SEEG/iEEG bipolar reference setup

💬 Support & Discussions
, ,
1

Dear all,

I have recently been dealing with SEEG preprocessing using MNE. I want to use bipolar reference (mne.set_bipolar_reference — MNE 1.8.0 documentation), but I am not sure what exactly I should do.

I noticed that set_bipolar_reference needs two required parameters: anode and cathode. So, do I have to build the bipolar pairs manually?

Assume we have two electrodes: A & OB.
contacts at A: A1 A2 A3 A5 A6 A7 A8 (missing A4)
contacts at OB: OB1 OB2 OB3 OB4 OB5 OB6 OB9 OB10 (missing OB7 OB8).

  1. How should I determine which contacts belong to anode/cathode? Is there any general principle?
  2. I knew that in a normal case, the new channel would be like A1-A2, A2-A3. But for missing contacts like A4/OB7 OB8 above, can we create the new channel A3-A5? How about OB6-OB9?

Thank you!
Kun

2

That function was originally designed with EOG in mind, where we usually have a horizontal and vertical setup. I’m not familiar with SEEG, so perhaps this makes no sense, but you can give lists of channels to the function to create bipolar pairs. So, if we would want A1-A2, A2-A3, A3-A5, we can do:

raw_ref = mne.set_bipolar_reference(
    raw,
    ["A1", "A2", "A3"],
    ["A2", "A3", "A5"],
    ch_name=["A1-A2", "A2-A3", "A3-A5"]
)
1 Like
3

It is advisable to make the bipolar montages in one direction, so A1-A2, A2-A3, etc. However, this gives you the same information as the other way around A8 - A7, the signals will only be inverted.

In my work, I skip the cases with missing contacts, but that is because the missing electrodes are used for intracranial stimulation. It mainly depends on your application and what you want to do with the data.

1 Like
4

@wmvanvliet @SBeumer thank you both! It’s really helpful.

5

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.