Yes the text file contain the necessary information and dtifit and bedpostx use these files for processing.
As a check, use FSLVIEW to quickly verify that the text files and the volumes are in fact in the same order (they should be if it's coming straight out of the converter program like dcm2niigui).
You should see a T2-like image for bvals=0, bvecs=[0 0 0] volumes and otherwise (directionally-weighted) for non-zero bvals/bvecs volumes.
Usually the first b0 volume is ok to use, but if the subject moved for that volume, then use the second or third (again, check using FSLVIEW).
Always good idea to check for motion-corruption/artifacts since a lot of artifact corrupted volumes can give misleading dtifit results.
Best,
Danny Kim
On 2015-09-22, at 12:37 PM, Jon Anderson wrote:
> Thanks a lot Danny for this detailed explanation. Actually this relieved me. I thought that I need to apply fslroi to extract all the b0 volumes then merge them together.
> Sounds perfect!
>
> So I need only the first b0 volume to start processing the DTI data.
>
> Actually what I still unable to figure out is how the "dtifit" tool will work while I have non diffusion volumes reside between the diffusion volumes. Is it because the information about the vectors and b values are already recorded in bval and bvecs as you said?
|