Hi all,

I’m trying to compute cortical thickness using mris_thickness (stable 5), but getting slightly different results from the recon-all 5.1.0-amd64-2.13 (mri_make_surfaces, stable 5).

Hinted from an old thread: https://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg05994.html, I set the switch “-N 20", but it returns different values: 50% of vertices have identical values with the cross-correlation of about 0.98 and RMS of 0.08 mm, which is of course very high, but a bit odd while I was expecting nearly identical results. When I checked the scatter plots, the mris_thickness returns a bit smaller values for many vertices than the standard result. Is this somewhat expected difference between mri_make_surfaces and mris_thickness?

Also I found that mri_thickness reads “?h.gray”, not “?h.pial”. As I assumed that it reads “?h.pial” and “?h.white” to compute thickness, so I just linked “?h.pial” to “?h.gray”. Is this assumed input for mris_thickness?

Best regards,