Dear FreeSurfers,

I would like to perform surface based analyses using PET but I am not sure I am doing things correctly. For each subject, I have reconstructed the 3DT1 MRI using recon-all –all.

Then I have coregistered the PETs with the anatomical using:

Bbregister  --s name  --mov pet.nii  --bold  --init-fsl  --reg  pet_2_T1_register.dat

Registration worked very well.

Now I  am trying to process the PET in order to do a surface based analysis of the PET signal.

I was thinking of running:

1)      mris_preproc --target fsaverage --hemi lh  --iv  subject1/pet/subject1_pet.nii  subject1/pet/pet_2_T1_register.dat  --iv  subject2/pet/subject2_pet.nii  subject2/pet/pet_2_T1_register.dat  --iv  subject3/pet/subject3_pet.nii  subject3/pet/pet_2_T1_register.dat (…)  --projfrac 0.5  --out  lh.pet.mgh

I am not sure this is the right command line. I am especially concerned with the registration file I should use. Here I used the PET àMRI registration file. Is this the right one to use. Or should I use a registration file from native space to fsaverage space. If so, how do I do that?

2)      mri_surf2surf --hemi lh --s fsaverage --fwhm 10 --cortex --sval  lh.pet.mgh --tval lh.pet.sm10.mgh

Any advice on the fwhm I should use for smoothing the data to the surface?

3)      mri_glmfit  --y  lh.fmz.sm10.mgh  --fsgd  fsgd.txt  --C X.mtx  --surf  fsaverage lh  --cortex  --glmdir lh.pet-age.glmdir

Can you please tell me if I got it right, especially regarding the registration file I should use with mris_preproc?

Thank you so much,

Leorah Freeman, MD, MSc