Re: [Icatb-discuss] GIFT - temporal sorting questions

[Marc S. <ms...@un...>]

hello

We have collected some pilot fMR data (n=4) and are using GIFT to look  
for patterns of robust activation.

We have extracted 20 components from the group data and have attempted  
to sort them using individualised (different) SPM.mat files for the  
n=4 subjects.

However, we have a couple of questions for the more experienced GIFT  
users:

1) when we temporally sort the components using individual SPM.mat  
files we occasionally end up with odd correlation values ( > 1 or <  
-1, see attached).  These values are derived from the text file that  
GIFT automatically generates while sorting components.

What does it mean if an individual subject has a correlation of this  
magnitude with one of the group derived components?
it is only meaningful to be looking at the overall group correlation  
value?

2) we are interested to know how the values for group correlations are  
calculated from the individual values, since the group correlations  
often vary significantly from the reported individual values?
Is there a reference to help us with this issue.

Thanks

Marc

-  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -
Marc Seal                   ms...@un...

Griffiths Research Fellow
  Melbourne Neuropsychiatry Centre
The University of Melbourne
c/o National Neuroscience Facility
161 Barry Street
Carlton South   3053
AUSTRALIA

ph: (+61 3) 8344 1861
fax: (+61 3) 9348 0469
www.psychiatry.unimelb.edu.au/mnc/
  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -  -