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.

Equalization of odor representations by a network of electrically coupled inhibitory interneurons

Abstract

Robustness of neuronal activity patterns against variations in input intensity is critical for neuronal computations. We found that odor representations in the olfactory bulb were stabilized by interneurons that were densely coupled to the output neurons by electrical and GABAergic synapses. This interneuron network modulated responses of output neurons as a function of stimulus intensity in two ways: it globally boosted responses to weak odors, but attenuated responses to strong odors, and it increased the sensitivity of some output neurons, but decreased the sensitivity of others. These effects are closely related to strategies used in engineering to increase dynamic range. Together, they maintained not only the mean, but also the distribution, of activity across the population of output neurons within narrow limits, which is important for pattern classification. Neuronal circuits in the olfactory bulb therefore stabilize combinatorial sensory representations against variations in stimulus intensity by generic mechanisms.

Your institute does not have access to this article

Relevant articles

Open Access articles citing this article.

Access options

Buy article

Get time limited or full article access on ReadCube.

$32.00

All prices are NET prices.

Figure 1: Functional properties of dlx4/6 neurons.
Figure 2: Dual coupling between dlx4/6 neurons and mitral cells.
Figure 3: Dense and widespread coupling between dlx4/6 neurons and mitral cells.
Figure 4: Effects of dlx4/6 neurons on mitral cell responses.
Figure 5: Effects of dlx4/6 neurons on response sensitivity: DRC.
Figure 6: Effects of dlx4/6 neurons on response sensitivity: DRS.
Figure 7: Effects of dlx4/6 neurons on mitral cell tuning.
Figure 8: Equalization of activity distributions by dlx4/6 neurons.

References

  1. McNaughton, B.L. & Morris, R.G.M. Hippocampal synaptic enhancement and information storage within a distributed memory system. Trends Neurosci. 10, 408–415 (1987).

    Article  Google Scholar 

  2. Christiani, N. & Shawe-Taylor, J. Introduction to Support Vector Machines and Other Kernel-Based Learning Methods (Cambridge University Press, Cambridge, 2000).

  3. Carandini, M. & Heeger, D.J. Normalization as a canonical neural computation. Nat. Rev. Neurosci. 13, 51–62 (2012).

    Article  CAS  Google Scholar 

  4. Luo, S.X., Axel, R. & Abbott, L.F. Generating sparse and selective third-order responses in the olfactory system of the fly. Proc. Natl. Acad. Sci. USA 107, 10713–10718 (2010).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  5. Olsen, S.R., Bhandawat, V. & Wilson, R.I. Divisive normalization in olfactory population codes. Neuron 66, 287–299 (2010).

    CAS  PubMed  PubMed Central  Google Scholar 

  6. Friedrich, R.W. & Korsching, S.I. Combinatorial and chemotopic odorant coding in the zebrafish olfactory bulb visualized by optical imaging. Neuron 18, 737–752 (1997).

    Article  CAS  PubMed  Google Scholar 

  7. Rubin, B.D. & Katz, L.C. Optical imaging of odorant representations in the mammalian olfactory bulb. Neuron 23, 499–511 (1999).

    Article  CAS  PubMed  Google Scholar 

  8. Wachowiak, M. & Cohen, L.B. Representation of odorants by receptor neuron input to the mouse olfactory bulb. Neuron 32, 723–735 (2001).

    Article  CAS  PubMed  Google Scholar 

  9. Leveteau, J. & MacLeod, P. Olfactory discrimination in the rabbit olfactory glomerulus. Science 153, 175–176 (1966).

    Article  CAS  PubMed  Google Scholar 

  10. Wilson, D.A. & Sullivan, R.M. Cortical processing of odor objects. Neuron 72, 506–519 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Gross-Isseroff, R. & Lancet, D. Concentration-dependent changes of perceived odor quality. Chem. Senses 13, 191–204 (1988).

    Article  Google Scholar 

  12. Cleland, T.A., Johnson, B.A., Leon, M. & Linster, C. Relational representation in the olfactory system. Proc. Natl. Acad. Sci. USA 104, 1953–1958 (2007).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Friedrich, R.W. & Laurent, G. Dynamic optimization of odor representations in the olfactory bulb by slow temporal patterning of mitral cell activity. Science 291, 889–894 (2001).

    Article  CAS  PubMed  Google Scholar 

  14. Friedrich, R.W. & Laurent, G. Dynamics of olfactory bulb input and output activity during odor stimulation in zebrafish. J. Neurophysiol. 91, 2658–2669 (2004).

    Article  PubMed  Google Scholar 

  15. Niessing, J. & Friedrich, R.W. Olfactory pattern classification by discrete neuronal network states. Nature 465, 47–52 (2010).

    Article  CAS  PubMed  Google Scholar 

  16. Friedrich, R.W. Information processing in the olfactory system of zebrafish. Annu. Rev. Neurosci. 36, 383–402 (2013).

    Article  CAS  PubMed  Google Scholar 

  17. Zerucha, T. et al. A highly conserved enhancer in the Dlx5/Dlx6 intergenic region is the site of cross-regulatory interactions between Dlx genes in the embryonic forebrain. J. Neurosci. 20, 709–721 (2000).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Zhu, P. et al. Optogenetic dissection of neuronal circuits in zebrafish using viral gene transfer and the Tet system. Front. Neural Circuits 3, 21 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Li, J. et al. Early development of functional spatial maps in the zebrafish olfactory bulb. J. Neurosci. 25, 5784–5795 (2005).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Bundschuh, S.T., Zhu, P., Zhang Schärer, Y.-P. & Friedrich, R.W. Dopaminergic modulation of mitral cells and odor responses in the zebrafish olfactory bulb. J. Neurosci. 32, 6830–6840 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Kiyokage, E. et al. Molecular identity of periglomerular and short axon cells. J. Neurosci. 30, 1185–1196 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Kosaka, T. & Kosaka, K. Tyrosine hydroxylase–positive GABAergic juxtaglomerular neurons are the main source of the interglomerular connections in the mouse main olfactory bulb. Neurosci. Res. 60, 349–354 (2008).

    Article  CAS  PubMed  Google Scholar 

  23. Aungst, J.L. et al. Centre-surround inhibition among olfactory bulb glomeruli. Nature 426, 623–629 (2003).

    Article  CAS  PubMed  Google Scholar 

  24. Liu, S., Plachez, C., Shao, Z., Puche, A. & Shipley, M.T. Olfactory bulb short axon cell release of GABA and dopamine produces a temporally biphasic inhibition-excitation response in external tufted cells. J. Neurosci. 33, 2916–2926 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Yaksi, E., Judkewitz, B. & Friedrich, R.W. Topological reorganization of odor representations in the olfactory bulb. PLoS Biol. 5, e178 (2007).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Zhu, P., Fajardo, O., Shum, J., Zhang Schärer, Y.-P. & Friedrich, R.W. High-resolution optical control of spatiotemporal neuronal activity patterns in zebrafish using a digital micromirror device. Nat. Protoc. 7, 1410–1425 (2012).

    Article  CAS  PubMed  Google Scholar 

  27. Tallini, Y.N. et al. Imaging cellular signals in the heart in vivo: cardiac expression of the high-signal Ca2+ indicator GCaMP2. Proc. Natl. Acad. Sci. USA 103, 4753–4758 (2006).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Kosaka, K. & Kosaka, T. Synaptic organization of the glomerulus in the main olfactory bulb: compartments of the glomerulus and heterogeneity of the periglomerular cells. Anat. Sci. Int. 80, 80–90 (2005).

    Article  PubMed  Google Scholar 

  29. Rickgauer, J.P. & Tank, D.W. Two-photon excitation of channelrhodopsin-2 at saturation. Proc. Natl. Acad. Sci. USA 106, 15025–15030 (2009).

    Article  PubMed  PubMed Central  Google Scholar 

  30. Bamann, C., Kirsch, T., Nagel, G. & Bamberg, E. Spectral characteristics of the photocycle of channelrhodopsin-2 and its implication for channel function. J. Mol. Biol. 375, 686–694 (2008).

    Article  CAS  PubMed  Google Scholar 

  31. Gradinaru, V. et al. Molecular and cellular approaches for diversifying and extending optogenetics. Cell 141, 154–165 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Pinching, A.J. & Powell, T.P. The neuropil of the glomeruli of the olfactory bulb. J. Cell Sci. 9, 347–377 (1971).

    CAS  PubMed  Google Scholar 

  33. Kosaka, T. & Kosaka, K. Intraglomerular dendritic link connected by gap junctions and chemical synapses in the mouse main olfactory bulb: electron microscopic serial section analyses. Neuroscience 131, 611–625 (2005).

    Article  CAS  PubMed  Google Scholar 

  34. Zhang, C. & Restrepo, D. Heterogeneous expression of connexin 36 in the olfactory epithelium and glomerular layer of the olfactory bulb. J. Comp. Neurol. 459, 426–439 (2003).

    Article  CAS  PubMed  Google Scholar 

  35. Zhang, C. & Restrepo, D. Expression of connexin 45 in the olfactory system. Brain Res. 929, 37–47 (2002).

    Article  CAS  PubMed  Google Scholar 

  36. Blumhagen, F. et al. Neuronal filtering of multiplexed odour representations. Nature 479, 493–498 (2011).

    Article  CAS  PubMed  Google Scholar 

  37. Sato, Y., Miyasaka, N. & Yoshihara, Y. Mutually exclusive glomerular innervation by two distinct types of olfactory sensory neurons revealed in transgenic zebrafish. J. Neurosci. 25, 4889–4897 (2005).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Yaksi, E. & Friedrich, R.W. Reconstruction of firing rate changes across neuronal populations by temporally deconvolved Ca2+ imaging. Nat. Methods 3, 377–383 (2006).

    Article  CAS  PubMed  Google Scholar 

  39. Lechner, H.A., Lein, E.S. & Callaway, E.M. A genetic method for selective and quickly reversible silencing of mammalian neurons. J. Neurosci. 22, 5287–5290 (2002).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Tan, E.M. et al. Selective and quickly reversible inactivation of mammalian neurons in vivo using the Drosophila allatostatin receptor. Neuron 51, 157–170 (2006).

    Article  CAS  PubMed  Google Scholar 

  41. Heeger, D.J. Normalization of cell responses in cat striate cortex. Vis. Neurosci. 9, 181–197 (1992).

    Article  CAS  PubMed  Google Scholar 

  42. Yaksi, E. & Wilson, R.I. Electrical coupling between olfactory glomeruli. Neuron 67, 1034–1047 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Huang, J., Zhang, W., Qiao, W., Hu, A. & Wang, Z. Functional connectivity and selective odor responses of excitatory local interneurons in Drosophila antennal lobe. Neuron 67, 1021–1033 (2010).

    Article  CAS  PubMed  Google Scholar 

  44. Olsen, S.R., Bhandawat, V. & Wilson, R.I. Excitatory interactions between olfactory processing channels in the Drosophila antennal lobe. Neuron 54, 89–103 (2007).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  45. Connors, B.W. & Long, M.A. Electrical synapses in the mammalian brain. Annu. Rev. Neurosci. 27, 393–418 (2004).

    Article  CAS  PubMed  Google Scholar 

  46. Kastner, D.B. & Baccus, S.A. Coordinated dynamic encoding in the retina using opposing forms of plasticity. Nat. Neurosci. 14, 1317–1322 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Völgyi, B., Deans, M.R., Paul, D.L. & Bloomfield, S.A. Convergence and segregation of the multiple rod pathways in mammalian retina. J. Neurosci. 24, 11182–11192 (2004).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  48. Liberman, M.C. Auditory-nerve response from cats raised in a low-noise chamber. J. Acoust. Soc. Am. 63, 442–455 (1978).

    Article  CAS  PubMed  Google Scholar 

  49. Barbour, D.L. Intensity-invariant coding in the auditory system. Neurosci. Biobehav. Rev. 35, 2064–2072 (2011).

    Article  PubMed  PubMed Central  Google Scholar 

  50. Marr, D. Simple memory: a theory for archicortex. Phil. Trans. R. Soc. Lond. B 262, 23–81 (1971).

    Article  CAS  Google Scholar 

  51. Tang, W. et al. Faithful expression of multiple proteins via 2A-peptide self-processing: a versatile and reliable method for manipulating brain circuits. J. Neurosci. 29, 8621–8629 (2009).

    CAS  PubMed  PubMed Central  Google Scholar 

  52. Kawakami, K. et al. A transposon-mediated gene trap approach identifies developmentally regulated genes in zebrafish. Dev. Cell 7, 133–144 (2004).

    Article  CAS  PubMed  Google Scholar 

  53. Karpova, A.Y., Tervo, D.G., Gray, N.W. & Svoboda, K. Rapid and reversible chemical inactivation of synaptic transmission in genetically targeted neurons. Neuron 48, 727–735 (2005).

    Article  CAS  PubMed  Google Scholar 

  54. Boyden, E.S., Zhang, F., Bamberg, E., Nagel, G. & Deisseroth, K. Millisecond-timescale, genetically targeted optical control of neural activity. Nat. Neurosci. 8, 1263–1268 (2005).

    CAS  PubMed  Google Scholar 

  55. Kwan, K.M. et al. The Tol2kit: a multisite gateway-based construction kit for Tol2 transposon transgenesis constructs. Dev. Dyn. 236, 3088–3099 (2007).

    Article  CAS  PubMed  Google Scholar 

  56. Distel, M., Wullimann, M.F. & Koster, R.W. Optimized Gal4 genetics for permanent gene expression mapping in zebrafish. Proc. Natl. Acad. Sci. USA 106, 13365–13370 (2009).

    Article  PubMed  PubMed Central  Google Scholar 

  57. Schärer, Y.P., Shum, J., Moressis, A. & Friedrich, R.W. Dopaminergic modulation of synaptic transmission and neuronal activity patterns in the zebrafish homolog of olfactory cortex. Front. Neural Circuits 6, 76 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  58. Pologruto, T.A., Sabatini, B.L. & Svoboda, K. ScanImage: flexible software for operating laser scanning microscopes. Biomed. Eng. Online 2, 13 (2003).

    PubMed  PubMed Central  Google Scholar 

  59. Suter, B.A. et al. Ephus: multipurpose data acquisition software for neuroscience experiments. Front. Neural Circuits 4, 100 (2010).

    Article  PubMed  PubMed Central  Google Scholar 

  60. Vinje, W.E. & Gallant, J.L. Sparse coding and decorrelation in primary visual cortex during natural vision. Science 287, 1273–1276 (2000).

    Article  CAS  PubMed  Google Scholar 

Download references

Acknowledgements

We thank G. Jacobson, I. Namekawa, G. Keller and F. Albeanu for comments on the manuscript and members of the Friedrich laboratory for fruitful discussions. We are grateful to K. Deisseroth (Stanford University) for DNA constructs containing NpHR3.0-YFP and ChR2-YFP, and to E. Callaway (Salk Institute) for DNA constructs containing AlstR. This work was supported by the Novartis Research Foundation, the European Molecular Biology Organization, the Human Frontiers Science Program and the Swiss National Science Foundation.

Author information

Authors and Affiliations

Authors

Contributions

R.W.F. and P.Z. conceived the study. P.Z. performed the experiments, except for those shown in Supplementary Figure 3a, which were performed by T.F. P.Z. and T.F. generated transgenic fish lines. P.Z. and R.W.F. constructed the equipment and analyzed the data. R.W.F. wrote the manuscript with help from P.Z. and T.F.

Corresponding author

Correspondence to Rainer W Friedrich.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Integrated supplementary information

Supplementary Figure 1 Pattern equalization and classification

(a) Basic effects of intensity on pattern recognition by models that linearly summate and threshold multiple inputs (perceptrons). The scheme shows two perceptrons (“target neurons”) that receive different inputs from a population of input units (e.g., “mitral cells”). A subset of these input units is active. Left: the perceptron whose connections match the input pattern generates a suprathreshold response whereas the other perceptron does not. The input pattern is therefore classified correctly. Center and right: pattern recognition becomes incorrect when the same input units are active at lower intensity (failure) or higher intensity (false positive). (b) This problem can be overcome by normalizing input intensities by their mean when the distribution of activity across the input population does not change with intensity. Such normalization can, for example, be achieved by global feed-forward inhibition onto all perceptrons. (c) Often, however, increasing the strength of an input not only increases the response of the most sensitive neurons but also recruits less sensitive neurons with different tuning properties. This effect is common in the periphery of most or all sensory systems. As a result, activity becomes more broadly distributed across the population, and population activity patterns cannot be properly equalized by normalizing to the mean. (d) Intensity-invariance of response patterns therefore requires an equalization of the mean and the distribution of activity patterns across the population but it is unclear how this can be achieved.

Supplementary Figure 2 Inward currents in mitral cells evoked by optical stimulation of dlx4/6-neurons are resistant to antagonists of neurotransmitter receptors.

Traces show currents recorded in mitral cells at −70 mV in response to wide-field optical stimulation of dlx4/6-neurons (dlx-itTA/Ptet-Chr2YFP transgenic fish). Each trace is an average over recordings from multiple mitral cells as indicated. Gray traces show baseline responses in the presence of different drugs; blue traces show responses after addition of further drugs. Drugs comprise antagonists of different neurotransmitter receptors: (a) ionotropic glutamate receptors (AMPA: NBQX; NMDA: APV; 2 fish); (b) GABAA (Gabazine) and GABAB (CGP54626) receptors (2 fish); (c) muscarinic (atropine) and nicotinic (mecamylamine) acetylcholine receptors (1 fish); (d) D1-like dopamine receptors (SKF 83566; 2 fish); (e) P2X purinergic receptors (TNP-ATP, iso-PPADS, PPADS; 4 fish). A previous study showed that optically evoked inward currents in mitral cells are not blocked by antagonists of D2-like dopamine receptors (Bundschuh et al., J. Neurosci. 32:6830-6840, 2012). The only drug that slightly decreased the inward current was an antagonist of D1-like dopamine receptors (d). This finding is consistent with an excitatory effect of dopamine released from short axon cells in mice that is mediated by D1-like receptors (Liu et al., J. Neurosci. 33:2916-2926, 2013). (f) Effect of CdCl2 on responses of two mitral cells, recorded in different fish, to two odors (overlay of two trials). CdCl2 inhibited spontaneous subthreshold membrane potential fluctuations and action potential firing, and abolished odor responses. These experiments confirm that CdCl2 (100 μM) blocked chemical synaptic transmission.

Supplementary Figure 3 Electrical coupling: further observations.

(a) To test whether gap junction inhibitors affect transmission at excitatory chemical synapses the olfactory tract was stimulated electrically (20 Hz, 10 pulses; black ticks) while voltage-clamp recordings were performed in neurons of the posterior zone of the dorsal telencephalon (Dp), a major target area of the olfactory bulb, as described (Blumhagen et al., Nature 479:493-498, 2011; Schärer et al., Front. Neural Circuits 6:76, 2012). Electrical stimulation evokes glutamatergic inputs from two sources (mitral cells and other Dp neurons) that arrive with different latencies. Top traces show currents recorded in one neuron at -70 mV before (dark green) and during (light green) exposure to gap junction inhibitors (CBX, 25 μM; 18α-GA and 18β-GA, 10 μM each; average over 5 trials). Bar graphs summarize effects of gap junction inhibitors on the peak current (left) and charge transfer (right; n = 5 Dp neurons from 1 fish). No significant change was detected. (b) Effect of gap junction inhibitors (CBX, 25 μM; 18α-GA and 18β-GA, 10 μM each) on the outward current measured in mitral cells at 0 mV, evoked by optical stimulation of dlx4/6-neurons (dlx-itTA/Ptet-ChR2YFP; averaged over n = 3 mitral cells from 1 fish). Bottom: enlargement of current onsets. AMPA and NMDA receptors were blocked throughout the experiment by NBQX and APV (50 μM each). Gap junction inhibitors reduced a late and slow component of the current but had only a minor effect on the initial fast component, indicating that GABAergic transmission was not affected. These results suggest that dlx4/6-neurons inhibit mitral cells by two pathways: (1) monosynaptic GABAergic connections and (2) one or more polysynaptic pathways that depend on gap junction coupling of dlx4/6-neurons to other glomerular layer interneurons such as periglomerular cells (see d). (c) Example of electrical coupling between a mitral cell (gray/black traces) and a dlx4/6-neuron (blue trace). Paired recordings were performed in the presence of CdCl2. Gray traces show membrane potential change evoked by current injection in 52 trials; black trace shows average. Blue trace is the averaged membrane potential response of the dlx4/6-neuron. (d) Example of electrical coupling between a dlx4/6-neuron and an unidentified juxtaglomerular interneuron in the presence of CdCl2. Traces were averaged over 93 trials (top) and 167 trials (bottom). In recordings from pairs of dlx4/6-neurons and mitral cells we noted that the time course of the voltage change in the recipient neuron sometimes deviated from a low-pass filtered version of the voltage change in the source neuron. In particular, the membrane potential change sometimes decreased over the duration of the injected current (Fig. 3a). This decrease could be caused by active conductances such as h-currents (Liu et al., J. Neurosci. 33:2916-2926, 2013). Alternatively, a rapid adaptation of electrical coupling can be mediated by connexins that form voltage-dependent gap junctions with rectifying properties such as connexin45 (Elenes et al., Biophys. J. 81:1406-1418, 2001). This connexin is expressed in the adult olfactory bulb, presumably by mitral cells (Zhang and Restrepo, Brain Res. 929:37-47, 2002).

Supplementary Figure 4 Responses of mitral cells in the lateral olfactory bulb to optical stimulation of sensory input to the medial olfactory bulb.

To further examine the spatial extent of synaptic interactions in the olfactory bulb we genetically targeted Chr2YFP to sensory neurons projecting to the medial olfactory bulb (zOMP-Chr2YFP; Sato et al., J. Neurosci. 25: 4889-4897, 2005; Blumhagen et al., Nature 479:493-498, 2011), optically stimulated the olfactory bulb with blue light, and recorded currents in mitral cells of the lateral olfactory bulb. In all mitral cells (n = 6), optical stimulation produced fast onset excitatory currents and slower inhibitory currents. Left: expression of Chr2YFP under the control of a promoter fragment from the olfactory marker protein (OMP) that targets expression to sensory neurons projecting to the medial olfactory bulb (zOMP-Chr2YFP transgenic fish; z-projection of multiphoton image stack). Green dot marks approximate position of mitral cells recorded in voltage clamp. A, anterior; M, medial. Right: currents recorded in mitral cells at holding potentials of +10 mV and -70 mV (averaged over n = 6 mitral cells from 3 fish). Blue line indicates optical stimulation (5 ms light flash).

Supplementary Figure 5 Analysis of inhibitory coupling between dlx4/6-neurons by paired recordings.

(a) Example of simultaneously recorded membrane potentials of a dlx4/6-neuron (gray: 73 individual traces; black: average) and a mitral cell (blue; averaged) in the absence of CdCl2. Traces are aligned on the peak of the initial action potential evoked in the dlx4/6-neuron (arrow; STA: spike-triggered average). A depolarizing membrane potential transient was observed in the mitral cell that coincides with the action potential in the dlx4/6-neuron, indicating that it is transmitted by electrical coupling. This transient is followed by a hyperpolarization. (b) Membrane potentials averaged over all pairs recorded in the absence of CdCl2 (n = 33 pairs in 8 fish). Arrow depicts the fast depolarizing membrane potential transient that coincided with the initial action potential in dlx4/6-neurons (a). (c) Distribution of mean membrane potential changes recorded in mitral cells (10 – 200 ms after onset of current step). (d) Membrane potential changes in mitral cells as a function of distance between somata, projected into the horizontal plane. Positive membrane potential changes at short distances were due to strong electrical coupling. In the absence of CdCl2, large depolarizing step currents (200 ms) were injected into the dlx4/6-neuron to produce a single action potential followed by a depolarization block. In the mitral cell, we frequently observed a depolarizing membrane potential transient at the onset of the current stimulus that was caused by electrical transmission of the action potential through gap junctions (a). Because this transient precluded the analysis of fast inhibitory postsynaptic potentials (IPSPs) locked to the onset of the current step we analyzed the mitral cell membrane potential during the last 190 ms of the current step. On average, the mitral cell membrane potential was slightly hyperpolarized compared to the 200 ms before the current step, confirming that dlx4/6 neurons inhibit mitral cells. The hyperpolarization was <1 mV in all pairs and usually <0.5 mV (c). These values underestimate the true inhibition because mitral cells were simultaneously depolarized by electrical coupling. Nevertheless, they indicate that the inhibition produced by a single dlx4/6 neuron is weak. Hyperpolarizing membrane potential changes were observed in 30 - 40% of the recorded pairs for distances up to 200 μm (c, d). Inhibitory coupling between dlx4/6-neurons and mitral cells is therefore dense, weak and widespread, as observed for electrical coupling.

Supplementary Figure 6 Optical measurement of mitral cell responses.

(a) Mitral cells loaded retrogradely with rhod-2 by injection of rhod-2-AM into the olfactory tract (z-projection of stack). (b) Raw calcium signals (ΔF/F) evoked by different concentrations of an amino acid odorant (Phe) in 54 mitral cells from 7 AlstR−/− fish (dlx-GFP). Responses of each mitral cell were measured before and during exposure to allatostatin (2 nM; AlstR−/− control). Mitral cells were sorted by increasing ChI (averaged over all stimuli for each mitral cell; as in Fig. 4a). Mitral cells showing the largest relative changes in response after addition of allatostatin are therefore found at the top (negative changes) and the bottom (positive changes). Allatostatin had no obvious effects on mitral cell responses. This result also demonstrates that mitral cell responses were highly reproducible in the absence of manipulations. (c) Estimated firing rates of the same mitral cells, sorted as in b. Firing rates were estimated from calcium signals by temporal deconvolution and calibration (Methods). (d) Estimated firing rates of 76 mitral cells from 7 dlx-itTA/Ptet-NpHR3.0YFP fish before and during hyperpolarization of dlx4/6-neurons by NpHR and blockade of gap junctions (GJ block; CBX [25 μM], 18α-GA and 18β-GA [10 μM each]). Mitral cells were sorted by ChI as in b and c. This manipulation reduced responses of some mitral cells, particularly at low odorant concentrations, but enhanced responses of other mitral cells, particularly at high odorant concentrations.

Supplementary Figure 7 Optogenetic inhibition of dlx4/6-neurons and multiphoton calcium imaging.

(a) Optical stimulation of a dlx4/6-neuron expressing NpHR3.0YFP (dlx-itTA/Ptet- NpHR3.0YFP fish) with yellow light for 15 s (overlay of three trials). Note strong and reversible hyperpolarization. Right: Mean amplitude of hyperpolarization (n = 5 dlx4/6-neurons from 1 fish). (b) Response of a dlx4/6-neuron expressing NpHR3.0YFP to depolarizing currents of increasing amplitude before and during exposure to yellow light (overlay of four trials). Right: mean firing rate (± s.e.m.) during current injection as a function of current amplitude, averaged over n = 5 dlx4/6-neurons from 2 fish. Optical stimulation of NpHR3.0YFP strongly suppressed action potential firing. (c) Schematic of multiphoton microscope with digital micromirror device (DMD) for optical stimulation during calcium imaging. Inset: timing schedule used for optical stimulation during the flyback of the laser (gray open circles) between successive lines (red filled circles). Full-field light pulses of 0.125 ms were generated by the DMD. Stimulation frequency (500 Hz) was determined by the line time of 2 ms.

Supplementary Figure 8 Effects of dlx4/6-neurons on mean mitral cell responses: additional results.

(a) Mean estimated firing rates (± s.e.m.) evoked by different concentrations of Phe before (gray) and during optogenetic inhibition of dlx4/6-neurons (NpHR3.0YFP, yellow light; n = 82 neurons from 12 fish) without gap junction blockade. Inset: line fits to normalized concentration-response curves. P = 0.0004, 0.036, 0.13, 0.63, 0.85 (left to right; Mann-Whitney U test); U = 2294, 2725, 2900, 3215, 3303; df = 164. (b) Same for hyperpolarization of dlx4/6-neurons by bath-application of allatostatin (2 nM; dlx-itTA/Ptet-AlstR-IRES-GFP fish; n = 140 neurons from 10 fish). P = 1.6 × 10–7, 0.002, 0.026, 0.063, 0.19 (left to right; Mann-Whitney U test); U = 6250, 7737, 8291, 8540, 8908; df = 280. (c) Mean estimated firing rates (± s.e.m.) evoked by different odorants before (gray) and during (blue) bath-application of allatostatin (2 nM) in dlx-itTA/Ptet-AlstR-IRES-GFP fish (n = 169 neurons from 13 fish). Odorants were sorted by the mean estimated firing rate. P = 0.029, 0.012, 0.012, 0.29, 0.58, 0.10, 0.87 (left to right; Mann-Whitney U test); U = 12322, 12028, 12026, 13331, 13780, 12780, 14128; df = 338. (d) Mean estimated firing rates (± s.e.m.) evoked by different concentrations of Phe before (light gray) and during (dark gray) bath-application of allatostatin (2 nM) in AlstR-/- fish (AlstR-/- control; n = 54 neurons from 7 fish). P = 0.37, 0.63, 0.51, 0.79, 0.59 (left to right; Mann-Whitney U test); U = 1311, 1379, 1350, 1415, 1369; df = 108. The slope of the normalized concentration-response function (insets) was increased by all manipulations but not in the AlstR-/- control. The AlstR-/- control showed a slight rundown but no change in the slope of the normalized concentration-response function. (e) Scatter plots on the left show ChIs of each individual mitral cell for responses to all different stimuli for each manipulation. Mitral cells were ranked by mean ChI. Colored lines show mean ChIs, colored shading shows SDs of ChIs for each mitral cell. Gray line and gray shading show the mean and SD of ChIs in AlstR-/- controls (lower right) for comparison. Colored bars on the right margin of the scatter plot show the mean and SD of ChIs pooled over the population (all neurons and stimuli). Histograms show the distribution of SDs of ChIs for individual mitral cells (SDMC,i), which reflect the heterogeneity of a manipulation's effect on responses of individual mitral cells to different stimuli. Bars on top of each histogram show the mean ± SD of the distribution. The gray line shows the SD of ChIs pooled over the population (SDMCpop), which represents the heterogeneity of a manipulation's effect across the population of mitral cells. For all manipulations, SDMC,i was clearly smaller than SDMCpop. Effects of manipulations on individual mitral cells were therefore less heterogeneous than effects on the whole population, indicating that effects on individual mitral cells were systematic. (f) Differences in ChIs between manipulations and AlstR−/− control (lower right in e) as a function of the mean ChI of individual mitral cells. For each data set (each manipulation and AlstR−/− control), mitral cells were ranked by mean ChI and grouped in 5 bins. The mean ChI in each bin for the AlstR−/− control was then subtracted from the mean ChI in the corresponding bins for each manipulation. The resulting curves therefore show the effect of each manipulation on ChIs, corrected for potential systematic experimental effects and stochastic effects resulting from sorting. Filled circles show statistically significant differences between the uncorrected ChIs for each manipulation and the AlstR-/- control (Mann-Whitney U test; P < 0.05; in most cases P < 0.01). A systematic increase of the corrected ChI was observed for all manipulations. This finding confirms that manipulations systematically decreased odor responses of some mitral cells (ChI < 0) but systematically increased responses of others (ChI > 0). *, P < 0.05; **, P < 0.01.

Supplementary Figure 9 Pharmacogenetic inhibition of dlx4/6 neurons by AlstR/allatostatin.

(a) Response of a dlx4/6 neuron expressing AlstR to depolarizing currents of increasing amplitude before and during exposure to allatostatin (2 nM; dlx-itTA/Ptet-AlstR-IRES-GFP fish; overlay of 12 trials). (b) Effect of allatostatin on input resistance (Rin), spontaneous firing rate, membrane potential (Vm) and action potential output evoked by current injection in different types of neurons under different conditions. Data shown in the left column were obtained in the presence of NBQX, APV, Gabazine and the gap junction inhibitors CBX, 18α-GA and 18β-GA in order to synaptically isolate dlx4/6-neurons. Binding of allatostatin to AlstR activates G-protein signaling that results in the opening of potassium channels. Consistent with this mechanism, allatostatin hyperpolarized dlx4/6 neurons expressing AlstR and reduced their excitability (top row). These effects of allatostatin were reversible (top left) and not observed in dlx4/6-neurons or other neurons that did not express AlstR (second and third row). Wilcoxon signed rank test, top left (n = 4 neurons): Rin: P= 0.016; W = 35; df = 8; Baseline rate: P= 0.0078; W = 36; df = 8; Vm: P = 0.0078; W = 36; df =8; Firing rate: P = 0.0078, 0.15, 0.0078, 0.0078, 0.11; W = 36, 29, 36, 36, 24; df = 8. Data obtained in allatostatin were compared to data obtained before wash-in (control) and after washout (wash), resulting in 2 × n = 8 comparisons. Top right (n = 16 neurons): Rin: P = 0.0097; W = 118; df = 16; Baseline rate: P = 0.039; W = 40; df = 16; Vm: P = 0.049; W = 106; df = 16; Firing rate: P = 0.039, 0.012, 0.015, 0.017, 0.013; W = 40, 103, 115, 114, 116; df = 16. Middle left (n = 4 neurons) and bottom left (n = 4 neurons): no statistical comparison was performed because n was too low. Middle right (n = 11): Rin: P = 0.76; W = 37; df = 11; Baseline rate: P = 0.37; W = 22; df = 11; Vm: P = 0.21; W = 18; df = 11; Firing rate: P = 0.37, 0.21, 0.58, 0.24, 0.70; W = 22, 18, 26, 8, 28; df = 11). Bottom right (n = 9): Rin: P = 0.65; W = 27; df = 9; Baseline rate: P = 0.50; W = 3; df = 9; Vm: P = 0.77; W = 23; df = 9; Firing rate: P = 0.50, 0.16, 0.13, 0.30, 0.73; W = 3, 3, 9, 13, 26; df = 9. Mitral cells were identified by intracellular fills. Interneurons comprise neurons in the glomerular layer that did not express the marker for dlx4/6-neurons and did not exhibit the characteristic morphology of mitral cells, as determined by intracellular fills. AlstR was co-expressed with GFP from a bicistronic construct (dlx-itTA/Ptet-AlstR-IRES-GFP; Methods). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Supplementary information

Supplementary Text and Figures

Supplementary Figures 1–9 and Supplementary Table 1 (PDF 2322 kb)

Source data

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Zhu, P., Frank, T. & Friedrich, R. Equalization of odor representations by a network of electrically coupled inhibitory interneurons. Nat Neurosci 16, 1678–1686 (2013). https://doi.org/10.1038/nn.3528

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/nn.3528

Further reading

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