Dear FSL Experts,
I had a quick question about the workflow that I'm doing right now. I have a 2x2 ANOVA design (Cases vs. Controls, Scanned pre and post intervention) resting state experiment. I used multi session temporal concatenation and then ran dual regression to get the differences between groups (cases > control were the only significant contrast). My question is in regards to fslcc when I am trying to compare my IC maps to reference networks so I have a standardized way to identify my networks. When I look at the IC maps of Cases + Controls from the multi session temporal concatenation, they don't really correlate with known resting state networks. However, if I run MELODIC over again with just the controls, and then compare them with known values it looks quite good...I'm a little confused as to why they are so different?
My question is as follows:
Does it make sense to use just the IC maps from the controls to just identify what the resting state networks are, but use the dual regression output (from the Multi session temporal concatenation of cases + controls) to highlight the differences between groups? (I.e.: cases > controls contrast)? http://journal.frontiersin.org/article/10.3389/fnhum.2013.00461/full
This paper seems to do just that, but I just wanted to make sure before I continue. Any help would be greatly greatly appreciated!
Thanks!
Pranish Kantak
|