Hi everyone, I am using dwifslpreproc to correct data from the HCP Aging project. Unfortunately this command takes ages on this data, I think particularly because of topup. Do you know of any way to speed this up? As far as I can see it is also only using one core.
Alternatively, do you know of a different tool that coul be used?
Tahnk you!
Peter
Hi Peter,
Yes, FSL Topup is single-threaded. If you are processing a large cohort, you can of course run a bunch of subjects in parallel.
In my experience, the largest factor affecting Topup’s run time is the number of b=0 images. You can get a significant speed-up if you select a subset of the “best” images. For example, in the Developing HCP I used a filter to select 4 (no. PEs) x 2 = 8 b=0 volumes (out of 20) that were least affected by motion or slice dropout.
Hope this helps, and equally curious about alternatives.
Cheers,
Daan
Thank you Daan, I will try to get rid of some of the b0 images.