Hi All,
I am working on a DTI analysis with some small brainstem seeds that are defined in MNI space. My question is whether it would be preferable to transform the seeds before submission to probtrackx or within it. I will have to calculate a good warp from MNI to FA space for either case to work well. But, I have tried doing the transformation and interpolation myself and while the placement of the seeds look okay, the interpolation leads to inconsistent ROI across subjects, leading to intersubject heterogeneity. This is primarily due to the fact that the seeds are: 1) small, 2) right next to each other, and 3) being transformed from 1mm3 to 2.5mm3. I know that I can alternatively provide probtrackx with the seeds and targets in MNI space, and a MNI-to-FA warp and invwarp. My worry is that in using this method my heterogeneity problems will persist but I will be unable to detect them at this point. Alternatively, probtrackx may utilize a method which aviods this issue entirely. So my question is:
Should I provide probtrackx with the warps and ROI in MNI and trust it?
OR
Should I provide probtrakx with the ROI in FA space (even though they're not great), but at least I know exactly what I'm getting?
Thanks,
Bryson
|