Wavelet Entropy based classification of depth of anesthesia

Abstract

Monitoring depth of anesthesia helps to prevent intraoperative awareness, reduces the doses of anaesthetic drugs and reduces the recovery time from anesthesia. Modern anesthesia monitors uses EEG signals to derive Depth of Anesthesia (DoA) measures, which decreases monotonically with increasing anesthetic drug levels. This study aims to measure DoA using Wavelet Analysis of EEG signals and classify them according to the DoA. Wavelet Entropy (WE) of the EEG signal is extracted as a measure of DoA from the EEG signals collected during the four phases of general anesthesia called awake, induction, maintenance and recovery. In order to find out the wavelet entropy, EEG signals during anesthesia were decomposed into its constituent frequency bands, then WE were calculated from the approximation and detail coefficients. Artificial Neural networks (ANN) were implemented using the extracted WE features of EEG signals as inputs and then classifying anesthetic depth as awake, light anesthesia, moderate anaesthesia and deep anaesthesia. Finally extracted WE is compared with BIS index, which is a commercially available DoA monitor.