Hi Mark,
Thanks for the reply. I ran an F-test over group mean and +ve score as you suggested and it didnt yield any active voxels. Wondering if i setup the demeaned scores correctly. I have 16 sessions pre and 16 sessions post and similarly 16x2 test scores. I used to mean of 32 scores to demean the data and added that to the paired design model. I assumed i dont need to multiply the post test scores by -1 . is that right? I also added the test scores in individual session's group mean analysis. To summarize..
pre mean result:
mean contrast(1 0) = active
+ve correl(0 1)= no active (or correlated) voxels
-ve correl(0 -1) = no active (or correlated) voxels
F-test over 1&2 = active voxels in some common and a few unique areas compared to mean contrast
post mean result:
mean contrast(1 0) = active
+ve correl(0 1)= no active (or correlated) voxels
-ve correl(0 -1) = no active (or correlated) voxels
F-test over 1&2 = active voxels in mostly common areas but lesser extent compared to mean contrast
pre vs post result:
pre>post (1 0 0..) = no difference ( voxels with significant difference in the analysis without test scores)
post>pre = no difference (no difference in analysis without test scores)
+ve correl = active (or correlated) voxels
-ve correl= active (or correlated) voxels
F-test over contrasts 1&3 = No active voxels
|