Linear and nonlinear EEG indexes in relation to the severity of coma

Glascow's coma scale (GCS) is a clinical standard for assessing the severity of coma in intensive care units (ICUs). The EEG monitoring can be on-line work, soon responds to the change in the brain wave, and allows long-term continuous monitoring of brain activity. In this paper, several EEG parameters, including spectral possession distribution and nonlinear description (approximate entropy and Lempel-Ziv complexity) were used to assess the capability of EEG indexes for the severity of coma. Our results demonstrated that all EEG parameters are moderately related to the GCS, with the Spearman statistical correlation of 0.62-0.71 between the spectral possession distribution and the GCS and 0.62-0.66 between nonlinear measures and the GCS. The moderate correlation between EEG parameters and the GCS implies the possibility of the EEG-derived index to reveal the neurological status of patients in coma.