Comparison of two linear models of dynamic cerebral autoregulation

The assessment of dynamic cerebral autoregulation response using changes in arterial blood pressure (ABP) as a stimulus is increasingly used. Transcranial Doppler ultrasonography measurements of middle cerebral artery velocity (MCAv) are often used in conjunction with ABP measurements using photoplethysmography (e.g. Finapres) to assess the response of the autoregulation mechanism. Two linear models of dynamic cerebral autoregulation have been developed independently. The first is an ARX model using the least-squares algorithm to fit the ABP and MCAv signals. The second is a flow dependent feedback mechanism controlling the pressure gradient across the MCA. Both models have been found to reproduce qualitatively similar results to those recorded in both thigh cuff and lower body negative pressure experiments, whereas the first model has also been used to analyse MCAv simulated using Ursino's physiological model. This paper assesses the ability of the two models to reproduce MCAv measurements from recordings of ABP from the same experiments.