The EEG signal in hepatic encephalopathy: from a time-domain statistical analysis towards fractional Brownian processes

The quantitative analysis of the EEG in hepatic encephalopathy is usually performed in the frequency domain; in this study we tested the effectiveness of a new statistical approach based on a time domain analysis (introduction of a scaling function and calculation of a dominance index) and on the modeling of the signal as a fractional Brownian process (calculation of the Hurst exponent). Sixty-five cirrhotic patients were assayed for hepatic function by the Child-Pugh score and for encephalopathy by psychometric testing and by digital EEG recording. The dominance index showed a significant correlation with the degree of hepatic failure and of vigilance. Moreover, a linear relationship was found between the dominance index and three psychometric tests. In conclusion, the dominance index proved to reflect hepatic encephalopathy and the Hurst exponent highlighted features of cerebral dysfunction that were not detectable by routine EEG examination.