The purpose of this investigation was to evaluate the new technique of direct cortical stimulation in conscious rats, using ramp generated pulse trains. In rats with electrodes permanently implanted in the frontoparietal cortex, twice daily stimulation for a fortnight, with ramp shaped trains of 2 msec duration, frequency 50 Hz, ramp time 200 sec and incrementing 5 microA per 0.1 sec yielded reproducible convulsive thresholds which could be repeated at short intervals over a period of time. Two different pharmacodynamic parameters viz Threshold for Localized Seizures (TLS) and Threshold for Generalised Seizure (TGS) could be distinguished. TLS is defined as the current (microA) required to elicit forelimb clonus or minimal restricted seizure activity and at the 14th session was found to be 650 +/- 91 microA (n = 14). TGS is defined as the current (microA), at which forceful clonic jerks and vigorous clonic activity without the tonic component occurred, and at the 14th session was found to be 1059 +/- 171 microA (n = 23). The effect of (doses, mg/kg, ip),phenytoin (PHY)10, carbamazepine (CBZ) 10 and 20; sodium valproate (SV) 300 and ethosuximide (ESM) 200 were studied on the TLS and TGS (n = 6). PHY, CBZ and SV significantly elevated both TLS and TGS, whereas ESM was ineffective. PHY and CBZ elevated thresholds within 0.5 hr and continued for 5 to 6 hr. For SV, significant elevation of TLS and TGS commenced at 5 to 10 min and lasted about 4 hr. Comparison of these results with conventional methods, suggests that this model provides a new dimension to profiling anticonvulsant compounds, with better extrapolation to clinical situations.