Brain imaging // Illustrative image © Pexels
Researchers from the University of Technology of Troyes (UTT) have developed a method aimed at early detection of Parkinson’s disease. This new process, which combines medical imaging and artificial intelligence, illustrates the advances made possible by this technology in the medical world.
Detecting Parkinson’s disease before the first clinical symptoms
Nearly 200,000 people in France suffer from Parkinson’s. The disease generally has a slow progression, with the first bothersome symptoms appearing only when 60 to 80% of dopaminergic neurons have already disappeared. Therefore, the diagnosis could in certain cases be made 10 to 20 years after the onset of the disease. In 2024, brain imaging will make it possible to confirm the presence of Parkinson’s disease at a stage described as advanced, when the first clinical symptoms have already appeared: hand tremors, slowing of movements, etc. The AMPIATI project, led by a team from UTT, aims to detect neurons that may have disappeared due to Parkinson’s in the preclinical phase, that is to say in the absence of apparent symptoms.
We believe our approach could improve the way Parkinson’s disease is diagnosed and treated. By combining medical imaging and artificial intelligence, we are opening new perspectives to improve and advance research into this debilitating disease.
Brain imaging and AI model
This method of early detection of Parkinson’s disease is based on medical imaging to detect the disappearance of dopamine-producing neurons before the appearance of the first clinical symptoms. To do this, researchers focused on a very specific area of the brain, the striatum, which notably ensures the control of movements in patients. By using specific image processing methods and models powered by artificial intelligence, it was possible to identify and extract biomarkers of Parkinson’s disease.
Evolution of the Parkinson’s disease biomarker. © University of Technology of Troyes
The illustrations in the diagram represent the evolution of the disease over time. Each step is represented by a specific shape. In stage 1, the biomarker is square, in stage 2, it takes the shape of a parallelogram, in stage 3, of a diamond, etc.
Using artificial intelligence and image processing methods, research is working to detect this biomarker in its rectangular state, which could indicate an early risk of disease development.
Our project aims to predict the development of the disease and offer personalized management from its first signs. The stated ambition of this research is not only to locate the initial biomarker but also to make it detectable by an innovative and economical medical device that could revolutionize the prevention and early diagnosis of Parkinson’s, thus making monitoring of neuronal health accessible to a larger population and reducing dependence on costly brain imaging.
The researchers now wish to adapt this method to a generalizable framework to facilitate the identification of precursor signals of Parkinson’s disease by medical teams.