arkapravo sen

Live concussion detection with electroencephalograms and machine learning

Started Aug. 5, 2024

Abstract or project description

The goal of Arkapravo's project is to develop a smart helmet capable of detecting when an individual experiences a concussion, for example from a motorcycle crash or in a football game. He is accomplishing this by both measuring the forces on the helmet via an accelerometer and reading the signal from a single-electrode electroencephalogram (EEG) recorded within the helmet. To classify the EEG recording as normal signal or concussion signal, he is training neural networks with published EEG data from individuals with and without concussions. He is optimizing his machine learning models by testing different model sizes and architectures. When completed, Arkapravo plans to synthesize all components of the project into a single system that he can demo at science and engineering fairs.