Clear displacement of not coregistrated images

Hello,

we are currently working with a data set of DTI images and use MRtrix for preprocesing.
Everything went quite fine, but we stumbled across a very strong displacement in the image before coregistration.
5tt_coreg_nocoreg
As one can see in the uploaded image, the overlay of the 5tt_nocoreg is extremely off place, but the coregistered one seems fine, as well as the boundaries in the next step.
DTIboundaries
My team and I are now a bit worried about how strong this displacement is, even tough it seems to be corrected successfully.
This is our first DTI project and in all example cases the offset is much smaller.
Could someone with a bit more experience tell us if this amount of displacement is OK, or if we should have a look at the raw data or our steps before?

Thanks a lot for your help!
Joshua

Hi Joshua,

I don’t think it’ll necessarily be anything to worry about, especially if this is the only subject where this occurs. But I agree it would be best to double-check what might have happened, this is quite a large displacement, and given that it’s mostly anterior-posterior, this is unlikely to be simple subject motion (difficult to picture them lifting their head by ~3cm within the confines of the head coil). It could be that the subject needed to be taken out of the scanner and re-positioned, or it could be indicative of some issue in the processing chain.

The best place to start would be to load both the T1 & DWI data directly from DICOM into mrview, and double-check the alignment in there. If the raw DICOM data are roughly in alignment, but no longer aligned at a later stage in the pipeline, then it would be good to figure out which stage is introducing the problem and fix that.

Cheers,
Donald.

1 Like

Hi Donald,

thanks a lot for your advice.
We went back to check the original data and as one can see below, there is already this big displacement.
displace_raw
Since the offset is already part of the raw data and is fixed by the coregistration, we assume it has to do with the data acquisition, even though we were assured the subject were not re-positioned. Hence, the reason for this displacement remains unknown.

Thanks again for your advice and best wishes,
Joshua

1 Like