Hi,
this might be a stupid question, but I've been at this point before and just don't get my head around it. But I think the answer is easy.
- We are working on animal data, so there is no MNI space.
- We use the dartel pipeline (1. old segment, 2. dartel import, 3. dartel warp) to create a tissue map template and flow-fields for each dataset.
- Now I want to normalize an original unsegmented dataset into the space of this tissue map template.
I see at least two posibilities:
1. Use dartel 'normalize to MNI Space' with the original images and the flow fields. Since I do not want this additional affine transformation to MNI Space I do not supply a template.
2. Use dartel 'Create Warped' - supply u-field and original image
3. Probably some fancy stuff with the utils/deformations tools
So if somebody could shed light on the following questions, I would be most gratefull:
- Are all these methods (especially 1+2) legit?
- I tested 1 and 2 and they give me simmilar but subtly different results.
- When I use 'dartel import' I have to supply the sn.mat files from the old segmentation (which contain a normalization). Then Dartel import creates resliced r...nii files. Is this transformation somehow included in the u-files that I later get? And if not, should I first run a Dartel Import on any dataset I want to normalize and then any of the methods above?
- Just a question for my understanding with human data: If dartel import applies the normalization parameters from the old-segment (sn.mat) and assuming this segmentation was done with a tissue-prior in MNI-space. Why would I need this addtional affine transformation to MNI-Space which is part of Dartel-> 'Normalize to MNI Space'?
Thank you very much for any hints!!!
Cheers,
Wolfgang
|