An attempt to predict drowsiness by Bayesian estimation

EEG (EEG-MPF, EEG-α/β), heart rate variability (RRV3), tracking error and subjective rating of fatigue (drowsiness) while performing a simulated driving task were measured to predict drowsiness on the basis of Bayesian estimation method. The relation between these measurements and drowsiness was analyzed. As a result, EEG-MPF tended to decrease with the increase of drowsiness. It tended that EEG-α/β, RRV3 and tracking error increased with the increase of drowsiness. Then, we have proposed a method that can predict drowsiness by applying Bayesian estimation method to physiological measurements such as EEG-MPF, EEG-α/β, RRV3. Bayesian estimation carries out a statistical inference using some kind of evidences or observations and calculating the probability that a hypothesis is true. As a result, we confirmed that the proposed method could to some extent predict the symptom of decreased consciousness (drowsiness).