This repository hosts the code for an alternative backend for plotting 2D-Data with MNE-Python.

The backend is based on pyqtgraph which uses Qt's Graphics View Framework for the plotting. Development started as a 2021's Google Summer of Code Project.

Install full MNE-Python with the instructions provided here or install minimal MNE-Python with

pip install mne matplotlib mne-qt-browser

or

conda install -c conda-forge mne-base matplotlib mne-qt-browser

The backend supports plotting for the following MNE-Python methods:

• mne.io.Raw.plot()
• mne.Epochs.plot()
• mne.preprocessing.ICA.plot_sources(raw)
• mne.preprocessing.ICA.plot_sources(epochs)

In the following example, we'll read M/EEG raw data from the MNE sample dataset and plot it using the qt-backend. (For mne-version >= 1.0 the qt-backend will be the default)

from pathlib import Path
import mne

sample_dir = mne.datasets.sample.data_path()
raw_path = Path(sample_dir) / 'MEG' / 'sample' / 'sample_audvis_raw.fif'
raw = mne.io.read_raw(raw_path)

mne.viz.set_browser_backend('qt')  # Enable mne-qt-browser backend if mne < 1.0
raw.plot(block=True)

If the plot is not showing, search for solutions in the troubleshooting section below.

This will use the mne-qt-browser for the current Python session. If you want to make this change permanent, so you don't have to use the set_browser_backend() each time after restarting Python, run the following line to modify your MNE configuration file:

import mne
mne.set_config('MNE_BROWSER_BACKEND', 'qt')

If you are running a script containing raw.plot() like

python example_script.py

the plot will not stay open when the script is done.

To solve this either change raw.plot() to raw.plot(block=True) or run the script with the interactive flag

python -i example_script.py

If the integration of the Qt event loop is not activated for IPython, a plot with raw.plot() will freeze. Do avoid that either change raw.plot() to raw.plot(block=True) or activate the integration of the event loop with

%gui qt5

Please report bugs and feature requests in the issues of this repository.

You can run a benchmark locally with:

pytest -m benchmark mne_qt_browser

To run tests, clone mne-python, and then run the PyQtGraph tests with e.g.:

pytest -m pgtest ../mne-python/mne/viz/tests

If you do not have OpenGL installed, this will currently raise errors, and you'll need to add this line to mne/conftest.py after the error:: line:

    ignore:.*PyOpenGL was not found.*:RuntimeWarning