Partial images can be particularly difficult. If the data starts out this far apart, it seems there is something off in the coordinates. Try "-giant_move" to overcome that and allow movement a long distance. Cost functions like lpc+ZZ, lpc and nmi may be used too (-cost ...). There are options for partial coverage, but that assumes movement mostly within the planes of the acquisition, and that doesn't appear to be the case. I would remove the deobliquing step except as a test to check for appropriate coordinates and use the original data to avoid interpolating the EPI data twice.