Abstract:Photon emission by single molecules is a random event with a well-defined distribution. This calls for event-based detection in single-molecule localization microscopy. The detector has the advantage of providing a temporal change in photons and emission characteristics within a single blinking period (typically, ~ 30 ms) of a single molecule. This information can be used to better localize single molecules within a user-defined collection time (shorter than average blinking time) of the event detector. The events collected over every short interval of time / collection time (~ 3 ms) give rise to several independent temporal photon distributions (tPSFs) of a single molecule. The experiment showed that single molecules intermittently emit photons. So, capturing events over a shorter period / collection time than the entire blinking period gives rise to several realizations of the temporal PSFs (tPSFs) of a single molecule. Specifically, this translates to a sparse collection of active pixels per frame on the detector chip (image plane). Ideally, multiple realizations of single-molecule tPSF give several position estimates of the single-molecules, leading to multiple tPSF centroids. Fitting these centroid points by a circle provides an approximate position (circle center) and geometric localization precision (determined by the FWHM of the fitted Gaussian) of a single molecule. Since the single-molecule estimate (position and localization precision) is directly driven by the data (photon detection events on the detector pixels) and the recorded tPSF, the estimated value is purely experimental rather than theoretical (Thomson's formula). Moreover, the temporal nature of the event camera and tPSF substantially reduces noise and background in a low-noise environment. The method is tested on three different test samples (1) Scattered Cy3 dye molecules on a coverslip, (2) Mitochondrial network in a cell, and (3) Dendra2HA transfected live NIH3T3 cells (Influenza-A model). A super-resolution map is constructed and analyzed based on the detection of events (temporal change in the number of photons). Experimental results on transfected NIH3T3 cells show a localization precision of ~ 10 nm, which is ~ 6 fold better than standard SMLM. Moreover, imaging HA clustering in a cellular environment reveals a spatio-temporal PArticle Resolution (PAR) (2.3lp × τ) of 14.11 par where 1 par=10 ~meter.second. However, brighter probes (such as Cy3) are capable of ~3.16 par. Cluster analysis of HA molecules shows > 81 % colocalization with standard SMLM, indicating the consistency of the proposed eventSMLM technique. The single-molecule imaging on live cells reveals temporal dynamics (migration, association, and dissociation) of HA clusters for the first time over $60$ minutes. With the availability of event-based detection and high temporal resolution, we envision the emergence of a new kind of microscopy that is capable of high spatio-temporal particle resolution in the sub-10 par regime.

Neural network‐assisted localization of clustered point spread functions in single‐molecule localization microscopy

Alternative deep learning method for fast spatial-frequency shift imaging microscopy.

Sub-Diffraction Dark Spot Localization Microscopy

Localization of Ultra-dense Emitters with Neural Networks

High-axial-resolution Single-Molecule Localization under Dense Excitation with a Multi-Channel Deep U-Net.

Deep learning-based spectroscopic single-molecule localization microscopy

Model-free machine learning-based 3D single molecule localisation microscopy

Learning Consistent Subcellular Landmarks to Quantify Changes in Multiplexed Protein Maps.

Event-based Single Molecule Localization Microscopy (eventSMLM) for High Spatio-Temporal Super-resolution Imaging

Multimodal registration of FISH and nanoSIMS images using convolutional neural network models

Deep Learning for Super-Resolution Localization Microscopy

Automated Nanoplasmonic Analysis of Spherical Nucleic Acids Clusters in Single Cells.

Multicolor localization microscopy by deep learning

Single molecule localization by $\ell_2-\ell_0$ constrained optimization

Identification and calibration of ultrabright localizations to eliminate quantification error in SMLM

SNSMIL, a Real-Time Single Molecule Identification and Localization Algorithm for Super-Resolution Fluorescence Microscopy.

Deep-learning-assisted Spectroscopic Single-Molecule Localization Microscopy Based on Spectrum-to-spectrum Denoising.

Deep localization of protein structures in fluorescence microscopy images

A coordinate-based co-localization index to quantify and visualize spatial associations in single-molecule localization microscopy

High-density Localization of Active Molecules Using Structured Sparse Model and Bayesian Information Criterion

Digital-SMLM for precisely localizing emitters within the diffraction limit