Irregular FODs being generated

We have recently upgraded our scan sequences to 32 channel multishell dwi with bvalues of 3,000.

However, I am running into an issue with strange FODs being created (see screenshot below).

I have run the usual preprocessing and processing steps (dwidenoise, dwifslpreproc, dwi2mask,dwi biascorrect, intensity normalisation, dwi2response with dhollander, upsampling (mrgrid regrid -vox 1.3), dwi2mask)

Then I ran dwi2fod with msmt_csd and the FODs were generated. My question is, is there another step that should be taken in order to properly generate the FODs?

Thanks for your help!

Welcome Aaron!

There’s definitely something unusual with your data; it’s not a matter of inserting a step to turn your FODs from “bad ones” into “good ones”. Your data need to be checked at each step of the pipeline to make sure that things make sense. I would also recommend running dwigradcheck to see if it indicates a potential underlying issue with your gradient table.