 |
 |
Hi Mark The gradient directions provided by your physicist should be identical to the bvecs, as long as there is no slice angulation in the acquisition. If you acquire oblique slices, then dcm2nii will apply a rotation to the gradient directions to account for the slice angulation. If you have b matrices for each DW volume, you should also be able to get the corresponding bvec entry using the principal eigenvector of the b matrix. Hope this helps Stam On 29 Oct 2013, at 12:26, Mark Postans <[log in to unmask]> wrote: > Hi FSL experts, > > I’ve acquired a series of 60 direction DWI’s with a 3T GE Signa scanner, and would like to use > BEDPOSTX and PROBTRACKX in FSL to analyse the data. > > Unfortunately, I’ve run into some problems generating the necessary bvec files, which I was hoping > you could help me with... > > I have basically been using Chris Rordens dcm2nii to convert my dicom files to nifti format, and to > produce the bval and bvec files for each participant. The bval files are correct (as are the nifti > images themselves), but the bvec files each contain nothing but a matrix full of zeros. > > In response to this, our MR physicist provided me with a copy of the gradient directions associated > with our 60 dir DTI acquisition sequence, however I’m not sure whether this is equivalent to the > bvecs required by FSL, and my understanding is that the bvecs should be slightly different for each > participant (?). > > Anyway, we typically use ExploreDTI to analyse our diffusion-MRI data, and whilst I can’t extract > subject-specific bvec files, I am able to obtain subject-specific b-matrices with this software. > > Our physicist says that he should be able to estimate the subject-specific bvecs that I require from > the b-matrix of each participant, but as he is unfamiliar with BEDPOSTX/PROBTRACKX and had > originally assumed that the list of standard gradient directions he had provided me with were > equivalent to the bvecs, he is unsure exactly what info needs to be specified in the bvec files that > I provide with each bval/nifti image. > > So....could you please advise on what info the bvec file should actually correspond to or reflect, > as this will help us to work out exactly what we need to do to derive subject-specific bvecs from > the b-matrix that I have for each of my participants? > > Many thanks in advance! > > > Mark Postans > Cardiff University Brain Research Imaging Center (CUBRIC) >