> I'm working with a dataset collected on a Phillips 3T scanner, and I'm currently putting together a diffusion/TBSS pipeline. The acquisition protocol has the B0 volume collected last out of the 60 volumes (instead of first, as listed in the TOPUP userguide example). Other coworkers of mine have written scripts to split and recombine the volumes with the B0 at the front, so as to fully follow the FSL practicals and userguide, and then written further scripts to change the bvecs and bvals so the 0s are at the front of the rows vs the end.
> Instead of shuffling everything around and scripting, is it an acceptable practice to extract the B0 volume using "fslroi dwidata nodif 60 1" instead of "fslroi dwidata nodif 0 1"? If I understand correctly, this would extract the B0 volume and would negate the need for to shuffle the volumes, and change the bvecs/bvals files.
I think I would do what your co-workers do. In principle there is a way to select a different reference volume than the first, but in practice that option has been tested a lot less so I would play it safe.
I should also say that if there is a way to instead make sure that the b0 is acquired first, that would be even better. Is there not a text file with b-values and b-vecs that can be edited on Philips systems? The ideal acquisition for running topup and eddy would be 3-b0-PE1 + b0-PE-1 dwi dwi dwi ...
Not also that if you only want to run FA-TBSS it is probably ok with a single b0. But if you want to complement that with for example MD-TBSS it would be more optimal to acquire more b0 (for example every 10 or 15 volumes).
> Please advise if this would work, or if I am heading astray.