Hi!

I'm looking to do some network analysis on brain wide cfos (activity marker protein) counts to better understand which regions are active during a behavior. I have 10 control and 10 experimental brains. Many analysis create networks from control and experimental groups and then compare each other. i was wondering would it be fair to bootstrap both groups and subtract both groups distributions to create a network from the difference between bootstrapped distributions of groups in order to have positive of negative correlations or edges which correspond to similar changing groups or inversely changing groups respectively?