Hi,
What we measure in diffusion MRI is signal attenuation compared to a baseline signal. The b0 volume in your case is the baseline signal that you use for the analysis.
Taking multiple b0s has been shown to minimise the variance of the estimated diffusivities via the DTI model. Have a look at Jones et al, MRM, 1999:
http://www.ncbi.nlm.nih.gov/pubmed/10467296
Cheers
Stam
On 3 May 2013, at 18:39, Nima Kasraie wrote:
> OK. Here's my first question on b0 and DTI:
>
> We've been doing a little work lately with concussion patients using FSL to compare their FA (fractional anisotropy) and MD (mean diffusivity) values to "normal" controls. In a number of articles which talk about this topic, people seem to be running their DTI with 60 directions, either 1 or 2 NEX, and (here is the question) ... multiple b0 values. Sometimes it is said to be multiple b0's other times they run the DTI with a single average, but run the b0 with multiple (up to 10) averages. First of all, why would this aid in DTI processing? Before running the data through tbss, we make a brain mask with the b0 image using FSL BET. Then we do eddy correction on the 4D DTI data set with FSL/FDT. After that we run the eddy-corrected data, bvec table, bval table, and the brain mask through DTIfit. Then we use the FA or MD maps in tbss. I'm not sure why having multiple b0s (or b0 with 10 NEX) would help. I assume we'd get more signal on the b0, but that wouldn't seem necessary for the brain mask. Does running multiple b0s improve the bvecs or bvals table?
>
|