Cerebral autoregulation is a complex physiological process composed of both fast and slow components that may respond differently to different rates and patterns of blood pressure variation. To assess the temporal nature of autoregulation, transcranial Doppler velocity recordings of the middle cerebral artery obtained over prolonged periods were compared with blood pressure recordings in 5 patients without cerebral disease and in 13 patients with intracranial pathological changes. Correlations between the velocity and pressure wave forms at various frequencies of variation were measured with systems analysis techniques. Patients with aneurysmal subarachnoid hemorrhage had high correlations indicating pressure-dependent flow and impaired autoregulation. Patients without cerebral disease had significantly lower correlations (P less than 0.01), indicating intact autoregulation. Examples of increasing correlations and correlations at new frequencies emerging as the clinical condition worsened are given. These preliminary examples suggest that the application of systems analysis techniques to velocity and pressure data allow measurement of the temporal nature of cerebral autoregulation.