Reproducibility of a ramping protocol to measure cerebral vascular reactivity using functional magnetic resonance imaging

Abstract

Though individual differences in arterial carbon dioxide and oxygen levels inherently exist, the degree of their influence on cerebral vascular reactivity (CVR) is less clear. We examined the reproducibility of BOLD signal changes to an iso‐oxic ramping PetCO2 protocol. CVR changes were induced by altering PetCO2 while holding PetO2 constant using a computer‐controlled sequential gas delivery (SGD) device. Two MRI scans, each including a linear change in PetCO2, were performed using a 3‐Tesla (3T) scanner. This ramp sequence consisted of 1 min at 30 mmHg followed by 4 min period during where PetCO2 was linearly increased from 30 to 50 mmHg, 1 min at 51 mmHg, and concluded with 4 min at baseline. The protocol was repeated at a separate visit with 3 days between visits (minimum). Intraclass correlation coefficients (ICC) and coefficients of variation (CV) were used to verify reproducibility. Eleven subjects (6 females; mean age 26.5 ± 5.7 years) completed the full testing protocol. Good reproducibility was observed for the within‐visit ramp sequence (Visit 1: ICC = 0.82, CV = 6.5%; Visit 2: ICC = 0.74, CV = 6.4%). Similarly, ramp sequence were reproducible between visits (Scan 1: ICC = 0.74, CV = 6.5%; Scan 2: ICC = 0.66, CV = 6.1%). Establishing reproducible methodologies for measuring BOLD signal changes in response to PetCO2 alterations using a ramp protocol will allow researchers to study CVR functionality. Finally, adding a ramping protocol to CVR studies could provide information about changes in CVR over a broad range of PetCO2.

ICA-based denoising strategies in breath-hold induced cerebrovascular reactivity mapping with multi echo BOLD fMRI

“…Regarding CVR reliability, the whole-brain average reliability of SE-MPR was comparable to long-term reliability (days or weeks apart) found in previous studies of CVR induced by BH ( Peng et al, 2019 ), by paced deep breathing ( Sousa et al, 2014 ), or by gas challenges ( Leung et al, 2016 ), and higher than that reported in other studies on BH induced CVR estimated with a non-lagged optimized P ET CO 2 hrf trace ( Lipp et al, 2015 ) or with Fourier modelling ( Pinto et al, 2016 ), and by gas challenges ( Dengel et al, 2017 ; Evanoff et al, 2020 ). Consequently, the reliability of CVR estimates obtained with the optimal combination dataset and conservative ME-ICA modelling approaches were found higher than those previously reported in the literature.…”

“…average CVR value across sessions and within subjects) ( Lipp et al, 2015 ; Magon et al, 2009 ). Both short and long term reliability of BH-induced CVR were found to be comparable to that of other non-invasive means of estimating CVR, such as resting state fMRI ( Liu et al, 2017 ), inhaled gas challenges ( Dengel et al, 2017 ; Evanoff et al, 2020 ; Leung et al, 2016 ), Fourier modelling of a BH task ( Pinto et al, 2016 ), and a paced deep breathing task ( Sousa et al, 2014 ).…”

ICA-based Denoising Strategies in Breath-Hold Induced Cerebrovascular Reactivity Mapping with Multi Echo BOLD fMRI

“…Consequently, the reliability of CVR estimates obtained with the optimal combination dataset and conservative ME-ICA modelling approaches were found higher than those previously reported in the literature. However, all strategies produced a reliability that was lower than the short-term (within-session) reliability reported in BH induced CVR (Peng et al, 2019), resting state based CVR (P. Liu et al, 2017), and gas challenge induced CVR (Leung et al, 2016), although lower intersession reliability in gas challenges has also been reported (Dengel et al, 2017;Evanoff et al, 2020). Note that the reliability observed in this study seems to be globally higher and spatially less variable than that reported in previous studies (Lipp et al, 2015;Sousa et al, 2014).…”

“…Regarding CVR reliability, the whole-brain average reliability of SE-MPR was comparable to longterm reliability (days or weeks apart) found in previous studies of CVR induced by BH (Peng et al, 2019), by paced deep breathing (Sousa et al, 2014), or by gas challenges (Leung et al, 2016), and higher than that reported in other studies on BH induced CVR estimated with a non-lagged optimized P ET CO 2 hrf trace (Lipp et al, 2015) or with Fourier modelling (Pinto et al, 2016), and by gas challenges (Dengel et al, 2017;Evanoff et al, 2020). Consequently, the reliability of CVR estimates obtained with the optimal combination dataset and conservative ME-ICA modelling approaches were found higher than those previously reported in the literature.…”

“…average CVR value across sessions and within subjects) (Lipp, Murphy, Caseras, & Wise, 2015;Magon et al, 2009). Both short and long term reliability of BH-induced CVR were found to be comparable to that of other non-invasive means of estimating CVR, such as resting state fMRI (P. Liu et al, 2017), inhaled gas challenges (Dengel et al, 2017;Evanoff et al, 2020;Leung, Kim, & Kassner, 2016), Fourier modelling of a BH task (Pinto, Jorge, Sousa, Vilela, & Figueiredo, 2016), and a paced deep breathing task (Sousa, Vilela, & Figueiredo, 2014).…”