# This week's homework is simply to use FreeSurfer to # reconstruct a brain, and send your TA a screenshot # of something pretty when the reconstruction finishes. # To work on your brain, get an interactive job going # on compute: ssh compute qsub -IX -q instruction -l walltime=24:00:00 # the extra -l walltime... part is so you can stay # logged in long enough for the long job to run. # Change to the /pkg/classes/psy5065 # directory (by typing cd and then the path). cd /pkg/classes/psy5065 # Change your shell type from bash to tcsh so things work: tcsh # Execute the FreeSurfer set-up file by typing: source SetUpFreeSurferPSY5065.csh # (That sets various paths and defaults for your # shell so the reconstructed data will go to the # right place, and FreeSurfer will know where to # find the libaries it needs.) # Make sure that your output directory is set correctly echo $SUBJECTS_DIR # if the answer is not /pkg/classes/psy5065/Anatomy # there's been an error along the way and nothing else # will work. Contact Cheryl or Erin for help. # Then start the reconstruction job for your brain: recon-all -subject [uniqueID] -i [path] -all # The things in square brackets are variables unique # to you: # uniqueID is a non-identifying name for the output data # path is the path to the first .dcm of your T1 anatomy # ... all our data is in /pkg/classes/psy5065/Data, so # you'll have to look in their to find the scan # you want to reconstruct. # Then wait 8-24 hours. # Then go back and follow instructions found at # freesurfer.net/fswiki/Tutorials # to use either Freeview or tkmedit or tksurfer (or both) to look at # your brain with the color overlay that indicates the # atlas labels. You'll find your reconstructed data in # /pkg/classes/psy5065/Anatomy (this output # location is determined by the SUBJECT_DIR shell # environment variable set by SetUpFreeSurferPSY5065.csh). # Email your instructor a pretty screenshot indicating success. # This is the image that will get printed on your class mug. # If you're not going to be in town over the summer, give Cheryl # your mailing address so she can send it to you ... # If you're having a hard time finding brain views you like, try # one of these: # 1) The do-everything-manually approach, looking just at brain # volumes and unlabeled surfaces # freeview # launch freeview # File --> Load Volume --> mri/orig.mgz # File --> Load Volume --> mri/aparc+aseg.mgz # ... in the controls on the left side, drop the opacity of # the aparc/aseg file, so you can see it coloring the # GM and labeling the different regions # File --> Load Surface --> surf/rh.pial # ... you'll see a green surface overlaid on your brain. The sulci # are labeled red. # 2) The cortex-centric method: just a pretty brain with all its labels. # Note that the command below is all one line, conceptually. The # backslash half-way through is an escape character, to get the # command shell to ignore the "enter" keypress half-way through the # command: freeview -f surf/rh.pial:annot=aparc.annot:name=pial_aparc:visible=0 \ surf/lh.pial:annot=aparc.a2009s.annot:name=pial_aparc_des:visible=0 # 3) Best of both worlds: brain with surface sticking out of it: freeview -v mri/orig.mgz mri/aparc+aseg.mgz:colormap:lut:opacity=0.4 \ -f surf/rh.white:annot=aparc.annot