Peter, Kayvan, Jianquan, and two of their colleagues at the University of Georgia have demonstrated the application of a”genetic algorithm” (GA) technique to adaptive optics (AO) to achieve a 4-fold increase in precision point-localization imaging through 50 microns of central nervous system tissue in fruit flies. The breakthrough, published in Optics Express, combines advances in machine learning derived from mathematical models of evolution with methods for correcting optical wavefront aberrations derived from astronomy. Imaging through layers of tissue produces distortions similar to imaging through atmospheric layers, and this “thickness” problem has been one of the obstacles to achieving super-resolution imaging of molecules within living cells. The GA technique allows the computer to sift rapidly through the thousands of variations in wavefront corrections generated by the adaptive optics algorithms and hone in on the “fittest” of these to prepare each of the image frames that will then be combined in the optical reconstruction process. This is a big advance for STORM imaging, which, because of its reliance on sequentially recording randomized intermittent light pulses from fluorophore-tagged molecules, delivers light intensity fluctuations too extreme for traditional optimization techniques to handle. Now that proof of concept has been achieved in a single 2D plane of central nervous system tissue, the team plans to move on to volumetric STORM imaging in which the image plane is stepped through the sample. This will bring new challenges, but the authors note that “the GA approach is well-suited to correct these slowly varying dynamic aberrations and astigmatism can also be dynamically added for 3D STORM.” See the full article and all images.