Hi Gabor - If you provide the DWI dicoms directly, tracula uses mri_convert to convert them to nifti, and mri_convert doesn't read directions from the header, it looks them up from: $FREESURFER_HOME/diffusion/mgh-dti-seqpack/

This should be fine assuming you use the same standard Siemens diffusion protocol that we use at MGH. If you're using anything different than that (including if you concatenate two DWI series into one), then you should set bvecfile, bvalfile, nb0. For more info see: http://www.freesurfer.net/fswiki/dmrirc

If you have doubts about your bvecs is you can check the primary eigenvectors from dtifit (dmri/dtifit_V1.nii.gz) to make sure they're pointing the way you'd expect.

Hope this helps, a.y

On Sat, 31 Mar 2012, Gabor Perlaki wrote:

Dear Freesurfers,

I have DTI images acquired on a Siemens 3T TIM Trio (software version VB17A).

1, Is TRACULA able to extract the correct b-values and b-vectors from the DICOM headers of Siemens IMA files or do I need to specify them?

2, I'm not sure whether the gradient directions are stored in the image- or in the scanner coordinate system in the headers of Siemens IMA files. Does anybody have any experience with this version of Syngo software (VB17A)? Does TRACULA need the b-vecs in the image or in the scanner coordinate system? Is it able to reorient the bvecs if they are stored in the wrong coordinate system in the header?

3, My DTI protocol (30 diffusion weighted directions and 1 b0) was repeated twice (in separate series, the second right after the first one, same orientation). Because I'd like to correct for eddy current and movement I don't want to average them on the scanner. Is it possible to put the 62 volume (30+30 diffusion weighted directions and 2 b0) into the same TRACULA evaluation in dicom format or do I need to convert them into a single niftii file?   Thanks a lot, Gabor