phase modulation of CA1 and PFC neuronal activity during this period was linked to accurate decisions, suggesting that this temporal modulation influences sensory-cued decision making.").

   Read the original source
7. 

   Version published to 10.1101/2020.06.08.140939v3 on bioRxiv

   Apr 8, 2022
8. 

   Version published to 10.1101/2020.06.08.140939v2 on bioRxiv

   Feb 5, 2021
9. 

   Version published to 10.1101/2020.06.08.140939v1 on bioRxiv

   Jun 9, 2020