I have noticed differences between MRIConvert and dcm2nii (from MRICron).
This could be the result of one of the programs rotating the bvecs to
account for an FOV rotation and the other one not doing this (this is most
likely why both are different from the bvecs from the sequence).
Differences in sign only if they are properly balanced could refer to the
bvec pointing 180 degrees in the opposite direction (which is effectively
the same). I have successfully converted Siemens 7T DICOM files with the
latest version of dcm2nii from MRICron, so you might ensure that you have
the latest version.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Jeff Eriksen
Sent: Thursday, April 01, 2010 6:54 PM
To: [log in to unmask]
Subject: [FSL] which bvecs to trust?
I am wondering which bvecs to use in our DTI processing. Per the FSL
documentation suggestions, I first tried using dcm2nii to convert DICOM to
nii, which seemed to produce usable bvec files. I soon discovered that it
could not create proper nii files using our Siemens 7T DICOM files. I then
tried MRIConvert, which has yet to fail converting any of our DICOM files
and has produced bvec viles that seems to work fine. However, I am confused
about the differences in the bvecs produced by these two programs, which are
both different from the generic bvecs our Siemens physicist has given us.
For the most part, the MRIConvert bvecs match the Siemens ones in value, but
not sign. Two or three of the components are reversed in sign. I do not see
how all three of these sets of bvecs can produce what visually look like
very similar results (FA and V1), when the apparent directions are
different.
1. What do other FSL users employ for DICOM to nii conversion?
2. Does anyone else out there use FSL for 7T DTI?
Thanks.
|