Hi Rachel,

Given how much overlap there is, it seems unlikely that there
was any problem with @auto_tlrc. The alignment is probably the
same in +orig space. Keep in mind that @auto_tlrc does nothing
to align the EPI and anatomy, it aligns the anatomy with an
anatomical template, and then performs the same transformation
to the EPI (if you ask it to).

So if the anat and EPI were in somewhat poor alignment in orig
space, then they will still be in a somewhat poor alignment in
tlrc space.

----

If the anatomical dataset was acquired after the EPI data, make
sure to use one of the last EPI volumes as a volreg registration
base, rather than the first. But if the EPI and anat were
acquired close in time, then the subject probably moved between
those scans.

In such a case, consider using align_epi_anat.py. Note that
afni_proc.py can also do this for you (add it to the processing
script) via the 'align' block, and maybe the -volreg_align_e2a
option.

- rick