Abstract
Confocal microscopy1 remains a major workhorse in biomedical optical microscopy owing to its reliability and flexibility in imaging various samples, but suffers from substantial point spread function anisotropy, diffraction-limited resolution, depth-dependent degradation in scattering samples and volumetric bleaching2. Here we address these problems, enhancing confocal microscopy performance from the sub-micrometre to millimetre spatial scale and the millisecond to hour temporal scale, improving both lateral and axial resolution more than twofold while simultaneously reducing phototoxicity. We achieve these gains using an integrated, four-pronged approach: (1) developing compact line scanners that enable sensitive, rapid, diffraction-limited imaging over large areas; (2) combining line-scanning with multiview imaging, developing reconstruction algorithms that improve resolution isotropy and recover signal otherwise lost to scattering; (3) adapting techniques from structured illumination microscopy, achieving super-resolution imaging in densely labelled, thick samples; (4) synergizing deep learning with these advances, further improving imaging speed, resolution and duration. We demonstrate these capabilities on more than 20 distinct fixed and live samples, including protein distributions in single cells; nuclei and developing neurons in Caenorhabditis elegans embryos, larvae and adults; myoblasts in imaginal disks of Drosophila wings; and mouse renal, oesophageal, cardiac and brain tissues.
🔬 Techniques
🔭 Microscopes
🧬 Organisms
✨ Fluorophores
🧪 Sample Preparation
🔬 Cell Lines
🏭 Microscope Brands
💻 Software Details
💾 Data Repositories
🏛️ Research Organizations (ROR)
Affiliated research institutions:
📊 Figures
Extended Data Fig. 1,
Instrument overview.
a) Photograph of instrument, highlighting objectives, cameras, sample area. b) CAD rendering of MEMS line scanner. c) Optical setup. Diode lasers are combined and passed through an acousto-optic tunab...
Extended Data Fig. 2,
Further characterization of imaging field.
a) 175 u00d7 175 u03bcm 2 imaging field, showing representative horizontal illumination lines across field, as measured in fluorescent dye. Also superimposed are example regions of interest (#1 - #9) ...
Extended Data Fig. 3,
a) Lateral maximum intensity projection image of fixed U2OS cells immunolabeled with mouse-anti-alpha tubulin, anti-mouse-biotin, streptavidin Alexa Fluor 488, marking microtubules (fused result after...
Extended Data Fig. 4,
Contrast and resolution is enhanced after multiview fusion or SIM.
a) Nanoscale imaging of organelles (mitochondrial outer membrane, cyan; and double stranded DNA, magenta) in expanded U2OS cells, to accompany Fig. 1f . Top: images, bottom: line profiles correspondin...
Extended Data Fig. 5,
Triple-view comparisons in adult C. elegans .
a) Axial views of whole fixed worm labeled with NucSpot Live 488, comparing raw views gathered by objectives A, B, C; triple-view reconstruction; and point-scanning confocal microscope. Arrows highlig...
Extended Data Fig. 6,
Triple-view comparisons in scattering tissue.
a) Schematic of kidney: approximate region where tissue was extracted. b) Four color triple-view reconstruction of mouse kidney slice. Lateral image at 4 u03bcm depth, highlighting glomerulus surround...
Extended Data Fig. 7,
Triple-view comparisons in Drosophila wing imaginal disks.
a) Schematic of larval wing disc, lateral (top) and axial (bottom) views, including adult muscle precursor myoblasts and notum. b) Lateral plane from triple-view reconstruction, 30 u03bcm from sample ...
Extended Data Fig. 8,
Live triple-view confocal imaging of Cardiomyocytes.
a) Cardiomyocytes expressing EGFP-Tomm20, labeled with MitoTracker Red CMXRos, imaged with triple-view CM in 2.03 s, every 20 s, for 100 time points. Lateral (top) and axial (bottom) maximum intensity...
Extended Data Fig. 9,
Neural network schematics.
a) Workflow for two-step deep learning procedure used in Fig. 2e u2013 i . Denoising neural networks are trained for views A, B, C, by using matched high and low SNR volumes derived from embryos paral...
Extended Data Fig. 10,
Triple-view 1D SIM methods.
a) Workflow for a single scanning direction and extension for multi-view 1D SIM. Five confocal images are acquired per plane, each with illumination structure shifted 2u03c0/5 in phase relative to the...
Extended Data Fig. 11,
Triple-view 1D SIM of C. elegans embryo.
a) Fixed C. elegans embryo (strain DCR6681) with tubulin immunolabeled with u03b1-alpha tubulin primary, u03b1-mouse-biotin, Streptavidin Alexa Fluor 568, imaged via lower View C (left) and triple-vie...
Extended Data Fig. 12,
Triple-view 1D SIM of cells.
a, b) Comparative triple-view SIM a) and instant SIM b) maximum intensity projections of the same fixed HEY-T30 cell embedded in collagen gel, labeled with MitoTracker Red CMXRos and Alexa Fluor 488 p...
Extended Data Fig. 13,
Expansion workflow and imaging results for C. elegans embryos, related to Fig. 3d .
a) Immobilization, permeabilization, fixation, immunostaining, and expansion takes approximately four days. b) Overlay of embryo stained with DAPI before and after expansion after 12-degree affine reg...
Extended Data Fig. 14,
Simulation of the deep learning method for isotropic in-plane super-resolution imaging.
a) An object (left column) consisting of lines and hollow spheres can be blurred to resemble diffraction-limited confocal input (middle column). Fourier transform of the raw input is shown (right colu...
Extended Data Fig. 15,
Isotropic in-plane super-resolution imaging of fixed cells.
a) Immunolabeled microtubules in fixed U2OS cells, from the same sample shown in Fig. 4b . Shown are raw input (View C); 1D SIM output after physics-based reconstruction with five images; deep learnin...
Extended Data Fig. 16,
Isotropic in-plane super-resolution imaging of living cells.
a) Lateral maximum intensity projections of Jurkat T cell expressing EMTB-3XGFP (yellow) and F-tractin-tdTomato (red), volumetrically imaged every 2 s, as imaged with raw line confocal input (View C, ...
Extended Data Fig. 17,
Multi-modality imaging enabled with the multiview line confocal system.
a) Different methods of combining data, enabling a highly versatile imaging platform. Left: Diffraction limited volumes acquired from views A (yellow), B (green), C (red) may be combined with joint de...
Extended Data Fig. 18,
Multiview super-resolution imaging of larval worm.
a) Maximum intensity projection of fixed L2 stage larval worm expressing membrane targeted GFP primarily in the nervous system, imaged in triple-view 2D SIM mode. Anatomy as highlighted. b) Higher mag...
Fig. 1
Multiview line confocal microscopy.
a) Concept. Diffraction-limited line illumination is serially scanned via three objectives (OBJ A/B/C), collecting resulting fluorescence. b) 100 nm beads, raw views and triple-view fusion. c) Triple-...
Fig. 2
Multiview confocal live imaging.
a) HCT-116 human colon carcinoma cells, MitoTracker Green (cyan) and LysoTracker Deep Red (magenta) labels, 1.15 s/three-views, every 10 s, 60 time points. Triple-view lateral, axial MIPs are shown fo...
Fig. 3
Multiview super-resolution microscopy.
a) Illumination geometry. Sparse, periodic line illumination is scanned in orthogonal directions (xu2019, y, zu2019), enhancing 3D resolution. b) Triple-view 1D SIM MIPs, fixed B16F10 mouse melanoma c...
Fig. 4
Deep learning enhances multiview super-resolution imaging.
a) Deep learning, joint deconvolution produce 2D super-resolution images from diffraction-limited input. b) CM volumes of immunolabeled microtubules before and after 2D SIM predictions. Inset: Fourier...
Figure images are served from the NIH/NLM PubMed Central Open Access Subset or Europe PMC; copyright remains with the publishers and authors.
💬 Discussion
0 commentsNo comments yet. Be the first to start a discussion!
Leave a Comment