Dear FSL admins,
I have the following multishell diffusion scheme. The diffusion scheme has 5 shells 142 directions and 3b0. bmax 5000.
Scanner Siemens Prisma.
Now the diffusion directions forms a hemisphere, the phase encoding is recorded in AP. For timing issues we could only record b0's with PA .
Now to run EDDY (eddy_openmp), I have to used " --data_is_shelled flag". I am using the fsl 5.0.9 with advance patch for EDDY.
However the direction scheme being hemispherical we are not sure how to rely on this EDDY result as stated on FSL EDDY wiki page.
The other approach is to use eddy_correct. For eddy_correct we have two issue .
1. For higher b-value (b5000) doing eddy_correct the brain is getting stretched.
2. There is considerable smoothing of DWI image specially for b=5000 shell during eddy_correct.
However eddy_openmp outputs doesn't suffer from either stretching or broadening.
Please advice what is best way for us to approach eddy current correction in our present protocol?
Does using eddy_openmp is good enough?
In case we go by eddy_correct how to get about this stretching and smoothing issue mentioned above?
Thanking you
sourajit
|