Hi Everyone,
I had a quick question regarding identifying RSNs. I've run Group ICA on a 27 subject sample (2x2 ANOVA design). I've ran it at 30 40 50 60 dimensionalities. I was hoping to compare the Melodic_IC.nii.gz files to cortical and subcortical parcellations using fslcc. Is this the right way to go about identifying networks? I used Yeo's 7 and 17 cortical parcellation. Does anybody know if there is a subcortical reference that is similar to Yeo's? Additionally, how would you go about picking which dimensionality to use? i think 30 looks pretty good, and I think some of the networks become fragmented at the very higher dimensionalities.
Any help would be appreciated!
Pranish
|