Dear Anderson,
thank you for invaluable comments.
I fixed my cluster-forming threshold to z- threshold and included lh.white.avg.area.mgh as second argument to -s option and now the cluster p-values (after 50000 permutations) of PALM and FreeSurfer are almost equal (up to 2 decimal places).
I would like to remind you of the issue in PALM code: I found that to correctly read the area file in PALM, the statement on line 131 of palm_takeargs.m (s=s+1) should be executed only after the opts.sa{s} = vararginx{a+3} is assigned.
Without this modification, the area file is not properly read.
Regards,
Antonin Skoch
|