[Marsbar-users] random effect analysis question

[LiuKarl <rem...@ho...>]

 
Hello, Matt
 
I have some questions of the random effect analysis. We used an emotional face recognization task in which subjects view fearful, neutral faces and cross hair. In the first level analysis, we got contrast images for each subject such as (fear-crosshair), (neutral-crosshair) and (fear-neutral). Since we have two groups: patient and healthy control, we further put the contrast images into random effect analysis. Now we want to move to ROI analysis:
 
1)      the manual says “import the random effect design into Marsbar, and run it using your ROI”. How to “import the random effect design into Marsbar”, just choosing the spm.mat file?
 
2)      if we extract the values, what exactly is it? I think it is mean beta estimates from the design matrix in the contrast image, but not the BOLD signal, right?
 
3)      How can we examine the % BOLD signal change if what we get is beta estimate? 
 
Thanks
 
Karl