Dwidenoise on 1b0+1DW reversed phase encoding direction volume

Hi MRtrx3 experts,

I’m writing to ask if you would recommend concatenate a 1b0+1DW revered PE direction volume and the 4 b0+38DW normal PE direction volume together to do dwidenoise together?

These two volumes are in separate scans next to each other for TopUp correction. However, because of the opposite phase encoding direction, the susceptibility-induced distortions are in the opposite direction, with one being stretched and the other condensed)

The dilemma is that: If I do dwidenoise separately on two volumes, 1b0+1DW won’t have enough information for noise estimation. If I concatenate them, the image intensity is different per voxel in the 4th dimension between the two scans considering the opposite susceptibility-induced distortion.

Thank you very much for your help!

All the best,


This issue has come up before, and I’m not sure we have a good answer to it… I think this is going to require some investigation before we can settle the matter, but someone will need to invest the time to do this, and that’s in very short supply these days.

Here are a few posts where the issue has been discussed:

1 Like