Evaluating similarities in 3D clusters across subjects

Hi Mike and all,

Say that I have extracted significant clusters (3D) of activity compared to some baseline, that are not time-locked to any event or stimulus. And I’ve done this for each subject.

So there is no obvious structure (like event locking) to where and when clusters will be identified across subjects.

Does anyone have suggestions for how I could look for commonalities in clusters of activity across subjects?

Many thanks,

Hi Ira. I suppose you could compute marginal sums over each dimension, like counting the number of frequency bins that are significant. Would something like that work?

Thanks! ill give it a go