Cerebrovascular reactivity (CVR) is used as an outcome measure of brain health. Traditionally, lower CVR is associated with ageing, poor fitness and brain-related conditions (e.g. stroke, dementia). Indeed, CVR is suggested as a biomarker for disease risk. However, recent findings report conflicting associations between ageing or fitness and CVR measures. Inconsistent findings may relate to different neuroimaging modalities used, which include transcranial Doppler (TCD) and blood-oxygen-level-dependant (BOLD) contrast magnetic resonance imaging (MRI). We assessed the relationship between CVR metrics derived from two common imaging modalities, TCD and BOLD MRI, within the same individuals and with expected significant differences (i.e. younger vs. older) to maximise the expected spread in measures.We conducted two serial studies using TCD- and MRI-derived measures of CVR (via inspired 5% CO2 in air). Study 1 compared 20 younger (24±7 years) with 15 older (66±7 years) participants, Study 2 compared 10 younger (22±2 years) with 10 older (72±4 years) participants. Combining the main measures across studies, no significant correlation (r=0.15, p=0.36) was observed between individual participant TCD- and BOLD-CVR measures. Further, these measures showed differential effects between age groups; with TCD-CVR higher in the older compared to younger group (4±1 vs. 3±1 %MCAv/mmHg PETCO2; p<0.05, Hedges’ g=0.75), whereas BOLD-CVR showed no difference (p=0.104, Hedges’ g=0.38). In Study 2 additional measures were obtained to understand the origin of the discrepancy: phase contrast angiography (PCA) MRI of the middle cerebral artery, showed a significantly lower blood flow (but not velocity) CVR response in older compared with younger participants (p>0.05, Hedges’ g=1.08). The PCA CVR metrics did not significantly correlate with the BOLD- or TCD-CVR measures. The differing CVR observations between imaging modalities were despite expected, correlated (r=0.62-0.82), age-related differences in resting CBF measures across modalities.Taken together, findings across both studies show no clear relationship between TCD- and BOLD-CVR measures. We hypothesize that CVR differences between imaging modalities are in part due to the aspects of the vascular tree that are assessed (TCD:arteries; BOLD:venules/veins). Further work is needed to understand the between-modality CVR response differences, but caution is needed when comparing CVR metrics derived from different imaging modalities.
Bibliographical noteFunding Information:
We would like to thank Dr. Matthew Ryan for assessing the ECGs and Dr. Molly Bright for early conversations about the design of the MRI CVR studies. We would also like to thank the Birmingham University Imaging Centre (BUIC) for generously providing additional MRI scanning time to support this study. Finally, thank you to Universitas 21 and Guarantors of Brain for travel grants that supported student technical training and conference presentation opportunities. Funding. This work was supported by The Physiological Society (Research Grant #444, 2014), a University of Birmingham Ph.D. studentship and a University of Birmingham-University of Nottingham Strategic Collaboration Grant.
© Copyright © 2021 Burley, Francis, Thomas, Whittaker, Lucas and Mullinger.
- brain vascular health
- cerebral blood flow
- cerebrovascular reactivity
- magnetic resonance imaging
- transcranial Doppler
ASJC Scopus subject areas
- Physiology (medical)