 |
 |
Dear all,
I have some questions about 2 sample t test in 2nd-level.
It's a 2 x 4 mixed design, I have 2 Groups (Normal Controls & Patients) and 4 Conditions ( a within-subject factor with 4 level ).
This design wants to compare the difference between Groups when both of them compare Condition A and B.
In other words, it tries to do this contrast: (G1A - G1B) - (G2A - G2B).
My problem is that SPM showed a error message "Please check your data, there are no significant voxels."
(I set Independence as Yes, Variance as Unequal)
I had read some discussion in SPM mail list and tried these ways.
I checked image data by CheckReg and mw_checkimages.m Marko wrote, and excluded some bad image.
( If I just compare difference between Conditions in one Group with one sample t test, it work)
I also tried to modify defaults.stats.fmri.ufp from 0.001 to 0.05 or 1, but the problem still exist, except I change
Variance Option from "Unequal" to "Equal".
I would be thankful for any help.
Best regards,
Chih-Hao