New intelligent microscope: self-adaptive, perfect imaging on curved surfaces

Modern biology relies on our ability to use microscopes to observe living cells. Recent advances in light microscopy allow cellular and subcellular imaging in model organisms such as vinegar flies, zebrafish, and mice.

One of the basic limitations of current technology is the toxicity associated with lighting, which affects the biological process being studied. So far, there have not been many solutions to this problem except that the brightness is reduced and the image quality is degraded.

In a new paper published in Light Science & Application, a team led by Drs. Loïc Le Goff and Frédéric Galland of the Fresnel Institute at the University of Aix-Marseille in France have developed a new type of smart microscope that can use learning strategies to automatically calculate where to send light in the most effective way to take interest in the structure of interest in the sample Perform imaging.

Left: Drawing curved biological tissue. The hexagons represent the fluorescent contours of the cells organized in the cell table. The tissue can be covered by a second epithelium, which can be ignored by the imaging process. Right: From several acquisitions (green dots), the microscope automatically estimates the tissue surface (red grid), and then the acquisition can be concentrated on this surface, or even only on the outline of fluorescent cells due to the propagation acquisition algorithm. Image source: Faris Abouakil et al.

The starting point of the project was to observe that most biological tissues have well-characterized structures. In particular, most embryos are organized into surfaces—cell sheets—curved in space.

General microscopes are not adapted to the operation of this architecture: they scan the laser focused on the entire 3D space containing the embryo, which is very inefficient in terms of the speed and amount of light irradiating the sample. The microscope developed by the Fresnel Institute automatically adjusts its scanning mode to adapt to the morphology of curved biological surfaces without having to know the surface in advance. Compared with traditional confocal microscopes, our smart scanning microscope reduces radiation by up to 100 times on the test sample.

This breakthrough technology is the result of close collaboration between data scientists, physicists, and biologists at the Fresnel Institute. This method opens up a new method for long-term imaging of the behavior of very fragile objects such as embryos and organoids. Interestingly, this technique can be very simply applied to many commercial microscopes in the imaging facilities of the Institute of Biology.