In a functional analysis, we have a set of external regressors (e.g. mean luminance of stimuli, size of stimuli, etc.). Should we normalize values before including them in the regressor files? If so, what is the recommended way of normalization ... z-scoring?

Thanks,
Reza