Dear FSL experts,
We have run a dual regression after group melodic in order to perform a resting state analysis. The mask created by melodic seems quite okay, but when doing the dual regression a lot of the parietal lobe and the cerebellum is cut off (so the mask simply looks too small). Our first guess is, that dual regression might not be able to deal with missing values (cases in which it was not able to cover the whole brain during scanning).
So the questions are:
(1) Might this be due to missing values?
(2) Is there a possibility to handle it (like melodic obviously does)?
Thanks for your help and kind regards,
Stefan Lenzen
PS: In the attachment you can find the mask created by dual regression.
|