Due to the aging population, Alzheimer's disease is on the rise, but it is not inevitable: nearly half of the risk depends on factors that can be acted upon…again, before taking action, most people will first want to know if they are at risk or not. Several methods for the early detection of people at risk attempt to help us answer this crucial question, including one blood test diagnostic aids with a detection rate of 90% (%). However, these tests, which rely on biological markers already present in the blood or cerebrospinal fluid, only confirm a diagnosis; they cannot predict the future onset of the disease. Researchers from Stanford University and New York University have developed an alternative method that could identify at-risk individuals several years before the disease: an artificial intelligence that analyzes brain activity and uses it to deduce the risk of developing Alzheimer's disease in the future. Their model, presented on November 4, 2026, in Science Advances, This could facilitate the care of these people in order to reduce their risk and thus avoid the disease, or at least mitigate its severity.
Learning to detect the differences between good and bad brain aging
The authors started from the observation that it is difficult to decree a brain “normality,” because there can be enormous variability between two people with good cognitive health. The simplest approach, they say, is rather to determine what is not within this normality, that is to say, the two extremes of cognitive health: on the one hand, seniors who already have dementia, and on the other, those who, on the contrary, age remarkably well, the “super-old” (superagers (in English). These are people over 75 years old who retain a memory comparable to, or even better than, the average healthy fifty-year-old.
English 
French