Automatic machine classification of patient anaesthesia levels using EEG signals

The authors explore the possibility of using EEG (electroencephalographic) signals for automatic machine classification of the level of anesthesia that a patient is in. EEG data obtained under different levels of anesthesia have been modeled as an AR (autoregressive) process for that purpose. It is shown that AR model order, the AR power spectral density, and the second and fourth moments of the probability density function of the EEG signals can be used for classifying the level of anesthesia into low, medium, and high levels