Axial to planar gradiometer transformation - #13196

Hello! 👋 Thanks for opening your first pull request here! ❤️ We will try to get back to you soon. 🚴

Looks like you're making some progress, let me know when you'd like some feedback!

Hi @larsoner! After a while, I finally found some time :)

I think this is a good moment for some feedback.

Issues / Questions I encountered:
1. Handling ch_type when interpolating from CTF → Neuromag

My initial understanding was that when we interpolate from CTF to Neuromag, the ch_type should already be known.

This is because, CTF systems have reference sensors that should not be interpolated and Neuromag systems have gradiometers and magnetometers, so we need to know the type of each channel.

That's why I added the ch_type parameter in the .txt montage files — to explicitly specify this.

2. Scope of interpolation: only gradiometers?
In relation to (1), I assumed we would only interpolate gradiometers (i.e., axial ↔ planar), and not magnetometers.

Based on this, I thought I would input in _map_eeg_and_meg_channels would use info_from and info_to objects that contain only the gradiometer channels. This was another reason for adding the ch_type parameter in the montage.

3. Limitations of make_dig_montage
All the montages are created via make_dig_montage.

From the documentation of make_dig_montage:

 .. note::
            For custom montages without fiducials, this parameter must be set
            to ``'head'``.

I understood that for my setup (no fiducials) I should use a custom montage.

However: running montage = make_dig_montage(ch_pos=ch_pos, coord_frame="head") does not parse the ch_type or ori and my channels are read as generic EEG labels, e.g., standard_montage.ch_names[0] = 'EEG #1' etc.

4. Custom _meg() function and related problems
Because of (3), I created a custom_meg()function to parse the .txt files.

But now I run into new problems. When I run interpolate_to() :

_validate_type(sensors, DigMontage, "sensors") leads to the error: AttributeError: 'CustomMontage' object has no attribute 'get_positions'
ch_pos = sensors.get_positions().get("ch_pos", {}) leads to the error: sensors must be an instance of DigMontage, got <class 'mne.channels._standard_montage_utils._meg.<locals>.CustomMontage'> instead

5. I added _meg() inside _standard_montage_utils but am I allowed to change such private function ?

6. I am trying to follow how interpolate_to() code style and structure is already built.

However, this forces us to diverge a bit from the standard interpolate_bads recipe

One idea I had would be to encapsulate some logic in a new function interpolate_to_meeg() inside interpolation.py. Just as interpolate_bads_meeg() works inside interpolation.py. But if we do that then we need to do the same thing for the EEG part of interpolate_to()

@contsili would it help if I pushed some commits here? Would be good to get this in!

@larsoner yes it would help! The implementation is working, the only thing to be done is see why the tests are failing and adjust the code.

Once you are done, I can chime in and finish whatever is left :)

@larsoner thank you for the finishing touches!

🎉 Congrats on merging your first pull request! 🥳 Looking forward to seeing more from you in the future! 💪

Hi all! Thanks for implementing this transformation! I just noticed that interpolate_to() resets sampling frequency, filter info, and other information which should instead be maintained across the interpolation step. I will open pull request containing a fix.

Dear all,
Thanks for the fix! I have reviewed the code and I was wondering whether the equivalent of ft_megplanar's sincos method is implemented in "interpolate_to" to generate topoplots like this. This synthetic planar gradient is routinely performed in CTF systems.
I came across issue #9609 and wanted to check if a sincos implementation exists anywhere in MNE-Python.
Thanks in advance,
Diego

Hi Diego,
As you probably saw yourself, the actual work is done by mne.forward._field_interpolation. _map_meg_or_eeg_channels() and as far as I see there is no sincos-like method. The axial to planar transformation achieved here switches from a CTF to a neuromag layout, which might actually change the sensor positions, so even there was a sincos method, it might not result in the same topoplot you get with ft_megplanar. I guess you can define an info object that contains sensor position informations in a way that result in a fieldtrip-like sincos planar CTF output and read in this info object before calling _interpolate_to_meg() instead of the default neuromag one - but others probably know better, e.g. @larsoner ?

No IIRC we only support sourceproject equiv in MNE-Python. I think this is mostly because we haven't seen a need for it yet. @chsquare you need it for your analyses and expect it to give a different (maybe better) result than minimum-norm based transformation?

Thank you @chsquare and @larsoner for your replies. In fieldtrip we routinely compute the synthetic planar gradient because the strongest field is usually situated above the neural sources. It's a kind of "crude" source reconstruction that aligns very well across participants. No worries if nothing is implemented yet!

So would a evoked.interpolate_to("planar") that mapped axial to planar grad pair work for you? I don't think we need the sincos method for this as minimum norm could still get us there

Thank you @larsoner ! I didn't know this option and I'll definitely check it out

It doesn't exist yet, but I could add it. Would it take care of your use case?

As long as the synthetic planar gradient is generated taking into account the neighboring axials, that would cover my use case perfectly. In any case, if you implement this axial-to-planar interpolation, I can compare it to the FieldTrip sincos equivalent myself making a pull request in the following months.
Thanks a lot!