BOLD‐based cerebrovascular reactivity vascular transfer function isolates amplitude and timing responses to better characterize cerebral small vessel disease

Abstract

Cerebrovascular reactivity (CVR) is a dynamic measure of the cerebral blood vessel response to vasoactive stimulus. Conventional CVR measures amplitude changes in the blood‐oxygenation‐level‐dependent (BOLD) signal per unit change in end‐tidal CO2 (PETCO2), effectively discarding potential timing information. This study proposes a deconvolution procedure to characterize CVR responses based on a vascular transfer function (VTF) that separates amplitude and timing CVR effects. We implemented the CVR‐VTF to primarily evaluate normal‐appearing white matter (WM) responses in those with a range of small vessel disease. Comparisons between simulations of PETCO2 input models revealed that boxcar and ramp hypercapnia paradigms had the lowest relative deconvolution error. We used a T2* BOLD‐MRI sequence on a 3 T MRI scanner, with a boxcar delivery model of CO2, to test the CVR‐VTF approach in 18 healthy adults and three white matter hyperintensity (WMH) groups: 20 adults with moderate WMH, 12 adults with severe WMH, and 10 adults with genetic WMH (CADASIL). A subset of participants performed a second CVR session at a one‐year follow‐up. Conventional CVR, area under the curve of VTF (VTF‐AUC), and VTF time‐to‐peak (VTF‐TTP) were assessed in WM and grey matter (GM) at baseline and one‐year follow‐up. WMH groups had lower WM VTF‐AUC compared with the healthy group (p < 0.0001), whereas GM CVR did not differ between groups (p > 0.1). WM VTF‐TTP of the healthy group was less than that in the moderate WMH group (p = 0.016). Baseline VTF‐AUC was lower than follow‐up VTF‐AUC in WM (p = 0.013) and GM (p = 0.026). The intraclass correlation for VTF‐AUC in WM was 0.39 and coefficient of repeatability was 0.08 [%BOLD/mm Hg]. This study assessed CVR timing and amplitude information without applying model assumptions to the CVR response; this approach may be useful in the development of robust clinical biomarkers of CSVD.