Distinct cell-type contributions and network topography of theta-nested gamma oscillations in the medial entorhinal cortex
Curation statements for this article:-
Curated by eLife
eLife Assessment
This study investigates the cellular mechanisms underlying theta-nested gamma oscillations in the medial entorhinal cortex; the experiments are rigorous, and the analyses and modeling provide potentially useful insights into cell-type-dependent circuit dynamics. However, the evidence supporting several key conclusions remains incomplete. The study is limited by conceptual constraints in the experimental design and a modeling approach that does not fully address underlying physiological mechanisms. Overall, this is a careful study that addresses how distinct neuronal populations in superficial MEC participate in theta-gamma coordination and provides new data linking cell-type-specific activity patterns to oscillatory network structure.
This article has been Reviewed by the following groups
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
- Evaluated articles (eLife)
Abstract
Theta-nested gamma oscillations in the medial entorhinal cortex (mEC) are essential for spatial coding and memory, but the underlying cellular mechanisms remain unclear. We combined optogenetics, whole-cell electrophysiology, intracellular voltage imaging, and local field potential (LFP) recordings in acute slices from CaMKIIα-ChR2 mice to investigate how excitation and inhibition shape theta-gamma coupling in layer II/III mEC. During theta-frequency stimulation, fast-spiking interneurons received strong gamma-frequency excitation and fired rhythmic bursts, whereas stellate and pyramidal neurons fired more sparsely and were dominated by gamma-frequency inhibition. This sparse firing could support the selective firing of grid cells. Excitatory post-synaptic currents in interneurons preceded inhibitory currents and LFP gamma by ∼3 ms, supporting a pyramidal-interneuron network gamma (PING) mechanism. Pyramidal neurons fired on the descending phase of the gamma cycle, whereas stellate cells and fast-spiking interneurons fired before and after the trough, respectively. Intracellular voltage imaging revealed network gamma synchronization among excitatory neurons at a population level, with topographic clustering of subthreshold membrane potentials, but not spike timing, while individual neurons often skipped gamma cycles. These findings identify the dominant role of reciprocal E-I interactions in generating theta-nested gamma oscillations and highlight distinct cell-type contributions to the temporal dynamics of the mEC. Further, a biophysically realistic computational model predicted gamma cycle skipping in stellate cells and burst firing in fast-spiking interneurons during PING. Our experimental and computational results provide mechanistic insight into how the intrinsic properties of mEC cell types generate oscillatory activity in a manner that could support grid cell function and spatial computation.
Article activity feed
-
eLife Assessment
This study investigates the cellular mechanisms underlying theta-nested gamma oscillations in the medial entorhinal cortex; the experiments are rigorous, and the analyses and modeling provide potentially useful insights into cell-type-dependent circuit dynamics. However, the evidence supporting several key conclusions remains incomplete. The study is limited by conceptual constraints in the experimental design and a modeling approach that does not fully address underlying physiological mechanisms. Overall, this is a careful study that addresses how distinct neuronal populations in superficial MEC participate in theta-gamma coordination and provides new data linking cell-type-specific activity patterns to oscillatory network structure.
-
Reviewer #1 (Public review):
Summary:
The question posed on cell-type-dependent relationships to theta-nested gamma rhythms is an important one. The authors use a variety of ontogenetic, imaging, electrophysiology, and computational techniques to show that reciprocal interactions between excitatory neurons and interneurons in the medial entorhinal cortex generate gamma oscillations. They measure LFP gamma, gamma power of postsynaptic currents in different neurons, spike phases with reference to LFP gamma, and spatial correlations of membrane potentials across a large population of neurons. Arguing (correctly) that gamma rhythm in this setting is generated through a pyramidal-interneuron network gamma (PING) mechanism, they demonstrate cell-type-specific differences in gamma phase-locking. While they show spatial dependencies of …
Reviewer #1 (Public review):
Summary:
The question posed on cell-type-dependent relationships to theta-nested gamma rhythms is an important one. The authors use a variety of ontogenetic, imaging, electrophysiology, and computational techniques to show that reciprocal interactions between excitatory neurons and interneurons in the medial entorhinal cortex generate gamma oscillations. They measure LFP gamma, gamma power of postsynaptic currents in different neurons, spike phases with reference to LFP gamma, and spatial correlations of membrane potentials across a large population of neurons. Arguing (correctly) that gamma rhythm in this setting is generated through a pyramidal-interneuron network gamma (PING) mechanism, they demonstrate cell-type-specific differences in gamma phase-locking. While they show spatial dependencies of sub-threshold voltages and even argue for topographic clustering, these could simply be reflections of the synchronous stimulation paradigm that they use.
Overall, I appreciate the methodology and rigor, but would have expected more from the study in terms of relevance to physiological stimulation conditions as well as in terms of mechanisms underlying the differences that they report here..
Strengths:
The authors are rigorous in how they conduct the experiments, report the data, and perform the analyses. The modeling respects the heterogeneities and is truthful to the experimental design. The conclusions on PING mechanisms are fine, but are not unexpected given the circuitry of the mEC.
Weaknesses:
The interpretation of the conclusions, while for the most part is fine, could have been better, especially given the conceptual limitations of the experimental design. The modeling part could have gone beyond simple descriptive matching and addressed mechanistic questions.
-
Reviewer #2 (Public review):
In this manuscript, the authors studied the cellular mechanism of theta-nested gamma oscillations in the medial entorhinal cortex (MEC) in vitro. The theta-nested gamma activity was induced by theta-modulated optogenetic stimulation of CaMKII+ neurons. In Figures 1 through 4, they describe the firing phase, synaptic input, and LFP-IPSC coupling of stellate cells, pyramidal cells, and interneurons. They then conducted voltage imaging, capturing the simultaneous activity of 41 cells, and found that subthreshold membrane potentials cluster in a weakly distance-dependent manner (Figure 5). The experiments and analysis are done rigorously for the most part.
However, the results described in Figures 1 to 4 are largely descriptive and highly similar to those in their recent publication, which utilized almost …
Reviewer #2 (Public review):
In this manuscript, the authors studied the cellular mechanism of theta-nested gamma oscillations in the medial entorhinal cortex (MEC) in vitro. The theta-nested gamma activity was induced by theta-modulated optogenetic stimulation of CaMKII+ neurons. In Figures 1 through 4, they describe the firing phase, synaptic input, and LFP-IPSC coupling of stellate cells, pyramidal cells, and interneurons. They then conducted voltage imaging, capturing the simultaneous activity of 41 cells, and found that subthreshold membrane potentials cluster in a weakly distance-dependent manner (Figure 5). The experiments and analysis are done rigorously for the most part.
However, the results described in Figures 1 to 4 are largely descriptive and highly similar to those in their recent publication, which utilized almost identical experiments. While the voltage imaging data during theta-nested gamma oscillations are novel, the authors report data from only a single experiment, leaving it unclear whether the results are reproducible. Furthermore, without a comparison to in vivo data, it remains unclear what novel insights this manuscript provides to advance our understanding of the cellular mechanisms underlying theta-nested gamma oscillations.
(1) The authors recently published another paper on the topic of theta-nested gamma oscillations in the MEC (Williams et al., eNeuro, 2026). In that study, they utilized a Thy1 promoter instead of the CaMKII promoter used here. The motivation for testing the CaMKII promoter in the current manuscript, as well as the novel insights expected from this experimental setup, remains unclear. Given that existing literature suggests inhibitory MEC cells play a critical role in theta activity (e.g., Gonzalez-Sulser et al., 2014)-implying that theta modulation should drive inhibitory rather than excitatory cells-the previous use of the Thy1 promoter appears closer to in vivo conditions than the CaMKII promoter used here.
The overall conclusion of the current manuscript is that excitatory-inhibitory (E-I) interactions dominate the generation of theta-nested gamma oscillations. However, in their previous eNeuro paper, the authors demonstrated that the interneuron network gamma (ING) mechanism can sustain gamma oscillations without excitatory synaptic transmission. It seems expected that excitatory cells would be involved when the optogenetic stimulation selectively drives excitatory cells. If CaMKII stimulation is less physiological and artificially forces the theta-nested gamma activity to rely on excitatory connections, this conclusion could be misleading. It may potentially describe a mechanism that is irrelevant to physiological processes in vivo. Please see my comment 3, which is related to this point.
In addition, Figures 1 and 2 heavily overlap with the authors' previous eNeuro publication. The differences in experimental settings and the motivation for performing almost identical experiments must be clearly articulated prior to these figures to avoid confusion. The authors must also justify why it is necessary to present such similar data, and explicitly point out the novel findings in the current paper compared to their previous work.
(2) Using voltage imaging to investigate theta-nested gamma oscillations is novel. However, the impact of the findings from this experiment appears minimal in the manuscript's current state. The most novel and interesting observation is likely presented in Figure 6, where the authors identified clustered voltage correlations. However, this appears to be an n=1 experiment, and these findings should be replicated at least in a few experiments. Furthermore, the manuscript lacks a discussion or interpretation of this observation, making it unclear whether the result is biologically meaningful. Please find specific suggestions regarding this point below.
(3) The authors' primary motivation for investigating the mechanisms underlying theta-modulated gamma oscillations is their potential role in grid cell firing. Therefore, it is critical that the mechanisms studied here in vitro accurately reflect in vivo processes. For this reason, greater effort should be made to better link this in vitro study with existing in vivo data. Numerous public in vivo datasets are available that detail the firing activity of putative principal cells and interneurons during exploratory behavior in mice. Intracellular recordings in awake animals have also been published, some of which the authors already cite. The data presented in Figures 1 and 4, for example, could be straightforwardly compared with those existing in vivo metrics. Furthermore, available in vivo silicon probe recordings could provide a reliable estimate of the spatial distribution of gamma-related spike activity. Such data should be compared with the voltage imaging results presented in this study.
This limitation connects back to the first point. In this manuscript, the authors tested a different method for inducing theta-nested gamma oscillations (via the CaMKII promoter) than in their recent eNeuro paper (via the Thy1 promoter). The outcomes of these two induction methods must be systematically compared against in vivo data to determine which approach aligns more closely with physiological conditions. Without such a comparison, the scientific justification for testing a different promoter in this study remains unclear.
-
Reviewer #3 (Public review):
Summary:
In this manuscript, Williams et al. combine optogenetics, whole-cell electrophysiology, local field potential recordings, large-scale voltage imaging, and computational modeling to investigate the cellular and circuit mechanisms underlying theta-nested gamma oscillations in superficial medial entorhinal cortex (mEC). The authors propose that fast-spiking interneurons receive strong gamma-frequency excitatory drive and provide rhythmic inhibition onto principal neurons, supporting a pyramidal-interneuron network gamma (PING) mechanism. They further report cell-type-specific differences in gamma phase locking, spatial clustering of subthreshold voltage signals, and a network model reproducing several observed features, including interneuron bursting and gamma-cycle skipping in excitatory neurons.
Stren…
Reviewer #3 (Public review):
Summary:
In this manuscript, Williams et al. combine optogenetics, whole-cell electrophysiology, local field potential recordings, large-scale voltage imaging, and computational modeling to investigate the cellular and circuit mechanisms underlying theta-nested gamma oscillations in superficial medial entorhinal cortex (mEC). The authors propose that fast-spiking interneurons receive strong gamma-frequency excitatory drive and provide rhythmic inhibition onto principal neurons, supporting a pyramidal-interneuron network gamma (PING) mechanism. They further report cell-type-specific differences in gamma phase locking, spatial clustering of subthreshold voltage signals, and a network model reproducing several observed features, including interneuron bursting and gamma-cycle skipping in excitatory neurons.
Strengths:
The study is technically sophisticated and addresses an important question in entorhinal circuit function. The combination of intracellular recordings, voltage imaging, and computational modeling is a clear strength.
Weaknesses:
Several key conclusions developed from experimental results require additional raw data, statistical support, clearer methodological description, and more cautious interpretation. The computational modeling focuses primarily on stellate cells, whereas the experimental results suggest an important role for pyramidal neurons in PING dynamics. This creates inconsistency between theory and experiments.
-
Author response:
We thank the editors and reviewers for their thoughtful comments. Below, we list our provisional responses to the reviewers’ major points:
On the rationale for CaMKIIα versus Thy1-driven stimulation and physiological relevance: We agree that we did not make clear the motivation for using CaMKIIα-driven stimulation, distinct from the Thy1-driven paradigm in our previous work (Williams et al., 2026). Using the Thy1 driver, both excitatory and inhibitory cells received direct theta drive. In contrast, CaMKIIα expression is largely restricted to principal neurons. Comparing these models lets us isolate a "driven I-cell" PING mechanism from the "E cell recovers first" mechanism relevant when interneurons are also directly driven.
Regarding physiological relevance, Gonzalez-Sulser et al. (2014) found that septal GABAergic …
Author response:
We thank the editors and reviewers for their thoughtful comments. Below, we list our provisional responses to the reviewers’ major points:
On the rationale for CaMKIIα versus Thy1-driven stimulation and physiological relevance: We agree that we did not make clear the motivation for using CaMKIIα-driven stimulation, distinct from the Thy1-driven paradigm in our previous work (Williams et al., 2026). Using the Thy1 driver, both excitatory and inhibitory cells received direct theta drive. In contrast, CaMKIIα expression is largely restricted to principal neurons. Comparing these models lets us isolate a "driven I-cell" PING mechanism from the "E cell recovers first" mechanism relevant when interneurons are also directly driven.
Regarding physiological relevance, Gonzalez-Sulser et al. (2014) found that septal GABAergic projections selectively and directly inhibit mEC interneurons, rather than exciting either principal cells or interneurons, implying that theta drive in vivo likely acts through rhythmic disinhibition of interneurons rather than direct excitation of any cell type. Neither the Thy1 nor the CaMKIIα paradigm reproduces this disinhibitory mechanism: both rely on excitatory optogenetic drive rather than rhythmic inhibition of interneurons, and replicating the natural drive (tonic excitatory tone plus rhythmic, interneuron-selective inhibition) is technically difficult in acute slices, which are largely quiescent without exogenous stimulation. We therefore view CaMKIIα and Thy1 as complementary approximations, each isolating a different circuit interaction. If forced to choose, we’d argue that the CaMKIIα is a better model of disinhibition of excitatory neurons. We will revise the Discussion regarding this point.
On reproducibility of the voltage imaging findings: We thank the reviewer for this comment and agree that clarification is warranted.
The voltage imaging dataset combines two levels of analysis with different sample sizes. The population-level firing and spike-correlation analyses (Fig. 5F–H) are pooled across multiple imaging sessions (n = 240 neurons). The spatial clustering analysis of subthreshold voltage correlations (Fig. 6, and the corresponding example traces in Fig. 5A–E) are drawn from a single representative recording session, as the reviewer correctly notes. We have voltage imaging data from 14 fields of view (1 FOV per slice) across 6 mice (240 neurons total; 3–41 neurons per FOV). In revision, we will extend the clustering and spatial-correlation analysis from Fig. 6 across sessions to assess whether the reported organization is reproducible, rather than relying on a single example. We will also revise the text to distinguish clearly which analyses are single-session versus pooled.
On restricting the computational model of excitatory neurons to stellate cells: We modeled stellate cells as the excitatory population because they are the principal cells reciprocally connected to fast-spiking PV+ interneurons (Fuchs et al., 2016), the interneuron class most directly implicated in theta-nested gamma. Pyramidal cells, by contrast, are primarily connected via 5-HT3a-positive interneurons (Fuchs et al., 2016), with the exception of a subset of "intermediate" pyramidal cells that do show reciprocal PV+ connectivity. Our model, which captures the full measured heterogeneity of stellate cell and PV+ interneuron intrinsic properties and their reciprocal connectivity, is, to our knowledge, the most biophysically constrained implementation of this specific microcircuit to date. Incorporating the PV+-connected intermediate pyramidal population is a natural next step. Because this refinement, which requires more experimental data, is nontrivial and beyond the scope of this study, we will note this explicitly as a limitation of the current model in the revised Discussion.
In vivo comparison (temporal/phase-locking): We agree that grounding our findings in existing in vivo data strengthens the study and will add these comparisons to the revision.
Our whole-cell recordings reproduce the temporal organization in vivo and provide further insights into cell-type differences between the principal cells. All cell types were strongly phase-locked to theta, while gamma phase-locking declined across successive spikes, with stellate cells decoupling after the first spike and pyramidal cells after the second. This earlier decoupling in stellate cells may contribute to their weaker theta rhythmicity reported in freely moving rats (Ray et al., 2014; Tang et al., 2014). In extracellular recordings from behaving mice, spike-train cross-correlation identifies putative monosynaptic excitatory connections (1–4 ms) from principal cells onto fast-spiking interneurons (Latuske et al., 2015); the excitation-to-inhibition offset we measured is of comparable magnitude, here resolved as a synaptic-current delay in electrophysiologically classified cell types.
We note that bursting and theta engagement have been assigned inconsistently across in vivo datasets. Bursty cells are preferentially classified as putative stellate by spikepattern classifiers (Latuske et al., 2015), while anatomically identified pyramidal cells are reported as the bursty, theta-rhythmic population in other work (Ebbesen et al., 2016). Because our cell-type assignments are based on subthreshold intrinsic properties (membrane sag, time constant) rather than spike patterning, our phase-locking results are independent of this classification ambiguity.
In vivo comparison (spatial organization): We agree high-density silicon-probe datasets are the appropriate reference here. To our knowledge, the anatomical distribution of gamma-locked spiking in superficial mEC has not been characterized in vivo. The highest-density available recordings (Gardner et al., 2022) analyze population activity in the decoded state rather than tissue coordinates, do not examine gamma, and are restricted to grid cells. We regard the dissociation we observe between spatially clustered subthreshold input and spatially distributed spiking as a principal advance of the present study, and as a testable prediction for future high-density recordings.
Ebbesen CL, Reifenstein ET, Tang Q, Burgalossi A, Ray S, Schreiber S, Kempter R, Brecht M. 2016. Cell Type-Specific Differences in Spike Timing and Spike Shape in the Rat Parasubiculum and Superficial Medial Entorhinal Cortex. Cell Reports 16:1005–1015. DOI: https://doi.org/10.1016/j.celrep.2016.06.057
Fuchs EC, Neitz A, Pinna R, Melzer S, Caputi A, Monyer H. 2016. Local and Distant Input Controlling Excitation in Layer II of the Medial Entorhinal Cortex. Neuron 89:194–208. DOI: https://doi.org/10.1016/j.neuron.2015.11.029
Gardner RJ, Hermansen E, Pachitariu M, Burak Y, Baas NA, Dunn BA, Moser M-B, Moser EI. 2022. Toroidal topology of population activity in grid cells. Nature 602:123–128. DOI: https://doi.org/10.1038/s41586-021-04268-7
Gonzalez-Sulser A, Parthier D, Candela A, McClure C, Pastoll H, Garden D, Sürmeli G, Nolan MF. 2014. Gabaergic projections from the medial septum selectively inhibit interneurons in the medial entorhinal cortex. Journal of Neuroscience 34:16739–16743. DOI: https://doi.org/10.1523/JNEUROSCI.1612-14.2014, PMID: 25505326
Latuske P, Toader O, Allen K. 2015. Interspike Intervals Reveal Functionally Distinct Cell Populations in the Medial Entorhinal Cortex. Journal of Neuroscience 35:10963–10976. DOI: https://doi.org/10.1523/JNEUROSCI.0276-15.2015
Ray S, Naumann R, Burgalossi A, Tang Q, Schmidt H, Brecht M. 2014. Grid-Layout and Theta-Modulation of Layer 2 Pyramidal Neurons in Medial Entorhinal Cortex. Science 343:891–896. DOI: https://doi.org/10.1126/science.1243028
Tang Q, Burgalossi A, Ebbesen CL, Ray S, Naumann R, Schmidt H, Spicher D, Brecht M. 2014. Pyramidal and Stellate Cell Specificity of Grid and Border Representations in Layer 2 of Medial Entorhinal Cortex. Neuron 84:1191–1197. DOI: https://doi.org/10.1016/j.neuron.2014.11.009
Williams B, Vedururu Srinivas A, Baravalle R, Fernandez FR, Canavier CC, White JohnA. 2026. Fast spiking interneurons autonomously generate fast gamma oscillations in the medial entorhinal cortex with excitation strength tuning ING–PING transitions. eneuro ENEURO.0452-25.2026. DOI: https://doi.org/10.1523/ENEURO.0452-25.2026
-
-
-
