Resolving T2 and T2' contributions to BOLD response speed and nonlinearity using 7T EPTI

Motivation: Non-linearities in BOLD are expected to be greater in draining veins, suggesting that attenuating macrovascular signals should increase temporal accuracy of fMRI. Removing venous contamination is challenging due to T2' effects during typical imaging readouts. Goal(s): To separate pure T2-weighted and T2'-contaminated BOLD components to characterize macrovascular non-linearities. Approach: We used single-shot echo-planar time-resolved imaging, reconstructed with varying T2' contamination levels, and estimated hemodynamic response properties in response to short- and long-duration stimuli. Results: Analyzing each signal contribution in the same dataset (i.e., same neuronal activity) showed that T2' contamination influences BOLD linearity, and extended the time-to-peak of responses by hundreds of milliseconds. Impact: We tested whether pure-T2 signals BOLD can enable more specific BOLD responses by simultaneously measuring pure-T2 and T2' contributions using spin-echo echo-planar time-resolved imaging. Pure-T2 weighting decreased the non-linearity of BOLD responses and reduced response time-to-peak by hundreds of milliseconds.