Print

---

Dear FSL experts,

I have a quick question that relates to the correct use of Dual_Regression inputs. 

According to the wiki, one needs to use the group-melodic_IC.nii.gz file as input-mask for the dual_regression . I was wondering if the script itself applies a threshold on these ICs (e.g. 75th percentile) before measuring the mean time series, or it uses a weighted mean on the ICs scores. 

My question is then, is it valid to use a binary group-ICA masks (obtained from a separate control-group of subjects) to perform stage1 of dual regression? Or shall I use the original whole brain ICs not thresholded? 

Best,

Valerio