Amelia,
If you really wanted to, you could split out the b0s from the DTI volume and
motion correct them first with rigid body registration (flirt with -dof 6,
affine registration is not needed because eddy current effects will be the
same for all of the b0 volumes), average them, and then merge the average
back to the front of the diffusion volume and run eddy_correct, however for
most datasets this would be overkill. If your data have fairly low SNR,
(for example high resolution in vivo monkey data), this kind of technique
can be helpful, however for standard human data it is fine to just use
eddy_correct in the default way.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of versacea
Sent: Friday, August 01, 2008 10:22 AM
To: [log in to unmask]
Subject: [FSL] multiple B0s
Dear FSL experts.
I have a new DTI dataset acquisition with 61 directions plus 7 b0s (each
about every 9 directions: 0, 10, 20, 30 40, 50, 60). I was wondering if
there is any way to enter the all 7 B0s in the eddy current correction.
For example:
if generally:
eddy_current dti#.nii.gz data.nii.gz 0
then, something like:
eddy_current dti.nii.gz data.nii.gz 0 10 20 30 40 50 60
Probably this is wrong, but must be a way to perform an optimized
correction. Any suggestion how to better pre-process my data would be really
appreciate!!
Thank you very much indeed
Best Regards
Amelia
|