I tried both -big_move and -cmass cmass and neither seems to do a good job. I'm going to upload some data -- can you take a look at it and try to see if you see something I'm doing wrong or if I can do something else?
Here's the version of the script I ran with this data:
align_epi_anat.py \
-big_move \
-epi2anat \
-anat 16001a_anat+orig \
-epi 16001x_fa_do+orig \
-epi_base 91 \
-master_epi 16001x_fa_do+orig \
-volreg_opts -Fourier -1Dfile 16001x_fa_motion.1D \
16001a is the anatomical from session 1. 16001x is one of the functionals from session 2 that has been deobliqued with 3dWarp.