[Freesurfer] interscan variability

External Email - Use Caution

Hi,

I am looking at various brain volumes in the aprarc and aseg stats generated by recon-all and SAMSEG> stats. Some of the subjects were scanned on Philip and some on GE, and some on GE at time point 1 and PH on time point 2.

Q1: Is there a measure in the aparc, aseg, and SAMSEG stats files that is fairly resistant to scanner variability? I clearly see eTIV for the same subject have 15% or more discrepancy between GE and PH. Interestingly, cortical thickness relative difference between PH and GE is 1.23% in controls.

Q2: Is there a way to harmonize between GE and PH values?

Many thanks,

AJ