Hi all,
Any advice on the following would be greatly appreciated:
a) Is there a way to explicitly specify a mask in Melodic ICA and
therefore limit the ICA analysis to a specific part of the brain?
b) How do I limit the number of output components in Melodic ICA? I have
tried switching off the Automatic dimentionality estimation and
explicitly specifying 15 output components. I still get more than 60
components.
and finally,
c) The more I smooth my fMRI data the more components I get out (a
kernel of 4mm on 2x2x2mm^3 data produces 67 components and a kernel of
10mm produces 176 components). This seems counter-intuitive. Any
suggestions?
I use Multi-session Tensor-ICA for a 10 subject group analysis with
Melodic 3.09.
Many thanks,
Emmanuel
________________________________________________________________________
E A Stamatakis PhD,
Division of Anaesthesia
School of Clinical Medicine
University of Cambridge
Box 93, Addenbrooke's Hospital
Hills Road
Cambridge, CB2 2QQ
U.K.
|