Brainstem auditory (BAEPs) and somatosensory evoked potentials (SEPs) have been shown to be useful in detecting brainstem or cortical dysfunction in neurological diseases and in combination with other methods to diagnose brain death (37,38). These neurophysiological methods are simple and easy to perform. BAEPs and SEPs can even be easily recorded in intensive care units and guarantee a standardized examination. Moreover, these methods require no extensive patient cooperation and are not heavily influenced by learning effects. The role of BAEPs in the evaluation and diagnosis of hepatic encephalopathy is not clear. BAEPs are obviously strongly influenced by the etiology of liver disease and are normal in viral hepatitis, but prolonged in alcoholic liver disease, Wilson's disease or in hepatic coma (8,12). Unfortunately, BAEPs were not compared to psychometric tests. There was no clear-cut differentiation between various hepatic encephalopathy-gradings. At present, the use of BAEPs in the detection of subclinical hepatic encephalopathy cannot be recommended, whereas in comatose patients BAEPs can be useful as a prognostic marker and for follow-up examinations (12). Recently, Pozessere et al. (12) examined 13 comatose patients with advanced coma stages (Glasgow coma scale 5-10) and recorded unspecific changes in their EEG tracings. In all cases of hepatic coma and in one intoxicated patient they found prolongation of interpeak latencies. In addition, in this small study the interpeak latencies correlated well with the clinical outcome of the patients. Only two studies were performed using SEPs to detect neurophysiological alterations in hepatic encephalopathy (32,33). The design as well as the results of these studies are quite different. Despite the small number of patients (n = 10), the prolongation of late components in 50% of patients with hepatic encephalopathy stage 0 could be a promising result (32). The value of SEPs in detecting subclinical hepatic encephalopathy is rather undefined. The fact that the generation of SEPs is due to an activation of complex structures of the central nervous system justifies the need for further investigations with this modality. The recording of visual evoked potentials requires much more methodological and technical effort than the recording of BAEPs or SEPs. The discrimination between pattern reversal (PVEP) and flashlight (FVEP) stimulation is highly important for the proper interpretation of the published data. Most of the studies were done using FVEPs, which are in particular clinically relevant for comatose patients (31). The conclusions of the authors using FVEPs (22-25) are not supported by the American Electroencephalographic Society (31).(ABSTRACT TRUNCATED AT 400 WORDS)