Hi Barbara:
1. The rotation component of the registration matrix from eddy_correct is extracted for each DWI and applied to the corresponding gradient vector.
2. dmri_pathstats is not a script, it's a C++ executable, so to edit it you have to download the source code, modify, and recompile. Alternatively, you can get the x, y, z coordinates of each voxel from the pathstats.byvoxel.txt file and get values at that voxel with something like: mri_info --voxel x y z dmri/dtifit_L3.nii.gz
a.y
On Fri, 17 Oct 2014, Barbara Kreilkamp wrote:
Dear all, Do you have suggestions/ideas for me please? Thank you, Barbara
-------- Original Message -------- Subject: TRACULA preprocessing: bvector rotation, negative eigenvalues Date: Tue, 07 Oct 2014 19:09:00 +0100 From: Barbara Kreilkamp bakk.hbg@gmail.com To: Freesurfer support list freesurfer@nmr.mgh.harvard.edu
Dear Freesurfers,
I am looking for information on how exactly the b-table gets rotated.
- It seems that eddy_correct runs once with 12 doF, and then the
bmatrix gets updated based on this registration. Would you please confirm?
- I know that my data has negative eigenvalues (most of them seem to
appear in the 'halo' around the images), and I would like to save L2, L3 voxel values within tracts in the pathstats.byvoxel.txt file. How may I edit dmri_pathstats to accommodate this? My aim is to identify MD, RD, FA, AD (L1) voxel values that were biased by a lot of noise.
Thank you very much, Best, Barbara