We recorded regional cerebral blood flow, somatosensory evoked potentials, and auditory evoked potentials in the thalamic relay nuclei (ventral posterior lateral nucleus and medial geniculate body) and in the somatosensory and auditory cortices during and after 1 hour of transient left middle cerebral artery occlusion in nine cats. Regional cerebral blood flow was also measured in the thalamocortical tracts of five of these cats. Additionally, the integrity of thalamocortical connections was tested by retrograde labeling of the thalamic nuclei with horseradish peroxidase in eight cats (three of which experienced no ischemia). Regional cerebral blood flow was severely reduced during middle cerebral artery occlusion in the left primary auditory cortex (8.5 ml/100 g/min) and in white matter pathways (6.4-7.6 ml/100 g/min). In contrast, regional cerebral blood flow did not change significantly in the somatosensory cortex or in either thalamic nucleus. Evoked potentials were abolished in both cortices but remained unchanged in the thalamic nuclei. Cortical somatosensory evoked potentials disappeared 5-8 minutes later than auditory evoked potentials. Recirculation after 1 hour of ischemia resulted in rapid and almost complete recovery (94%) of somatosensory evoked potentials and little recovery (18.4%) of auditory evoked potentials. We conclude that in the auditory pathway both cortical and fiber tract ischemia are (perhaps synergistically) responsible for dysfunction, while in the somatosensory cortex evoked potentials are abolished due to white matter ischemia. The delayed disappearance and better recovery of somatosensory than of auditory evoked potentials indicate that ischemic tolerance is higher in fiber tracts than in cortex. (Stroke 1990^1:923-928) N umerous experimental studies have concentrated on the effect of ischemia on the functional and morphologic integrity of the cerebral cortex.1 -6 In many investigations, evoked potentials (EPs) recorded from specific cortical areas have served as the paradigm of function, and disturbances of EPs were usually interpreted as indicating cortical disorder. 1 -36 -9 In many clinical cases, however, strokes cause capsular or supracapsular lesions in the white matter, resulting in neurologic deficits not related to cortical destruction. As demonstrated by x-ray computed tomography and even better by magnetic resonance imaging, ischemic damage can be confined in such cases to the white matter or can From the