Hello AFNI gurus,

I have multiple oblique, partial coverage EPI datasets, that I am trying to align to a structural dataset, something that align_epi_anat.py has a very hard time doing, even with various options turned on, as suggested elsewhere on this message board. I find that doing an initial alignment of the EPI dataset using the nudge plugin in Define Datamode, usually makes align_epi_anat.py perform much better.

The only problem with this approach is that manually aligning like this produces clipping in the EPI dataset, that is, I lose part of the EPI data when I move to far from the starting position. This happens both in the AFNI viewer, and when using 3drotate, and it happens even when the -noclip option is used. I think it happens when I move off the original grid of the EPI dataset, and I have gotten around it by resampling the EPI data to the structural grid, prior to doing the manual alignment.

This strategy, however, produces a very large EPI dataset, which causes unnecessary overhead further along my analysis pipeline. Is there a way to apply six transformation parameters, like those produced by the nudge plugin, to an EPI dataset without causing this clipping? Or is there a way to reduce the dimensions of the EPI dataset following the resampling and initial alignment, getting rid of all the "unnecessary" voxels?

I realize that I should be able to accomplish the alignment using align_epi_anat.py alone, and it is possible that I could, but I have tried a wide range of different combinations of options with no luck so far. I also feel that even if align_epi_anat.py could resolve my problem, my current strategy should be something that was possible to do in AFNI.

Thanks!

Peter Kohler