Transcranial Doppler ultrasound validation of BOLD-fMRI cerebral blood flow relationship

A precise understanding of the interplay between cerebral blood flow (CBF) and blood oxygen level-dependent (BOLD) fMRI signals is essential for advancing cerebrovascular research. Although calibrated BOLD approaches often rely on arterial spin labelling (ASL) to estimate CBF, alternative validation using transcranial Doppler ultrasound (TCD) has not been explored. This study aims to determine whether a simplified hemodynamic model and linear regression can accurately characterize the relationship between TCD-derived CBF velocity and BOLD-fMRI responses during a ramp CO₂ stimulus. We hypothesized that both models would provide robust fits within the moderate partial pressure of end-tidal carbon dioxide (PETCO₂) and BOLD signal ranges tested.

Twenty-five healthy participants underwent two sessions. In session 1, right middle cerebral artery velocity (MCAv) was acquired using clinical TCD. In session 2, 3 T BOLD-fMRI data were collected. Both sessions used a ramp PETCO₂ protocol with deep breaths followed by 5% and 10% CO₂. Data processing included motion correction, spatial smoothing, fieldmap correction, high-pass filtering, and PETCO₂ alignment with smoothed MCAv (MCA $\bar{v}$ ) and BOLD signals from the right parietal lobe. A simplified hemodynamic model and linear regression were applied to assess the MCA $\bar{v}$ -BOLD relationship, with model performance evaluated by R².

Final analysis included 21 participants. The hemodynamic model produced consistent fits (R² ≥ 0.69). Linear regression showed strong agreement between MCA $\bar{v}$ and BOLD (R² = 0.759).

Both modeling approaches successfully linked TCD-derived MCA $\bar{v}$ and BOLD-fMRI responses during hypercapnia. These findings support the use of TCD as a complementary surrogate for CBF in BOLD calibration and cerebrovascular research.