Models of Cerebrovascular Reactivity in BOLD-fMRI and Transcranial Doppler Ultrasound.

The ability of cerebrovasculature to respond to meet tissue demands is vital for normal brain function and health. Cerebrovascular reactivity (CVR), a measure of the responsiveness of cerebrovasculature to vasoactive stimuli, is a valuable tool for evaluating cerebrovascular health. While CVR is commonly assessed using transcranial Doppler ultrasound (TCD), which measures blood velocity, or MRI-based techniques such as blood oxygenation level-dependent (BOLD) imaging, which reflect changes in blood oxygenation, direct comparisons between these modalities remain limited, particularly with stimuli that induce a large dynamic range. Because both methods capture hypercapnia-induced vascular changes, we hypothesised that their CVR metrics may be correlated. This study evaluates inter-modality correlations of CVR using TCD and BOLD-fMRI extracted from the MCA territory (parietal lobe) during a ramped hypercapnic protocol and different modelling strategies. Linear correlations across broad PETCO2 ranges validated the utility of linear CVR modelling in capturing repeatable metrics using TCD and MRI. A four-parameter sigmoid model revealed significant inter-modality variability in span and bound parameters, improved by fixing these parameters and focusing on slope and inflection point, which enhanced the correlations between modalities. These results support the reliability of linear CVR modelling within narrow vasoactive response ranges in healthy subjects and propose a simplified two-parameter sigmoid model as an effective framework for characterising non-linear CVR dynamics. This work adds to the sparse literature on inter-modality CVR comparisons and indicates which CVR metrics are comparable between TCD and BOLD-fMRI, emphasising CVR as a useful tool for assessing cerebrovascular health in research and clinical contexts.