Dear FSL Experts,
Hi,
I wanna use dual_regression approach in order to do between-groups comparison. Reading your poster, my impression was that in the first stage you concatenate all the data (I mean from the 2 groups of subjects (patients and controls) that you have) and then apply concat-ICA to find "group ICA maps". Is it true or instead you apply concat-ICA seperately on each group of subjects (controls together and patients together)? In the first senario you are using the varibility of the whole dataset to extract networks whereas in the second senario it is coming just from just one group and you assume from the beginning that the controls may share the same group_ICA maps but it is different from the group_ICAmaps that patients may share.
I would appreciate your clarification on this.
Best,
Mona
|