Tung-Yen Wu (吳東彥)

My research interest lies in developing optical microscopy techniques to decipher neuronal coding in the mouse cerebellum. Since neurons encode information collectively, it is desirable to image large populations of neurons, requiring a field of view (FOV) on the millimeter scale. Furthermore, because neurons are distributed in three dimensions within the cerebellum, volumetric imaging with a depth of approximately 200 μm is essential. However, expanding the FOV and achieving volumetric sampling typically come at the cost of reduced volume imaging rates. To effectively capture calcium signals, a volume rate exceeding 10 Hz is necessary.

My current work focuses on extending the FOV of our two-photon volumetric microscope to the millimeter scale while maintaining a high volume rate of 10–30 Hz. The large FOV is enabled by the COUSA objective, and rapid volumetric scanning is achieved by synchronizing a resonant galvo–galvo pair with a TAG lens for x–y–z scanning. The resulting sampling trajectory follows a Lissajous pattern, due to the sinusoidal motion of both the resonant galvo and TAG lens. This non-uniform sampling introduces downsampling artifacts, which we address using deep learning algorithms to reconstruct complete volumetric images.