Posted by Gary Strangman on November 13, 2002 at 14:47:56:
I'm afraid that the AFNI doc, FAQ and message board haven't been able to answer a volume-registration- and anatomy-parent-related problem I've encountered. I have a set of subjects, each of whom are scanned on 2 separate days. For each subject, I believe I followed a reasonable procedure for motion correcting and co-registering across sessions ...
1) motion correct all functional runs in session1 (3dvolreg on epi files)
2) co-register session1 functionals to session1 anatomy (3drefit on anat)
3) co-register session2 functionals to session2 anatomy (3drefit on anat)
4) co-register session2 anatomy to session1 anatomy (3dvolreg -twopass)
5) motion correct and shift functional runs in session2 to the session1
frame of reference (3dvolreg -rotparent -gridparent)
I'm happy with the result ... each session has both anatomy and functional datasets that are co-registered with each other and with the corresponding datasets in the other session. The problem is that now I'm not allowed to set the transformation markers for the transformed session2 anatomy dataset. That is, when I select the session2 anatomy that's in the session1 space and click "Define Markers", the "Allow Edits" button is grayed out. From all the doc, it seems that this is anatomy-parent related, but I tried a number of (what I thought were) rational anatomy parent changes to no avail. It's not clear from the 3dinfo and HEAD files how the parents are set (or inferred) when the transformed session2 anatomy BRIK is created (step 4 above). I can set markers for the original session2 anatomy, and the transformed session2 anatomy gets transformed with it (i.e., inappropriately). I've even tried making the transformed session2 anatomy dataset start with 'aa', so it would be found first, but that didn't help. It's probably quite obvious what I'm missing here ... unfortunately not so to me.
Gary Strangman