A PyTorch model that recognizes epilepsy states from brain-signal data, with 95% accuracy.
EEG records the electrical activity of the brain through electrodes on the scalp. This project is a machine learning model, written in Python with PyTorch, that learns to recognize epilepsy states from that data. It reaches 95% accuracy.
I built it as a member of the NCSU Neurotech Club’s software team, where I’ve been since August 2025. [ADD: the dataset you used, how you split and trained, and what you’d improve next.]
The club has a research partnership with Synaptive, a company making dry EEG electrodes. I’m designing the protocol for a study that tests their electrodes, designed for different hair types, against other electrodes, with recordings taken on an OpenBCI board. Better signals in, better models out.