Abstract:Two-photon microscopy has been widely used in neuroscience due to its deeper penetration depth compared with confocal microscopy. However, as imaging depth increases, its imaging quality will also degrade dramatically because of the accumulated aberration and scattered light. Here we report a high-resolution two-photon microscopy deep into the scattering medium, by structure illuminating the sample in the focal volume and demodulating the fluorescent signal thereafter. Our results show that compared with conventional two-photon microscopy the maximum imaging depth increases by hundreds micron in the mouse brain, and even in the surficial layer the imaging resolution is improved by around 15%.Two-photon microscopy has been widely used in neuroscience due to its deeper penetration depth compared with confocal microscopy. However, as imaging depth increases, its imaging quality will also degrade dramatically because of the accumulated aberration and scattered light. Adaptive optics is one of the most powerful tools to recover diffraction-limited resolution deep into the brain. However, it takes additional time for wavefront aberration measurement and compensation. Here we report a high-resolution two-photon microscopy deep into the scattering medium, by structure illuminating the sample in the focal volume and demodulating the fluorescent signal thereafter. The special structure-illumination pattern is design to make the focal point scanning sinusoidally only in the small focal volume. Here we first theoretically analyze the image formation of the structure illuminated two-photon microscopy (SITP) with D-shaped pupils and one annular one circular pupils, respectively. Their intensity points spread functions are investigated and the three-dimensional (3D) optical transfer functions (OTF) are provided. Then, the inertial-free scanning capability of SITP is demonstrated. After that, the ability of rejecting out-of-focus background and reducing noise is demonstrated. Our experimental results show that compared with conventional two-photon microscopy the maximum imaging depth increases by hundreds micron in the mouse brain, and even in the surficial layer the imaging resolution is improved by around 15%. Moreover, no additional time will be taken during to optical aberration and scattered light compensation / rejection.

Deep learning based tomographic phase microscopy with blind structured illumination

Structured Illumination Microscopy with Modified Joint Richardson-Lucy Deconvolution

High Resolution Tomographic Phase Microscopy

Adaptive Optics For Structured Illumination Microscopy Based On Deep Learning

Structure illuminated two-photon microscopy for high resolution imaging deep into the brain.

Deep Tissue Focusing Based on Machine Learning

GPU-based Deep Convolutional Neural Network for Tomographic Phase Microscopy with ℓ1 Fitting and Regularization

Accelerated Phase Shifting for Structured Illumination Microscopy Based on Deep Learning

Deep learning enables structured illumination microscopy with low light levels and enhanced speed

Deep‐MSIM: Fast Image Reconstruction with Deep Learning in Multifocal Structured Illumination Microscopy (Adv. Sci. 27/2023)

Deep learning for tomographic image reconstruction

System Model Enabling Fast Tomographic Phase Microscopy with Total Variation Regularisation.

Deep learning-based quantitative phase microscopy

Computational interference microscopy enabled by deep learning

Recognizing three-dimensional phase images with deep learning

Limited Angle Tomography for Transmission X-Ray Microscopy Using Deep Learning

Deep Learning Diffuse Optical Tomography

Unsupervised deep learning enables blur-free super-resolution in two-photon microscopy

High-throughput intensity diffraction tomography with a computational microscope

High-resolution limited-angle phase tomography of dense layered objects using deep neural networks