Hello all,
Our lab is interested in comparing the effects of different registration methods at the second level. To do this, we have scripted our own code to create output directories identical to that of FSL. We noticed that featregapply seems to be able to tell if registration has been run, and hoped that by creating our own "reg_standard" folder and copying in our registered images using the same names as the FSL default output, this would trick the command into not running for the second level.
However, it somehow seems to know that it did not in fact run on these folders, and attempts to apply registration. This errors out, as we don't have the registration matrices in the correct format for FSL.
Does anyone know if there's some sort of flag hidden somewhere in one of the files that we can set so featregapply will not attempt to run on the second level? I ran it on a tru fsl directory and looked to see if any files were modified by it, but other than the log and the reg_standard directory, everything seemed untouched.
One additional thing: Our script doesn't output all the usual files for the "reg" folder, which FSL looks for to make sure registration did run. To trick it, we just created empty dummy files in the reg directory. The Dummy files we use are:
example_func2standard.mat
example_func2standard.nii.gz
example_func2standard_warp.nii.gz
highres2example_func.mat
highres2highres_jac.nii.gz
highres2standard_linear.nii.gz
highres2standard.mat
highres2standard.nii.gz
highres2standard_warp.nii.gz
highres_head_to_standard_head.log
standard2example_func.mat
standard2highres.mat
If somehow featregapply looks at these, we could probably supply them in their real form.
(FSL version 5.0.7)
|