Generalized Eigenvalue De-Artifacting Instrument (GEDAI)

Authors: Tomas Ros, Yingqi Huang, Takfarinas Medani, Raymundo Cassani

This tutorial introduces the generalized eigenvalue de-artifacting instrument (GEDAI) algorithm which is an unsupervised EEG denoising method based on the leadfield filtering.

Introduction

EEG signals may be considered to be a mixture of electrical activities from a brain “signal” (sub)space, and one containing different types of non-cerebral noise or “artifacts”. This mixture may be “unmixed” by linear decomposition techniques (e.g. PCA or ICA) into separate "components" with individual source locations and respective time-courses. However, although PCA and ICA leverage statistical properties within mixed data to recover underlying sources, they are "blind" source separation methods, functioning without a priori knowledge of the original signals or their mixing process. GEDAI combines theoretical knowledge of the brain’s “signal” subspace with generalized eigenvalue decomposition (GEVD) to automatically separate brain and artifact components. Here, a theoretical model of EEG generation is used as an estimate of the brain’s “noise-free” subspace. An overview of This is shown in the Panel A of the figure below.

Panels B, C and D show these steps at more detail. Panel B: each data covariance matrix (dataCOV) is decomposed into source components with GEVD. Panel C: The GEVD uses a fixed theoretical reference matrix (refCOV) across all epochs, based on the leadfield matrix of an EEG forward model. Panel D: To determine the optimal threshold separating brain and artifactual subspaces, output EEG data is evaluated using the Signal & Noise Subspace Alignment Index (SENSAI). This is done by respectively maximizing and minimizing the subspace similarities of the retained "signal" and removed "noise" with the refCOV.

Install

Being a Brainstorm plugin, GEDAI plugin can be installed, updated and removed directly from the Brainstorm GUI. For further information, see the plugins tutorial.

• From the main window go to Plugins > Artifacts > gedai > Install.

  

• A message will appear saying, "Plugin gedai is not installed on your computer. Download the latest version of GEDAI now?" click 'Yes'.

• The plugin will be downloaded and installed automatically. Once installed, you will see a confirmation message.

  

• The README file will appear; after thoroughly reviewing it, click on the "I agree" button to confirm your acceptance of the plugin's terms and conditions.

By following these steps, you will successfully install the GEDAI plugin.

Using GEDAI

Now that GEDAI has been installed as a plugin, we can use it on the example data that is provided by the GEDAI team.

• Download the example file empirical_NOISE_EOG_EMG.set in the GEDAI repository.
  Direct download link

• Select from the menu File > Create new protocol Name the new protocol TutorialGEDAI and select the options:

  • Yes, use protocol's default anatomy, No, use one channel file per acquisition run.

• Switch now to the Functional data view of your database contents (), then right-click on the newly created TutorialGEDAI node > New subject > Subject01.

Keep the default options defined for the protocol.

• Importing data
• Install GEDAI plugin
• Running GEDAI process Description of the process GUI and each of their elements
• Results

On epilepsy EEG

• We can use GEDAI on the raw EEG data in the Epilepsy tutorial

  https://neuroimage.usc.edu/brainstorm/Tutorials/Epilepsy

Additional documentation

Related tutorials

• Artifact detection

• Artifact cleaning with SSP

Articles

* Ros, T., Férat, V., Huang, Y., Colangelo, C., Kia, S. M., Wolfers, T., Vulliemoz, S., & Michela, A. (2025).
Return of the GEDAI: Unsupervised EEG denoising based on leadfield filtering.
bioRxiv : The Preprint Server for Biology.