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.

  • Article
  • Published:

Pathway-specific reorganization of projection neurons in somatosensory cortex during learning

Nature Neuroscience volume 18, pages 1101–1108 (2015)Cite this article

Subjects

Abstract

In the mammalian brain, sensory cortices exhibit plasticity during task learning, but how this alters information transferred between connected cortical areas remains unknown. We found that divergent subpopulations of cortico-cortical neurons in mouse whisker primary somatosensory cortex (S1) undergo functional changes reflecting learned behavior. We chronically imaged activity of S1 neurons projecting to secondary somatosensory (S2) or primary motor (M1) cortex in mice learning a texture discrimination task. Mice adopted an active whisking strategy that enhanced texture-related whisker kinematics, correlating with task performance. M1-projecting neurons reliably encoded basic kinematics features, and an additional subset of touch-related neurons was recruited that persisted past training. The number of S2-projecting touch neurons remained constant, but improved their discrimination of trial types through reorganization while developing activity patterns capable of discriminating the animal's decision. We propose that learning-related changes in S1 enhance sensory representations in a pathway-specific manner, providing downstream areas with task-relevant information for behavior.

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: Behavior correlates of learning a texture discrimination task.
Figure 2: Active whisking increases stick-slip events.
Figure 3: Changes in touch-related neurons across sessions.
Figure 4: M1P neuron activity reliably encodes whisker kinematic features across learning.
Figure 5: S2P neurons increase trial type discrimination during learning.
Figure 6: S2P neurons shift preference toward P1200-related trials.
Figure 7: S2P neurons display decision-related activity in late phases of learning.

Similar content being viewed by others

References

  1. Pais-Vieira, M., Lebedev, M.A., Wiest, M.C. & Nicolelis, M.A. Simultaneous top-down modulation of the primary somatosensory cortex and thalamic nuclei during active tactile discrimination. J. Neurosci. 33, 4076–4093 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  2. Zagha, E., Casale, A.E., Sachdev, R.N., McGinley, M.J. & McCormick, D.A. Motor cortex feedback influences sensory processing by modulating network state. Neuron 79, 567–578 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Gutierrez, R., Simon, S.A. & Nicolelis, M.A. Licking-induced synchrony in the taste-reward circuit improves cue discrimination during learning. J. Neurosci. 30, 287–303 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Pantoja, J. et al. Neuronal activity in the primary somatosensory thalamocortical loop is modulated by reward contingency during tactile discrimination. J. Neurosci. 27, 10608–10620 (2007).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Fanselow, E.E. & Nicolelis, M.A. Behavioral modulation of tactile responses in the rat somatosensory system. J. Neurosci. 19, 7603–7616 (1999).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Aronoff, R. et al. Long-range connectivity of mouse primary somatosensory barrel cortex. Eur. J. Neurosci. 31, 2221–2233 (2010).

    Article  PubMed  Google Scholar 

  7. Alloway, K.D., Lou, L., Nwabueze-Ogbo, F. & Chakrabarti, S. Topography of cortical projections to the dorsolateral neostriatum in rats: multiple overlapping sensorimotor pathways. J. Comp. Neurol. 499, 33–48 (2006).

    Article  PubMed  Google Scholar 

  8. Chen, J.L., Carta, S., Soldado-Magraner, J., Schneider, B.L. & Helmchen, F. Behavior-dependent recruitment of long-range projection neurons in somatosensory cortex. Nature 499, 336–340 (2013).

    Article  CAS  PubMed  Google Scholar 

  9. Sato, T.R. & Svoboda, K. The functional properties of barrel cortex neurons projecting to the primary motor cortex. J. Neurosci. 30, 4256–4260 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Yamashita, T. et al. Membrane potential dynamics of neocortical projection neurons driving target-specific signals. Neuron 80, 1477–1490 (2013).

    Article  CAS  PubMed  Google Scholar 

  11. Smith, A.C. et al. Dynamic analysis of learning in behavioral experiments. J. Neurosci. 24, 447–461 (2004).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Boubenec, Y., Shulz, D.E. & Debregeas, G. Whisker encoding of mechanical events during active tactile exploration. Front. Behav. Neurosci. 6, 74 (2012).

    Article  PubMed  PubMed Central  Google Scholar 

  13. Arabzadeh, E., Zorzin, E. & Diamond, M.E. Neuronal encoding of texture in the whisker sensory pathway. PLoS Biol. 3, e17 (2005).

    Article  PubMed  PubMed Central  Google Scholar 

  14. von Heimendahl, M., Itskov, P.M., Arabzadeh, E. & Diamond, M.E. Neuronal activity in rat barrel cortex underlying texture discrimination. PLoS Biol. 5, e305 (2007).

    Article  PubMed  PubMed Central  Google Scholar 

  15. Wolfe, J. et al. Texture coding in the rat whisker system: slip-stick versus differential resonance. PLoS Biol. 6, e215 (2008).

    Article  PubMed  PubMed Central  Google Scholar 

  16. Zuo, Y., Perkon, I. & Diamond, M.E. Whisking and whisker kinematics during a texture classification task. Phil. Trans. R. Soc. Lond. B 366, 3058–3069 (2011).

    Article  Google Scholar 

  17. Madisen, L. et al. A robust and high-throughput Cre reporting and characterization system for the whole mouse brain. Nat. Neurosci. 13, 133–140 (2010).

    Article  CAS  PubMed  Google Scholar 

  18. Horikawa, K. et al. Spontaneous network activity visualized by ultrasensitive Ca2+ indicators, yellow Cameleon-Nano. Nat. Methods 7, 729–732 (2010).

    Article  CAS  PubMed  Google Scholar 

  19. Curtis, J.C. & Kleinfeld, D. Phase-to-rate transformations encode touch in cortical neurons of a scanning sensorimotor system. Nat. Neurosci. 12, 492–501 (2009).

    CAS  PubMed  PubMed Central  Google Scholar 

  20. de Kock, C.P., Bruno, R.M., Spors, H. & Sakmann, B. Layer- and cell type–specific suprathreshold stimulus representation in rat primary somatosensory cortex. J. Physiol. (Lond.) 581, 139–154 (2007).

    Article  CAS  Google Scholar 

  21. O'Connor, D.H., Peron, S.P., Huber, D. & Svoboda, K. Neural activity in barrel cortex underlying vibrissa-based object localization in mice. Neuron 67, 1048–1061 (2010).

    Article  CAS  PubMed  Google Scholar 

  22. Fee, M.S., Mitra, P.P. & Kleinfeld, D. Central versus peripheral determinants of patterned spike activity in rat vibrissa cortex during whisking. J. Neurophysiol. 78, 1144–1149 (1997).

    Article  CAS  PubMed  Google Scholar 

  23. Jadhav, S.P., Wolfe, J. & Feldman, D.E. Sparse temporal coding of elementary tactile features during active whisker sensation. Nat. Neurosci. 12, 792–800 (2009).

    Article  CAS  PubMed  Google Scholar 

  24. Xu, N.L. et al. Nonlinear dendritic integration of sensory and motor input during an active sensing task. Nature 492, 247–251 (2012).

    Article  CAS  PubMed  Google Scholar 

  25. Gambino, F. & Holtmaat, A. Spike-timing-dependent potentiation of sensory surround in the somatosensory cortex is facilitated by deprivation-mediated disinhibition. Neuron 75, 490–502 (2012).

    Article  CAS  PubMed  Google Scholar 

  26. Guic, E., Carrasco, X., Rodriguez, E., Robles, I. & Merzenich, M.M. Plasticity in primary somatosensory cortex resulting from environmentally enriched stimulation and sensory discrimination training. Biol. Res. 41, 425–437 (2008).

    PubMed  Google Scholar 

  27. Kuhlman, S.J., O'Connor, D.H., Fox, K. & Svoboda, K. Structural plasticity within the barrel cortex during initial phases of whisker-dependent learning. J. Neurosci. 34, 6078–6083 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. House, D.R., Elstrott, J., Koh, E., Chung, J. & Feldman, D.E. Parallel regulation of feedforward inhibition and excitation during whisker map plasticity. Neuron 72, 819–831 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Gambino, F. et al. Sensory-evoked LTP driven by dendritic plateau potentials in vivo. Nature 515, 116–119 (2014).

    Article  CAS  PubMed  Google Scholar 

  30. Huber, D. et al. Multiple dynamic representations in the motor cortex during sensorimotor learning. Nature 484, 473–478 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Wiest, M.C., Thomson, E., Pantoja, J. & Nicolelis, M.A. Changes in S1 neural responses during tactile discrimination learning. J. Neurophysiol. 104, 300–312 (2010).

    Article  PubMed  PubMed Central  Google Scholar 

  32. Petreanu, L. et al. Activity in motor-sensory projections reveals distributed coding in somatosensation. Nature 489, 299–303 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Lee, S., Kruglikov, I., Huang, Z.J., Fishell, G. & Rudy, B. A disinhibitory circuit mediates motor integration in the somatosensory cortex. Nat. Neurosci. 16, 1662–1670 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Nienborg, H. & Cumming, B.G. Decision-related activity in sensory neurons reflects more than a neuron's causal effect. Nature 459, 89–92 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Romo, R., Lemus, L. & de Lafuente, V. Sense, memory and decision-making in the somatosensory cortical network. Curr. Opin. Neurobiol. 22, 914–919 (2012).

    Article  CAS  PubMed  Google Scholar 

  36. Diamond, M.E. & Arabzadeh, E. Whisker sensory system: from receptor to decision. Prog. Neurobiol. 103, 28–40 (2013).

    Article  PubMed  Google Scholar 

  37. Romo, R., Hernandez, A., Zainos, A., Lemus, L. & Brody, C.D. Neuronal correlates of decision-making in secondary somatosensory cortex. Nat. Neurosci. 5, 1217–1225 (2002).

    Article  CAS  PubMed  Google Scholar 

  38. Langer, D. et al. HelioScan: a software framework for controlling in vivo microscopy setups with high hardware flexibility, functional diversity and extendibility. J. Neurosci. Methods 215, 38–52 (2013).

    Article  PubMed  Google Scholar 

  39. Grewe, B.F., Voigt, F.F., van't Hoff, M. & Helmchen, F. Fast two-layer two-photon imaging of neuronal cell populations using an electrically tunable lens. Biomed. Opt. Express 2, 2035–2046 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Chen, J.L., Pfaffli, O.A., Voigt, F.F., Margolis, D.J. & Helmchen, F. Online correction of licking-induced brain motion during two-photon imaging with a tunable lens. J. Physiol. (Lond.) 591, 4689–4698 (2013).

    Article  CAS  Google Scholar 

  41. Knutsen, P.M., Derdikman, D. & Ahissar, E. Tracking whisker and head movements in unrestrained behaving rodents. J. Neurophysiol. 93, 2294–2301 (2005).

    Article  PubMed  Google Scholar 

  42. Conte, W.L., Kamishina, H. & Reep, R.L. Multiple neuroanatomical tract-tracing using fluorescent Alexa Fluor conjugates of cholera toxin subunit B in rats. Nat. Protoc. 4, 1157–1166 (2009).

    Article  CAS  PubMed  Google Scholar 

  43. Dombeck, D.A., Khabbaz, A.N., Collman, F., Adelman, T.L. & Tank, D.W. Imaging large-scale neural activity with cellular resolution in awake, mobile mice. Neuron 56, 43–57 (2007).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. Birdwell, J.A. et al. Biomechanical models for radial distance determination by the rat vibrissal system. J. Neurophysiol. 98, 2439–2455 (2007).

    Article  PubMed  Google Scholar 

  45. Green, D.M. & Swets, J.A. Signal Detection Theory and Psychophysics (Wiley, 1966).

Download references

Acknowledgements

We thank A. Iqbal for assistance with data analysis, and A. Ayaz, L. Egolf, S. Hofer, T. Mrsic-Flogel and S. Sachidhanandam for comments on the manuscript. This work was supported by a Swiss National Science Foundation (SNSF) grant (310030-127091 to F.H.), the Swiss SystemsX.ch initiative (project 2008/2011-Neurochoice to F.H. and B.L.S.), a SNSF Sinergia grant to F.H. and B.L.S. (CRSII3_147660/1), an AMBIZIONE grant from the SNSF to D.J.M. (PZ00P3_143232), a Forschungskredit of the University of Zurich (grant 541541808 to J.L.C.), and a fellowship from the US National Science Foundation, International Research Fellowship Program (grant 1158914 to J.L.C.).

Author information

Author notes

  1. David J Margolis

    Present address: Present address: Department of Cell Biology and Neuroscience, Rutgers University, New Brunswick, New Jersey, USA.,

Authors and Affiliations

  1. Brain Research Institute, University of Zurich, Zurich, Switzerland

    Jerry L Chen, David J Margolis, Atanas Stankov, Lazar T Sumanovski & Fritjof Helmchen

  2. Neuroscience Center Zurich, University of Zurich/ETH Zurich, Zurich, Switzerland

    David J Margolis & Fritjof Helmchen

  3. Institute of Neuroinformatics, University of Zurich/ETH Zurich, Zurich, Switzerland

    Atanas Stankov

  4. Brain Mind Institute, EPFL, Lausanne, Switzerland

    Bernard L Schneider

Authors
  1. Jerry L Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. David J Margolis
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Atanas Stankov
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Lazar T Sumanovski
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Bernard L Schneider
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Fritjof Helmchen
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

J.L.C., D.J.M. and F.H. designed the study. J.L.C. and D.J.M. performed the experiments. J.L.C., A.S., L.T.S. and F.H. performed data analysis. B.L.S. contributed viral reagents. J.L.C. and F.H. wrote the paper.

Corresponding author

Correspondence to Fritjof Helmchen.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Integrated supplementary information

Supplementary Figure 1 Individual animal performance.

a, State-space model identification of learning onset and completion for example animal. Left panel shows a learning curve estimate (red line) determined from incorrect (gray marks) and correct responses (black tick) of the animal shown above. Black lines indicate the 95% confidence interval. The trial representing learning onset represents the trial in which the lower 95% confidence bound of the estimated learning curve exceeds and remains above 50% correct response probability for the remainder of the experimental time course. Middle panel shows the level of certainty the ideal observer has that the animal’s performance is better than chance. The trial representing the completion of learning represents the trial in which the ideal observer obtained a 0.95 level of certainty (blue line) that the animal’s performance is better than chance at each trial for the remainder of the experimental time course. Right panel shows animal performance (d’) overlaid with pre-stimulus lick rate and pre-stimulus whisking amplitude along with training phases as defined by learning onset and completion indicated. b, Animal performance (d’) overlaid with pre-stimulus lick rate and pre-stimulus whisking amplitude along with defined training phases for remaining animals in the study.

Supplementary Figure 2 Individual animal hit and false alarm rate across training.

Supplementary Figure 3 Whisking behavior during learning.

a, Time course of 7-10 Hz whisking power aligned to first touch across different training periods (solid line, mean; shared area, s.e.m.) b, Cumulative distribution of pre-stimulus whisking amplitude across naive and expert trials. c, Whisking range of principal whisker across the initial 1-s period of touch. d, Mean speed of principal whisker across the initial 1-s period of touch (error bars, s.e.m. [c,d]; n = 7 animals, 13,715 trials [a,b]; 5 animals, 4,606 trials [c,d]).

Supplementary Figure 4 Additional whisker kinematics analysis.

a, Trial-by-trial correlation of curvature change and the number of stick-slip events during the initial 1-s touch period across textures and training. b, Expert animal performance on trials with or without principal whisker stick-slips events. Grey line indicates individual animals. Black line indicates mean. c, ROC analysis for P100 vs.P1200 by curvature change or stick-slip events according to whisking amplitude. Red lines represents 95th percentile of shuffled performance. * indicates discrimination above chance. d, Maximum curvature change across the first second after touch onset for P100 and P1200 textures in non-training sessions. e, Stick-slip events across the first second after touch onset for P100 and P1200 textures in non-training sessions. Error bars, s.e.m.; *P < 0.05.

Supplementary Figure 5 In vivo two-photon calcium imaging during behavior.

a, In vivo images of YC-Nano140-expressing neurons at three imaging depths. b, Single trial calcium responses of identified S2P (red), M1P (blue), or UNL (blue) touch neurons and neuropil (NP) indicated in [a] from naive animal during texture discrimination training. Focus alternated between assigned imaging depths on a trial-by-trial basis. Orange shaded area indicates period of texture touch. Texture and decision for each trial are indicated below.

Supplementary Figure 6 Example of touch neuron classification.

a, Single trial responses of example UNL and S2P neuron aligned to first touch (dotted line) sorted by trial number. Corresponding session number and d’ are also shown. b, Cross-correlation analysis of calcium signals with touch across different time lags for neurons in [a] for session indicated with (*). UNL neuron shown is classified as “non-touch”. S2P neuron shown is classified as “touch”. Correlations with positive time lags indicate that calcium signals follow touch onset whereas correlations with negative time lags indicate that calcium signals precede touch onset. Shaded grey region indicate time lags for classifying touch neurons. Shaded orange regions indicate 95% c.i. from bootstrap test. c, Example concatenated calcium signals from the non-touch [upper] and touch [lower] neurons along with corresponding touch vectors at varying lags. Bottom rows: Scatter plot of ΔR/R values for all points of the calcium trace sorted according to touch (T) and non-touch (NT) periods for different lag times with regression lines shown. d, Behaviour classification of individual neurons (rows) across sessions (columns) for one animal. Session type is indicated below. Only neurons imaged across all sessions are shown. e, Distribution of classified neurons across sessions for UNL neurons pooled across animals. f, Distribution of touch neurons from active neuron pool. Larger fraction of touch neurons are seen in S2P neurons during expert sessions relative to other cell types. g, Session-to-session (S-to-S) variability in touch response during training periods compared to control non-trained animals for UNL neurons. h, Distribution of UNL touch neurons in pre-training sessions and their classification in post-training sessions (‘fate’, left bars). Distribution of touch neurons in post-training sessions and their classification in pre-training sessions (‘history’, right bars). Only neurons imaged across both sessions are shown. Error bars: s.d. from bootstrap test [e, f, h], s.e.m. [g]. n = 707 neurons in 7 training animals; *P < 0.05. ** P < 0.02.

Supplementary Figure 7 Calcium responses to whisker kinematics.

a, Whisker kinematics and calcium signals of two example touch neurons with preference for maximum curvature change (left) or stick-slip events (right). For each neuron, individual trials (rows) are shown with texture presented, first second of calcium signals after initial touch, along with corresponding kinematic parameters (maximum curvature change, cumulative stick-slip events) binned according to imaging frame duration. Trials are sorted in descending order according to mean ΔR/R. Cross correlation values (R) of calcium transients against each kinematic parameter are also displayed and large R values are highlighted with a red box. Average trial responses binned by kinematic features are shown below. b, Average calcium responses of touch neuron across cell types to stick-slip events aligned to first touch (dotted line) in naive and expert trials. c, Average calcium responses of touch neuron across cell types to maximum curvature changes aligned to first touch (dotted line) in naive and expert trials. d, Relative change in calcium transient response peak amplitude with increasing stick-slip events (left) or maximum curvature change (right) for UNL neurons in naive and expert trials. Changes are quantified relative to ΔR/R amplitudes for zero stick-slip events and the smallest |ΔΚ|max bin, respectively. e, Scatter plot of single neuron response correlations to maximum curvature change and stick-slip events for all cell types in naive and expert trials. Linear regressions are shown (dotted lines). f, Touch neuron correlation of mean calcium responses to kinematic features in naive and expert trials across UNL neurons. Error bars and shaded area: s.e.m. n = 3,373 trials, 307 touch neurons, 5 animals.

Supplementary Figure 8 Additional ROC analysis of neurons across learning.

a, Fraction of active neurons discriminating hit vs. correct rejection trials above chance determined by single-cell ROC analysis. More S2P neurons can discriminate hit vs. correct rejection than M1P neurons. Error bars: s.d., * P < 0.05, bootstrap analysis. n = 215 active neurons, 7 animals. b, Distribution of neurons discriminating go vs. no go (P100 vs. P1200) trials across training for UNL neurons obtained by sliding window single cell ROC analysis. Trial times are aligned and normalized according to learning onset and completion (dotted white line). Neurons are ranked according to fraction of trials discriminating (values indicated by color plot). Fraction of neurons discriminating above chance is indicated by solid white line, corresponding to 95th percentile of distribution from permutation test of trial labels. Fraction of neurons discriminating above chance for pre- and post-training sessions is shown on both sides, respectively. c, Mean change in fraction of trials discriminating relative to naive phase during training for UNL neurons. d, Standard deviation in fraction of trials discriminating P100 vs. P1200 for neurons across learning. An increase variance in discriminability was observed in S2P neurons. e, Fraction of trials discriminating P100 vs. P1200 across cell types ranked according to the top 99th, 90th, and 75th percentile. Top ranked S2P neurons increased their discriminability during learning.

Supplementary Figure 9 Calcium responses to P100 and P1200 trials.

a, Average calcium responses of touch neuron across cell types to P100 vs. P1200 aligned to first touch (dotted line) in naive trials. Deconvolved traces representing estimated mean changes in firing rate (ΔFR) are shown below the calcium traces. Arrows indicate mean reaction time in expert sessions for hit trials. b, Average calcium responses of touch neuron across cell types to P100 vs. P1200 aligned to first touch (dotted line) in expert trials. Deconvolved traces representing estimated changes in firing rate (ΔFR) are shown below the calcium traces. Arrows indicate overall mean reaction time. c, Peak touch neuron calcium response across UNL neurons for P100 vs. P1200 trials in naive and expert session. d, Cumulative distribution of peak calcium response for UNL touch neurons for P100 vs. P1200 trials in naive and expert sessions. e, Response preference for P100 vs. P1200 trials in UNL touch neurons lost after naive sessions or gained during expert sessions. f, Response preference for P100 vs. P1200 textures of neurons responding to touch at least one naive and expert session. No consistent change in response preference was observed for these neurons. Error bars and shaded area: s.e.m. n = 411 touch neurons [a-b], 215 touch neurons [f], 7 animals.

Supplementary Figure 10 Calcium responses to false alarm and correct rejection trials.

Average touch neuron calcium response to false alarm vs. correct rejection in late learning to expert trials. Deconvolved traces representing estimated mean changes in firing rate (ΔFR) are shown below the calcium traces. Arrows indicate mean reaction time in expert sessions for hit trials.

Supplementary information

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Chen, J., Margolis, D., Stankov, A. et al. Pathway-specific reorganization of projection neurons in somatosensory cortex during learning. Nat Neurosci 18, 1101–1108 (2015). https://doi.org/10.1038/nn.4046

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/nn.4046

This article is cited by

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