Supplementary material to the scientific publication "HArtMuT - Modeling eye and muscle contributors in neuroelectric imaging" in Journal of Neural Engineering 19(6), 2022.

Correctly localizing and identifying ICA patterns as brain or non-brain sources relies on head models that so far only take brain sources into account.
We developed the Head Artefact Model using Tripoles (HArtMuT), a volume conduction head model, that includes, besides brain sources, eyes and muscles that can be modeled as single dipoles, symmetrical dipoles, and tripoles.
We compared HArtMuT with the EEGLAB standard head model on their localization accuracy and residual variance (RV) using firstly, a Finite Element Model (FEM) as ground truth, and secondly, real-world data of mobile participants, and found that HArtMuT improves localization for all sources, especially non-brain.

For more details, feel free to visit https://www.hartmut.berlin.

The final HArtMuT (BEM- and FEM-models) including cortical and artefactual sourcemodels, leadfields and labels.

A HArtMuT parcellation atlas for the purpose of labeling within source localization routines.

Artefact model warping for individual head geometries.

HArtMuT is already part of SEEREGA and UnfoldSim.jl, where it can be used to simulate artefacts in synthetic EEG data.
We are currently working on integrating HArtMuT into standard neuoscience pipelines for source localization: OpenMEEG, FieldTrip toolbox, DIPFIT (EEGLAB plug-in) planned. Details will follow soon.

If you find HArtMuT useful for your research, please consider citing our related paper:

@article{Harmening_2022,
      author  = {Harmening, Nils and
                 Klug, Marius and
                 Gramann, Klaus and
                 Miklody, Daniel},
      title   = {HArtMuT - Modeling eye and muscle contributors in neuroelectric imaging},
      year    = {2022},
      doi     = {10.1088/1741-2552/aca8ce},
      journal = {Journal of Neural Engineering}
      volume  = {19},
      number  = {6},
      pages   = {066041},
}