Dear Jan,
> I have a subject who moved substantially between blip up/down (AP/PA (y-axis)) acquisitions for DTI. The movement is in the z-direction, approximately 10 mm, otherwise individual acquisitions are of good quality and still include entire brain, 2 mm isotropic, same number of slices, without significant movement within each acquisition.
>
> When I perform topup and applytopup, using the b02b0.cnf file it cuts off a large part of the cerebral vertex, even though this is not missing on either acquisition.
that is quite strange and shouldn’t happen.
>
> Any advice on how to correct this properly?
>
> I noted that using topup ... -iout test.nii.gz, that in the -iout output file, the second B0 image (that is being aligned to the first one) has been moved correctly (caudally by about the 10 mm), but that the bottom slices that were there on the original have been "added" onto the top of the head.
This was a “problem” in the —iout output of earlier versions, but should have been fixed by now. Are you using the latest version of FSL/topup?
> I suspect that this move is likely accounting for some of the subsequent difficulty. When I "pad the images" using fslroi and add a few empty slices on the top of the head, the problem is solved, e.g. the slices are still added on top of the image, but now the vertex of the corrected data.nii.gz file after running the entire sequence of steps does not have the significant error and I have a full brain to work with.
>
> I noted that this also fixes the times when there is less pronounced (but present) movement between the 2 acquisitions, in those instances a single slice or so on the bottom gets cut half way off.
In general it is the case that data has to be present in both acquisitions for it to be retained in the final (or iout) output, and that will always be the case. From your description above it sounds like you think that in some cases data is present in both acquisitions, but that it is still not present in the final output. Is that right?
If so, could you maybe send me a screenshot showing this? Please send it directly to me so as not to send big files to everyone on the mailbase.
Jesper
>
> Is there a more elegant way to fix this, am I missing a step in the topup process that normally would take care of this?
>
> Jan
|