Hi Whitney,

You found a required cluster size of 7 voxels. But those voxels
are 3.75 x 3.75 x 6 mm, which is 84.375 mm^3, each. If you wish
to translate that volume to a 1 mm^3 grid, then use a minimum
volume of 85 voxels (mm^3), not 7. However, I don't think it's
appropriate to compute the minimum cluster size with large voxels,
but then apply the clustering on your small voxels.

I think you should run AlphaSim with Full_Mask_Controls+tlrc as
the mask, and -rmm 1.01 (since those would be small voxels).
Then run your 3dmerge command using the same 1.01 mm radius.

---

To add a little confusion, you might compute the blur values in
another way, too. Consider running '3dFWHMx -detrend' on the
error time series for each subject, and work with the average
blur values (they should not vary much).

If you don't have the error time series, you can either compute
it with 3dDeconvolve's input minus the fitts, if you have it, or
else just use the input dataset to 3dDeconvolve.

Then you can just work with that one set of blur values for any
AlphaSim computation.

- rick