Autoregulation maintains cerebral blood flow (CBF) almost constant in the face of changes in arterial blood pressure (ABP). Tests for impairment of this process using only spontaneous fluctuations in ABP, without provoking large variations, are of great clinical interest, and a range of different approaches have previously been applied. Extending earlier work based on linear filters, we propose a simple parametric method using a first order finite impulse response filter. We evaluate the method on ABP and CBF velocity [(CBFV), from trancranial Doppler ultrasound] signals collected in 60 patients with stenosis or occlusion of the carotid arteries. Data were collected during the inspiration of ambient air, a 5% CO2/air mixture, and finally the return to ambient air. Equivalent data were collected in 15 normal subjects. The filters estimated from the data segments with constant inspiratory pCO2 showed the expected high-pass characteristic, which was reduced during hypercapnia and also in patients. Highly significant correlation between the filter parameters and cerebrovascular reactivity (percent increase in CBFV per unit change in end-tidal pCO2) gives further evidence that the filters reflect autoregulation. The method allows simple parametrization of the dynamic autoregulatory responses in CBFV, and the analysis of short (1 min) data segments.