AI has learned to see the aging of cells

The study was conducted by a team from NYU Langone Medical Center. They used a high-precision image of cells and trained the AI to recognize the first signs of aging: when the nucleus grows, loses its roundness, seals appear in it and it turns lighter. All these are signals that the cell has begun to lose its youth.

Cells age not only with age. Damage, illness, and even stress can cause them to stop dividing and start releasing substances that signal: “I broke down.” Such cells influence wound healing and the development of age—related diseases, from cancer to heart problems.

In order for the AI to learn to see this, the scientists created a laboratory model of aging: cells were subjected to chemical stress, as if they were aging in the body. The AI analyzed exactly how the core changes during this process. This is how the nuclear morphometric pipeline (or NMP) appeared, a tool that evaluates the level of aging on a scale from -20 to +20.

It sounds dry, but it works amazing. For example, it was used to distinguish the cells of young mice from old ones with high accuracy. And also track how muscles recover after injury — and where senescent cells appear at this moment.

In one experiment, scientists noticed: healthy animals without injury have almost no old cells. But once a muscle is damaged, such cells are activated to help with healing. And then they disappear. Exactly and on time. In an aging body, this mechanism works worse — and now we can measure it.

Technology opens the way to new ways of treatment. In the future, it will be possible to accurately track where and when tired cells appear and target them. For example, with the help of senolytics— drugs that kill senescent cells. This is a real breakthrough for longevity and recovery medicine. After all, in order to fight aging, you must first learn to see it.