Dear Reut,
> i was wondering what kind of multiple comparison correction is being
> done in randomise (using TFCE for example). I didn't find the answer
> in the randomise manual, so can someone please explain the kind of
> multiple comparison test being performed or refer to papers
> concerning the subject?
randomise builds a distribution of maximum tfce statistics. I.e. for
each permutation the tfce statistic is calculated everywhere in the
brain and the largest of those is save into the "histogram" that
constitutes the empirical distribution.
For an intuitive explanation of why this is the right thing to do see
e.g.
http://www.fmrib.ox.ac.uk/fslcourse/lectures/randomise.pdf
noting that there is no principal difference between maximum t-stat,
maximum tfce-stat, maximum cluster size etc etc.
Good luck Jesper
|