Hello,
I’m running FreeSurfer recon-all on multiple T1-weighted scans. For every subject, the cortical surfaces (lh.white, lh.pial) are clearly shifted relative to the underlying volume — they sit too high above the cortex when viewed together with brainmask.mgz or orig.mgz. The surface shapes look reasonable, but their global position does not match the anatomy.
The T1w NIfTI files come directly from DICOM conversion (no preprocessing).
What could cause this kind of global surface displacement, and what should I check to fix it?