Hello everybody,
I have used gift toolbox for group analysis of resting state fMRI data. 
I used component labeling utility to form a subset of components that are actually resting state network (RSN). However using the correlation coefficient this utility gives is not enough and I have to analyze the power spectrum as well. I figure choosing RSNs is not a simple job. 
Anybody knows what are the different methods of differentiating RSNs from noise? 
Can fuzzy inference system be used to solve this problem?

Thanks