Dear FSL experts,
I was wondering how to generate design and contrast matrix to perform a two-sample unpaired t-test with controlling for patient-specific scores using randomise.
I have two subject groups: patients and healthy controls. I would like to get rid of the medication effect in patients group.
Is it possible to regress out patient-specific score from patients dataset, without affecting healthy controls?
If possible, how design and contrast matrix should look like?
Is the following example correct?
Should I demean the patient-specific score within patients group only or across all groups?
HC PT score
-------------------
1 0 0
1 0 0
...
1 0 0
1 0 0
0 1 X1
0 1 X2
...
0 1 Xn-1
0 1 Xn
-------------------
Thank you for your help in advance.
Best regards,
Hiroki Togo
|