Hello-
I would like to directly compare asymmetries in FMRI activation in surface space based on the Greve et al pipeline presented at HBM this year. All of our anatomical data have been processed with FS v4.5, but the FMRI data is in AFNI volumes. I am very new to the freesurfer program so any corrections on how to achieve cross-hemisphere registrations would be greatly appreciated.
1) We transformed the functional data to anatomical space using bbregister.
2) We then registered each subject’s T1 data to the contralateral hemisphere of fsaverage and made templates of the contra (?h.con.sphere.reg ) and ipsilateral (?h.sphere.reg ) registrations
Registration step
mris_register -curv surf/rh.sphere {$FREESURFER_HOME}/average/lh.average.curvature.filled.buckner40.tif surf/rh.CON.sphere.reg
mris_register -curv surf/lh.sphere {$FREESURFER_HOME}/average/rh.average.curvature.filled.buckner40.tif surf/lh.CON.sphere.reg
template step
make_average_subject --out CON_TEMPLATES --surf-reg CON.sphere.reg --subjects M87150222_20110311 M87138477_20080429 ..
make_average_subject --out IPS_TEMPLATES --surf-reg sphere.reg --subjects M87150222_20110311 M87138477_20080429 ..
3) I then re-registered RH data to the new contralateral template (study-specific) and LH data to the new ipsilateral template to get everything in the same sphere space
mris_register -curv surf/rh.sphere CON_TEMPLATES/rh.reg.template.tif surf/rh.finalTEM.sphere.reg
mris_register -curv surf/lh.sphere IPS_TEMPLATES/lh.reg.template.tif surf/lh.finalTEM.sphere.reg
4) Finally, I apply the final registration template to aux data (thickness, FMRI) to bring everything into a common space that can be directly tested for hemispheric asymmetries
mris_preproc --hemi rh --target CON_TEMPLATES --surfreg finalTEM.sphere.reg --meas thickness --out final.thickness.nii --s ${subject}
mris_preproc --hemi lh --target IPS_TEMPLATES --surfreg finalTEM.sphere.reg --meas thickness --out final.thickness.nii --s ${subject}
Please let me if this pipeline makes sense and follows what was proposed at the HBM meeting. Also are the iterations mentioned in the poster necessary (ie make additional templates and re-reregister) when the sample is relatively large (N = 100)?
Thanks!
Andrew