With all due respect, I have to disagree.
Usually there's no problem realigning sessions from different days; and while there are definitely intrinsic difficulties---e.g., EPI distortion is very unlikely to be identical on both days---not realigning the sessions doesn't help in that regard. If T1-weighted struturals are used, the researcher can always align between sessions by aligning the T1s and using the intermodal coregistration routine to align the EPIs to the T1s.
Furthermore, if the researcher wants to bring only one image per subject to the group level, putting different days/sessions in different design matrices will then necessitate extra work (e.g. using ImCalc) or statistical care (paired t-test). While none of that is truly difficult, it amounts to extra bookkeeping, and I don't see benefits that justifies the cost.
|