Scientists are using light-field microscopy to observe fast biological processes by capturing large 3D images. The technique allows researchers to track and measure extraordinarily fine movements, such as neuronal signals in a fish brain, at very high speeds. But the image quality if often lacking, and it can take days for the massive amount of data to be converted into a viewable format. A technique previously used called light-sheet microscopy homes in on a 2D plane of a given sample producing images that are quicker to process, but the data is not as comprehensive as the latter.
Combining artificial intelligence (AI) algorithms with two cutting-edge microscopy techniques, researchers at the European Molecular Biology Laboratory have shortened image processing time from days to mere seconds, while ensuring the resulting images are clear and precise. The team took advantage of the benefits of light field microscopy and light-sheet microscopy to train AI algorithms how to create an accurate 3D picture of the sample. The study was published in the journal Nature Methods.
‘"If you build algorithms that produce an image, you need to check that these algorithms are constructing the right image," explains Anna Kreshuk, the EMBL group leader. In the study, the researchers used light-sheet microscopy to make sure the AI algorithms were working, Kreshuk says. "This makes our research stand out from what has been done in the past."’
In the future, this approach could be modified to work with various types of microscopes and widen the range of specimens for biologists to observe. The researchers plan to explore whether their new method can be applied to larger species, namely mammals.
Photo: A representation of a neural network provides a backdrop to a fish larva’s beating heart. Credit: Tobias Wuestefeld