Abstract

Info

Title:: Sample entropy and wavelet entropy of electroencephalogram for monitoring the depth of anesthesia

Author(s):: DING Zhengmin1; XIONG Dongsheng1; CHEN Yuke2; ZHANG Xing’an3; DOU Jianhong3; CHEN Yayu3; 1. Department of Biomedical Engineering, School of Material Science and Engineering, South China University of Technology, Guangzhou 510006, China; 2. Department of Equipment, General Hospital of Guangzhou Military Command of PLA, Guangzhou 510010, China; 3. Department of Anesthesia, General Hospital of Guangzhou Military Command of PLA, Guangzhou 510010, China

Keywords:: ; depth of anesthesia; electroencephalogram; sample entropy; wavelet entropy

Abstract:: Objective To research the characteristics of the electroencephalogram (EEG) signals of patients under general anesthesia, and to compare the performances of sample entropy and wavelet entropy algorithms in monitoring the depth of anesthesia, including classification accuracy, calculation complexity and calculation time. Methods Based on the characteristics of nonlinearity and instability of EEG signals, two kinds of nonlinear dynamics analysis methods, namely sample entropy algorithm and wavelet entropy algorithm, were used to extract the characteristics of the EEG signals of 30 patients under general anesthesia. The sample entropy and wavelet entropy of the EEG signals of patients under different anesthesia states (including waking state, light anesthesia and moderate anesthesia) were also compared with variance analysis. Results The sample entropy and wavelet entropy of the EEG signals under different states was significantly different. Moreover, the change threshold of sample entropy was larger than that of wavelet entropy. Conclusion Both sample entropy and wavelet entropy algorithms can be used as effective indicators for monitoring the depth of anesthesia, but when classification accuracy, calculation complexity and calculation time are taken into consideration, sample entropy algorithm is better than wavelet entropy algorithm.