Dear Anastasia & Michael,
Thank u for your answers (I attached my last email and your answers at the end of this email.
I still want to specify my question. If I only want to extract FA values using the following script (see script below), and I know that an angle was applied during the dMRI scans, should I use the rotated gradient matrix ?
When I first analyzed 20 subjects with the gradient matrix that is published for a T3 Philips magnet (I also read the gradients from the dicom header - they were the same) I got reasonable results - but now I don't know what to think of them... __________________________________________________________________________________
source ~/Desktop/fs_script
dt_recon --i /usr/local/freesurfer/subjects/Anatomy0207/DTI/I160.dcm --s Anatomy0207 --o /usr/local/freesurfer/subjects/Anatomy0207/DTI --b /usr/local/freesurfer/subjects/Anatomy0207/DTI/bvals.bval /usr/local/freesurfer/subjects/Anatomy0207/DTI/bvecs.bvec
To check results, run:
tkregister2 --mov /usr/local/freesurfer/subjects/Anatomy0207/DTI/lowb.nii --reg /usr/local/freesurfer/subjects/Anatomy0207/DTI/register.dat --surf
tkregister2 --s fsaverage --surf white --reg /usr/local/freesurfer/subjects/Anatomy0207/DTI/mask-tal.nii.reg --mov /usr/local/freesurfer/subjects/Anatomy0207/DTI/mask-tal.nii
tkregister2 --s fsaverage --surf white --reg /usr/local/freesurfer/subjects/Anatomy0207/DTI/fa-tal.nii.reg --mov /usr/local/freesurfer/subjects/Anatomy0207/DTI/fa-tal.nii
freeview -v /usr/local/freesurfer/subjects/Anatomy0207/DTI/fa.nii /usr/local/freesurfer/subjects/Anatomy0207/DTI/adc.nii
freeview -v /usr/local/freesurfer/subjects/Anatomy0207/mri/brain.mgz /usr/local/freesurfer/subjects/Anatomy0207/DTI/fa.nii:reg=register.dat:colormap=heat:heatscale=.2,.2,1
tkmedit Anatomy0207 orig.mgz -aux brain.mgz -seg wmparc.mgz -reg /usr/local/freesurfer/subjects/Anatomy0207/DTI/register.dat -overlay /usr/local/freesurfer/subjects/Anatomy0207/DTI/fa.nii -fthresh 0.2 -fmax 1
mri_vol2vol --mov /usr/local/freesurfer/subjects/Anatomy0207/DTI/lowb.nii \ --targ /usr/local/freesurfer/subjects/Anatomy0207/mri/wmparc.mgz \ --inv --interp nearest --o /usr/local/freesurfer/subjects/Anatomy0207/mri/wmparc2diff.mgz \ --reg /usr/local/freesurfer/subjects/Anatomy0207/DTI/register.dat --no-save-reg
mri_vol2vol --mov /usr/local/freesurfer/subjects/Anatomy0207/DTI/lowb.nii \ --targ /usr/local/freesurfer/subjects/Anatomy0207/mri/aparc+aseg.mgz \ --inv --interp nearest --o /usr/local/freesurfer/subjects/Anatomy0207/mri/aparc+aseg2diff.mgz \ --reg /usr/local/freesurfer/subjects/Anatomy0207/DTI/register.dat --no-save-reg
freeview -v /usr/local/freesurfer/subjects/Anatomy0207/DTI/fa.nii \
/usr/local/freesurfer/subjects/Anatomy0207/mri/aparc+aseg2diff.mgz:colormap=lut
mri_mask /usr/local/freesurfer/subjects/Anatomy0207/DTI/fa.nii \ /usr/local/freesurfer/subjects/Anatomy0207/mri/wmparc2diff.mgz \ /usr/local/freesurfer/subjects/Anatomy0207/DTI/fa-masked.mgz
freeview -v /usr/local/freesurfer/subjects/Anatomy0207/DTI/fa-masked.mgz /usr/local/freesurfer/subjects/Anatomy0207/mri/aparc+aseg2diff.mgz:colormap=lut
mri_segstats --seg /usr/local/freesurfer/subjects/Anatomy0207/mri/wmparc2diff.mgz --ctab $FREESURFER_HOME/FreeSurferColorLUT.txt --i /usr/local/freesurfer/subjects/Anatomy0207/DTI/fa-masked.mgz --sum /usr/local/freesurfer/subjects/Anatomy0207/stats/all_stats_fa-masked
__________________________________________________________________________________________________________________
It'd be nice if things were that simple, but!
Although the pre-processing is done with FSL tools, tractography is done with locally written code. Some gradient sets may appear to work with FSL, but they won't work for us. For example, some dicom converters may perform a L-R flip as preparation for feeding the gradient vectors into FSL. We need the unflipped vectors because, instead of flipping vectors, we convert both vectors and images to LAS orientation. This ensures that they work both for FSL and for us, and that you can view things correctly in both freeview and fslview, without the subject looking like it's hanging upside down from its bat cave or some such awkward arrangement.
In the end, the best way of knowing that the gradient table is correct is to run one example subject through and view the eigenvectors of the tensors as lines to see if they're pointing in the right direction. (Even though the tensors themselves aren't used for tractography, they're good for troubleshooting.)
On Mon, 21 Jul 2014, Harms, Michael wrote:
FS uses FSL tools under the hood, so you want to use the bvecs that are rotated into the axes of the scan.
cheers, -MH
--? Michael Harms, Ph.D.
Conte Center for the Neuroscience of Mental Disorders Washington University School of Medicine Department of Psychiatry, Box 8134 660 South Euclid Ave. Tel: 314-747-6173 St. Louis, MO ?63110 Email: mharms@wustl.edu
From: Rotem Saar saar.rotem@gmail.com Reply-To: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Date: Monday, July 21, 2014 1:54 PM To: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] Diffusion analysis qs - gradient matrix
Dear freesurfer experts,
For diffusion analysis in freesurfer - I want to know which gradient
matrix
freesurfer expects.
The thing is - when I look at the gradient matrix created using a MATLAB script that reads the directions from the dicom header - I get one table, but the bvecs file created in MRIconvert (for converting dicoms to nifti) looks different.
This is the situation when an angle is applied during the acquisition of
the
dMRI scans, which means that each subject has its own gradient matrix.
So my question is which table should I use for diffusion analysis in freesurfer? The one that was created from the dicom header or the one created for NIFTI format using MRIconvert ??
Thanks
Rotem Saar-Ashkenazy
Department of Brain and Cognitive Science Ben Gurion University of the Negev, Beer-Sheva, 84105
School of Social Work Ashkelon Academic College, Ashkelon, 78211
Israel