Dear Dr. Michiru
> As I do not like to golobal normalise our fMRI data, I wonder if it is
> valid that I omit to perform any grand mean normalisation in creating
> adjusted mean images and in statistical analysis with the random effect
> model.
There is nothing wrong with this. However, if there are severe global
confounds it will reduce the sensitivity of your analysis. I would do
both and compare the results.
I hope this helps - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|