Classification of Depth of Anesthesia Using Continuous Wavelet Transform and Entropy Measures

The aim of this paper is to classify depth of general anesthesia (DOA) using electroencephalogram (EEG) signalwhich measures the brain electrical activity. Awareness during anesthesia is probably the most helpless and terrifyingfeeling in the world and should be prevented during general anesthesia. To end this, we propose to obtain thecontinuous wavelet transfrom (CWT) of EEG signal in different sub-bands of delta and alpha bands. After that,approximate entropy and sample entropy are obtained from absoulte of CWT of each sub-band and feature vector isconstructed. Finally, support vector machine (SVM) with linear kernel is used to classify the epochs of EEG signal intodifferent DOA levels. Results on clinical data show that proposed method can efficiently determine DOA level.

Keywords:

Depth of anesthesia , continuous wavelet transform , approximate entropy , sample entropy