Development of AUC diagrams for analysis of interhemispheric asymmetry of amplitude-frequency characteristics of EEG to detect delayed cerebral ischemia induced by non-traumatic subarachnoid hemorrhage

Abstract

The article describes a new type of AUC diagrams intended for the analysis of interhemispheric asymmetry of amplitude-frequency characteristics of electroencephalograms (EEG) of patients with subarachnoid hemorrhage, as well as a new type of head maps named maps of interhemispheric asymmetry of EEG. AUC diagrams are a new statistical tool for identifying regularities in biomedical signals. The idea of AUC diagrams is to visually represent the dependence of the area under the ROC curve (AUC) when comparing data samples from the bounds of the range of given characteristic of this data, for example, frequency or amplitude, etc. The article demonstrates that this principle of data analysis allows us to identify some signs of postoperative complications that may occur in patients undergoing intensive care unit. It is known that the signs of such complications are changes in the amplitude and frequency of EEG oscillations in the neurophysiological frequency ranges delta, theta, alpha, and beta; however, amplitude changes can be caused by other reasons including the state of sleep and exposure to pharmacological drugs. Changes in the amplitude caused by postoperative complications can be revealed by analysis of the interhemispheric asymmetry of the patient's EEG. The developed type of AUC diagrams and interhemispheric EEG asymmetry maps help to automate such EEG analysis. The effectiveness of the developed statistical tools was demonstrated by the analysis of data in two patients with clinically confirmed delayed cerebral ischemia induced by subarachnoid hemorrhage.