So I understand that when doing mri_glmfit you want to use smoothed surface maps to comply with random field theory while controlling for multiple comparisons (as discussed here: MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg67902.html). 

But I was also wondering whether we have to smooth surface maps when using mri_anatomical_stats to extract anatomical measurements? I see that mri_anatomical_stats has the option to smooth by # of iterations but is there a way to know how many # of iterations would equal let's say 15 FWHM? 

Best,

Julia Shin