search

yeatmanlab / AFQ

Automated Fiber Quantification
74 stars 52 forks source link

Making mrtrix3 comments more generic #33

Closed grotheer closed 5 years ago

grotheer commented 5 years ago

On this branch, I edited the AFQ_mrtrixInit function, so that it is compatible with ACT even outside of multi-shell acquisition. Specifically:

1) AFQ_mrtrixInit now allows the user to compute the 5tt image outside of multi-shell, as this is needed for ACT as well 2) AFQ_mrtrixInit now allows the user to specify the anatFolder that contains the aparc+aseg.mif (usually FREESUFER_HOME/subject/mri), rather than having to copy this file to session 3) AFQ_mrtrix_5ttgen now also computes the gmwmi.mif, which is needed for ACT and only takes seconds to create 4) AFQ_mrtrix_build_files was edited so that it includes all files created in accordance with the edits above

jyeatman commented 5 years ago

@grotheer I made a bunch of edits to this. Partially I wanted to make sure it still worked properly when not using 5tt. The way it was written it crashed when the freesurfer anat wasn't available. Secondarily I changed some of the defaults by merging another open pull reqest. I am slightly worried that this might cause some issues. Can you re-check on your data? Also, looking at line 355 of AFQ_Create, does this call need to also include the potential for the 5tt files? I am not using 5tt so that is something for you to check

grotheer commented 5 years ago

@jyeatman I ran a test subject through my pipeline and the output looks good. From my side this would be good to go. However, I am not using AFQ_Create. Rather, I am calling in AFQ_mrtrixInit directly. Is does seem like AFQ create might also need to be edited, if users want to use it to run ACT outside of multishell. The code won't crash as it is, though, and this may be a very specific scenario which may or may not apply to other user. I'm not very familiar with the inner workings of AFQ_Create, so I'm hesitant to touch it.