Hi Shaunna,
The AFNI folks probably are also getting ready for the Neuroscience conference, but another approach you might take is to apply the .mat file and write the functionals out in structural grid using SPM. You might try searching the SPM mailing list archives for reorient.m. I'm thinking this might write r*.img versions of the volumes, which can be imported directly into AFNI with 3dANALYZEtoAFNI (make sure you use -orient flag).
Donna Hanlon