 |
 |
Hi gui
1) It could be a good idea, it could improve your detection of correlation a little, although it wouldn't be guaranteed. I'm not familar with these sort of studies: might it be a bit difficult to interpret correlations?
2) Two successive gaussian smooths of n & m mm FWHM are equivalent to a single gaussian filter of sqrt(n^2+m^2). However, applying the second filter after Feat's modelling is not the same as doing the full filtering in the preprocessing stage.
Cheers,
Eugene
--
Centre for Functional MRI of the Brain (FMRIB)
University of Oxford
John Radcliffe Hospital, Headington OX3 9DU Oxford UK
Ph: +44 (0) 1865 222 523 Fax: +44 (0) 1865 222 717
--
2010/1/13 Gui Xue <[log in to unmask]>Hi- i am running analysis to see the correlation between brain structure (FSLVBM) and brain activation (FEAT cope).
the functional data (time series) has been smoothed with a 5mm FWHM filter before statistic analysis.
for the FSLVBM, people suggest 9 to 10 or bigger smooth filter.
my question is (1) in order to do voxel-wise correlation between VBM and COPE, do them have to have the same degree of smooth?
if so (2) how much more smooth should i add to the COPE given it's been previous smoothed?
thanks,
Gui