Dear All,
I am working on dual regression analysis, however my sample sizes are uneven (30 controls and 400 patients). I noticed that the resting state default mode network map for the controls is not like a previous analysis (the control map is asymetric). Could the patients resting state results be biasing the controls from the initial melodic ica decomposition ? Apart from seed based correlation analysis, are there any other options to remedy this ?
Thank you.
dominic
|