Automatic Sleep Scoring Using Keras

Abstract: Sleep scoring is a diagnosis tool used for medical research. Using electroencefalography (EEG) a researcher observes the electrical activity in the brain and classifies the EEG into different stages. The goal of this project is to develop a tool for automatic classification of sleep data from rodents, one of the most common test subjects in modern medical research. EEG data from AstraZeneca is used to train a neural network, developed with Keras. Some augmentation of the data is done to increase the accuracy. The data is classified into different sleep stages with 91% accuracy.