 |
 |
Hi Pascal Yes that is caused by the bvecs file indicating a single direction. (note that MD is also affected by that, not just FA, and the effect on MD will depend on what the true bvecs are). First, check that the different volumes have indeed been acquired with different or the same directions, i.e. whether it is a problem with the bvecs or with the data. You can for example open the 4D file in fslview and look through the volumes to see how contrast changes. For example do regions of dark and bright signal in the corpus callosum change location?, etc. If it looks like the bvecs are wrong (lots of contrast across volumes) then you might want to talk to whoever wrote the dicom converter (e.g. Chris Rorden if you are using dcm2nii) Cheers, Saad On 10 Jun 2016, at 23:25, Pascal Tétreault <[log in to unmask]<mailto:[log in to unmask]>> wrote: Hi, I am using dtifit to generate FA maps to be able to run TBSS. The FA maps generated are all voxels with a value of 1, but the MD maps seems fine. I am using the following command line: dtifit -k data.nii.gz -o dti -m mask.nii.gz -r bvecs -b bvals My data.nii.gz is a 4D file, eddy corected, 31 volumes, 1st volume is the b0 and the others the 30 directions. My mask is a standard binarized mask with only 1 and 0. bvals = 0 1000 1000 1000 ... (30x1000) My bvecs is most likely my problem, it is a standard 3x31 values, but all the values from one line after the 0 are the same: e.g.: 0 -0.998676 -0.998676 -0.998676 -0.998676... 0 0.041493 0.041493 0.041493 0.041493... 0 -0.0304004 -0.0304004 -0.0304004 -0.0304004... I tried different dicom to nifti software to convert my dicom files, but the bvecs files are always like that. Any idea what might went wrong?! And if my bvecs file is not the problem, what can cause the all 1s FA map? Thanks a lot. Pascal Pascal Tétreault, Ph.D Postdoctoral fellow Dr. Christian Beaulieu's lab Biomedical Engineering, Faculty of Medicine University of Alberta