Computing connectivity with ROI yields the same cluster. Reasons?

Submitted by JH89 on

Hi everyone,

after preprocessing I computed functional connectivity of one cluster (one node of the DMN predefined in the toolbox: x = -6 y = 52 z = -2). I did this for 20 subjects. Afterwards I computed a t-test versus zero (SPM contrast: 1) based on the connectivity map of each participant.

 

However, this yields only one very large significant cluster, namely the very ROI I had selected originally (-6 52 -2). Is there any reasonable explanation for this finding? I actually expected to find some more clusters of the DMN that should covary with the DMN-node I picked.

 

 

Thanks for any suggestions

Jonas

YAN Chao-Gan

Sat, 11/04/2017 - 03:25

If you didn't set a high threshold, the whole brain may connect to a single cluster if you didn't perform global signal regression.