Hi, I googled this issue on adwarp vs. @auto_tlrc and I'd like to follow up on this post. A few questions, if you don't mind:

-When it comes to transforming an epi +orig dataset into +tlrc space, are there any advantages to using adwarp vs @auto_tlrc? I know from the posts above that adwarp does linear transformation while auto_tlrc does 5th order polynomial (or other transformations, based on your -rmode input), but is there any particular reason you would want to choose one over the other?

-I get basically the same cluster sizes/locations at the same thresholds after transforming my data using either method, but the transformed data from @auto_tlrc is bigger, with more voxels included outside of the brain. I linked images at the bottom of this message (sorry, the "attachments" option didn't work for me!)

Any idea why this might be? When I do parametric analyses across subjects, these "halo" voxels get masked out by using a group mask.

How I got this data:
1. orig anat was transformed into tlrc space using "@auto_tlrc -base TT_N27+tlrc -input anat_final.s01+orig -no_ss"
2a. for @auto_tlrc transformation of epi data: "@auto_tlrc -apar anat_final.s01+tlrc -input s01.epi+orig -no_ss -dxyz 3"
2b. for adwarp transformation of epi data: "adwarp -apar anat_final.s01+tlrc -dpar s01.epi+orig -dxyz 3"

Thanks in advance,
Lisa

https://www.dropbox.com/s/guzkxmmcozypml2/ax_adwarp.jpg
https://www.dropbox.com/s/fchck7nqx4e32j8/ax_autotlrc.jpg



Edited 2 time(s). Last edit at 11/21/2014 05:02PM by Daniel Glen.