Benzodiazepine receptor ligands are suggested to play a role in the pathogenesis of hepatic encephalopathy (HE). Accumulation of these ligands in brain was suggested to explain in part the notion of"increased GABAergic tone," the rational for which arose initially from reports of a beneficial effect of the selective benzodiazepine antagonist flumazenil in HE patients. It was suggested on the basis of the effect of flumazenil in human HE that liver failure may result in alterations of the density and/or affinity of the benzodiazepine-associated GABA-A receptor site. Subsequent controlled-clinical trials showed that fumazenil had a transient beneficial effect in only a subpopulation of HE patients. In contrast to the antagonists, partial inverse agonists of the benzodiazepine receptor have unequivocal beneficial effects on behavioral and electro-physiological performance in all experimental models of HE studied so far. Benzodiazepine-associated GABA-A receptors have consistently been demonstrated to be unaltered in both human and experimental HE. Contrary to initial reports, the so-called "endogenous benzodiazepines" do not appear to be significantly related to the pathogenesis of HE. On the other hand, nonbenzodiazepine GABA-A receptor complex modulators, such as neuro-steroids, recently identified in brain in human and experimental HE, may provide a new mechanistic basis for this disorder and lead to novel treatments for human HE.