ABSTRACT Optical imaging holds tremendous promise in our endeavor to understand brain functions. The major challenges for optical brain imaging are depth and speed. Due to optical aberrations and tissue scattering, the penetration depth and imaging speed of optical microscopy in the brains (e.g., mouse) is limited. The constraints in depth and speed make large-scale, deep imaging of mouse brain activity out of reach of current imaging techniques. Hardware adaptive optics (AO) has proven to be valuable for in vivo brain imaging with two-photon microscopy (2PM), and will have even larger impact for deep brain 3-photon microscopy (3PM); however, existing AO techniques require iterative optimization using multiphoton excited fluorescence signal. While adequate for imaging relatively shallow regions of the brain (< 1 mm deep), iterative optimization is impractical with ultra-deep imaging since the fluorescence signal deceases exponentially with imaging depth. The required integration time to obtain the necessary signal-to-noise ratio for iterative optimization becomes prohibitively long. In this program, we will leverage the advantages provided by computational adaptive optics (CAO) in optical coherence microscopy (OCM), specifically the strong OCT signal (from linear backscattering), and parallel computing on a high-end graphics processing unit (GPU), to provide orders of magnitude speed-up for the sensing of sample-induced aberrations throughout a volume of interest. A long-wavelength OCM system and CAO aberration sensor will be utilized to drive a hardware AO system to correct depth-dependent aberrations, push 3PM imaging depth beyond currently demonstrated limits, and increase imaging speed by at least one order of magnitude. Additionally, by combining with recently developed adaptive excitation laser technology, we will achieve approximately 300-fold increase in photon budget, which will enable truly unprecedented 3PM imaging speed and depth. Successful completion of this program will enable rapid aberration sensing at the depth range of 1 to 2 mm, and will open the exciting new opportunity of recording the neural activity of the dentate gyrus of adult mice through the intact brain. With its high-speed and deep-tissue aberration sensing capability, and zero additional photobleaching and phototoxicity, our novel method for real-time aberration sensing and correction is ideally positioned to transform our ability for large-scale, deep recording of mouse and adult zebrafish brain activity. This imaging approach will significantly extend the information available for neuroscience studies on individual cell-cell as well as circuit interactions that underlie normal and diseased brain function. The technology developed by this proposal will be applicable to imaging in other biological systems where large-scale, deep imaging at high spatiotemporal resolution is needed.

NARRATIVE The goal of this research program is to develop and demonstrate a novel aberration sensor to address the most significant hurdle to multiphoton microscopy of neural activity deep in the mouse and adult zebrafish brains. We will synergistically combine computational and hardware adaptive optics on a long-wavelength multimodal platform, and offload the time-consuming and cumbersome hardware-based aberration sensing step in aberration-corrected three-photon microscopy, to provide orders of magnitude speed-up for the sensing of aberrations throughout a volume of interest. Our approach for aberration-corrected three-photon microscopy will be used to demonstrate the deepest and fastest activity recording in the intact brain of awake and behaving mice, and throughout the adult zebrafish brain.