ǧ����Ƶ

Co-authors Hari Shroff (HHMI Janelia) and Patrick La Rivière (University of Chicago) collaborate in the ǧ����Ƶ Whitman Center.

A in Nature Biotechnology from the details two practical ways to improve the axial, or z, resolution of 3D structured illumination microscopy, a technique to see inside living cells pioneered by former Janelia Group Leader Mats Gustafsson, who died in 2011. 

In 3D-SIM and other fluorescence microscopy techniques, the axial resolution of the image is often blurred. This means researchers can clearly see details in two dimensions, on the x and y planes, but details in the third dimension, on the z plane, are fuzzy. Previous attempts to resolve this issue were difficult to implement.  

A project led by Xuesong Li, a postdoc in the Shroff Lab, developed two ways to practically deal with the problem. In one method, a mirror is added to the microscope to create an additional beam of light, changing the interference pattern, and enabling finer, sharper resolution along the z axis.  

The second method, which uses deep learning, blurs the sharp x and y axes to look like the blurry z axis and then trains a neural network to reverse these blurry images. The network then uses that information to un-blur the z axis.