We recorded somatosensory evoked potentials in pentobarbital-anesthetized rats before and after middle cerebral artery occlusion. Trigeminal (vibrissae), median (forelimb), and sciatic (bind limb) nerve stimuli produced consistent, robust, and sharply localized responses in the trigeminal, forelimb, and hind limb regions of the somatosensory cortex of 18 rats. These regions are situated at sequentially greater distances from the center of infarcts produced by middle cerebral artery occlusion. In eight rats, occlusion 1-2 mm below the rhinal fissure abolished somatosensory evoked potentials in all three cortical region within minutes. Positive wavelets preceding the primary cortical response were also diminished by the occlusion, suggesting that ischemia affected the thalamocortical white matter. Four of these eight rats did not show histologically apparent ischemic involvement of the hind limb cortical region at 3 hours after occlusion; sciatic nerve evoked potentials recovered substantially in all four rats, and the amplitudes exceeded baseline (129±30% at 1 hour, 173±33% at 3 hours) in three of the four rats. Three of the eight rats did not have gross ischemic involvement of the forelimb cortical region; median nerve evoked potentials recovered fully in all eight rats, but the amplitudes did not exceed baseline. All eight rats had evidence of ischemic damage in the trigeminal cortex; no rat showed full recovery in this region, and all but one had trigeminal evoked potentials that were <20% of baseline amplitudes by 3 hours after occlusion. Two rats had large infarcts involving the hind limb and forelimb cortical regions and were studied for 24 hours; these rats did not recover evoked potentials in the three regions at 24 hours, although one rat transiently recovered hind limb evoked potentials to 83% of baseline amplitude at 3 hours. Amplitude of the hind limb somatosensory evoked potentials recorded in the contralateral hemisphere doubled after occlusion. Thus, middle cerebral artery occlusion causes widespread loss of somatosensory evoked potentials in the hind limb, forelimb, and trigeminal cortical regions of the ischemic hemisphere of rats. These regions show distinctly different temporal patterns of recovery. Our findings suggest that cortical evoked potentials may be useful for monitoring ischemic damage in these cortical regions after middle cerebral artery occlusion (Stroke 1990^1 89 and causes infarcts largely localized to the cerebral cortex.:The rat primary somatosensory cortex has three anatomically and physiologically distinct regions: the hind limb area (HL), the forelimb area (FL), and the parietal 1 area (Parl).1112 Parl is situated most laterally and receives projections mainly from the cutaneous mechanoreceptors of the mystacial vibrissae, the head, and the neck. HL is located most medially, and FL lies between HL and Parl. Peripheral nerve stimulation produces localized field potentials in the somatosensory cortex. The typical ischemic infarct produced by MCAO encroaches on these regions....