mrswolf / rasrmatlab

A Matlab Toolbox for correcting EEG artifacts using Riemannian Artifact Subspace Reconstruction.

Riemannian Artifact Subspace Reconstruction Matlab Toolbox

This repository contains code for an EEGLAB [0] plugin which can be used to correct EEG artifacts using Riemannian geometry. The method is an adaptation of the original ASR method published by Mullen et al. [1]. For details of its workings, please refer to Blum et al. [3]. The rASR toolbox uses the Manopt toolbox for computations involving manifolds and Riemannian geometry, please note that in the current implementation an adapted nonlinear_eigenspace function is used and included in the rASR toolbox, while the rest of manopt is used from the toolbox which must be in the Matlab path. This is only a temporary solution and will be coded more sustainably soon.

Usage

To use the rASR cleaning, simply call this plugin like the original ASR plugin after adding it to the Matlab path:

addpath('rASRfolder/', '-begin');    
% call clean_rawdata
EEG = clean_rawdata(EEG, arg_flatline, arg_highpass, arg_channel, arg_noisy, arg_burst, arg_window);

The rASR folder can also be put directly into the EEGLAB folder: eeglab/plugins/rASR/, this is where EEGLAB looks for the functions to clean the data.

The function will show a message in the command window of Matlab, informing that the cleaning is done with Riemannian geometry now.

Dependencies

The Riemannian geometry functions are implemented in the Manopt toolbox [2] which needs to be in Matlab's path as well.

Support

Please drop me a mail if you have any questions, fork this work to improve it or open an issue in this repository if you have any problems.

Paper and Analysis Code

rASR has been published in a peer-reviewed journal: https://www.frontiersin.org/articles/10.3389/fnhum.2019.00141/full?&utm_source=Email_to_authors_&utm_medium=Email&utm_content=T1_11.5e1_author&utm_campaign=Email_publication&field=&journalName=Frontiers_in_Human_Neuroscience&id=421678.

The analysis code for the paper can be found in the folder analysis_scripts_frontiers. Please also note the readme file in there. Data will be linked shortly.

References

[0] https://sccn.ucsd.edu/eeglab/index.php
[1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4710679/
[2] https://www.manopt.org/index.html
[3] https://www.frontiersin.org/articles/10.3389/fnhum.2019.00141/full?&utm_source=Email_to_authors_&utm_medium=Email&utm_content=T1_11.5e1_author&utm_campaign=Email_publication&field=&journalName=Frontiers_in_Human_Neuroscience&id=421678

License

This software is published under the MIT license.