External Email - Use Caution
Hello
My interest is to create an aparc+aseg file with a combination of a cortical parcellation with a custom cortical atlas and the subcortical grey matter components from the standard aseg....basically i want to mask and extract signal from multiple areas in a series of pipelines in python. The commands below are called from python, but edited to make it easy to read.
About 50% of the time the code below works fine, but the other 50% it drops two cortical parcellations. Interestingly it is always the same two parcellations, both at the end of the left hemisphere list, that get dropped when making aparc+aseg. My only guess is that something about the aparc and aseg combination is causing them to be dropped.
I have read the annot files it makes and the parcellations are made in the mris_ca_label step and making the aparc+aseg. I have viewed them in freeview and they are not there when viewed in the LUT for this atlas and I have tried omittng and excluding all forms of mri_binarize -replace commands (added to try to eliminate segmentations of the ventricle and white atter which i am trying to mask out of a resting state image), and this doesnt make a difference in the cases where it is failing. I have tried using the cases where it worked to train a new gcs file without success. I have also rune mri_aparc2aseg with and without the ribbon option, with all varieties of aseg files...etc etc.
If there is something obviously wrong please let me know. Alternately, what is the best way to get aparc+aseg.mgz to print a list of what areas made it in there..in that case i can likely create a way to make sure the lables it is using for an adjacency matrix atleast match the rows and columns accurately..i tried using segstats but this wasnt always an accurate list as it sometimes created a mask with areas not on any LUT's i have and thus i have been unable to view them in freeview or the segstats list and figure out what is going wrong
the commands i invoke from python are below...thanks in advance
recon-all -subjid sub1 -autorecon1 -openmp 6 recon-all -subjid sub1 -autorecon2 -openmp 6 'recon-all -subjid sub1 -sphere -surfreg -jacobian_white -avgcurv -open mp6
mris_ca_label -l $SUBJECTS_DIR/sub1/label/lh.cortex.label -aseg $SUBJECTS_DIR/sub1/mri/aseg.presurf.mgz sub1 lh lh.sphere.reg $SUBJECTS_DIR/sub1/parc/lh.hcp-mmp_6p0.gcs lh.aparc.annot mris_ca_label -l $SUBJECTS_DIR/sub1/label/rh.cortex.label -aseg $SUBJECTS_DIR/sub1/mri/aseg.presurf.mgz sub1 rh rh.sphere.reg $SUBJECTS_DIR/sub1/parc/rh.hcp-mmp_6p0.gcs rh.aparc.annot recon-all -subjid sub1 -pial -cortribbon -hyporelabel -openmp 6 mri_binarize --i mri/aseg.presurf.hypos.mgz --o mri/aseg.mgz --replace 2 0 --replace 4 0 --replace 5 0 --replace 7 0 --replace 14 0 --replace 15 0 --replace 24 0 --replace 30 0 --replace 31 0 --replace 41 0 --replace 43 0 --replace 44 0 --replace 46 0 --replace 62 0 --replace 63 0 --replace 77 0 --replace 85 0 --replace 251 0 --replace 252 0 --replace 253 0 --replace 254 0 --replace 255 0') mri_aparc2aseg --s sub1 mri_binarize --i mri/aparc+aseg.mgz --o mri/aparc+aseg.mgz --replace 41 0