Nonlinear analysis of EEG in septic encephalopathy

Abstract

Electroencephalograms (EEG) were recorded in fourteen patients who experienced a severe septic encephalopathy (SE). EEG analysis included visual inspection, spectral analysis and a recently developed nonlinear analysis (the Kaplan test). All EEGs showed decreased fast activity and an increase of slow wave activity on visual inspection. There was a nonsignificant trend of negative correlation between the spectral EEG analysis and the severity of the acute systemic illness (based on the sum score of 14 variables known as APACHE II score) (standard coefficient = -0.43, p = 0.118). However, a much more pronounced and significant negative correlation was observed between the Kaplan test and the APACHE II score (standard coefficient = -0.94, p = 0.005). The EEG abnormalities seen in these patients were independent of the sedation level. Neither the EEG parameters, nor the APACHE II score, predicted outcome. Nonlinear analysis is more powerful than spectral analysis to extract clinical relevant information from EEGs in patients who experience a severe SE. The nonlinear EEG analysis suggest that brain dynamics in SE may be characterized by a shift into a fundamentally different level of cortical information exchange which can be summarized in nonlinear terminology as a loss of deterministic structure in the EEG.