On 11/06/2017 12:03 PM, lanbo Wang wrote:
Dear experts,
I have two questions about hemisphere analysis:
- When I run left-right hemisphere paired t-test, can I add age as
covariate? If use fsgd to add covariate, to this paired t-test analysis, how to make the fsgd table?
use this one-group, one-covariate example https://surfer.nmr.mgh.harvard.edu/fswiki/Fsgdf1G1V
- I want to reorganize the hemisphere from left-right to
symptom-nosymptom according to body symptom side record and compare different between symptom hemisphere and nosymptom hemisphere. Can I use freesurfer to do it?
Yes, though it is a little tricky with multiple ways, each complicated in its own way. 1. Create your own design matrix. You can use the one created by mri_glmfit to start. Assuming you want symptomHemi-nosymptomHemi, then, in each subject whose symptomHemi is rh, multiply its line in the design matrix by -1. Then pass this design matrix to mri_glmfit with --X instead of passing an FSGD file. 2. Load the output of mris_preproc into matlab, eg, y = MRIread('y.mgh');, then change the sign as in #1 above, eg, y.vol(:,:,:,10) = -y.vol(:,:,:,10); and so on for each applicable subject. Then save the data with MRIwrite(y,'new.y.mgh');, then run mri_glmfit as normal with the new file. 3. Run mris_preproc for each subject separately to generate an lh.lh-rh.thickness.sm00.subject10.mgh file. Then change the sign as in #1 above with fscalc lh.lh-rh.thickness.sm00.subject10.mgh mul -1 -o lh.lh-rh.thickness.sm00.subject10.mgh Then run mri_concat to concatenate all the subjects together in the same order as they are listed in the FSGD file, then use this stack as input the mri_glmfit
Thanks for your patient and help.
All best, Lanbo Wang
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer