mkanalysis-sess seems to default to a FWHM of 20 for the smooting of the ACF (in version 5.1). I was wondering whether it makes sense to use the same FWHM for volume and surface based analyses of functional data in FSFAST? I seem to remember that the autocorrelations differ between, e.g., grey and white matter, which might be a reason to use a different FWHM on the surface?