Analysis of anesthesia stages based on the EEG entropy estimation

New algorithm for anesthesia depth analysis using EEG signal is presented. The algorithm is intended for the use in anesthesia depth monitors in the course of surgical operations. The suggested algorithm is based upon the combination of the following three approaches: signal randomness analysis with the use of approximate entropy, power spectrum analysis and analysis of specific signal changes that take place at the state of deep anesthesia. The algorithm was tested with the use of real ECG recordings obtained in the course of surgical operations and demonstrated good performance. The software package realizing this algorithm is used in an anesthesia depth monitor prepared to the batch production. Further efforts for the algorithm improving should be directed to the increase of the algorithm robustness to noises.