Abstract
The P300 complex was derived from the electroencephalogram (EEG) as subjects mentally counted infrequent large checkerboard visual stimuli, presented randomly among frequent small checkerboard stimuli. Use of low contrast (10%) stimuli and four midline scalp electrodes, facilitated separation of cognitive and sensory components and enabled the P300 complex to be resolved into three distinct components--N200, P3a, and P3b. In 20 healthy adult subjects normative data were established and the P3a and P3b components were shown to depend on cognitive function. In 19 age-matched cirrhotic patients without overt hepatic encephalopathy (HE) the EEG and visual evoked potentials (VEPs) were normal, but latencies of P3a and/or P3b were prolonged in 9. Prolonged latencies were not associated with an abnormal number connection test. Ten additional age-matched cirrhotic patients without overt HE, who were alcohol, drug, and caffeine free, were randomized to receive flumazenil (1 mg) and placebo intravenously, double-blind. After flumazenil or placebo, latencies of P3a and P3b and psychometric test results did not change significantly. These findings suggest that in cirrhotic patients without overt HE (i) impaired cognitive sensory function may occur in the absence of abnormalities of a standard psychometric test, the EEG, or VEPs, and (ii) increased latencies of P3a and P3b may constitute a component of subclinical HE, which is not mediated by increased brain levels of central benzodiazepine receptor agonist ligands.