- Sommeil
- 2026
EEG-based Alzheimer’s detection: abstract selected for Recent Advances in Machine Learning for Healthcare event.
Intro :
On November 25-26, 2025, the Grand Palais hosted Adopt AI, bringing together more than 25,000 participants and 500 speakers across major topics in artificial intelligence and healthcare. Within this global event, Bioserenity took part in the AI for Health, spotlighting how AI is transitioning from research to real-world clinical applications.
In short :
During the poster presentation we will address the challenges of costly and limited-access Alzheimer’s diagnosis by exploring the use of electroencephalography combined with artificial intelligence. We tuned BioSerenity-E1, our EEG foundation model, to detect Alzheimer’s disease from EEG signals, using one of the largest Alzheimer’s EEG cohorts available to date. The model was then evaluated on a fully independent external dataset, demonstrating strong generalization and its ability to learn transferable neurophysiological signatures. This work illustrates the potential of foundation models to overcome limitations related to EEG variability and opens the way toward more accessible diagnostic tools.