Creating 3D topographic map using only EEG data and sensor positions.

Support & Discussions
1
  • MNE version: e.g. 1.11.0
  • operating system: e.g. Windows 11

I want to plot a 3D topomap using only EEG data and sensor position (64 channels). How to implement? I have no MEG data. Any resources will be highly appreciated. Thanks in advance.

2

I’m not sure we currently support that. I was initially thinking to call mne.make_field_map followed by mne.viz.plot_evoked_field, but we would need some sort of 3D mesh of a standard head to draw the topomap on, and we don’t have a good one for this purpose (the “fsaverage” template looks weird without a nose :wink:)

3

One could use the head from the “sample” dataset I suppose… but beware that this may be inaccurate.

import mne

# Use the "kiloword" EEG dataset for this example. It has 32 channels.
path = mne.datasets.kiloword.data_path()

# Read EEG epochs.
epochs = mne.read_epochs(path / "kword_metadata-epo.fif")

# Set channel positions based on the 10-5 system.
epochs.set_montage("standard_1005")

# Compute the ERP.
evoked = epochs.average()

# Use the "sample" head to draw the topomap on.
sample_path = mne.datasets.sample.data_path()
trans = sample_path / "MEG/sample/sample_audvis_raw-trans.fif"

# Create the 3D topomap.
fieldmap = mne.make_field_map(
    evoked,
    subject="sample",
    subjects_dir=sample_path / "subjects",
    trans=trans,
)
evoked.plot_field(fieldmap)

image
image1252Ă—1348 152 KB

4

Thanks a ton! It works. I had to apply simple scaling in my case, as my data was not in volts. However, I do see some discrepancies between the topomap and the 3D field map; they are not exactly similar. I may not fully understand the underlying process for generating the 3D field map. Could you enlighten me?

Screenshot 2026-01-06 160927
Screenshot 2026-01-06 160927720Ă—916 104 KB

5

they are not similar, and I warned you it may be inaccurate, because the standard EEG electrodes positions follow a perfect sphere. However, the head is not a perfect sphere. The EEG positions get projected somewhat awkwardly on the head of this particular subject, which also probably doesn’t match the head of the actual subject. I think that, and of course the fact that a different color map is used, explains most of the discrepancy.

1 Like
6

I think a word of caution is in order here: do not trust the 3d topomap we are cobbling together here. Do not use this for a figure in a scientific publication.

1 Like