Compressed EEG Pattern Analysis for Critically Ill Neurological-Neurosurgical Patients

Abstract

Recent advances in continuous electroencephalogram (EEG) monitoring with digital EEG acquisition, storage, and quantitative analysis allow uninterrupted assessment of cerebral cortical activity in critically ill neurological-neurosurgical patients. Early recognition of worsening brain function can prove of vital importance as one can initiate measures aimed to prevent further brain damage. Although continuous EEG monitoring provides adequate spatial and temporal resolution and is able to continuously assess brain function in these critically ill patients, it requires a trained electroencephalographer to interpret the massive amounts of data generated. This limitation impedes the widespread use of EEG in assessing real-time brain function in critically ill patients. Here, we demonstrate the utility of a novel method of automated EEG analysis that segments and extracts EEG features, classifies and groups them according to various patterns, and then presents them in a compressed fashion. This permits real-time viewing of several hours of EEG on a single page. Examples are presented from three patients, two with recurrent seizures and one with diagnosis of subarachnoid hemorrhage. These patients illustrate the ability of this novel method to detect important real-time physiological changes in brain function that could enable early interventions aimed to prevent irreversible brain damage.