Next-generation optical microscopy for imaging living biological systems

Optical microscopy has been a cornerstone in scientific discovery, essential to modern biomedical research and medical diagnostics. In our lab, we strive to contribute to the next generation of optical microscopy by integrating advances in hardware, physical simulations, and computational approaches. By refining imaging precision across spatiotemporal dimensions, we aim to enhance our understanding of living biological systems. Our goal is to explore new possibilities in imaging, hoping to gain deeper, faster, and more diverse insights into the complex processes that govern life.

We are particularly interested in imaging the brain, as it is the key regulator of health and the source of human intelligence. Our hope is to develop more precise diagnostic and discovery tools for neuroscience and neurology.

Examples of past works include:

  • Multicolor Three-Photon Microscopy: Multicolor three-photon fluorescence imaging with single-wavelength excitation deep in mouse brain. Science advances, 7(12), eabf3531. (2021). [link]

  • Adaptive Optics Enhanced Three-Photon Microscopy: Three-photon adaptive optics for mouse brain imaging. Frontiers in neuroscience, 16, 880859. 2022. [link]

  • Simultaneous Three-Photon and Optical Coherence Microscopy: Simultaneous multimodal three-photon and optical coherence microscopy of the mouse brain in the 1700 nm optical window in vivo. bioRxiv, 2023-09. (2023). [link]

  • Short-Wave Infrared One-Photon Confocal Microscopy: Short-wave infrared confocal fluorescence imaging of deep mouse brain with a superconducting nanowire single-photon detector. ACS Photonics, 8(9), 2800-2810. (2021). [link]

Computational optical microscopy and sensing

Light, as a non-invasive probing tool, offers significant potential for advancing observation techniques, particularly in the biomedical field. In our lab, we focus on enhancing the use of photons in optical microscopy and sensing through the development of innovative algorithms and computational models. These efforts aim to improve the quality of information obtained and to extract valuable insights through computation, enhancing resolution, sensitivity, and imaging speed in both imaging and sensing applications.

We approach this direction by developing computational frameworks to guide the co-design of the system with both hardware and software, and then implementing the design in the lab to explore real-world applications.

Examples of past works include:

  • Unrolled Neural Network-Guided Wavefront-Shaping Computational Microscopy: Physics-based neural network for non-invasive control of coherent light in scattering media. Optics Express 30, no. 17 (2022): 30845-30856. 2022. [link]

  • Computational Microscopy and Sensing of the Brain (Perspective): Neurophotonics beyond the surface: unmasking the brain’s complexity exploiting optical scattering. Neurophotonics 11, no. S1 (2024): S11510-S11510. 2024. [link]

  • Computational Resolution-Enhanced Fluorescence Microscopy: Replica-assisted super-resolution fluorescence imaging in scattering media." arXiv preprint arXiv:2404.19734 (2024). [link]

  • Computational wavefront sensing: “Closed-loop wavefront sensing and correction in mouse brain enabled by computed optical coherence microscopy”, Biomedical Optics Express, 12(8), 4934-4954, (2021). [link]

Neuromorphic (brain-inspired) optical information processors and computation

We aim to overcome the bottleneck of information processing in the digital domain for data acquired by optical systems by developing neuromorphic (brain-inspired) optical information processors. Our research focuses on creating optical systems that emulate the brain’s computational principles to enhance the efficiency and effectiveness of optical data processing. By designing and implementing these brain-inspired architectures, we seek to achieve high-performance optical processing in terms of speed, energy consumption, etc. We are particularly interested in developing systems for optical data science and optical machine learning for in situ smart data processing.

Examples of past works include:

  • Nonlinear Optical Encoder with Linear Optics: "Nonlinear optical encoding enabled by recurrent linear scattering." Nature Photonics (2024): 1-9.. [link]

  • Large-Scale Nonlinear Photonic Computation: "Large-scale photonic computing with nonlinear disordered media." Nature Computational Science (2024): 1-11. [link]

  • Large-scale Reconfigurable Optical Neural Network: “Hardware-efficient, large-scale reconfigurable optical neural network (ONN) with backpropagation” SPIE Photonics West, AI and Optical Data Sciences IV, (2023) [link]

AI for Optics

The impact of AI on science is profound, driving advancements across various fields. We focus on applying AI specifically to optics, designing and tailoring state-of-the-art AI algorithms and machine learning techniques. Our aim is to enhance optical system design, improve imaging processes, and optimize data analysis. By integrating these advanced AI methods, we seek to push the boundaries of optical science and enable innovations in imaging, data interpretation, and information processing.

Example past work:

  • Neural Implicit Representation Guided Diffusion Model to Enhance Optical Microscopy: MicroDiffusion: Implicit Representation-Guided Diffusion for 3D Reconstruction from Limited 2D Microscopy Projections." In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 11460-11469. (2024). [link]

  • Unrolled Neural Network-Guided Wavefront-Shaping Computational Microscopy: Physics-based neural network for non-invasive control of coherent light in scattering media. Optics Express 30, no. 17 (2022): 30845-30856. (2022). [link]

Optical tools into the fundamental sciences and medical clinic

Integrating optical tools into fundamental sciences and medical clinics offers transformative potential. We focus on advancing optical technologies to enhance both basic research and clinical applications. By developing cutting-edge optical systems and techniques, we aim to improve scientific discoveries and clinical diagnostics, bridging the gap between fundamental science and practical medical use.

Example of past works:

  • Label-free In Vivo Deep Brain Imaging: “In vivo label-free confocal imaging of the deep mouse brain with long-wavelength illumination.” Biomedical Optics Express 9, no. 12: 6545-6555. 2018. [link]

  • Label-free Operando optical imaging of charging battery: “Three-dimensional operando optical imaging of single particle and electrolyte heterogeneities inside Li-ion batteries.” Nature Nanotechnology. (2023). [link]