Skip to content

autoreject/autoreject

Repository files navigation

autoreject

CircleCI GitHub Actions Codecov PyPI Conda-Forge

This is a library to automatically reject bad trials and repair bad sensors in magneto-/electroencephalography (M/EEG) data.

https://autoreject.github.io/stable/_images/sphx_glr_plot_auto_repair_001.png

The documentation can be found under the following links:

Installation

We recommend the Anaconda Python distribution and a Python version >= 3.10. We furthermore recommend that you install autoreject into an isolated Python environment. To obtain the stable release of autoreject, you can use pip:

pip install -U autoreject

Or conda:

conda install -c conda-forge autoreject

If you want the latest (development) version of autoreject, use:

pip install https://github.com/autoreject/autoreject/archive/refs/heads/main.zip

To check if everything worked fine, you can do:

python -c 'import autoreject'

and it should not give any error messages.

Below, we list the dependencies for autoreject. All required dependencies are installed automatically when you install autoreject.

  • mne (>=1.5.0)
  • numpy (>=1.21.2)
  • scipy (>=1.7.1)
  • scikit-learn (>=1.0.0)
  • joblib
  • matplotlib (>=3.5.0)

Optional dependencies are:

  • openneuro-py (>= 2021.10.1, for fetching data from OpenNeuro.org)

Quickstart

The easiest way to get started is to copy the following three lines of code in your script:

>>> from autoreject import AutoReject
>>> ar = AutoReject()
>>> epochs_clean = ar.fit_transform(epochs)  # doctest: +SKIP

This will automatically clean an epochs object read in using MNE-Python. To get the rejection dictionary, simply do:

>>> from autoreject import get_rejection_threshold
>>> reject = get_rejection_threshold(epochs)  # doctest: +SKIP

We also implement RANSAC from the PREP pipeline (see PyPREP for a full implementation of the PREP pipeline). The API is the same:

>>> from autoreject import Ransac
>>> rsc = Ransac()
>>> epochs_clean = rsc.fit_transform(epochs)  # doctest: +SKIP

For more details check out the example to automatically detect and repair bad epochs.

Bug reports

Please use the GitHub issue tracker to report bugs.

Cite

[1] Mainak Jas, Denis Engemann, Federico Raimondo, Yousra Bekhti, and Alexandre Gramfort, "Automated rejection and repair of bad trials in MEG/EEG." In 6th International Workshop on Pattern Recognition in Neuroimaging (PRNI), 2016.

[2] Mainak Jas, Denis Engemann, Yousra Bekhti, Federico Raimondo, and Alexandre Gramfort. 2017. "Autoreject: Automated artifact rejection for MEG and EEG data". NeuroImage, 159, 417-429.