Dear FSL experts:
I have 2 group of subjects to compare, one is controls and the other is patients.
I merged all subjects nii data, patients first in the order, using fslmerge -t .
Then, I create design.mat and design.con using Text2Vest. With EV1 for patient group, EV2 for control group ,EV3 for covariate need to remove.
The design.mat is this:
/NumWaves 3
/NumPoints 36
/Matrix
1 0 2141.25
1 0 14581.25
1 0 -27758.75
1 0 2522.25
1 0 -25247.75
1 0 -31483.75
1 0 -22558.75
1 0 10636.25
1 0 17171.25
1 0 2526.25
1 0 -5238.75
1 0 26846.25
1 0 14786.25
1 0 129666.25
1 0 -283.75
1 0 -522.75
1 0 -14333.75
1 0 16446.25
1 0 22125.25
1 0 -18903.75
1 0 -23613.75
1 0 -10723.75
1 0 1361.25
1 0 -8298.75
1 0 -25648.75
0 1 -12633.75
0 1 20090.25
0 1 8384.25
0 1 -27645.75
0 1 -11823.75
0 1 -23349.75
0 1 22811.25
0 1 12341.25
0 1 -1737.75
0 1 -15153.75
0 1 -17475.75
and design.con is this:
/NumWaves 3
/NumPoints 2
/Matrix
1 -1 0
-1 1 0
for EV3, I demeaned before input into design.mat.
Now my problem is, randomise finished with no warning or error, by keep telling me this:
Critical Value for: result_tfce_tstat1 is: 0
And the output images always blank. The output files are tfce_corrp_tstat1.nii.gz, tfce_p_tstat1.nii.gz, tfce_tstat1.nii.gz.
Only this one : tstat1.nii.gz is not blank.
My randomise cmd is :
randomise -i merged -o result -d design.mat -t design.con -T -R --uncorrp -m GreyMask.nii.gz
I checked many times, read the wiki about randomise and GLM, but I really don't know where the problem is.
Please, NEED some help!
By the way, the same data using PALM integrated in DPABI_STAT_TOOL did not have this problem.
|