Manchester researchers have developed an AI system that “cleans” brain signals in real time on wearable devices, a breakthrough that could transform epilepsy care and pave the way for smarter, more portable healthcare.
Wearable brain-monitoring devices, known as EEGs, allow patients to track conditions like epilepsy without surgery. However, these devices often pick up background noise—known as “artifacts”—which can make readings less accurate.
To solve this problem, a team led by Dr. Mahdi Saleh designed a deep learning model that filters out unwanted noise from EEG signals in real time. The model runs on small, low-power wearable devices, providing clearer, more reliable readings without sending sensitive brain data to remote servers.
This is the first demonstration that advanced AI for EEG monitoring can function on compact, wearable devices. While each device involves trade-offs between speed and power consumption, the research shows that real-time, portable, and secure brain monitoring is achievable.
Dr. Saleh said: “This turns hospital-level monitoring into an everyday reality, helping people with epilepsy now and unlocking a future of smarter healthcare.”
Beyond epilepsy, the innovation could support next-generation health devices and brain-computer interfaces, where privacy, portability, and real-time performance are critical.
The project was supported by the EIC Pathfinder RELIEVE Project.
Related topic：
