Incomplete comodulogram and stats issues when using PAC()

[bonokat]

Dear Tensorpac team,

I am a newbie in computing PAC and using Tensorpac. I am using Pac(idpac=(pac_method, surr_method, norm_method), f_pha=phase, f_amp=f_amp) function and compute stats as suggested in one of the examples on 30 electrodes of interest.

I have tried different methods in idpac parameter; I have 30 channes and the data.shape for each channels is (734, 501).

# either theta-lower gamma of alpha-upper gamma coupling
phase = (3, 9, 3, .2) OR (6, 14, 3, .2)
f_amp = (30, 50, 2, 0.5) OR (50, 80, 2, 0.5)

When doing this I have encountered a couple of issues:
1. Incomplete comodulograms
When computing MI (which is the primary thethod I'd like to use for my data) with idpac=(2,2,4) I am getting incomplete comodulograms and stats maps:
Fp1


As you can see, there's also an issue with statistical output (everything is significant what does not happen when I use other methods)
It happens only for MI and HR methods, all other methods gave complete maps. The data and the script were the same.

---

The rest of the questions are about the data computed using idpac=(1,2,4)
2. Inconsistencies between PAC and stats maps
For some electrodes the statistically significant activations do not correspond to the maps with PAC activity (I am getting significant values where PAC values were pretty weak):
F4

3. All significance in one place
For both theta- and alpha-gamma PAC I am only getting significant values for the leftmost values along the x-axis. I don't know whether they are true or whether it is some strange sort of edge effect. Here are a couple of examples:
AF8 - theta

FC1 - theta

FC4 - alpha

C5 - aplha

Could you please suggest what might be causing these issues?

Thank you!

Best,
Kat