Hello,
In doing a group PPI, I see that in the results there is a significant
cluster at the seed region I chose (in this example, BA9). Does this
result make sense? [This is from an fMRI experiment; task vs. control]
From this page (http://www.fmrib.ox.ac.uk/Members/joreilly/what-is-ppi)
in Figure 1, it is stated:
"Voxels in which activity is equally correlated with the seed region
timecourse all the time will not show any correlation with the PPI
regressor."
Also in an email from Dr. Penny
(https://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind04&L=SPM&D=0&P=1605413):
"if the beta value for I is significantly non-zero there is a
psycho-physiological interaction ie. the psych variable changes the
correlation between source and sink voxels."
The psych variable shouldn't change the correlation between the seed
region and itself.....
I followed the steps in the manual (1. Get VOI in seed region, adjusted
for main effect of task; 2. Do the PPI; 3. Specify + estimate design, w/
regressors PPI.ppi, PPI.Y, PPI.P). I would have thought that the PPI.Y
regressor would explain the variance in the seed region...
Any thoughts?
Thanks,
Chris
|