In this investigation a newly developed direct cortical stimulation technique was evaluated for measurement of anticonvulsant efficacy in rats. The kinetics of drug action for carbamazepine, phenytoin, valproate, phenobarbital, ethosuximide and oxazepam were studied in conjunction with their pharmacokinetics. Motor cortex stimulation with a ramp-shaped pulse train allowed successive determination of a threshold for localized seizure activity (TLS) and for generalized seizure activity (TGS). For each drug the time course of effect was followed in individual animals. Differential effects on the pharmacodynamic parameters were seen. Phenytoin and carbamazepine clearly elevated the TGS. However, phenytoin did not affect TLS and carbamazepine only marginally. Valproate increased both TLS and TGS to the same extent. Phenobarbital and oxazepam elevated both thresholds, but the effect on TGS was more pronounced. Ethosuximide had little effect on both thresholds. Comparison with other animal models suggested that elevation of TLS reflects an effect on seizure initiation, whereas elevation of TGS above TLS reflects an effect on seizure propagation. All drugs exhibited a nonlinear relationship between plasma concentration and anticonvulsant efficacy, without ceiling of anticonvulsant intensity at the highest concentrations. The effective concentration range of most compounds coincided with the "therapeutic" range in humans. The direct cortical stimulation technique is useful for preclinical monitoring of anticonvulsant efficacy with most antiepileptic drugs because it allows detection of both qualitative and quantitative differences. In addition the model is particularly useful for time course studies.