Fellow SPMers,
I'm trying to pull multiple image types from a longitudinal data collection effort [e.g., each subject w/ 3 timepoints of fMRI/BOLD, ASL, DTI(FA/ADC), FLAIR, & T1] into registration for eventual voxel-wise, repeated-measures analyses. Earlier we were attempting this using SPM5 and conventional methods, but I'd like to investigate how to accomplish complex higher order registration with DARTEL and SPM8. Is it simply a matter of coregistering:estimate all of the various image types with the T1 at each time point, then using DARTEL and applying the warping transforms to the other image types (some of which would not be the same voxel dimensions at the 1x1x1 T1)? Assuming this is the case, then how about bringing the time point 1, 2, and 3 image groupings into registration for each subject (e.g., Time1 > Time2 > Time3 high-dimensional warp registration)? And, if that wasn't enough, what about bringing everything after intra- and inter-time point registration into MNI space. I think I have a general idea of the logical ordering needed, but as with anything SPM the best solution is most likely to be found here amongst fellow imaging comrades. Any pointers, suggestions, sympathy, etc. would be most appreciated.
Regards,
Jeff Browndyke
Duke Univ. Med. Center
|