Rick and Ziad,
Thanks for the great information. I wondered if just converting to NIfTI would avoid this issue.
With regard to the issue of the +tlrc and which normalized space my data is in, as part of the processing protocol here, FSL's FLIRT is used to normalize, using the MNI152T1 template resampled to a isotropic 4X4X4 mm voxel size.
Thanks again.