Hi,
If the peaks are truly distinct -- and statistically significant -- then you might have to raise the initial cluster-forming threshold (e.g., from z=2.3 to z=3.1) to separate them. You could also try voxelwise inference, which will tend to be much more conservative, but will give you more spatial precision.
Cheers,
David
On Dec 2, 2011, at 11:26 AM, Leslie Engineering wrote:
> Hello!
>
> I am trying to subdivide thalamic activation from an fMRI dataset. Some regions are rather close and I believe both are activating. Is there a method for testing for two peaks within a cluster?
>
> Thanks so much,
> Leslie
>
|