Hi, all.

I suspect the problem is similar to one that I am experiencing with CONN/SPM12, in which some participants longitudinal scan data is not unwarping and realigning properly during the early preprocessing stage.  

I haven't tested it yet, but I suspect for some participant imaging data that the 0,0,0 centering function in CONN is impeding proper registration over timepoints?  

I thought initially the problem might be where the origin is located when the images come off our scanner, so I reoriented the origin to the AC along the AC-PC line and then coregistered (estimate only) the corrected origin from the baseline scan to the subsequent scans for a participant.  Everything looked great and in reasonable alignment prior to starting CONN/SPM12 preprocessing, but I continue to receive clearly problematic results for some participants.  The fact that it doesn't screw up for all participants has me stumped.  

See attached as a gross example of the problem with realignment over scans/time.

Any thoughts or possible solutions?

Warm regards,

Jeff