Thanks Ben. My biggest fear in trying a new analytic technique is knowing enough to be dangerous. Your response makes sense. I am actually only interested how "connectivity" of DMN with the rest of the brain varies with my behavioral variable. Therefore I would take the dr_stage2_ic###.nii.gz results that correspond with the DMN component number identified in my group ICA and run a glm analysis with my mean centered behavioral variable as a covariate (e.g., explanatory variable)? The script would look like:
randomise -i dr_stage2_ic###.nii.gz -o glm_output -d behavioral_variable.mat -t behavioral_variable.con -m <mask_image> -n 5000 -D -T
What would I use for the "mask image"? I suppose I could mask the group DMN ICA component but I think that the dr_stage2_ic### 4D file has a 3D image for the individual component that corresponds to the group DMN component. Would I need a mask? Thanks in advance for your time and expertise.
Best,
Ross
|