Hi Daniel,
Thanks for explaining about the oblique data set sitting in a cardinal grid....this makes alot of sense now. The T1 after running 3dAllineate looks like it is in good alignment with the EPI. When I deobliqued the aligned image is when things looked severely out of alignment (which makes sense given what you said that I should not have to deoblique).

Just curious for 3dWarp -card2oblique would this command work even if the original T1 is also oblique (10.44 deg) at a different angle than the EPI (12.14 deg) image and not necessarily cardinal in nature? If so, then I assume the T1 would be rewritten to be in the same angle as the EPI scan (but not necessarily aligned depending on the original alignment between the two scans), and then I can run Freesurfer on the T1 and then use bbregister or 3dAllineate to register the T1 to the last timepoint of the EPI (since we discard first few timepoints for saturation effects). Do you think this would take care of the warnings?

Secondly, I want to make a generalized pipeline to handle datasets that are not taken at an oblique angle. Would keeping this step (3dWarp -card2oblique) be valid if all datasets are taken with no angles and just produce a T1 file which is exactly the same as the input, if the EPI is not aquired with oblique angles, or would this step still intoduce some type of interpolation etc to the outputl file?

Thanks,
Ajay