afni.nimh.nih.gov

Where does the 6 come from in "3dClustSim -fwhm 6"?
The initial 3dFWHMx command should output FWHM
parameters along with the ACF ones, and I would expect
it to be less than 6.

That suggests that 6 was your applied blur, in which case
3dFWHMx is giving an underestimate of the blur. Please
add a mask to your 3dFWHMx command, so that it is
not computed outside the brain. It might be good to
follow an afni_proc.py example, for example,
AFNI_data6/FT_analysis/s15.proc.FT.uber.

- rick