> I'm calculating the lateralization index and I wonder how to realize:
> (1) variable statisctical thresholds across subjects - for each
> individual the selected threshold should correspond to a fixed number
> of activated voxels in certain regions.
> (2) mean activation of the 5% of voxels showing the highese level of
> activation in certain ROI.
> (3) % signal change analysis with a given threshold for inclusion
> (i.e. only the voxels that exceed the given activation level are
> included to calculate).
I would recommend the excellent Marco Wilke LI toolbox - it does
everything you want within the robust bootstrap framework ...