Mne-python: Plotting MNI transformed ECoG

[timonmerk]

For decoding analysis across patients it becomes relevant to mainly work with MNI transformed coordinates. Using this tutorial https://mne.tools/stable/auto_tutorials/misc/plot_ecog.html, instead of the native FreeSurfer file the 'fsaverage' MNI brain can thus be used:

sample_path = mne.datasets.sample.data_path()
subject = 'fsaverage'
subjects_dir = sample_path + '\\subjects'

I installed MNE according to https://mne.tools/dev/install/mne_python.html, and using the fsaverage and run through the tutorial https://mne.tools/stable/auto_tutorials/misc/plot_ecog.html.
At the line mne.viz.set_3d_view(fig, 200, 70, focalpoint=[0, -0.005, 0.03]) I get the following error:
"AttributeError: module 'mne.viz.backends._notebook' has no attribute '_set_3d_view'"

My mne.sys_info() looks as follows:
(base) C:\Users\ICN_admin\Documents\WRITE_BERLIN_BIDS>python -c "import mne; mne.sys_info()"
Platform: Windows-10-10.0.18362-SP0
Python: 3.7.4 (default, Aug 9 2019, 18:34:13) [MSC v.1915 64 bit (AMD64)]
Executable: C:\Users\ICN_admin\Anaconda3\python.exe
CPU: AMD64 Family 23 Model 8 Stepping 2, AuthenticAMD: 64 cores
Memory: 127.9 GB

mne: 0.21.0
numpy: 1.19.2 {blas=D:\a\1\s\numpy\build\openblas_info, lapack=D:\a\1\s\numpy\build\openblas_lapack_info}
scipy: 1.5.2
matplotlib: 3.1.1 {backend=Qt5Agg}

sklearn: 0.22.1
numba: 0.45.1
nibabel: 3.0.1
cupy: Not found
pandas: 0.25.1
dipy: Not found
mayavi: 4.7.2
pyvista: 0.26.1 {OpenGL 4.5.0 NVIDIA 432.00 via GeForce RTX 2070 SUPER/PCIe/SSE2}
vtk: 9.0.1
PyQt5: 5.9.2

Here the complete error message:

[larsoner]

@GuillaumeFavelier can you look into the bug here?

A separate issue is how we should make it clear how to use MNI-coordinate sEEG/ECoG data. @adam2392 do you think it would make sense to take the tutorial idea I proposed in #8364 and make the dataset we use have electrode coordinates in MNI space? It might allow us to show how to use MNI-coordinate data and sEEG data in the same tutorial.

[GuillaumeFavelier]

It seems that _set_3d_view() is not exposed in the notebook 3d backend so I opened https://github.com/mne-tools/mne-python/pull/8392

However, even using the standard pyvista backend, I got the following which I believe is a regression:

...
Setting channel info structure...
Creating raw.info structure...
Reading 0 ... 21657  =      0.000 ...    21.670 secs...
Plotting 0 ecog locations
...

v0.21.0 | master
--|--
|

[GuillaumeFavelier]

I tracked it to https://github.com/mne-tools/mne-python/pull/8190, let me know if you can reproduce @larsoner

[larsoner]

8190 just exposed a bug from #7747, see fix in #8393

[adam2392]

A separate issue is how we should make it clear how to use MNI-coordinate sEEG/ECoG data. @adam2392 do you think it would make sense to take the tutorial idea I proposed in #8364 and make the dataset we use have electrode coordinates in MNI space? It might allow us to show how to use MNI-coordinate data and sEEG data in the same tutorial.

I've been a bit mIa but I will try to start the series of PRs sometime. Just a disclaimer I'm not too familiar with using freesurfer brains yet so might need help along the way. Will begin with say a tutorial on "loading in SEEG data" that more or less looks EXACTLY like the ecog example.