Single-molecule localization microscopy (SMLM) can be used to resolve subcellular structures and achieve a tenfold improvement in spatial resolution compared to that obtained by conventional fluorescence microscopy. However, the separation of single-molecule fluorescence events that requires thousands of frames dramatically increases the image acquisition time and phototoxicity, impeding the observation of instantaneous intracellular dynamics. Here we develop a deep-learning based single-frame super-resolution microscopy (SFSRM) method which utilizes a subpixel edge map and a multicomponent optimization strategy to guide the neural network to reconstruct a super-resolution image from a single frame of a diffraction-limited image. Under a tolerable signal density and an affordable signal-to-noise ratio, SFSRM enables high-fidelity live-cell imaging with spatiotemporal resolutions of 30 nm and 10 ms, allowing for prolonged monitoring of subcellular dynamics such as interplays between mitochondria and endoplasmic reticulum, the vesicle transport along microtubules, and the endosome fusion and fission. Moreover, its adaptability to different microscopes and spectra makes it a useful tool for various imaging systems.

We combine a generative adversarial network (GAN) with light microscopy to achieve deep learning super-resolution under a large field of view (FOV). By appropriately adopting prior microscopy data in an adversarial training, the neural network can recover a high-resolution, accurate image of new specimen from its single low-resolution measurement. Its capacity has been broadly demonstrated via imaging various types of samples, such as USAF resolution target, human pathological slides, fluorescence-labelled fibroblast cells, and deep tissues in transgenic mouse brain, by both wide-field and light-sheet microscopes. The gigapixel, multi-color reconstruction of these samples verifies a successful GAN-based single image super-resolution procedure. We also propose an image degrading model to generate low resolution images for training, making our approach free from the complex image registration during training data set preparation. After a well-trained network has been created, this deep learning-based imaging approach is capable of recovering a large FOV (~95 mm²) enhanced resolution of ~1.7 μm at high speed (within 1 second), while not necessarily introducing any changes to the setup of existing microscopes.

Combining information from multispectral images into a fused image is informative and beneficial for human or machine perception. Currently, multiple photodetectors with different response bands are used, which require complicated algorithms and systems to solve the pixel and position mismatch problem. An ideal solution would be pixel-level multispectral image fusion, which involves multispectral image using the same photodetector and circumventing the mismatch problem. Here we presented the potential of pixel-level multispectral image fusion utilizing colloidal quantum dots photodiode array, with a broadband response range from X-ray to near infrared and excellent tolerance for bending and X-ray irradiation. The colloidal quantum dots photodiode array showed a specific detectivity exceeding 10¹² Jones in visible and near infrared range and a favorable volume sensitivity of approximately 2 × 10⁵ μC Gy⁻¹ cm⁻³ for X-ray irradiation. To showcase the advantages of pixel-level multispectral image fusion, we imaged a capsule enfolding an iron wire and soft plastic, successfully revealing internal information through an X-ray to near infrared fused image.

Hemodynamic shear force has been implicated as modulating Notch signaling–mediated cardiac trabeculation. Whether the spatiotemporal variations in wall shear stress (WSS) coordinate the initiation of trabeculation to influence ventricular contractile function remains unknown. Using light-sheet fluorescent microscopy, we reconstructed the 4D moving domain and applied computational fluid dynamics to quantify 4D WSS along the trabecular ridges and in the groves. In WT zebrafish, pulsatile shear stress developed along the trabecular ridges, with prominent endocardial Notch activity at 3 days after fertilization (dpf), and oscillatory shear stress developed in the trabecular grooves, with epicardial Notch activity at 4 dpf. Genetic manipulations were performed to reduce hematopoiesis and inhibit atrial contraction to lower WSS in synchrony with attenuation of oscillatory shear index (OSI) during ventricular development. γ-Secretase inhibitor of Notch intracellular domain ( NICD ) abrogated endocardial and epicardial Notch activity. Rescue with NICD mRNA restored Notch activity sequentially from the endocardium to trabecular grooves, which was corroborated by observed Notch-mediated cardiomyocyte proliferations on WT zebrafish trabeculae. We also demonstrated in vitro that a high OSI value correlated with upregulated endothelial Notch-related mRNA expression. In silico computation of energy dissipation further supports the role of trabeculation to preserve ventricular structure and contractile function. Thus, spatiotemporal variations in WSS coordinate trabecular organization for ventricular contractile function.