Hi,
  I recently saw the thread below regarding correlations between groups and I was wanting to run this approach with my own data.

https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=fsl;dc11d1db.1305

However, I got something that confused me. For my design I have Age as a covariate and behavioral scores as the variable of interest making my design look like this.

Con  Patient   Age_Con   Age_Patient   Behav_Con   Behav_Pat
1           0          Age             0                  Behav            0
1           0          Age             0                  Behav            0
0           1          0                Age                0                  Behav
0           1          0                Age                0                  Behav


Age and Behavioral performances were both demeaned.
I previously ran a correlation analysis within my patient group (essentially same design but only EV2, EV4, EV6, and no Controls) and since this is a new approach for me I wanted to double check it with my previous result.  Therefore, I set up my contrast as:

0  0  0  0  0  1
0  0  0  0  -1  1
0  0  0  0  1  -1

Now the results I got from the group comparisons (aka. contrast 2 and three) agreed with my hypothesis.  However, the first contrast that I believed would replicate my within group comparison yielded 2613 more significant voxels than previously seen.  I would like to chalk this up to variability with permutations but I feel its a bit much to assume that.  I am using 5000 permutations in my randomise script.

Thank you,
Jeremy