Procedurally, you would create stacks (eg, mris_preproc) of the thickness and the PET data, then you would use one as the --y input to mri_glmfit and the other as --pvr input (in addition to your fsgd file or design matrix). The PVR is per-voxel regresssor, so it will create a design matrix for each voxel using the data found in the pvr.

Theoretically, you might run into problems because the partial volume effect will interact with thickness, so, depending on the nature of the PET tracer,  it may be hard to know whether you are seeing a true effect or an artifact or it could make a true effect disappear.

On 3/5/19 5:56 PM, Boris Rauchmann wrote:

     External Email - Use Caution

Dear FS community,

I would like to assess how PET tracer uptake is related to thickness changes in AD patients. Is it possible to generate difference maps for both methods and run a GLM? Or would you suggest another approach?

Thanks, Boris

---

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