Breath-hold BOLD fMRI without CO2 sampling enables estimation of venous cerebral blood volume: potential use in normalization of stimulus-evoked BOLD fMRI data

Biondetti, Emma, Chiarelli, Antonio Maria, Germuska, Michael ORCID: https://orcid.org/0000-0003-0580-4350, Lipp, Ilona, Villani, Alessandro, Caporale, Alessandra S., Patitucci, Eleonora, Murphy, Kevin ORCID: https://orcid.org/0000-0002-6516-313X, Tomassini, Valentina ORCID: https://orcid.org/0000-0002-7368-6280 and Wise, Richard G. ORCID: https://orcid.org/0000-0003-1700-2144 2024. Breath-hold BOLD fMRI without CO2 sampling enables estimation of venous cerebral blood volume: potential use in normalization of stimulus-evoked BOLD fMRI data. NeuroImage 285 , 120492. 10.1016/j.neuroimage.2023.120492

PDF - Published Version Download (4MB)

Abstract

BOLD fMRI signal has been used in conjunction with vasodilatory stimulation as a marker of cerebrovascular reactivity (CVR): the relative change in cerebral blood flow (CBF) arising from a unit change in the vasodilatory stimulus. Using numerical simulations, we demonstrate that the variability in the relative BOLD signal change induced by vasodilation is strongly influenced by the variability in deoxyhemoglobin-containing cerebral blood volume (CBV), as this source of variability is likely to be more prominent than that of CVR. It may, therefore, be more appropriate to describe the relative BOLD signal change induced by an isometabolic vasodilation as a proxy of deoxygenated CBV (CBVdHb) rather than CVR. With this in mind, a new method was implemented to map a marker of CBVdHb, termed BOLD-CBV, based on the normalization of voxel-wise BOLD signal variation by an estimate of the intravascular venous BOLD signal from voxels filled with venous blood. The intravascular venous BOLD signal variation, recorded during repeated breath-holding, was extracted from the superior sagittal sinus in a cohort of 27 healthy volunteers and used as a regressor across the whole brain, yielding maps of BOLD-CBV. In the same cohort, we demonstrated the potential use of BOLD-CBV for the normalization of stimulus-evoked BOLD fMRI by comparing group-level BOLD fMRI responses to a visuomotor learning task with and without the inclusion of voxel-wise vascular covariates of BOLD-CBV and the BOLD signal change per mmHg variation in end-tidal carbon dioxide (BOLD-CVR). The empirical measure of BOLD-CBV accounted for more between-subject variability in the motor task-induced BOLD responses than BOLD-CVR estimated from end-tidal carbon dioxide recordings. The new method can potentially increase the power of group fMRI studies by including a measure of vascular characteristics and has the strong practical advantage of not requiring experimental measurement of end-tidal carbon dioxide, unlike traditional methods to estimate BOLD-CVR. It also more closely represents a specific physiological characteristic of brain vasculature than BOLD-CVR, namely blood volume.

Item Type:	Article
Date Type:	Publication
Status:	Published
Schools:	Medicine Physics and Astronomy Psychology Cardiff University Brain Research Imaging Centre (CUBRIC)
Additional Information:	License information from Publisher: LICENSE 1: URL: http://creativecommons.org/licenses/by/4.0/, Start Date: 2023-12-06
Publisher:	Elsevier
ISSN:	1053-8119
Date of First Compliant Deposit:	11 December 2023
Date of Acceptance:	6 December 2023
Last Modified:	11 Dec 2023 10:30
URI:	https://orca.cardiff.ac.uk/id/eprint/164662

Actions (repository staff only)

Edit Item