• Starts: 2:30 pm on Monday, December 9, 2024
  • Ends: 4:00 pm on Monday, December 9, 2024

ECE PhD Prospectus Defense: Guorong Hu

Title: Advancing the Computational Miniature Mesoscope with Structured Illumination and Differentiable Optics

Presenter: Guorong Hu

Advisor: Professor Lei Tian

Chair: Professor Tianyu Wang

Committee: Professor Lei Tian, Professor Jerome C. Mertz, Professor Tianyu Wang, Professor Abdoulaye Ndao

Google Scholar Profile: https://scholar.google.com/citations?user=4upjOSYAAAAJ&hl=en

Abstract: Widefield fluorescence microscopy is widely used in biological sciences due to its simplicity and versatility. However, its bulky optics hinder our understanding of how populations of neurons and underlying circuits give rise to complex behaviors in freely behaving animals, which is a key endeavor of systems neuroscience. To overcome this challenge, miniaturized versions of widefield fluorescence microscopy have been advanced in the past decade, including the Computational Miniature Mesoscope (CM²) developed by our lab.

Although our device achieves micron-scale resolution across a centimeter-scale field of view (FOV) through the joint design of optics and computation, there is still room for improvement. First, the illumination module cannot reject fluorescence background, which affects the observation of neuronal signals when imaging highly scattering brain tissue. Second, the aperture-splitting strategy employed in the imaging module is not an efficient way of using collected photons. The objective of all my work is to tackle these two problems and further advance the CM².

My prior work focuses on miniaturizable HiLo for effective background rejection. HiLo attains optical sectioning by computationally combining two images: one with uniform illumination and the other with structured illumination. However, the latter, commonly used in the traditional HiLo, employs speckle, which is generated by a laser source and is difficult to miniaturize. Here, we propose a new structured illumination strategy that uses only a diffuser and an LED source to create random caustic patterns. This structured illumination module bypasses the need for a laser and thus facilitates the miniaturization of the HiLo.

My proposed aims are to: (1) build a 2 × 2 imaging module with commercially available micro-lenses for a better aperture-splitting strategy; (2) optimize this imaging module with differentiable ray tracing for improved imaging quality; and (3) design a HiLo-based miniaturized illumination module for the CM². Once these aims are fulfilled, our device is expected to provide the neuroscience community with a more powerful tool to address their novel hypotheses concerning brain function in freely behaving animals.

Location:
PHO 339