Catherine:
You shouldn't get this error. The only problem you might see going
from SPM to AFNI at this stage should be possibly the brain
oriented wrong.
I don't know whether any of your preprocessing steps might have
made the brain image look less brainlike.
to3d should work for time series data. It should also work if you've
already got statistical volumes, but 3dcopy might be easier for these.
I'd be glad to look at your data if you'd like to upload a volume. I'd
definitely need to see the SPM volume you're trying to import, then
maybe your original SPM volume (pre-pre-processing) and your
AFNI volume you generated that doesn't look brainlike. If you
don't upload an AFNI volume, what to3d command did you use?
rich