Dependence of gradient-echo and spin-echo BOLD fMRI at 4 T on diffusion weighting

Diffusion weighting and spin-echo (SE) acquisitions can be used to help improve the spatial localization of BOLD fMRI at the cost of reduced acquisition rates and lower signal-to-noise ratio (SNR). To evaluate these costs, SE and gradient-echo (GE) data were acquired at 4 T at five diffusion weightings ranging from b = 0 to 1110 s/mm(2) using a robust visual stimulus. The data showed reduced functional contrast when diffusion weighting was applied. As the amount of diffusion weighting increased, the functional contrast initially dropped sharply, and then remained relatively constant for diffusion weightings above 15 s/mm(2) for SE and 30 s/mm(2) for GE data. GE functional BOLD contrast was attenuated to 94.0 +/- 10.1, 87.6 +/- 12.2, 86.4 +/- 8.8 and 83.3 +/- 20.6% of the non-diffusion-weighted GE contrast for diffusion weightings of 15, 30, 200 and 1,110 s/mm(2). The non-diffusion-weighted SE contrast greatly reduced to 19.3 +/- 3.3% of the non-diffusion-weighted GE contrast, demonstrating the large activation attenuation of a SE acquisition. The SE contrast was further reduced to 10.0 +/- 3.6, 9.0 +/- 2.5 and 8.6 +/- 2.0% of the non-diffusion-weighted GE contrast for the 15, 30 and 200 s/mm(2) diffusion-weighted data. These results suggest that only a small amount of diffusion weighting is necessary to suppress the vascular contribution and spin-echo imaging should only be used if there is adequate statistical power available or accurate localization is critical.