I want to normalise either PSDs or signals in order to compare theta in old or young mice. I thought about doing z-score normalisation of the LFP and then calculate PSD. The PSDs look good with all the peaks in the right place. Thanks a lot!
This is the same as removing the DC globally and then scaling the signals, so there's no intrinsic problem there in my opinion.
But if you compare different conditions, then it is a relative power spectrum, so a change in a higher-frequency band could be driven by a change in a lower-frequency band.