Title: Non-invasive and noise-robust light focusing using confocal wavefront shaping

Abstract: One of the hardest barriers to our ability to see inside biological tissue is the fact that light is highly aberrated when scattering by the tissue sub-components. One of the promising approaches for overcoming such aberrations is wavefront-shaping, where one modulates the incoming and/or the outgoing wavefront in a way that will allow it to focus into one spot despite scattering in the tissue. Wavefront modulations are specific to the tissue sample being imaged and need to be estimated based on a non-invasive feedback from a camera collecting back-scattered light. Such modulations have been successively estimated using feedback from strong fluorescent beads which have been manually added to a sample. However, in a real biomedical application, such feedback should be provided by the fluorescent components of the tissue itself, whose emission is orders of magnitude lower than the one provided by beads. When such a low number of photons is spread over multiple sensor pixels, the image is highly susceptible to noise, and the feedback signal required for previous algorithms cannot be detected. In this work we suggest a wavefront shaping approach that works with a confocal modulation of both illumination and imaging arms. The advantage of this approach is that as part of the optimization, aberrations are corrected in the optics, before the detector. Hence the low photon budget can be directed into a single sensor spot and detected with high SNR. We derive the optimization problem from mathematical principles and show why it favors modulations that actually correct the aberrations and focus all light into one spot. We successfully demonstrate wavefront-shaping correction on EGFP neurons sliced out of a mouse brain, despite scattering through thick tissue.