- Sommeil
- 2026
Our pediatric sleep research to be presented in Paris at the International Conference on Signal, Image Processing and Pattern Recognition
Intro :
Our abstract, “Adapting a General-Purpose EEG Foundation Model for Pediatric Sleep Staging: Development of a Channel-Agnostic Multimodal Framework”, has been accepted for poster presentation at the International Conference on Signal, Image Processing and Pattern Recognition, held in Paris in February 2026. This conference focused on advances in signal and image processing, from fundamental research to applied developments.
In short :
At this conference, Colin Flipo, intern in our data science team, will present our work on sleep stage classification. This step is essential for assessing sleep quality and diagnosing sleep disorders, yet remains time-consuming and highly dependent on expert interpretation. In pediatric populations, the challenge is even greater due to strong age-related neurophysiological variability. In this context, we adapted our foundation model, BioSerenity-E1, to design a framework capable of handling multimodal configurations and efficiently processing pediatric polysomnography data. Despite being initially pretrained on non sleep-specific data, the model shows promising performance, demonstrating the ability of EEG foundation models to transfer to specialized use cases such as pediatric sleep analysis.