PSD log transformation

Thanks Marc.Lalancette, got your point.

What about this formula:
PSD value minus mean PSD value across subjects, divided by standard deviation of PSD values across subjects.