 |
 |
Hi, I am working with some diffusion data that was already acquired and the voxels are not isotropic. I am going to do tractography and for that reason I need to make the data isotropic before doing the pre-processing. The voxel size is: 1x1x4 I extracted the b0 (first volume) and used flirt to register the DWI to the b0 forcing it to give the result with isotropic voxels. I registered to the b0 because at this point I dont want to transform the DWI to any space, I just want it to have isotropic voxels: 3dTcat -prefix b0.nii.gz DWI.nii.gz'[0]' flirt -in DWI.nii.gz -omat DWI_resampled.mat -out DWI_resampled.nii.gz -applyisoxfm 1 -ref b0.nii.gz My question is, what do I do with the bvecs/bvals? My understanding is that if I apply any transformation to the data I have to transform the vectors too. I know about the script fdt_rotate_bvecs, but that is only to rotate the vectors after the eddy distortions, not after re-sampling. Eitherway, I looked at the code and I modified it to input my DWI_resampled.mat instead of the eddy log file. But my DWI_resampled.mat is just the identity matrix, so my bvecs look the same after rotating. When I run dtifit with the original data and then the resampled data (just for quality control, I will not use the FA/V1, I only want tractography), using the same bvecs/bvals in both dtifit, I do get much better results with the resampled data. But I want to make sure it is ok to continue using the same vectors even after I modified the DWI? I was surprised to see that dtifit worked with the original ones. Thank you, ######################################################################## To unsubscribe from the FSL list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1