Direct access to optical aberration information in fluorescence laser scanning microscopy using detector arrays

In the recent years, numerous adaptive optics techniques have emerged to address optical aberrations in fluorescence microscopy imaging. However, many existing methods involve complex hardware implementations or lengthy iterative algorithms that may induce photo-damage to the sample. Our study proposes an innovative approach centered around a novel detector array capable of potentially capturing the probed sample in a single acquisition. Our solution is gentle on the sample and applicable to any laser scanning microscope equipped with a detector array. We demonstrate that the multi-dimensional dataset obtained using the detector array inherently encodes information about optical aberrations. Finally, we propose a convolutional neural network approach to decode these optical aberrations in real-time and with high accuracy, establishing the foundation for a new class of adaptive optics laser-scanning microscopy methods.