Dear SPM'ers
The general random effects theory tells that alpha(i) ~ i.i.d [0,
sigma**2(alpha)], each i.
In Andrew Holmes's RFXposter on page 2 it is written that
" to extend inference to population subjects are chosen at random from
population so alpha(i):s are random effects
alpha(i) ~ N [alpha, sigma**2(alpha)] "
[ **=power , ()= subscript]
Why the mean of Andrew's expression is not 0?
Could Andrew or somebody else comment on this?
Thanks,
Esa Wallius
Turku PET Centre, Finland
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|