Dear everyone,
I would like to Dartel-normalise some data onto a pre-existing Dartel template (the IXI555 included in CAT12), however, I would prefer the final data to be close to another template image (the MNI152_T1_1mm.nii as included in FSL, seems to be the asymmetric version of the 6th gen MNI atlas). For the latter, I only have a T1 image, no GM/WM files, but I also have a T1 image for the final iteration of the Dartel template set (and I could create one based on my subjects). Accordingly, I resliced the MNI152_T1_1mm.nii to match the dimensions of the "T1 IXI Dartel" image. Then I went with a Run Dartel (existing Templates) step and forwarded the two images, resulting in a u_ Dartel flow field. To normalise my subjects' data I would then go with the Deformations utility and a composition of the individual flow fields and the just created one.
I have also thought of applying the obtained Dartel flow field from above onto the IXI Dartel template set so that it matches my target space right from the beginning, and use this adjusted set to obtain individual Dartel flow fields.
Now I would be curious to hear other opinions on pros / cons or alternative approaches (if it makes sense at all to discuss these, as they might depend on the particular data sets). E.g. for the second approach there seems to be a disadvantage, as the Dartel template set would be altered, likely resulting in some minor interpolation errors; for the first approach there might be larger interpolation errors due to the composition of the two flow fields (?) at "the end" of the preprocessing.
Best
Helmut
|