Science

VIDEO. Camille’s chronicle: digital twins or mimic reality through computers

CJAMY. Camille Gaubert’s column is broadcast daily in the program “C Jamy”, presented by Jamy Gourmaud from Monday to Friday at 5 pm on France 5.

Mimicking reality with a mathematical model is the principle behind the concept of digital twins. Thanks to this work, it is possible to safely simulate an engine failure in an airplane, to accelerate the aging of a cell, and even to model an individual patient, with its own characteristics! The climax of personalized medicine.

Computer models that mimic the real world

Digital twins are computer duplicates of a real living object or being. For example, you can create a digital twin of an airplane or a hospital, so that you can safely simulate an engine failure or a change in ventilation. “The goal is to obtain an object close to reality from a mathematical model, which we can thus modify without damaging it, and make it rotate faster.“, explains Philippe Gesnouin, in charge of technology transfer at the National Institute for Research in Digital Sciences and Technologies (Inria). For example, the digital twin of a nuclear power plant makes it possible to simulate a cut in the cooling circuit, something impossible to be achieved in the real world. Time control is another essential advantage of this technology: “some of my colleagues are working on the evolution of bacteria and viruses, to understand the mechanisms of their evolution“, illustrates Philippe Gesnouin. With this type of model, they can virtually accelerate this process, and pass a few million years in a few hours.

The complexity of the digital twin applied to health

But it is in the health field that digital twins have great hopes. Thanks to this technology, it is possible to produce clones in the form of a computer model making it possible to test surgical procedures, predict the effectiveness of treatments, or adapt the dosages! It is the pinnacle of personalized medicine. She is a technologist of great complexity, which requires modeling a very large number of interactions, points out Philippe Gesnouins. For example, to make a heart twin, it is necessary to model the interactions in the molecules of actin and myosin, which allow muscle contraction, but also a set of mechanical, rheological, flow and electrophysiological phenomena. The idea is to be able to personalize this generic model with the data of a particular patient, such as his cardiac ultrasounds for example.

Use the digital twin in real time, during an operation

And that changes everything. The surgeon, for example, will be able to prepare his operation not on a standard heart, but on the heart of his patient. This will make it possible to gain in precision and to intervene early if necessary, explains Philippe Gesnouin. The patient’s digital twin can even be used during the operation to get a better view of their condition in real time. This technique has already been used during a heart operation at the Parisian hospital Lariboisière. The digital twin goes even further than artificial intelligence (AI), which uses large databases to compile statistics. For example, AI can predict that you have a 76% chance of developing type 2 diabetes, but it won’t be able to explain why. But the digital twin, which also incorporates AI but is not limited to it, it will be able.

One exception, however: our brain, which we do not understand enough to be able to model. “It is a major body which still largely escapes our modeling capacities”, explained to Sciences and the Future Dominique Chapelle, Inria research director. Patience!

Back to top button