Dear FSL Experts,
We are collecting multicenter DTI data. We are using two identical DTI repetitions with opposite phase encoding directions (A>P and P>A) so that we can do topup in FSL to correct for susceptibility-induced distortions.
However, we have noticed that on several Siemens scanners that we use, two identical repetitions have slightly different bvec values. For example, a vector would be (-0,61672 0,536649 -0,5759) in the A>P repetition and (-0,62927 0,521621 -0,57613) in the P>A repetition. These differences are typically not greater than .03 and they mostly affect x and y, not z.
I would really appreciate advice on where this difference comes from and how to deal with it. I guess it is not correct to do applytopup with the slr option, averaging two repetitions into one, since then we don't know which bvec file to take into further analysis. I have tried applytopup with the jac option (correcting the two repetitions separately and then concatenating them). But I am not sure if that's a good solution, because in our good datasets (the ones that have identical bvecs in A>P and P>A repetitions), slr seemed superior to jac.
Any help is greatly appreciated!
Best regards,
Svetlana Malyutina, Ph. D.
Research Fellow, Neurolinguistics Laboratory,
National Research University Higher School of Economics, Moscow
|