Dear SPM experts,
To investigate which brain areas are functionally connected to a specific brain area of interest, I set up a PPI analysis as described in the SPM manual (i.e., I extracted the eigenvariate for a 6mm sphere around a voxel of interest in the inferior frontal sulcus, calculated the PPI.mat using the batch editor, and defined a new GLM with PPI.ppi, PPI.Y, and PPI.P as regressors). (I am also interested in the psychophysiological interaction, that is why I set up a PPI in the first place, however, my problem here only regards the functional connectivity over the whole experiment, regardless of experimental condition.) As far as I understand, looking at the PPI.Y regressor (the contrast 0 1 0) should show brain regions correlating with activations in the area of interest. However, the resulting SPMs only show one huge cluster including more or less the whole brain and peaking around the area of interest (with P set to .05 FWE). The peak at the area of interest of course makes sense (the brain area should correlate highest with itself), but the uniform activation of the rest of the brain seems very strange.
Any ideas what might have gone wrong? (The other contrasts, i.e., PPI.ppi and PPI.P look quite reasonable.)
Thank you for your help,
Moritz
|