Dear freesurfer experts,

This is a follow up question from a previous post

http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg47480.html

I have created a sig.mgh from my own stats code and need  to also generate fwhm.dat file to run the monte carlo sim, which is estimated from the residuals.

What are the specifics and dimension of this residual file I need to generate for mris_fwhm? 

Is this a residual of model fit at each vertex across all subjects (so [number of subjects x number of verticies] matrix in .mgh file)? or variance or standard deviation of the residuals?

Perhaps pointing to an example file in the glmdir directory (created when running glmfit) will help me understand the process (for example  is rstd.mgh or rvar.mgh is being used for mris_fwhm --i or something else?). 

Thank you for your help!