autoreject

|CircleCI|_ |GitHub Actions|_ |Codecov|_ |PyPI|_ |Conda-Forge|_

.. |CircleCI| image:: https://circleci.com/gh/autoreject/autoreject/tree/main.svg?style=shield&circle-token=:circle-token .. _CircleCI: https://circleci.com/gh/autoreject/autoreject

.. |GitHub Actions| image:: https://github.com/autoreject/autoreject/actions/workflows/test.yml/badge.svg .. _GitHub Actions: https://github.com/autoreject/autoreject/actions/workflows/test.yml

.. |Codecov| image:: http://codecov.io/github/autoreject/autoreject/coverage.svg?branch=main .. _Codecov: http://codecov.io/github/autoreject/autoreject?branch=main

.. |PyPI| image:: https://badge.fury.io/py/autoreject.svg .. _PyPI: https://badge.fury.io/py/autoreject

.. |Conda-Forge| image:: https://img.shields.io/conda/vn/conda-forge/autoreject.svg .. _Conda-Forge: https://anaconda.org/conda-forge/autoreject/

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

.. image:: https://autoreject.github.io/stable/_images/sphx_glr_plot_auto_repair_001.png :width: 400

The documentation can be found under the following links:

• for the stable release <https://autoreject.github.io/stable/index.html>_
• for the latest (development) version <https://autoreject.github.io/dev/index.html>_

.. docs_readme_include_label

Installation

We recommend the Anaconda Python distribution <https://www.anaconda.com/>_ and a Python version >= 3.9. 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 <https://openneuro.org>_)

Quickstart

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

.. code:: python

>>> 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:

.. code:: python

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

We also implement RANSAC from the PREP pipeline <https://doi.org/10.3389/fninf.2015.00016>_ (see PyPREP <https://github.com/sappelhoff/pyprep>_ for a full implementation of the PREP pipeline). The API is the same:

.. code:: python

>>> 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 <https://autoreject.github.io/stable/_images/sphx_glr_plot_auto_repair_001.png>_.

Bug reports

Please use the GitHub issue tracker <https://github.com/autoreject/autoreject/issues>_ 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 <https://hal.archives-ouvertes.fr/hal-01313458/document>_." 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 <http://www.sciencedirect.com/science/article/pii/S1053811917305013>_". NeuroImage, 159, 417-429.