Hello,

I am interested in quantitating the extent of orbitofrontal cortex voxel activation elicited by our task in each individual subject. I managed to use 3dclust successfully to dump out the activated clusters corresponding to each of four contrasts (15, 17,21,23) in a functional map (entitled zantvsfbkregdsbpf)

with the following commands in a script:

3dclust -dxyz=1 -1thresh 3.3 -1noneg -1dindex 15 -1tindex 15 1.01 6 zantvsfbkregdsbpf+tlrc > voxelcountrvnant
3dclust -dxyz=1 -1thresh 3.3 -1noneg -1dindex 17 -1tindex 17 1.01 6 zantvsfbkregdsbpf+tlrc > voxelcountpvnant
3dclust -dxyz=1 -1thresh 3.3 -1noneg -1dindex 21 -1tindex 21 1.01 6 zantvsfbkregdsbpf+tlrc > voxelcountrvnfbk
3dclust -dxyz=1 -1thresh 3.3 -1noneg -1dindex 23 -1tindex 23 1.01 6 zantvsfbkregdsbpf+tlrc > voxelcountnvpfbk

with the -dxyz=1 clause, I get the raw voxel count as the first output column. This yields several clusters where the z-t value > 3.3 (p <.001), only a handful of which are in the ofc and of interest to us.

I would like to use 3dmerge to auto-create a mask of aggregate ofc activation (based on the group map) and *filter the 3dclust output* so that only the clusters that overlap in part the mask (defined by the group map of ofc activation) get dumped out for each individual subject. Then I can simply sum the voxels from these selected clusters to get a rough measure of "ofc activation".

Could someone please tell me if and how I could do this? What I could add to the script, for example? It seems I'm already applying a contrast t-value threshold, could I add an additional mask-defined fim threshold value as well?