Hi, sorry for the delay.

On 6/7/2019 3:33 AM, Avi Gharehgazlou wrote:

        External Email - Use Caution        

Hi experts,

 

Will you please confirm below command lines I run to examine between group differences in LGI? I have 3 questions listed.

 

I can’t run the below line, it gives an error:  

mris_preproc --fsgd TDvsNDDageSAsex.fsgd --cache-in lh.pial_lgi.fwhm0.fsaverage --target fsaverage --hemi lh --meas pial_lgi --out lh.pial_lgi.0.mgh

mris_preproc --fsgd TDvsNDDageSAsex.fsgd --cache-in rh.pial_lgi.fwhm0.fsaverage --target fsaverage --hemi rh --meas pial_lgi --out rh.pial_lgi.0.mgh

 

Therefore, I run the below line instead - although I know it’s for uncached data and I have already ran qcache for my subjects (using: recon-all -s 101031_01 -qcache -measure pial_lgi)

 

mris_preproc --fsgd ASDvsADHD.txt --target fsaverage --hemi lh --meas pial_lgi --out lh.pial_lgi.mgh

mri_surf2surf --hemi lh --s fsaverage --sval lh.pial_lgi.mgh --fwhm 0 --cortex --tval lh.pial_lgi.0.mgh

# I run above line separately for right hemisphere

1.       Do I have to write “cortex” in above command line or not?

No, not with --fwhm 0. In fact, you should not need to run that command at all.

2.       Is it okay to run this command line considering that qcache has already been run on all my subjects?

Yes

3.       I’ve ran qcache (and above command line) multiple times on each subject in my data. Is this ok? Or does this mess up anything?

No, it is fine.

To test for main effect of group, I ran below line:

mri_glmfit --y lh.pial_lgi.0.mgh --fsgd ASDvsADHD.txt dods --C betweengroupdifferencecontrollingforageSAsex.mtx --C agebydiagnosisinteraction.mtx --surf fsaverage lh --cortex --glmdir lh.ASDvsADHD.glmdir

mri_glmfit --y rh.pial_lgi.0.mgh --fsgd ASDvsADHD.txt dods --C betweengroupdifferencecontrollingforageSAsex.mtx --C agebydiagnosisinteraction.mtx --surf fsaverage rh --cortex --glmdir rh.ASDvsADHD.glmdir

 

To correct for multiple comparisons using monte carlo:

mri_glmfit-sim --glmdir lh.ASDvsADHD.glmdir --cache 1.3 abs --cwp 0.05 --2spaces

mri_glmfit-sim --glmdir rh.ASDvsADHD.glmdir --cache 1.3 abs --cwp 0.05 --2spaces

Those commands are technically fine, but you should not use a threshold of 1.3 with MC correction as the false positive rates will be highly elevated. If you want to use 1.3, then you should use a permutation test. Otherwise, you should use at least 3.0


I'd really appreciate your soon reply, 


Thanks so much for all your support, 


Avi


_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer