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:

Event-based vision sensor for fast and dense single-molecule localization microscopy

Nature Photonics volume 17, pages 1105–1113 (2023)Cite this article

Subjects

Abstract

Single-molecule localization microscopy (SMLM) enables crucial insights into cellular structures and processes to be revealed at the single-molecule level. However, SMLM is often hampered by limited temporal resolution and the fixed frame rate of the acquisition. Here we present a new approach to SMLM data acquisition and processing based on an affordable event-based sensor. This type of sensor reacts to changes in light intensity, rather than integrating photons during the exposure time of each frame. Each pixel works independently and returns a signal only when an intensity change is detected. Compared with video acquisition using traditional electron-multiplying charge-coupled device or scientific complementary metal–oxide–semiconductor cameras, the event-based sensor provides higher temporal resolution and throughput on the positions of blinking molecules. We demonstrate event-based SMLM super-resolution imaging on biological samples with spatial resolution on a par with the performance of electron-multiplying charge-coupled device or scientific complementary metal–oxide–semiconductor cameras, while registering only the on and off switching of blinking molecules. We use event-based SMLM to perform very dense single-molecule imaging, where frame-based cameras experience major limitations.

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

Fig. 1: Principle of event-based detection.
Fig. 2: Characterization of Eve-SMLM.
Fig. 3: Imaging of fixed COS-7 cells labelled with AF647 against α-tubulin (see Methods).
Fig. 4: High-density bioimaging.

Similar content being viewed by others

Data availability

Dataset examples are available at https://github.com/Clement-Cabriel/Evb-SMLM.git. More datasets are available from the corresponding author upon reasonable request.

Code availability

Processing codes are available at https://github.com/Clement-Cabriel/Evb-SMLM.git. The repository contains the codes for reading files and converting them into frame stacks, as well as the single-molecule localization code. More codes will be added in the future. Datasets (fluorescent beads or AF647 dyes) are available at the same repository to test the codes.

References

  1. Betzig, E. et al. Imaging intracellular fluorescent proteins at nanometer resolution. Science 313, 1642–1645 (2006).

    ADS  Google Scholar 

  2. Hess, S. T., Girirajan, T. P. K. & Mason, M. D. Ultra-high resolution imaging by fluorescence photoactivation localization microscopy. Biophys. J. 91, 4258–4272 (2006).

    ADS  Google Scholar 

  3. Rust, M. J., Bates, M. & Zhuang, X. Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM). Nat. Methods 3, 793–796 (2006).

    Google Scholar 

  4. Huang, B., Wang, W., Bates, M. & Zhuang, X. Three-dimensional super-resolution imaging by stochastic optical reconstruction microscopy. Science 319, 810–813 (2008).

    ADS  Google Scholar 

  5. Pavani, S. R. P. et al. Three-dimensional, single-molecule fluorescence imaging beyond the diffraction limit by using a double-helix point spread function. Proc. Natl Acad. Sci. USA 106, 2995–2999 (2009).

    ADS  Google Scholar 

  6. Bourg, N. et al. Direct optical nanoscopy with axially localized detection. Nat. Photonics 9, 587–593 (2015).

    ADS  Google Scholar 

  7. Cabriel, C. et al. Combining 3D single molecule localization strategies for reproducible bioimaging. Nat. Commun. 10, 1980 (2019).

    ADS  Google Scholar 

  8. Maynard, S. A. et al. Identification of a stereotypic molecular arrangement of endogenous glycine receptors at spinal cord synapses. eLife 10, e74441 (2021).

    Google Scholar 

  9. Khater, I. M., Nabi, I. R. & Hamarneh, G. A review of super-resolution single-molecule localization microscopy cluster analysis and quantification methods. Patterns 1, 100038 (2020).

    Google Scholar 

  10. Verdier, H. et al. A maximum mean discrepancy approach reveals subtle changes in α-synuclein dynamics. Preprint at bioRxiv https://doi.org/10.1101/2022.04.11.487825 (2022).

  11. Lampe, A., Haucke, V., Sigrist, S. J., Heilemann, M. & Schmoranzer, J. Multi-colour direct STORM with red emitting carbocyanines. Biol. Cell 104, 229–237 (2012).

    Google Scholar 

  12. Zhang, Z., Kenny, S. J., Hauser, M., Li, W. & Xu, K. Ultrahigh-throughput single-molecule spectroscopy and spectrally resolved super-resolution microscopy. Nat. Methods 12, 935–938 (2015).

    Google Scholar 

  13. Valades Cruz, C. A. et al. Quantitative nanoscale imaging of orientational order in biological filaments by polarized superresolution microscopy. Proc. Natl Acad. Sci. USA 113, E820–E828 (2016).

    Google Scholar 

  14. Kinkhabwala, A. et al. Large single-molecule fluorescence enhancements produced by a bowtie nanoantenna. Nat. Photonics 3, 654–657 (2009).

    ADS  Google Scholar 

  15. Lerner, E. et al. Toward dynamic structural biology: two decades of single-molecule Förster resonance energy transfer. Science 259, eaan1133 (2018).

    Google Scholar 

  16. Bouchet, D. et al. Probing near-field light–matter interactions with single-molecule lifetime imaging. Optica 6, 135–138 (2019).

    ADS  Google Scholar 

  17. Blanquer, G. et al. Relocating single molecules in super-resolved fluorescence lifetime images near a plasmonic nanostructure. ACS Photonics 7, 393–400 (2020).

    Google Scholar 

  18. Koenderink, A. F., Tsukanov, R., Enderlein, J., Izeddin, I. & Krachmalnicoff, V. Super-resolution imaging: when biophysics meets nanophotonics. Nanophotonics 11, 169–202 (2021).

    Google Scholar 

  19. Oleksiievets, N. et al. Fluorescence lifetime DNA-PAINT for multiplexed super-resolution imaging of cells. Commun. Biol. 5, 38 (2022).

    Google Scholar 

  20. Liao, F., Zhou, F. & Chai, Y. Neuromorphic vision sensors: principle, progress and perspectives. J. Semicond. 42, 013105 (2021).

    Google Scholar 

  21. Gallego, G. et al. Event-based vision: a survey. IEEE Trans. Pattern Anal. Mach. Intell. 44, 154–180 (2022).

    Google Scholar 

  22. van de Linde, S. et al. Direct stochastic optical reconstruction microscopy with standard fluorescent probes. Nat. Protoc. 6, 991–1009 (2011).

    Google Scholar 

  23. Sharonov, A. & Hochstrasser, R. M. Wide-field subdiffraction imaging by accumulated binding of diffusing probes. Proc. Natl Acad. Sci. USA 103, 18911–18916 (2006).

    ADS  Google Scholar 

  24. Jungmann, R. et al. Single-molecule kinetics and super-resolution microscopy by fluorescence imaging of transient binding on DNA origami. Nano Lett. 10, 4756–4761 (2010).

    ADS  Google Scholar 

  25. Nieuwenhuizen, R. P. J. et al. Measuring image resolution in optical nanoscopy. Nat. Methods 10, 557–562 (2013).

    Google Scholar 

  26. Culley, S. et al. Quantitative mapping and minimization of super-resolution optical imaging artifacts. Nat. Methods 15, 263–266 (2018).

    Google Scholar 

  27. Wade, O. K. et al. 124-color super-resolution imaging by engineering DNA-PAINT blinking kinetics. Nano Lett. 19, 2641–2646 (2019).

    ADS  Google Scholar 

  28. Donehue, J. E., Wertz, E., Talicska, C. N. & Biteen, J. S. Plasmon-enhanced brightness and photostability from single fluorescent proteins coupled to gold nanorods. J. Phys. Chem. C 118, 15027–15035 (2014).

    Google Scholar 

  29. Dertinger, T., Colyer, R., Iyer, G., Weiss, S. & Enderlein, J. Fast, background-free, 3D super-resolution optical fluctuation imaging (SOFI). Proc. Natl Acad. Sci. USA 106, 22287–22292 (2009).

    ADS  Google Scholar 

  30. Geissbuehler, S., Dellagiacoma, C. & Lasser, T. Comparison between SOFI and STORM. Biomed. Opt. Express 2, 408 (2011).

    Google Scholar 

  31. Hugelier, S. et al. Sparse deconvolution of high-density super-resolution images. Sci. Rep. 6, 21413 (2016).

    ADS  Google Scholar 

  32. Speiser, A. et al. Deep learning enables fast and dense single-molecule localization with high accuracy. Nat. Methods 18, 1082–1090 (2021).

    Google Scholar 

  33. Helmerich, D. A. et al. Photoswitching fingerprint analysis bypasses the 10-nm resolution barrier. Nat. Methods 19, 986–994 (2022).

    Google Scholar 

  34. Gerasimaitė, R. et al. Blinking fluorescent probes for tubulin nanoscopy in living and fixed cells. ACS Chem. Biol. 16, 2130–2136 (2021).

    Google Scholar 

  35. Mahecic, D. et al. Event-driven acquisition for content-enriched microscopy. Nat. Methods 19, 1262–1267 (2022).

    Google Scholar 

  36. Alvelid, J., Damenti, M., Sgattoni, C. & Testa, I. Event-triggered STED imaging. Nat. Methods 19, 1268–1275 (2022).

    Google Scholar 

  37. Gómez-García, P. A., Garbacik, E. T., Otterstrom, J. J., Garcia-Parajo, M. F. & Lakadamyali, M. Excitation-multiplexed multicolor superresolution imaging with fm-STORM and fm-DNA-PAINT. Proc. Natl Acad. Sci. USA 115, 12991–12996 (2018).

    ADS  Google Scholar 

  38. Gu, L. et al. Molecular resolution imaging by repetitive optical selective exposure. Nat. Methods 16, 1114–1118 (2019).

    Google Scholar 

  39. Cnossen, J. et al. Localization microscopy at doubled precision with patterned illumination. Nat. Methods 17, 59–63 (2019).

    Google Scholar 

  40. Jouchet, P. et al. Nanometric axial localization of single fluorescent molecules with modulated excitation. Nat. Photonics 15, 297–304 (2021).

    ADS  Google Scholar 

  41. Gu, L. et al. Molecular-scale axial localization by repetitive optical selective exposure. Nat. Methods 18, 369–373 (2021).

    Google Scholar 

  42. Thorsen, R. Ø., Hulleman, C. N., Rieger, B. & Stallinga, S. Photon efficient orientation estimation using polarization modulation in single-molecule localization microscopy. Biomed. Opt. Express 13, 2835–2858 (2022).

    Google Scholar 

  43. Izeddin, I. et al. Wavelet analysis for single molecule localization microscopy. Opt. Express 20, 2081–2095 (2012).

    ADS  Google Scholar 

  44. Ovesný, M., Křížek, P., Borkovec, J., Švindrych, Z. & Hagen, G. M. ThunderSTORM: a comprehensive ImageJ plug-in for PALM and STORM data analysis and super-resolution imaging. Bioinformatics 30, 2389–2390 (2014).

    Google Scholar 

  45. Stallinga, S. and Rieger, B. The effect of background on localization uncertainty in single emitter imaging. In Proc. 2012 9th IEEE International Symposium on Biomedical Imaging (ISBI) 988–991 (IEEE, 2012).

  46. Edelstein, A. D. et al. Advanced methods of microscope control using μManager software. J. Biol. Methods 1, e11 (2014).

Download references

Acknowledgements

We acknowledge C. Hubert as well as O. Thouvenin for their help with conceiving the project, as well as for loan of the hardware. We also thank R. M. Córdova-Castro, V. Scolari, A. Coulon, S. Lévêque-Fort and E. Fort for fruitful discussions and feedback. We are very grateful to G. Lukinavičius for providing the HMSiR-tubulin affinity probe used for the live-cell experiments. Finally, we thank A. Tchouprina for the useful technical discussions about the working principle and characterization of the sensor. This work has received support under the programme ‘Investissements d’Avenir’ launched by the French Government. We also acknowledge financial support by the French Agence Nationale de la Recherche, project ABC4M under reference ANR-20-CE45-0023 (I.I. and C.C.) and project SP-Tunnel-OHG under reference ANR-20-CE24-0021 (C.C.). This work was also supported by a grant from Région Ile-de-France, DIM-ELICIT, OPI project (I.I. and C.C.).

Author information

Authors and Affiliations

  1. Institut Langevin, ESPCI Paris, Université PSL, CNRS, Paris, France

    Clément Cabriel & Ignacio Izeddin

  2. Sorbonne Université, INSERM, CNRS, Institut de la Vision, Paris, France

    Tual Monfort

  3. National Hospital of Ophthalmology (CHNO) of Quinze-Vingts, INSERM-DGOS CIC 1423, Paris, France

    Tual Monfort

  4. DHNS, Inserm U1195, Université Paris-Saclay, Le Kremlin-Bicêtre, France

    Christian G. Specht

Authors
  1. Clément Cabriel
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Tual Monfort
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Christian G. Specht
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Ignacio Izeddin
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

C.C., I.I. and T.M. conceived the project and C.C. led its development, supervised by I.I. The optical setup and processing algorithms were designed by C.C., who also performed the acquisitions, processing and data analysis. C.G.S. prepared the fixed cell samples and C.C. performed the immunolabelling. C.C. wrote the original draught of the manuscript, which was reviewed by C.G.S. and I.I.

Corresponding authors

Correspondence to Clément Cabriel or Ignacio Izeddin.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Photonics thanks Fernando Stefani, Ian Dobbie and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Display of the motion of fluorescent beads on a coverslip at different time bases.

An acquisition is performed on 40 nm fluorescent beads on a coverslip while the stage is translated (bottom left to top right) to simulate molecule motion. From the same event list, frames are generated at Δt = 10 ms (left) and Δt = 1 ms (right). It should be noted that a small time bin (Δt = 1 ms) efficiently samples the dynamic to reveal two opposed lobes (leading and trailing edges) at the cost of a relatively modest number of events. On the contrary, increasing the time bin to Δt = 10 ms provides a large number of events, spoiled by motion blur as well as a loss of resolution. We also present in Supplementary Videos 2 and 3 a similar experiment where the motion is resolved down to 100 μs and 2.5 ms respectively. This illustrates that, while the value of the time bin has to be matched to the phenomenon under investigation, no knowledge or assumption is required prior to the acquisition (contrary to scientific camera-based approaches), as the selection of Δt is done as part of the processing workflow. This could even be used to perform a multi-timescale processing on the same dataset to extract both slow and fast processes. Scale bars: 5 μm.

Extended Data Fig. 2 Comparison of the SMLM images obtained with center of mass and Gaussian fitting localizations with the EBS and the EMCCD.

Acquisitions were done on fixed COS-7 with AF647 labelling against α-tubulin and imaged in standard density dSTORM with the event-based sensor and the EMCCD camera simultaneously using a 50:50 beamsplitter in the detection path (corresponding to the data displayed in Fig. 3a–b). a Eve-SMLM image displayed in Fig. 3a,b Zoom on the region indicated with a green square in a and comparison of the different localization methods for each sensor. In particular, the Gaussian fitting and center of mass SMLM images are very similar for Eve-SMLM. Scale bar: 5 μm (a), 1 μm (b).

Extended Data Fig. 3 SMLM FRC resolution measurements on fixed COS-7.

The cells are labeled with AF647 against α-tubulin and imaged in standard density dSTORM with the event-based sensor and the EMCCD camera simultaneously using a 50:50 beamsplitter in the detection path (corresponding to the data displayed in Fig. 3a). Resolution maps were obtained using the NanoJ-SQUIRREL Fiji plugin, ref. 26 in the main text. Scale bar: 5 μm.

Extended Data Fig. 4 Dependence of the frame-based localization performance on the integration time in the high density regime.

The SMLM results obtained in Fig. 4b–e from the sCMOS data with DNA-PAINT labelling and Gaussian fitting localization were reprocessed by summing subsequent raw frames into slices of different sizes (2, 4 and 8). This effectively corresponds to integration times of 60, 120 and 240 ms (with the minor addition of some thermal noise from the camera, but in the high density regime the resolution is not limited by the signal to noise ratio, but by the density of PSFs instead). a Raw frames corresponding to integration times of (left to right) 30 ms (raw image stack), and 60 ms, 120 ms and 240 ms (slices of 2, 4 and 8 frames). Cyan arrows highlight differences of PSF density. b Corresponding 2D SMLM images over the whole field of view. c FRC resolution measurement on the region indicated with a dashed rectangle in b. Note that, while the signal to noise ratio increases by summing frames, the density of PSFs also increases, which leads to a noticeable degradation of the performances in terms of resolution and structure sampling. Scale bars: 5 μm.

Extended Data Fig. 5 Frames extracted from the acquisition presented in Fig. 4g (high density dSTORM).

The acquisition was performed on fixed COS-7 cells labeled with AF647 against α-tubulin with a 50:50 beamsplitter over 250 seconds in a dense regime where the PSFs exhibit noticeable overlap, resulting in a degradation of the frame-based localization performance (see Fig. 4f–h). a Single 30 ms exposure frame taken from the EMCCD blinking movie. b Single frame (corresponding to a slightly different instant in the acquisition as a) generated from all the positive events detected in Δt = 10 ms (this frame is used only for the PSF detection). See Supplementary Note 7 for a more detailed explanation of this spatiotemporal resampling of the signal enabled by the event-based detection. Scale bars: 1 μm.

Extended Data Fig. 6 ThunderSTORM localization results in the high density regime.

To ensure that our custom-written camera-based localization code does not induce unexpected changes of performance in the high density regime, we reprocessed the high density data presented in Fig. 4f, h using the widely used ThunderSTORM ImageJ plugin, ref. 2 in the Methods Section. a Full 2D SMLM image obtained with (left) the event-based localization (Gaussian fitting), (center) the frame-based localization with our custom code (Gaussian fitting), and (right) the frame-based localization with ThunderSTORM (Gaussian fitting). b Zoom on the region indicated with a rectangle in a. Cyan and yellow arrows highlight differences of resolution and image fidelity, respectively. c FRC maps corresponding to the images presented in a. Note that while ThunderSTORM images provide better sampling than our custom code (probably due to the post-processing filters used in our code to discard bad fits), it remains far behind the event-based results in terms of both resolution and image fidelity. Besides, ThunderSTORM images exhibit some hot spots due to the high PSF density-they are associated with important local reductions of the resolution. Overall, the EMCCD resolution obtained in FRC with ThunderSTORM are only marginally better than that obtained with our custom code (90 nm versus 92 nm). Scale bars: 5 μm (a, c), 1 μm (b).

Extended Data Fig. 7 EMCCD localization results obtained in the high density regime with center of mass calculation.

The localization images and FRC-assessed resolution maps are displayed as in Fig. 4f–h, including the EMCCD data processed with a center of mass algorithm. a Event-based high density dSTORM image displayed in Fig. 4f. b Zoom on the area indicated with a rectangle in a. The SMLM images are shown for different localization conditions: from the event-based sensor data with Gaussian fitting (left) and from the EMCCD data with Gaussian fitting (center), and center of mass calculation (right). Cyan and yellow arrows highlight differences of resolution and image fidelity, respectively. c FRC resolution maps calculated on the area displayed in a for the different localization methods tested. All the localization was done using our custom code. Scale bars: 5 μm (a, c), 1 μm (b).

Extended Data Fig. 8 Influence of the size of the fitting area on the localization performance in the standard density regime.

Acquisition on fluorescent beads deposited on a coverslip (same acquisition as in Fig. 2b) were processed with different sizes of the area used for the Gaussian fitting. The fitting area is defined by its half size in pixels and one pixel is 67 nm. The localization uncertainty (left) and the number of events detected per PSF (right) are plotted as a function of the total number of photons in the PSF. While the number of events seems to slightly increase for larger fitting areas, the precision is almost identical for 4, 6 and 8 pixels, and slightly better than for 3 pixels. These results are very consistent with the simulations (see Supplementary Fig. 10 right).

Extended Data Fig. 9 Influence of the time window on the event-based localization performance in the high density regime (with center of mass localization).

a 2D SMLM images of the same high density dataset processed with four different time window values (that is different values of δTstart and δTend, see Supplementary Note 4): [-7.5 ms, +7.5 ms], [-15 ms, +15 ms], [-30 ms, +30 ms] and [-60 ms, +60 ms] (left to right). b Zoom on the region indicated with a rectangle in a for the different time windows. c FRC resolution maps obtained with the different time windows on the full field shown in a. d Histogram of the timestamps of the events relative to the time of the frame where the PSF was detected (with a frame time bin of 5 ms, see Supplementary Note 4 for more details about the processing workflow). e Mean resolutions obtained for each time window on the full field presented in a and on the Region Of Interest (ROI) presented in b. f Numbers of events detected for each time window on the full field and the ROI. Note that, in the range we investigated, the time window seems to have little influence on the resolution. The optimal value is found to be 30 ms, which is the value that was used in the manuscript (Fig. 4f–h). This can be understood from the event arrival time histogram, which exhibits a relatively narrow peak with a FWHM of 12 ms. While increasing the time window until 30 ms helps gathering more of the signal, any increase above seems to mostly result in higher noise. Scale bars: 5 μm (a, c), 1 μm (b).

Supplementary information

Supplementary Information

Supplementary Table 1, Notes 1–7, Figs. 1–12, Methods and descriptions of Videos 1–5.

Supplementary Video 1

Movie of fluorescent beads under time-modulated excitation This video is a slowed down reconstruction of the blinking of 40 nm fluorescent beads deposited on a coverslip and excited with a 10 Hz modulated square signal. The acquisition and the region of interest are the same as in Fig. 1e. We reconstructed the blinking movie on a time base of 2 ms. For each frame i, all of the events within the time interval [i × 2 ms, (i + 1) × 2 ms] are displayed without filtering (except for the video compression). Each positive or negative event is counted as +1 and −1, respectively. The pixel size is 67 nm in the object plane. The total duration of the video is 400 ms, and the display frame rate is 10 frames per second, which corresponds to a slow factor of 50. The colour scale is the same as in Figs. 1e and 2e, with the minimum (blue) and the maximum (white) display values set to –3 and +3, respectively. Black denotes zero events. The total width of the field is 8.5 μm. Note that the brightest pixels at the centre of the PSF are the first to respond, whereas the outer pixels respond more slowly.

Supplementary Video 2

Fluorescent bead motion resolved at 100 μs. This video is a slowed down reconstruction of the signal produced by 40 nm fluorescent beads deposited on a coverslip when the stage is translated from the bottom to the top to simulate molecule motion, similar to Extended Data Fig. 1. Frames were generated from the positive events with a time bin Δt = 100 μs, which is inaccessible to the vast majority of scientific cameras. For each frame i, all of the positive events within the time interval [i × 100 μs, (i + 1) × 100 μs] are displayed without filtering (except for the video compression). Each event is counted as +1. The pixel size is 67 nm in the object plane. The total duration of the video is 30 ms, and the display frame rate is 20 frames per second (that is, 2 ms per second of playback), which corresponds to a slow factor of 500. The colour scale is the same as in Figs. 1e and 2e and Extended Data Fig. 1, with the maximum (white) display value set to 2. Black denotes zero events. The total width of the field is 19.0 μm. As a comparison, we also reconstructed the same acquisition at 2.5 ms (see Supplementary Video 3), on which motion blur is clearly visible. Note that the signal and time response are sufficient to resolve the motion down to 100 μs, which is not accessible to the vast majority of scientific cameras.

Supplementary Video 3

Fluorescent bead motion reconstructed at 2.5 ms. This video is the same as Supplementary Video 2, but reconstructed with a time bin Δt = 2.5 ms (that is, in the typical range of maximum frame rates that can be achieved by scientific cameras with a reasonable crop of the region of interest). The total duration of the video is 30 ms, and the display frame rate is 0.8 frames per second (that is, 2 ms per second of playback), which corresponds to a slow factor of 500. The colour scale is the same as in Figs. 1e and 2e and Extended Data Fig. 1, with the maximum (white) display value set to 8. Note that the motion blur makes it impossible to resolve the motion properly, unlike Supplementary Video 2, which is reconstructed at Δt = 100 μs.

Supplementary Video 4

dSTORM blinking movie. This video is a real-time reconstruction of the blinking of AF647-labelled tubulin in fixed COS-7 cells. The acquisition and the region of interest are the same as in Fig. 3d, left. We reconstructed the blinking movie for display convenience on a time base of 20 ms. For each frame i, all of the events within the time interval [i × 20 ms, (i + 1) × 20 ms] are displayed without filtering (except for the video compression). Each positive or negative event is counted as +1 or −1, respectively. The pixel size is 67 nm in the object plane. The total duration of the video is 15 s, and the display frame rate is 50 frames per second (that is, real speed). The colour scale is the same as in Figs. 1e and 2e, with the minimum (blue) and the maximum (white) display values set to −6 and +6, respectively. Black denotes zero events. The total width of the field is 28.5 μm. Note the sparsity of the signal returned compared with cameras, where data are returned from all pixels even if they contain only background.

Supplementary Video 5

Time-resolved high-density SMLM imaging of tubulin in a living cell. The video was obtained by imaging tubulin labelled with HMSiR in live cells using the EBS (see Supplementary Methods). The molecules were localized and the frames were generated by binning the detected molecules in a sliding-window manner for the sake of visibility: the frames are separated by 10 s and each frame contains the molecules localized in the time range [ti, ti + 30 s], where ti = i × 10 s is the timestamp of frame i. The total width of the field is 34 μm. The white arrow indicates one particular filament that displays clearly distinguishable spatial reorganization over time.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Cabriel, C., Monfort, T., Specht, C.G. et al. Event-based vision sensor for fast and dense single-molecule localization microscopy. Nat. Photon. 17, 1105–1113 (2023). https://doi.org/10.1038/s41566-023-01308-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41566-023-01308-8

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