Dear Sam,
> Is there a corect way in which to test whether an activation is focal
> (say <20 voxels) or diffuse ( >80 voxels)? I am thinking of something
> along the lines of looking for changes in peak z-scores at different
> levels of smoothing, but I am not sure how valid tihs would be.
I think the simplest approach would be to operationally define 'focal' in
term of a spatial extent threshold and use the p values based on
spatial extent to say whether it is diffuse (large) or not, conditional
on the peak height being significant (i.e. it is an activation).
Multi-scale signatures of different sorts of clusters be be useful in
terms of characterizing the activation and I think J-B's group has done
some work on this.
I hope this helps - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|