Enhancing confocal microscopy with AI

Improving processes in the lab can enable researchers to spend their time on other tasks and increase the accuracy of their findings, including when imaging. In this Q&A, Dr Hari Shroff from the US National Institutes of Health (NIH) discusses his recent study, published in Nature where confocal microscopy was enhanced via artificial intelligence (AI) and hardware modifications.

Why is it important to enhance imaging techniques?

In general, there is a disconnect between what you would like to see and what you can see with an optical microscope. When you image anything with fluorescent dyes, bleaching of the dyes ultimately limits the time course of the experiment. As a fluorescence microscopy tool developer, we are always trying to find ways to increase resolution or speed while ideally decreasing the illumination dose supplied to the sample. If this can be achieved, more can be seen; the more that you can infer what is happening at the subcellular scale, the more you can know – for example – about how a drug might act.

We chose to focus on confocal microscopes as they are pretty ubiquitous; almost every lab has access to one, even if they do not have one in their lab. The defining characteristic of the confocal microscope is…