Hello,
I have a question about parametric modulation and orthogonalisation of GLM regressors. We have a fMRI-EEG dataset and have measured alpha and beta power response to simple stimuli. By demeaning the trial-by-trial alpha and beta power we get two regressors which can be used along with a main effect regressor to find the correlation of power with BOLD response. What we would like to do is model in a single GLM the alpha variability, beta variability, and also any variability that is common between alpha and beta.

I understand that orthogonalisation of one regressor (A) wrt another (B) removes from A the commonality between the two leaving that in B.  Is it possible in some way to use orthogonalisation (or a similar method) to extract the common factor between alpha and beta while maintaining the separate elements of the two?

Thank you,

Ross