EPFL's LogSync implantable chip decodes brain-network states and triggers stimulation at under 200 microwatts

EPFL's Integrated Neurotechnologies Lab built LogSync to address a key limitation of deep brain stimulation: most systems use fixed settings or single-region signals, while psychiatric disorders involve distributed brain networks. LogSync extracts oscillation strength and cross-region synchrony biomarkers, then uses machine learning to classify cognitive and emotional states and trigger timed stimulation, consuming under 200 microwatts. Published in the IEEE Journal of Solid-State Circuits, the work was funded by a CHF 6.6 million NIMH grant. It was validated on human recordings and rat experiments at the University of Minnesota, with applications beyond psychiatry envisioned.

✨ AI summary of the linked source — not editorial content.

Explore the topic

• AI