Project Goal

The primary goal of this project is to develop a system that can automatically interpret an electroencephalography (EEG) test, thereby improving the quality and efficiency of a physician’s diagnostic capabilities. The demand for monitoring systems based on EEG signals is growing rapidly, as they are increasingly being used for preventive diagnostic procedures. The recent emergence of a comprehensive big data resource, the TUH EEG database, has created a unique opportunity to apply state of the art machine learning algorithms to this problem.

We propose the application of deep learning to classify EEGs and automatically generate time-aligned markers indicating areas of interest for physicians, enabling real-time alerting and automatic generation of a physician’s report. This will also enable mining of vast archives of EEG data to discover new ways to analyze and interpret EEGs. Finally, a large searchable database of marked up EEGs will be an invaluable resource for training medical students.