Nonstationary multivariate modeling of cerebral autoregulation during resting state and hypercapnia (1184.10)

Abstract

We examined the time‐varying characteristics of cerebral autoregulation and hemodynamics during resting state and hypercapnia by using recursively estimated multivariate (two‐input) models which quantify the dynamic effects of mean arterial blood pressure (ABP) and end‐tidal CO2 tension (PETCO2) on middle cerebral artery blood flow velocity (CBFV). Experimental measurements of spontaneous variations of these signals were obtained from thirteen healthy subjects under normal, free‐breathing conditions. Beat‐to‐beat values of ABP and CBFV, as well as breath‐to‐breath values of PETCO2 were also obtained in 8 female subjects during baseline and sustained euoxic hypercapnia. The multiple‐input, single‐output linear models used to describe the relationship between ABP, PETCO2 and CBFV were based on the Laguerre expansion technique. In order to account for the different dynamics associated with each input, the model parameters were updated using a recursive least squares scheme with constant and adaptive multiple forgetting factors. The results reveal the presence of nonstationarities that are more pronounced in the very low frequency range. By comparing one‐input (MABP) and two‐input (MABP and PETCO2) models, our results point out that the incorporation of PETCO2 as an additional input yields less time‐varying estimates of dynamic pressure autoregulation obtained from single‐input (MABP‐CBFV) models, suggesting the important role of PETCO2 and the possible shortcomings of assessing dynamic autoregulation using such models.