Ah ha! I think you have caught it for me, Peter!
It appears I have NOT warped the masks to standard space... and in fact, I am wondering whether I have used the wrong data to create the group mask?
I have used the results from my data pre-processing steps; that is, the functional data (after having corrected it for motion, normalized signal, and concatenated the runs then run 3dAutomask on that). However, should I be using the standardized anatomical data? I am unclear on this; searching for how to create group masks for multiple comparison correction I have come across Gang's site (http://afni.nimh.nih.gov/sscc/gangc/mcc.html), which appears to me to suggest that a mask created for 3dDeconvlove could be used (so that would be from my functional data). I was also directed to your site, Peter (http://blog.cogneurostats.com/?p=196), which indicates to use the anatomical data. However, I have recently found a reply from Rick (http://afni.nimh.nih.gov/afni/community/board/read.php?1,64326,64327#msg-64327) indicating either dataset could be used (i.e., the EPI data that's fed into 3dDeconvolve, or the anatomical data). Could you confirm whether I am, indeed, OK to use the existing functional masks that I have?
If yes, to be clear, to correctly create the average group mask to be used for multiple comparison correction the steps are (and I'm going with your first suggestion simply because I have used it before during spatial normilization for my individual analysis):
@auto_tlrc –apar run1-T1_ref_avg152T1+tlrc –input mask_sub1+orig –suffix _avg152T1 –OK_maxite –dx 2 –dy 2 –dz 2
→ Output: mask_sub1_avg152T1+tlrc.{HEAD|BRIK}
(where -apar is the dataset created from Usage 1 of @auto_tlrc)
3dMean –prefix avg_group_mask mask_sub1_avg152T1+tlrc mask_sub2_avg152T1+tlrc mask_sub3_avg152T1+tlrc…
→ Output: avg_group_mask+orig.{HEAD|BRIK}
Then this should give me the correct, spatially normilized, group mask to use for subsequent steps?
