Permutation Lempel-Ziv complexity measure of electroencephalogram in GABAergic anaesthetics

Monitoring the brain state in anaesthesia is crucial for clinical doctors. In this study, we propose a novel nonlinear method, the permutation Lempel-Ziv complexity (PLZC) index, which describes the complexity in the electroencephalographic (EEG) signal to quantify the effect of GABAergic anaesthetics on brain activities.We applied the PLZC to two EEG data sets that were recorded under sevoflurane and propofol anaesthesia. The results are compared with traditional mean value-based Lempel-Ziv complexity (LZC), permutation entropy (PE), composite PE index (CPEI), response entropy (RE), state entropy (SE) and bispectral index (BIS) or SynchFastSlow (SFS, derived from BIS). Pharmacokinetic/pharmacodynamic (PK/PD) modelling and prediction probability (Pk) were used to assess the performance of the proposed method for tracking GABAergic anaesthetic concentrations.We found that PLZC correlates closely with the anaesthetic drug effect. When applied in sevoflurane anaesthesia, the coefficient of determination R2 between the PLZC values and the sevoflurane effect site concentrations was (0.90 ± 0.07), mean  ±  standard deviation), and the prediction probability Pk was (0.85 ± 0.04). These values were higher than those for the other indices. While in propofol anaesthesia, the value of R2 between PLZC and the effect site concentrations was (0.89 +/- 0.07), and the Pk was (0.86 +/- 0.28), which were close to those for CPEI but better than those for the others.PLZC based on electroencephalogram signals can be used as a new index to characterize the depth of anaesthesia. This index outperformed LZC, PE, CPEI, RE, SE, and BIS or SFS in tracking drug concentration changes during GABAergic anaesthetics.PLZC is a potentially superior method for applications in intra-operative monitoring.