Hi,
I'm not sure about #1, I know it's different than the way that Siemens
scanners measure them, which requires flipping one of the axses (reversing
the sign of the row). I just correct the bvecs based on how the diffusion
lines should look in particular anatomical regions, e.g.
http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0709&L=FSL&P=R29794&I=-3
#2 While that might be true for things like having the signs reversed for a
whole row, it definitely wouldn't be true if you randomly rearranged the
gradients (which I assume you don't mean).
#3 Not 100% sure about this one, but so long as the FA is the same, the TBSS
should be the same, as it does not depend on the orientation of the
diffusion, just it's magnitude.
#4 Tractography will not work properly unless your bvecs produce diffusion
directions that agree with the anatomy. I'm not sure why MRIConvert
produces different bvecs for the same sequence, but we typically use the
bvecs file we get from our MR physicist.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Sheu
Sent: Thursday, December 27, 2007 8:19 PM
To: [log in to unmask]
Subject: [FSL] Bvecs Questions
Hello:
Four questions about bvecs.
1. Does FSL handle or take bvec files that are measured against the magnet
coordinates (diffusion gradients in the XYZ directions) or measured
relative to the imaging plane (diffusion gradients in the RL AP FH
directions when the plane is oblique).
2. Is it true that FA values should be the same theoretically in
regardless of the bvec files that we put in?
3. Does the accuracy of TBSS analysis rely on the correctness of bvecs
(all images were taken in the transverse plane)?
4. Does the accuracy of tractography rely on the correctness of bvecs?
(all images were taken in the transverse plane)?
The reason I ask this question is because I got 10 different versions of
bvecs for each subject by using MRIConvert with our early subjects.
(However, it shows consistency with more recent subjects) I do not know
if it is because the scanner software problem or MRIConvert could not
extract the correct value. I applied the same bvec file (in XYZ
direction) retrieved from our imaging physicist to all subjects and the
DTI images look fine, but I am still worried.
Our imaging physicist told me that the FA value should be in theory the
same in regardless of the bvec files. However, I am curious and would
like to get a second opinion.
Kindly advise. Thank you.
Sheu
|