Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Letter
  • Published:

Prefrontal neuronal assemblies temporally control fear behaviour

Nature volume 535, pages 420–424 (2016)Cite this article

Subjects

Abstract

Precise spike timing through the coordination and synchronization of neuronal assemblies is an efficient and flexible coding mechanism for sensory and cognitive processing1,2,3,4,5,6. In cortical and subcortical areas, the formation of cell assemblies critically depends on neuronal oscillations, which can precisely control the timing of spiking activity7,8. Whereas this form of coding has been described for sensory processing and spatial learning9,10,11,12, its role in encoding emotional behaviour remains unknown. Fear behaviour relies on the activation of distributed structures, among which the dorsal medial prefrontal cortex (dmPFC) is known to be critical for fear memory expression13,14,15,16. In the dmPFC, the phasic activation of neurons to threat-predicting cues, a spike-rate coding mechanism, correlates with conditioned fear responses and supports the discrimination between aversive and neutral stimuli14,17,18,19. However, this mechanism does not account for freezing observed outside stimuli presentations, and the contribution of a general spike-time coding mechanism for freezing in the dmPFC remains to be established. Here we use a combination of single-unit and local field potential recordings along with optogenetic manipulations to show that, in the dmPFC, expression of conditioned fear is causally related to the organization of neurons into functional assemblies. During fear behaviour, the development of 4 Hz oscillations coincides with the activation of assemblies nested in the ascending phase of the oscillation. The selective optogenetic inhibition of dmPFC neurons during the ascending or descending phases of this oscillation blocks and promotes conditioned fear responses, respectively. These results identify a novel phase-specific coding mechanism, which dynamically regulates the development of dmPFC assemblies to control the precise timing of fear responses.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on Springer Link
  • Instant access to full article PDF
Buy now

Prices may be subject to local taxes which are calculated during checkout

Figure 1: dmPFC PNs participate in neuronal assemblies during freezing.
Figure 2: Temporal matching of 4 Hz oscillations and dmPFC assemblies during freezing.
Figure 3: Four-hertz oscillations organize dmPFC PNs into assemblies during freezing.
Figure 4: Phase-specific optogenetic inhibition of dmPFC PNs controls freezing.

Similar content being viewed by others

References

  1. Gire, D. H. et al. Temporal processing in the olfactory system: can we see a smell? Neuron 78, 416–432 (2013)

    Article  CAS  Google Scholar 

  2. Friedrich, R. W., Habermann, C. J. & Laurent, G. Multiplexing using synchrony in the zebrafish olfactory bulb. Nature Neurosci. 7, 862–871 (2004)

    Article  CAS  Google Scholar 

  3. Igarashi, K. M., Lu, L., Colgin, L. L., Moser, M. B. & Moser, E. I. Coordination of entorhinal-hippocampal ensemble activity during associative learning. Nature 510, 143–147 (2014)

    Article  ADS  CAS  Google Scholar 

  4. Pastalkova, E., Itskov, V., Amarasingham, A. & Buzsáki, G. Internally generated cell assembly sequences in the rat hippocampus. Science 321, 1322–1327 (2008)

    Article  ADS  CAS  Google Scholar 

  5. Huxter, J., Burgess, N. & O’Keefe, J. Independent rate and temporal coding in hippocampal pyramidal cells. Nature 425, 828–832 (2003)

    Article  ADS  CAS  Google Scholar 

  6. O’Neill, J., Senior, T. J., Allen, K., Huxter, J. R. & Csicsvari, J. Reactivation of experience-dependent cell assembly patterns in the hippocampus. Nature Neurosci. 11, 209–215 (2008)

    Article  Google Scholar 

  7. Harris, K. D., Csicsvari, J., Hirase, H., Dragoi, G. & Buzsáki, G. Organization of cell assemblies in the hippocampus. Nature 424, 552–556 (2003)

    Article  ADS  CAS  Google Scholar 

  8. Buzsáki, G. Neural syntax: cell assemblies, synapsembles, and readers. Neuron 68, 362–385 (2010)

    Article  Google Scholar 

  9. Dupret, D., O’Neill, J., Pleydell-Bouverie, B. & Csicsvari, J. The reorganization and reactivation of hippocampal maps predict spatial memory performance. Nature Neurosci. 13, 995–1002 (2010)

    Article  CAS  Google Scholar 

  10. Huxter, J. R., Senior, T. J., Allen, K. & Csicsvari, J. Theta phase-specific codes for two-dimensional position, trajectory and heading in the hippocampus. Nature Neurosci. 11, 587–594 (2008)

    Article  CAS  Google Scholar 

  11. Richmond, B. J., Optican, L. M., Podell, M. & Spitzer, H. Temporal encoding of two-dimensional patterns by single units in primate inferior temporal cortex. I. Response characteristics. J. Neurophysiol. 57, 132–146 (1987)

    Article  CAS  Google Scholar 

  12. Laurent, G., Wehr, M. & Davidowitz, H. Temporal representations of odors in an olfactory network. J. Neurosci. 16, 3837–3847 (1996)

    Article  CAS  Google Scholar 

  13. Corcoran, K. A. & Quirk, G. J. Activity in prelimbic cortex is necessary for the expression of learned, but not innate, fears. J. Neurosci. 27, 840–844 (2007)

    Article  CAS  Google Scholar 

  14. Courtin, J. et al. Prefrontal parvalbumin interneurons shape neuronal activity to drive fear expression. Nature 505, 92–96 (2014)

    Article  ADS  Google Scholar 

  15. Vidal-Gonzalez, I., Vidal-Gonzalez, B., Rauch, S. L. & Quirk, G. J. Microstimulation reveals opposing influences of prelimbic and infralimbic cortex on the expression of conditioned fear. Learn. Mem. 13, 728–733 (2006)

    Article  Google Scholar 

  16. Do-Monte, F. H., Quiñones-Laracuente, K. & Quirk, G. J. A temporal shift in the circuits mediating retrieval of fear memory. Nature 519, 460–463 (2015)

    Article  ADS  CAS  Google Scholar 

  17. Burgos-Robles, A., Vidal-Gonzalez, I. & Quirk, G. J. Sustained conditioned responses in prelimbic prefrontal neurons are correlated with fear expression and extinction failure. J. Neurosci. 29, 8474–8482 (2009)

    Article  CAS  Google Scholar 

  18. Chang, C. H., Berke, J. D. & Maren, S. Single-unit activity in the medial prefrontal cortex during immediate and delayed extinction of fear in rats. PLoS ONE 5, e11971 (2010)

    Article  ADS  Google Scholar 

  19. Likhtik, E., Stujenske, J. M., Topiwala, M. A., Harris, A. Z. & Gordon, J. A. Prefrontal entrainment of amygdala activity signals safety in learned fear and innate anxiety. Nature Neurosci. 17, 106–113 (2014)

    Article  CAS  Google Scholar 

  20. Karalis, N. et al. 4-Hz oscillations synchronize prefrontal–amygdala circuits during fear behavior. Nature Neurosci. 19, 605–612 (2016)

    Article  CAS  Google Scholar 

  21. Livneh, U. & Paz, R. Amygdala-prefrontal synchronization underlies resistance to extinction of aversive memories. Neuron 75, 133–142 (2012)

    Article  CAS  Google Scholar 

  22. Dzirasa, K. et al. Lithium ameliorates nucleus accumbens phase-signaling dysfunction in a genetic mouse model of mania. J. Neurosci. 30, 16314–16323 (2010)

    Article  CAS  Google Scholar 

  23. Dzirasa, K., Fuentes, R., Kumar, S., Potes, J. M. & Nicolelis, M. A. Chronic in vivo multi-circuit neurophysiological recordings in mice. J. Neurosci. Methods 195, 36–46 (2011)

    Article  Google Scholar 

  24. Fujisawa, S. & Buzsáki, G. A 4 Hz oscillation adaptively synchronizes prefrontal, VTA, and hippocampal activities. Neuron 72, 153–165 (2011)

    Article  CAS  Google Scholar 

  25. Tort, A. B. et al. Dynamic cross-frequency couplings of local field potential oscillations in rat striatum and hippocampus during performance of a T-maze task. Proc. Natl Acad. Sci. USA 105, 20517–20522 (2008)

    Article  ADS  CAS  Google Scholar 

  26. Ito, J. et al. Whisker barrel cortex delta oscillations and gamma power in the awake mouse are linked to respiration. Nature Commun. 5, 3572 (2014)

    Article  ADS  CAS  Google Scholar 

  27. Lesting, J. et al. Patterns of coupled theta activity in amygdala-hippocampal-prefrontal cortical circuits during fear extinction. PLoS ONE 6, e21714 (2011)

    Article  ADS  CAS  Google Scholar 

  28. Stujenske, J. M., Likhtik, E., Topiwala, M. A. & Gordon, J. A. Fear and safety engage competing patterns of theta-gamma coupling in the basolateral amygdala. Neuron 83, 919–933 (2014)

    Article  CAS  Google Scholar 

  29. Dejean, C. et al. Neuronal circuits for fear expression and recovery: recent advances and potential therapeutic strategies. Biol. Psychiatry 78, 298–306 (2015)

    Article  Google Scholar 

  30. Markram, H. et al. Interneurons of the neocortical inhibitory system. Nature Rev. Neurosci. 5, 793–807 (2004)

    Article  CAS  Google Scholar 

  31. Somogyi, P., Freund, T. F. & Cowey, A. The axo-axonic interneuron in the cerebral cortex of the rat, cat and monkey. Neuroscience 7, 2577–2607 (1982)

    Article  CAS  Google Scholar 

  32. Sik, A., Penttonen, M., Ylinen, A. & Buzsáki, G. Hippocampal CA1 interneurons: an in vivo intracellular labeling study. J. Neurosci. 15, 6651–6665 (1995)

    Article  CAS  Google Scholar 

  33. Kvitsiani, D. et al. Distinct behavioural and network correlates of two interneuron types in prefrontal cortex. Nature 498, 363–366 (2013)

    Article  ADS  CAS  Google Scholar 

  34. Wolff, S. B. et al. Amygdala interneuron subtypes control fear learning through disinhibition. Nature 509, 453–458 (2014)

    Article  ADS  CAS  Google Scholar 

  35. Stark, E. et al. Inhibition-induced theta resonance in cortical circuits. Neuron 80, 1263–1276 (2013)

    Article  CAS  Google Scholar 

  36. Peyrache, A., Benchenane, K., Khamassi, M., Wiener, S. I. & Battaglia, F. P. Principal component analysis of ensemble recordings reveals cell assemblies at high temporal resolution. J. Comput. Neurosci. 29, 309–325 (2010)

    Article  Google Scholar 

  37. Berens, P. CircStat: a MATLAB toolbox for circular statistics. J. Stat. Softw. 31, 1–21 (2009)

    Article  Google Scholar 

Download references

Acknowledgements

We thank the members of the Herry laboratory for comments on the manuscript, and K. Deisseroth and E. Boyden for sharing material. Microscopy was done in the Bordeaux Imaging Center of the CNRS-INSERM and Bordeaux University, a member of France BioImaging. This work was supported by grants from the French National Research Agency (ANR-10-EQPX-08 OPTOPATH; LABEX BRAIN ANR 10-LABX-43, LABEX TRAIL ANR 10-LABX-57), the European Research Council (ERC) under the European Union’s Seventh Framework Program (FP7/2007-2013)/ERC grant agreement number 281168, and the Conseil Regional d’Aquitaine.

Reviewer Information Nature thanks A. Holmes, P. O’Donnell and the other anonymous reviewer(s) for their contribution to the peer review of this work.

Author information

Author notes

  1. Cyril Dejean, Julien Courtin and Nikolaos Karalis: These authors contributed equally to this work.

Authors and Affiliations

  1. INSERM, Neurocentre Magendie, U862, 146 Rue Léo-Saignat, Bordeaux, 33077, France

    Cyril Dejean, Julien Courtin, Nikolaos Karalis, Fabrice Chaudun, Hélène Wurtz, Thomas C. M. Bienvenu & Cyril Herry

  2. Université de Bordeaux, Neurocentre Magendie, U862, 146 Rue Léo-Saignat, Bordeaux, 33077, France

    Cyril Dejean, Julien Courtin, Nikolaos Karalis, Fabrice Chaudun, Hélène Wurtz, Thomas C. M. Bienvenu & Cyril Herry

  3. Friedrich Miescher Institute for Biomedical Research, Maulbeerstrasse 66, Basel, CH-4058, Switzerland

    Julien Courtin

  4. Faculty of Medicine, Ludwig-Maximilians Universität München, Planegg-Martinsried, 82152, Germany

    Nikolaos Karalis

Authors
  1. Cyril Dejean
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Julien Courtin
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Nikolaos Karalis
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Fabrice Chaudun
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Hélène Wurtz
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Thomas C. M. Bienvenu
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Cyril Herry
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

F.C., J.C., C.D., N.K., T.C.M.B. and H.W. performed the experiments and analysed the data. J.C., C.D. and C.H. designed the experiments and wrote the paper.

Corresponding authors

Correspondence to Cyril Dejean or Cyril Herry.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Extended data figures and tables

Extended Data Figure 1 CS+-triggered and spontaneous occurrence of conditioned freezing responses.

a, Representative examples of conditioned freezing behaviour periods recorded 24 h after auditory fear conditioning (post-FC session). Freezing epochs occurred either independently of any sensory stimulation (thick grey lines) or were induced by CS+ presentations (thick black lines). b, Averaged percentage of freezing behaviour recorded 24 h after auditory fear conditioning (post-FC session) within or outside CS+ presentations (n = 7 mice, paired t-test, within versus outside CS+, **P < 0.01). c, Averaged percentage of freezing episodes initiated outside or within CS+ presentations and restricted or outlasting CS+ presentations (n = 7 mice). Error bars, mean ± s.e.m.

Extended Data Figure 2 Separation of putative PNs and putative interneurons.

a, Location and depth of recording sites in the dmPFC and number of neurons recorded (n = 7 mice). b, Top: superimposed waveforms recorded from three different units. Bottom: spikes originating from individual units were sorted using three-dimensional principal component analysis. c, Among the population of neurons recorded, 86% were classified as putative projection neurons (PN, red circles, n = 290) and 14% as putative interneurons (IN, blue circles, n = 48) using an unbiased unsupervised cluster separation algorithm based on three electrophysiological properties: firing frequency, spike half width (SHW) and spike area under waveform peak. Inset, average waveform of a representative PN and interneuron illustrating the methodology used to quantify SHW and the spike segment used to calculate the spike area under waveform. d, Top: representative cross-correlogram performed between a PN and interneuron showing a short-latency, presumably monosynaptic, excitatory interaction (no inhibitory interaction identified among all significant cross-correlograms identified). Bottom: representative cross-correlogram between an interneuron and a PN showing a short-latency, possibly monosynaptic, inhibitory interaction (no excitatory interaction identified among all significant cross-correlograms identified). Reference events correspond to the spikes of the presynaptic neuron (dashed line at time 0, bins of 0.5 ms). Red and blue circles represent PNs and interneurons, respectively. e, Box plot of the firing rate of PNs (n = 290 in 7 mice) and interneurons (n = 48 in 7 mice). For each box plot, the middle, bottom and top lines correspond to the median, lower quartile and upper quartile firing rates. Blue and red dots below and above the boxes correspond to outlier data points. f, Corresponding firing frequency statistics for PNs and interneurons.

Extended Data Figure 3 Isolation of dmPFC neuronal assemblies.

a, Top: example of a binarized rate histogram matrix computed from 55 PNs recorded in a single animal. Bottom: examples of correlation matrices computed for particular time bins from the same 55 PNs (bin size: 150 ms). b, Corresponding example of principal component scores for the same population of 55 PNs and freezing behaviour as a function of time. For the displayed component (first principal component), the distribution of scores is different for freezing and no freezing periods. This shows that freezing episodes are associated with a specific profile of activity for the simultaneously recorded neurons taken as a population. c, Averaged Mahalanobis distance between neuronal population profiles correlated with freezing and no freezing epochs as a function of the length of the sliding window used to produce the binarized rate histograms (see Methods). The optimal separation for neuronal population profiles correlated with freezing and no freezing epochs was obtained with sliding window lengths between 150 and 250 ms (n = 7 mice, Friedman repeated-measures one-way ANOVA on ranks test, *P < 0.05). Error bars, mean ± s.e.m. d, Corresponding example distribution of principal component scores for neuronal population profiles correlated with freezing (black dots) and no freezing (white dots) epochs (Wilcoxon signed rank test, P < 0.05). e, Corresponding example of freezing probability as a function of neuronal population principal component score. Freezing probability associated with each score is compared with surrogate data (black dots) for which freezing intervals have been shuffled. Plain and open red dots represent neuronal population profiles that are significantly over-represented compared with chance during freezing and no freezing periods, respectively (one sample t-test, surrogate data versus actual freezing probability for each score, for scores equal to or above 1.15, freezing probability superior to chance level and P < 0.05). The red dotted line represents the boundary between neuronal population profiles correlated with freezing and no freezing epochs. f, Top: coactivation matrix averaged over freezing epochs from the same 55 dmPFC PNs recorded in a single animal. The strength of the coactivation between pairs of neurons is expressed as the percentage of coactivation compared with chance. Bottom: surrogate data. Coactivation matrix averaged over freezing epochs from the same 55 dmPFC PNs recorded in a single animal for which spike trains have been shuffled 50 times. g, Average percentage of ANs (AP index > 1) and ONs (AP index < 1) to freezing-activated neuronal assemblies (n = 7 mice, ON: 59.6 ± 4.5%; AN: 40.4 ± 4.5%). h–m, Left: example of freezing probability as a function of neuronal population principal component score for individual mice as in e. Right: normalized coactivation matrix averaged over freezing predicting epochs from dmPFC PNs recorded in individual animals (n = 29, 40, 46, 26, 45 and 49 neurons). The strength of the coactivation between pairs of neuron is expressed as the percentage of coactivation compared with chance. Error bars, mean ± s.e.m.

Extended Data Figure 4 CS+-evoked firing activity of dmPFC PNs.

a, Averaged firing rate during freezing and no freezing periods for ANs and ONs (n = 7 mice, two-way ANOVA, factor 1: AN versus ON, P < 0.05; factor 2: freezing versus no freezing, P > 0.05; F1 × F2: P > 0.05; Student–Newman–Keuls post-hoc test within freezing, **P < 0.01, *P < 0.05). b, Average neuronal assembly rate during no freezing periods and during freezing episodes within or outside CS+ presentations (n = 7 mice, Friedman repeated-measures one-way ANOVA on ranks test, no freezing versus freezing (inside or outside CS+), P < 0.05; Student–Newman–Keuls post-hoc test, *P < 0.05). c, Z-score-normalized peri-stimulus time histograms of individual dmPFC ANs (n = 117) and ONs (n = 172) during freezing episodes. d–f, Left: mean z-score of CS+-evoked significant excitatory (d), unchanged (e) or significant inhibited (f) neuronal responses for dmPFC AN and ON. Grey and pink shaded areas correspond to mean ± s.e.m. Right: corresponding percentage of AN and ON (n = 7 mice, χ2 test, P > 0.05 in all cases; CS+-evoked excitation: AN: 38.5%, ON: 32.6%; CS+-evoked no changes: AN: 45.3%, ON: 55.8%; CS+-evoked inhibition: AN: 16.2%, ON: 11.6%). g, The probability of observing assembly activation is not affected by CS+ (repeated-measures one-way ANOVA on ranks, P > 0.05). Error bars, mean ± s.e.m.

Extended Data Figure 5 Histological controls and closed-loop stimulation technique.

a, Representative micrograph used for dmPFC GFP expression assessment. b, Representative dmPFC PV interneurons with distinct dendritic arborizations were infected. Left: multipolar neuron with a round soma, corresponding to a putative basket cell. Right: interneuron with an ovoid soma at the layer 1–2 border, displaying asymmetric and tufted dendrites. Some branches extend towards the pia, bend and follow the pial surface. This neuron corresponds to the typical description of axo-axonic (chandelier) cells in mPFC. c, Top: representative raw (black trace) and filtered (grey trace, 3–6 Hz) dmPFC LFP recorded in PV-IRES-Cre mice during 4 Hz oscillations (top) or in CAMKII-Cre mice outside 4 Hz oscillations infected with ChR2 (bottom) upon light activation (30 ms square pulses; black line, pulse start; red bar, pulse length). d, e, Averaged LFP power spectra obtained in ChR2-infected mice and GFP controls during stimulated (coloured traces) and non-stimulated (grey traces) periods when the stimulation was applied in the ascending (ChR2, n = 7 mice; GFP, n = 7 mice) or descending (ChR2, n = 7 mice; GFP, n = 7 mice) phase. f, Averaged 3–6 Hz LFP power for ChR2 and GFP mice during stimulation in the ascending or descending phase of 4 Hz oscillations expressed as a percentage of no stimulation periods (ChR2, n = 7 mice; GFP, n = 7 mice; two-way ANOVA, factor 1: GFP versus ChR2, P = 0.848; factor 2: descending versus ascending, P = 0.879; F1 × F2, P = 0.618). g, h, Averaged LFP power spectra obtained in ArchT infected mice and GFP controls during stimulated (coloured traces) and non-stimulated (grey traces) periods when the stimulation was applied in the descending (ArchT, n = 5 mice; GFP, n = 7 mice) or ascending phase (ChR2, n = 5 mice; GFP, n = 4 mice). i, Averaged 3–6 Hz LFP power for ArchT and GFP mice during stimulation in the descending or ascending phase of 4 Hz oscillations expressed as a percentage of no stimulation periods (ArchT, n = 5 mice; GFP, n = 7/4 mice; two-way ANOVA, factor 1: GFP versus ChR2, P = 0.872; factor 2: descending versus ascending, P = 0.183; F1 × F2, P = 0.5). Error bars, mean ± s.e.m.

Extended Data Figure 6 Phase-specific optogenetic inhibition of dmPFC PNs.

a, Twenty-four hours after conditioning, levels of freezing in wild type (n = 7) and GFP (n = 7) control animals where similar (2 min baseline, four CS+ and four post-CS+ periods of 30 s averaged, two-way ANOVA, factor 1: wild type versus GFP, P > 0.05; factor 2: time, P < 0.001; F1 × F2, P < 0.01, Student–Newman–Keuls post-hoc test within wild type and GFP, ***P < 0.001). b, Corresponding time-resolved changes in freezing behaviour around CS+ offset for wild type and GFP control animals (two-way ANOVA, factor 1: wild type versus GFP, P > 0.05; factor 2: time, P > 0.05; F1 × F2, P > 0.05). c, Comparison of GFP control levels of freezing during (CS+) and 30 s after CS+ (post-CS+) on days 2 (n = 7), 3 (n = 7) and 4 (n = 4) (2 min baseline, four CS+ and four post-CS+ periods of 30 s averaged; two-way ANOVA, factor 1: day, P > 0.05; factor 2: time, P < 0.001; F1 × F2, P > 0.05; Bonferroni post-hoc test between epochs, **P < 0.01, ***P < 0.001). d, Timeline scheme for the optogenetic stimulation protocol on day 2 with stimulation targeting the ascending phase of 4 Hz oscillation cycle (top) and day 3 with stimulation targeting the descending phase of 4 Hz oscillation cycle (bottom). e, f, Comparison of freezing levels between CS+ 1 and CS+ 2–4 for epochs 10 s preceding, during and 30 s after CS+ during optogenetic stimulation (pre-CS+, CS+ and post-CS+ respectively) on day 2 (e) and day 3 (f). On day 2, ascending phase stimulation after CS+ 1 induced a decrease of freezing levels during subsequent CS+ epochs as no more difference was observed between pre CS+, CS+ and post CS+ for CS+ 2–4 compared with CS+ 1 (n = 7 mice, repeated-measures two-way ANOVA, factor 1: CS+ 1 versus CS+ 2–4, P > 0.05; factor 2: epoch, P < 0.05; F1 × F2, P > 0.05; Student–Newman–Keuls post-hoc test within factor 1, *P < 0.05 compared with both pre- and post-CS+ epochs). On day 3, descending phase stimulation after CS+ 1 induced an increase in freezing levels during pre-CS+ epochs as no more difference was observed between pre-CS+, CS+ and post-CS+ for CS+ 2–4 compared with CS+ 1 (n = 7 mice, repeated-measures two-way ANOVA, factor 1: CS+ 1 versus CS+ 2–4, P > 0.05; factor 2: epoch, P < 0.05; F1 × F2, P < 0.05; Student–Newman–Keuls post-hoc test within factor 1, *P < 0.001 compared with both CS+ and post-CS+ epochs). g, Averaged freezing behaviour evoked by CS+ 1 presentation on day 3 if day 2 stimulation was in the ascending phase (GFP n = 7, ChR2 with day 2 ascending n = 7, Mann–Whitney rank sum test, P = 0.259).

Extended Data Figure 7 Phase specificity of the close loop stimulation technique.

a, Representative distribution of stimulation pulses in relation to the phase of the 4 Hz oscillation (black line, filtered LFP 3–6 Hz) when the stimulation was applied in the ascending phase of the oscillation (bins of 18°). b, Representative cross-correlogram between two dmPFC PNs (neurons A and B) participating to neuronal assemblies during (blue bars), or outside (grey bars) optogenetic inhibition of PV interneurons. Note the reduction in joint firing activity when the stimulation was on. Reference events correspond to the spikes of the neuron A (dashed line at time 0, bins of 5 ms). c, Firing frequency of dmPFC neurons A (top) and B (bottom) without optogenetic stimulation (grey bars, bins of 15 ms) or when the pulse was applied in the ascending phase of the 4 Hz oscillation (blue bars, bins of 15 ms) (black line, filtered LFP 3–6 Hz). d, Corresponding joint peri-stimulus time histograms performed between the same dmPFC neurons A and B when the pulse was applied in the ascending phase of the 4 Hz oscillation (top) or without stimulation (bottom). Note the inhibition of the co-firing activity when the stimulation was on. e, Firing probability per degree of identified PV+ interneurons as a function of 4 Hz oscillations phase (n = 14 PV+ interneurons). f, Cumulative firing probability per degree for individual PV+ interneurons (open circles) during the ascending or descending phase of 4 Hz oscillations and corresponding averaged cumulative firing probability (black bars, n = 14 PV+ interneurons, Wilcoxon signed rank test, P = 0.733). g, i, Top: firing of a PV+ interneuron in a mouse expressing ChR2 or ArchT in PV+ interneurons in the dmPFC in response to blue or yellow light (light-pulse duration 30 ms; 150 trials). Bottom: mean peri-stimulus time histograms of all PV+ interneurons displaying significant light-evoked excitation or inhibition (n = 10 PV+ from five mice infected with ChR2; n = 4 PV+ from two mice infected with ArchT, light-pulse duration, 30 ms). Bins of 5 ms. h, j, Top: firing of a PN in a mouse expressing ChR2 or ArchT in PV+ interneurons in the dmPFC in response to blue or yellow light activation of PV+ interneurons (light-pulse duration, 30 ms; 150 trials). Bottom: mean peri-stimulus time histograms of all PNs displaying significant light-evoked inhibition or excitation (n = 160 PNs from seven mice infected with ChR2; n = 117 PNs from five mice infected with ArchT; light-pulse duration, 30 ms). Bins of 5 ms. Error bars, mean ± s.e.m.

Extended Data Figure 8 Effect of phase-specific PV manipulation on dmPFC co-firing activity.

a, b, Top: corrected pairwise co-firing for pairs of dmPFC neurons recorded in mice infected with ChR2 or GFP during stimulation in the ascending (a, ChR2, n = 2,877 pairs; GFP, n = 2,067 pairs) or descending (b, ChR2, n = 2,067 pairs; GFP, n = 2,282 pairs) phase of 4 Hz oscillations. Bottom: averaged delta-corrected pairwise co-firing activity (delta = ChR2 minus GFP) as a function of 4 Hz phase (one sample t-test with Bonferroni correction, hypothetical mean = 0, **P < 0.01, ***P < 0.001). c, d, Top: corrected pairwise co-firing for pairs of dmPFC neurons recorded in mice infected with ArchT or GFP during stimulation in the ascending (a, ArchT, n = 2,389 pairs; GFP, n = 2,542 pairs) or descending (b, ArchT, n = 3,122 pairs; GFP, n = 2,282 pairs) phase of 4 Hz oscillations. Bottom: averaged delta-corrected pairwise co-firing activity (delta = ArchT minus GFP) as a function of 4 Hz phase (one-sample t-test with Bonferroni correction, hypothetical mean = 0, *P < 0.05, **P < 0.01, ***P < 0.001). Error bars, mean ± s.e.m.

Extended Data Figure 9 Phase-specific optogenetic disinhibition of dmPFC PNs enhanced fear behaviour.

a, Timeline scheme for the optogenetic stimulation protocol on day 3 with stimulation targeting the descending phase of 4 Hz oscillation cycle (top) and day 4 with stimulation targeting the ascending phase of 4 Hz oscillation cycle (bottom). b, d, Comparison of freezing levels between CS+ 1 and CS+ 2–4 for epochs 10 s preceding, during and 30 s after CS+ during optogenetic stimulation of PV-IRES-Cre mice infected within the dmPFC with ArchT (pre-CS+, CS+ and post-CS+ respectively) on day 3 (b) and day 4 (d). On day 3, descending phase stimulation after CS+ 1 had no effect on freezing levels during subsequent CS+ epochs before, during or after CS+ (n = 5 ArchT mice, repeated-measures two-way ANOVA, factor 1: CS+ 1 versus CS+ 2–4, P > 0.05; factor 2: epoch, P < 0.05; F1 × F2, P > 0.05; Student–Newman–Keuls post-hoc test within factor 1, **P < 0.01 compared with both pre- and post-CS+ epochs, *P < 0.05 comparing pre-CS+ with CS+ epochs). On day 4, ascending phase stimulation after CS+ 1 had no effect on freezing levels during subsequent CS+ epochs before, during or after CS+ (n = 5 ArchT mice, repeated-measures two-way ANOVA, factor 1: CS+ 1 versus CS+ 2–4, P > 0.05; factor 2: epoch, P < 0.05; F1 × F2, P < 0.05; Student–Newman–Keuls post-hoc test within factor 1, ***P < 0.001 compared with both CS+ and post-CS+ epochs, *P < 0.05 comparing pre-CS+ with CS+ epochs). c, e, Left: time-resolved changes in freezing behaviour during light-mediated inhibition of dmPFC neuronal assemblies in the descending (c) or ascending (e) phase of the 4 Hz oscillation in PV-IRES-Cre mice infected within the dmPFC with ArchT (descending phase: first CS+, n = 5; ascending phase: first CS+, n = 5) or GFP (descending phase: first CS+, n = 7; ascending phase: first CS+, n = 4; descending phase: two-way ANOVA, factor 1: ArchT versus GFP, P > 0.05; factor 2: time, P > 0.05; F1 × F2, P > 0.05; ascending phase: two-way ANOVA, factor 1: ArchT versus GFP, P < 0.01; factor 2: time, P < 0.01; F1 × F2, P < 0.01). Shaded area represents the period during which the light stimulation was applied. Right: averaged freezing behaviour during the stimulation in the descending (c) (ArchT n = 5; GFP n = 7, Mann–Whitney test, ArchT versus GFP, P > 0.05) or ascending phase (e) (ArchT n = 5; GFP n = 4, Mann–Whitney test, ArchT versus GFP, *P < 0.05). Error bars, mean ± s.e.m.

Extended Data Figure 10 Functional model of phase-specific coding for freezing behaviour.

This schematic illustrates how, within a 4 Hz cycle, the phase specificity and relative influence of dmPFC assembly and non-assembly neurons (AN, ON, respectively) could represent a functional signal for downstream structures implicated in the expression of freezing behaviour. a, In normal freezing conditions, the prevalence and restriction of AN activation in the ascending phase combined with the monotonic phase distribution of ONs provides downstream structures with a signal-to-noise ratio (SNR) over an entire 4 Hz cycle that allows freezing expression. b, During optogenetic PN inhibition this SNR is altered. Top: when stimulating in the ascending phase, the bulk of AN activity over a cycle is shut down. This dramatically reduces SNR and prevents freezing expression. Bottom: when stimulating in the descending phase, AN activity is largely preserved over a cycle while that of ONs is diminished to a large extent. This strongly increases SNR and exacerbates freezing expression. c, During optogenetic PN disinhibition the SNR is also altered. Top: when stimulating in the ascending phase, the AN activity over a cycle is promoted compared with non-ensemble activity. This increases SNR and enhances freezing expression. Bottom: when stimulating in the descending phase, both assembly and non-assembly neuron activities over a cycle are increased. This does not affect SNR and freezing remains similar to normal conditions. In this model the dmPFC can be seen as an emitter whose signal transmission conditions the SNR for the receiver and the further expression of fear behaviour. Our optogenetic dissection of phase specificity for dmPFC neuronal assemblies shows that SNR can be manipulated to either.

Supplementary information

Supplementary Information

This file contains a Supplementary Discussion relating to (i) the role of PV interneurons in mediating 4 Hz oscillations, (ii) the induction of fear behaviour through manipulation of 4 Hz oscillations, (iii) the induction of an aversive behavioural state through the manipulation of 4 Hz oscillations generate and (iv) 4 Hz oscillations as a physiological signature of fear memories and orchestrator a phase-specific encoding of fear behaviour. (PDF 156 kb)

PowerPoint slides

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Dejean, C., Courtin, J., Karalis, N. et al. Prefrontal neuronal assemblies temporally control fear behaviour. Nature 535, 420–424 (2016). https://doi.org/10.1038/nature18630

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/nature18630

This article is cited by

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing