Automatic labeling of EEG electrodes using combinatorial optimization

Abstract

An important issue in electroencephalography (EEG) experiments is to measure accurately the three dimensional (3D) positions of electrodes. We propose a system where these positions are automatically estimated from several images using computer vision techniques. Yet, only a set of undifferentiated points are recovered this way and remains the problem of labeling them, i.e. of finding which electrode corresponds to each point. This paper proposes a fast and robust solution to this latter problem based on combinatorial optimization. We design a specific energy that we minimize with a modified version of the Loopy Belief Propagation algorithm. Experiments on real data show that, with our method, a manual labeling of two or three electrodes only is sufficient to get the complete labeling of a 64 electrodes cap in less than 10 seconds. However, it is shown to be robust to missing electrodes in the reconstructed data.

EEG–MRI Co-registration and Sensor Labeling Using a 3D Laser Scanner

“…This can be achieved during the digitization step by acquiring the sensors' coordinates in a given order. As this is a time costly and operator dependent procedure, several automatic techniques have also been developed: the combinatorial optimization based algorithm, 27 recognition of colored stickers in photogrammetric images, 4 and our method that uses a priori information on the relative positions of sensors. 15 In this work, we focus on the automatic co-registration of EEG sensors (included in a cap) with MRI data in a clinical context of high resolution EEG.…”

Semi-Automated and Direct Localization and Labeling of EEG Electrodes Using MR Structural Images for Simultaneous fMRI-EEG

“…Functional imaging in clinical applications (e.g., epileptic foci detection), neurofeedback and brain-computer interfaces (BCI) can greatly benefit from an accurate representation of the spatial location of the electrodes. Many methods have been proposed so far (Steddin and Botzel, 1995;Yoo et al, 1997;Le et al, 1998;Russel et al, 2005;Koessler et al, 2007;Péchaud et al, 2007;Marino et al, 2016;Butler et al, 2017;Clausner et al, 2017;Fleury et al, 2019;Homölle and Oostenveld, 2019;Taberna et al, 2019). However, most approaches require laborious manual intervention to prepare for the experiment or the use of special digitization devices.…”

“…The process of spatial localization of EEG electrodes involves two major steps: (1) correctly localizing and obtaining the 3D coordinates of each electrode, and (2) distinguishing each electrode by finding its proper label. There have been attempts to solve these two steps of the problem ( De Munck et al, 1991 ; Steddin and Botzel, 1995 ; Yoo et al, 1997 ; Le et al, 1998 ; Koessler et al, 2007 ; Péchaud et al, 2007 ); however, there are only a few approaches that have automated this process of mapping electrodes. In the following paragraphs, we summarize the existing manual, semi-automated, and automated methods of EEG localization and labeling.…”

“…The combinatorial optimization and self-calibration is a different semi-automated approach to localizing electrode positions ( Péchaud et al, 2007 ). In this method, the 3D coordinates of each electrode are reconstructed with a photogrammetry-based method of ten different pictures of the subject’s head from various angles.…”

“…Our MR-based method provides a direct way of solving this two-step problem of localization and labeling of the electrodes in a simultaneous EEG/MRI setup. With our approach, the user does not have the need for additional equipment (e.g., digitizers and cameras) or the need for specialized MR sequences (e.g., UTE sequences) to solve this two-step problem ( Steddin and Botzel, 1995 ; Yoo et al, 1997 ; Le et al, 1998 ; Russel et al, 2005 ; Koessler et al, 2007 ; Péchaud et al, 2007 ; Marino et al, 2016 ; Butler et al, 2017 ; Clausner et al, 2017 ; Fleury et al, 2019 ; Homölle and Oostenveld, 2019 ; Taberna et al, 2019 ). We take advantage of the fact that standard EEG electrodes can be seen in the standard high-resolution MRI structural images.…”