Neural interfaces are revolutionizing medicine by creating direct connections between the brain and technology. These advancements have already restored hearing with implants, reduced tremors in Parkinson’s patients, and even provided basic vision for the blind. In the future, these technologies could help people who can’t speak communicate easily or even allow humans and machines to work together in entirely new ways.
For these systems to keep advancing, they need to process complex data quickly and accurately. This requires sophisticated machine learning models, which need a lot of information from devices that can capture signals from the brain. To make these systems work well, the devices collecting brain data need to be very small, energy-efficient, and able to handle tons of data at once.
Another challenge is that these systems will need to work in “closed-loop” ways, meaning they should be able to respond automatically to brain signals in real time. However, the electrical signals used to help stimulate the brain can sometimes interfere with the data being collected, making it harder to get clear results. Most current devices either do a great job collecting data but can’t handle this real-time feedback, or they handle feedback poorly and waste a lot of energy.
This research focused on solving these issues by designing a device that is super compact, energy-efficient, and can collect brain data without interference from these electrical signals. The study also looked into improving treatment for chronic symptoms after strokes, as part of the SCATMAN project, and aimed to create systems that could work with lots of channels for better rehabilitation.
The result of this work is a new device that collects brain data with high precision, using the smallest and most energy-efficient design ever made. Even with the electrical interference from brain stimulation, this new device works reliably, making it a step forward in the development of future neurorehabilitation systems.
6/2/2026 13:00 - 15:00
ESAT Aula C, B91.300
