Cerebral blood flow (CBF) was measured in a series of patients with ischemic cerebrovascular disease using xenon-133 inhalation and single photon emission computer tomography. The spontaneous course of CBF changes in a consecutive series of stroke patients was evaluated. A quite heterogeneous pattern of flow changes was observed: In patients with large cortical/subcortical infarcts, extensive hypoperfused areas were observed, often significantly larger than the corresponding hypodense lesion on the CT scan. Smaller CT lesions caused relatively smaller flow changes. Patients with lacunar infarcts showed only a discrete reduction of CBF, but comprising most of the ipsilateral hemisphere. Repeated CBF studies in the chronic phase showed, that the clinical improvement commonly noted in stroke patients is not related to a CBF increase. On the contrary, the CBF lesions tended to become somewhat larger and more demarcated even in cases where the finding of a normal angiogram and a transient state of hyperemia suggested a dissolution of the intracerebral embolus. The pathogenetic mechanisms for these persisting low flow areas in CT intact structures was discussed. One possibility was a selective neuronal cell damage in the peri-infarct areas caused by the ischemic insult. Such lesions would leave the structures macroscopically intact, but decrease both the metabolic demands and CBF. However, this interpretation finds little support in recent microscopic neuropathological studies in man. A more likely possibility was then considered to be disconnection (diaschisis) where the reduced flow is due to a decreased neuronal function caused by undercutting of afferent or efferent nervefibers. A crossed cerebellar diaschisis was observed in all patients with major infarcts in the forebrain. These findings were observed already in the acute phase, but persisted quite unchanged throughout the subacute and chronic phases. The patients with lacunar infarcts showed cerebellar diaschisis in the acute phase only, suggesting that a transient suppression of remote areas is possible too. In order to differentiate between permanent flow changes caused by a functional impairment and a possible hemodynamic component, CBF was measured before and after administration of a potent cerebral vasodilator, acetazolamide (Diamox). In normal cases tested with Diamox, an even CBF increase is noted throughout the hemispheres, while the cerebral metabolic rate for oxygen remains stable. In patients having a severe stenosis or occlusion of the internal carotid artery, this vasodilatory stress test will identify the patients having poor collateral capacity via the circle of Willis.(ABSTRACT TRUNCATED AT 400 WORDS)