Dear SPMer,
I have a question relating to the statistical analyses of my study. For the study I have 2 different groups of subjects (G1 & G2) and 3 different conditions during the fMRI (C1, C2 and Control) In the first level analyses I did 4 t-tests for each subject (con01: C1>Control; con02: C2>Control; con03: C1>C2 and con04: C2>C1).
The contrast vectors were as following: [1 0 -1] for Con01; [0 1 -1] for con02; [1 -1 0] for Con03 and [-1 1 0] To do the 2nd Level (between group) analyses I did a 2-sample t-test for each contrast (con01, con02, con03, con04) in which I added the respective con-images of both groups. After this calculation I have the option to decide whether I want to see Group 1 > Group2 OR Group2 >Group1.
And now my problem: The results of the 2 contrasts Con03(C1>C2) G1>G2 AND Con04(C2>C1) G2>G1 were exactly the same (i.e. they showed exactly the same activated areas in the glass brain and in the list of activated regions), as was the case for both in the oposite direction (Con03(C1>C2) G2>G1 AND Con04(C2>C1) G1>G2. This does not make sense to me.
Do you have any idea how I could analyze or rework either the data groups or the experimental design so to get more reasonable statistical results?
Thanks
Lara
|