Blood oxygenation level-dependent (BOLD) total and extravascular signal changes and ΔR2* in human visual cortex at 1.5, 3.0 and 7.0 T.

The characterisation of the extravascular (EV) contribution to the blood oxygenation level-dependent (BOLD) effect is important for understanding the spatial specificity of BOLD contrast and for modelling approaches that aim to extract quantitative metabolic parameters from the BOLD signal. Using bipolar crusher gradients, total (b = 0 s/mm(2) ) and predominantly EV (b = 100 s/mm(2) ) gradient echo BOLD ΔR(2)* and signal changes (ΔS/S) in response to visual stimulation (flashing checkerboard; f = 8 Hz) were investigated sequentially (within < 3 h) at 1.5, 3.0 and 7.0 T in the same subgroup of healthy volunteers (n = 7) and at identical spatial resolutions (3.5 × 3.5 × 3.5 mm(3)). Total ΔR(2)* (z-score analysis) values were -0.61 ± 0.10 s(-1) (1.5 T), -0.74 ± 0.05 s(-1) (3.0 T) and -1.37 ± 0.12 s(-1) (7.0 T), whereas EV ΔR(2)* values were -0.28 ± 0.07 s(-1) (1.5 T), -0.52 ± 0.07 s(-1) (3.0 T) and -1.25 ± 0.11 s(-1) (7.0 T). Although EV ΔR(2)* increased linearly with field, as expected, it was found that EV ΔS/S increased less than linearly with field in a manner that varied with TE choice. Furthermore, unlike ΔR(2)*, total and EV ΔS/S did not converge at 7.0 T. These trends were similar whether a z-score analysis or occipital lobe-based region-of-interest approach was used for voxel selection. These findings suggest that calibrated BOLD approaches may benefit from an EV ΔR(2)* measurement as opposed to a ΔS/S measurement at a single TE.