Lag-optimized BOLD cerebrovascular reactivity derived from breathing task data has a stronger relationship with baseline cerebral blood flow

Abstract

Cerebrovascular reactivity (CVR) is an important indicator of cerebrovascular health and is commonly studied with the Blood Oxygenation Level Dependent functional MRI (BOLD-fMRI) response to a vasoactive stimulus. There is theoretical and empirical evidence to suggest that baseline cerebral blood flow (CBF) modulates the BOLD signal amplitude, and that baseline CBF may influence BOLD-CVR estimates. We address how some pertinent data acquisition and modelling choices affect the relationship between baseline CBF and BOLD-CVR: whether BOLD-CVR is modelled with breathing task data or just resting-state data, and whether BOLD-CVR amplitudes are optimized for hemodynamic lag effects. For the relationship between baseline CBF and BOLD-CVR, we assessed both between-subject correlations of average GM values and within-subject spatial correlations across cortical regions. Our results suggest that a simple breathing task addition to a resting-state scan, alongside lag-optimization within BOLD-CVR modelling, can improve BOLD-CVR correlations with baseline CBF, both between- and within-subjects, likely because these CVR estimates are more physiologically accurate. We report positive coupling between baseline CBF and BOLD-CVR, both between and within subjects; the physiological explanation of this positive coupling is unclear, and future research with larger sample sizes and more tightly controlled vasoactive stimuli is needed. Understanding how baseline vascular physiology relates to dynamic cerebrovascular processes will bring further insights into what drives between and within subject participant variability in BOLD-CVR measurements and related measurements of cerebrovascular function. These insights are particularly relevant when interpreting results in populations with altered vascular and/or metabolic baselines or impaired cerebrovascular reserve.

Hemodynamic timing in resting-state and breathing-task BOLD fMRI

“…However, within ethical restrictions, analytic derivatives not included in this publication may be supplied upon request. The visual instructions for the breathing tasks, displayed during scanning, were created with PsychoPy code (Stickland, Bright, Moia, & Zvolanek, 2022) and the hemodynamic timing estimates were produced with the Rapidtide Toolbox v2.0.9 (Frederick, 2016-2022). Specific details on how these coding repositories were used, as well as other primary analysis code for this project, have been organized into this GitHub repository: github.com/BrightLab-ANVIL/Gong_2022.…”

Bayesian Modelling Approaches for Breath-Hold Induced Cerebrovascular Reactivity

“…Non-invasive magnetic resonance imaging (MRI) has emerged as a promising tool for characterising anatomical and haemodynamic changes in the brain. In particular, functional magnetic resonance imaging (fMRI) based on blood oxygen level-dependent (BOLD) contrast acquired during a breath-hold (BH) task is a robust method to derive CVR maps in response to vasoactive stimuli (Pinto et al, 2021; Stickland et al, 2022; Zvolanek et al, 2023). The BH task leads to an elevation of the partial pressure of CO 2 ( PaCO 2 ) in the blood, which triggers the dilation of blood vessels and leads to an increase in CBF, that can be measured with BOLD fMRI.…”