Folks,
I am trying to understand how SPM99 handles fitting an AR(1) model to
the average residual power spectrum for the purposes of estimating intrinsic
serial correlations. Specifically, I am curious how the routine handles
models in which Fourier components have been included as covariates. In such
cases, the power in the residuals corresponding to the modeled frequencies
would be zero. Is this of any concern for the accuracy of the AR(1) fit? I
would imagine that the presence of these zero values would tend to bias the
AR(1) model downward.
Thanks,
Geoff
--
Geoffrey Karl Aguirre, M.D., Ph.D.
University of Pennsylvania Center for Cognitive Neuroscience
3815 Walnut Street Fax: (215) 898-1982
Philadelphia, PA 19104-6196 mailto:[log in to unmask]
http://ccn.upenn.edu/~aguirre
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|