 |
 |
Thanks again, Anderson! I think I was trying to do the latter. So I do have to do everything in native space for the first round, including segmentation? And then plug it into Feat again, using the original bold image? So I wouldn't need to do anything with featregapply or put everything in standard space? If I wanted to do the former, how would I go about that? Which is better? I've also tried to do the dual_regression (after editing line 162 as advised). At first I tried fslmerge -t then fslmaths -Tmean as advised on the Glm wiki for multiple runs in randomise. But then the output is a 3d image that won't work with dual_regression - all the outputs are blank. All of my scans have 184 volumes, but some subject have more runs than others (each subject has between 1-5 runs). Do I need to have them all have the same number of volumes to plug into randomise? Or can I just fslmerge -t without fslmaths? Otherwise, what's the best way to get a 4d image with 184 volumes for an averaged dataset for 1 subject? Should I fslsplit each run and then fslmerge/fslmaths with the other runs and then fslmerge again? Thanks for all your help so far!