TRF condition average baseline noramlization

tl;dr: How to perform a condition average baseline (z-score) to TFR data?

Hi,
I want to do a within-subject cluster permutation analysis to assess the difference in activity between two experimental conditions. I believe that these experimental conditions have an overall different baseline activity, so in order to compare their time-frequency spectrum, I need to perform a condition average baseline (as discussed here)

I can’t tell from the EpochsTFR documentation if:
(1) The baseline application is done per epochs separately. In this case, I’m not sure how to use “apply_baseline” in order to perform a condition average baseline.
(2) It used the baseline period of the epoch’s mean. (In this case, I can just apply the baseline over the concatenated epochs data of the 2 conditions, right?)

I will really appriciated help with figuring out which of the two options is true, and how to perform a condition average baseline.

Thank you,
Sharon
MNE version: 1.1.1