A Principal Component Analysis Based Method for Estimating Depth of Anesthesia

This paper proposes a novel approach to estimating level of unconsciousness based on Principal Component Analysis (PCA). The Electroencephalogram (EEG) data was captured in both Intensive Care Unit (ICU) and operating room. Different anesthetic drugs, including propofol and isoflurane were used. Assuming the central nervous system as a 20-tuple source, the window length of 20 seconds is applied to electroencephalogram (EEG). The mentioned window is considered as 20 nonoverlapping mixed-signals (epoch). The PCA algorithm and more precisely Eigenvector Decomposition (EVD) is applied to these twenty 1-second length epochs, and the related eigenvalues were extracted. Largest remaining (LRE) and smallest remaining eigenvalue (SRE) reveal a sensible behavior due to changing depth of anesthesia (DOA). The correlation between LRE and DOA was measured with Prediction probability (Pk)- The same was done for SRE and DOA. The results show the superiority of SRE than LRE in predicting DOA in the case of ICU and isoflurane. Conversely, the results reveal the superiority of LRE than SRE in propofol induction. Moreover, the result of LRE indicates no obvious diference between ICU and the drugs, while in the case of SRE, the result of ICU was better than that of drugs. Finally, a mixture model containing both LRE and SRE could predict DOA as well as Relative Beta Ratio (RBR), which expresses the high capability of the proposed PCA based method in estimating DOA.