Recent studies in our laboratory have shown that PCP (phencyclidine) and D-amphetamine induce a cognitive deficit in rats, in a paradigm of potential relevance for the pathology of schizophrenia. Atypical, but not classical antipsychotics and the anticonvulsant, lamotrigine have been shown to prevent a selective reversal learning deficit induced by PCP. In contrast, only haloperidol reversed the D-amphetamine-induced deficit. The present study aimed to explore the ability of two anticonvulsants with differing mechanism of action, valproate and phenytoin to attenuate the cognitive defi cits induced by PCP and D-amphetamine in the reversal learning paradigm. PCP at 1.5 mg/kg and D-amphetamine at 0.5 mg/kg both produced a selective and significant reduction in performance of the reversal phase with no effect on the initial phase of the task in female-hooded Lister rats. Valproate (25-200 mg/kg) and phenytoin (25-50 mg/kg) had no effect on performance when administered alone. Valproate (100-200 mg/kg), whose principle action is thought to be the enhancement of GABA transmission, was unable to prevent the cognitive deficit induced by either PCP or D-amphetamine. Conversely, phenytoin (50 mg/kg), a use-dependent sodium channel inhibitor, significantly prevented the deficit induced by PCP, but not D-amphetamine. These results add to our earlier work with lamotrigine, and suggest that sodium channel blockade may be a mechanism by which some anticonvulsant drugs can prevent the PCP-induced deficit. These data have implications for the use of anticonvulsant drugs in the treatment of cognitive or psychotic disorders.