Localization of coherent sources by simultaneous MEG and EEG beamformer

XDL-ESI: Electrophysiological Sources Imaging via explainable deep learning framework with validation on simultaneous EEG and iEEG

Electroencephalography (EEG) or Magnetoencephalography (MEG) source imaging aims to estimate the underlying activated brain sources to explain the observed EEG/MEG recordings. Solving the inverse problem of EEG/MEG Source Imaging (ESI) is challenging due to its ill-posed nature. To achieve a unique solution, it is essential to apply sophisticated regularization constraints to restrict the solution space. Traditionally, the design of regularization terms is based on assumptions about the spatiotemporal structure of the underlying source dynamics. In this paper, we propose a novel paradigm for ESI via an Explainable Deep Learning framework, termed as XDL-ESI, which connects the iterative optimization algorithm with deep learning architecture by unfolding the iterative updates with neural network modules. The proposed framework has the advantages of (1) establishing a data-driven approach to model the source solution structure instead of using hand-crafted regularization terms; (2) improving the robustness of source solutions by introducing a topological loss that leverages the geometric spatial information applying varying penalties on distinct localization errors; (3) improving the reconstruction efficiency and interpretability as it inherits the advantages from both the iterative optimization algorithms (interpretability) and deep learning approaches (function approximation). The proposed XDL-ESI framework provides an efficient, accurate, and interpretable paradigm to solve the ESI inverse problem with satisfactory performance in both simulated data and real clinical data. Specially, this approach is further validated using simultaneous EEG and intracranial EEG (iEEG).

Multi-Modal Electrophysiological Source Imaging With Attention Neural Networks Based on Deep Fusion of EEG and MEG

The process of reconstructing underlying cortical and subcortical electrical activities from Electroencephalography (EEG) or Magnetoencephalography (MEG) recordings is called Electrophysiological Source Imaging (ESI). Given the complementarity between EEG and MEG in measuring radial and tangential cortical sources, combined EEG/MEG is considered beneficial in improving the reconstruction performance of ESI algorithms. Traditional algorithms mainly emphasize incorporating predesigned neurophysiological priors to solve the ESI problem. Deep learning frameworks aim to directly learn the mapping from scalp EEG/MEG measurements to the underlying brain source activities in a data-driven manner, demonstrating superior performance compared to traditional methods. However, most of the existing deep learning approaches for the ESI problem are performed on a single modality of EEG or MEG, meaning the complementarity of these two modalities has not been fully utilized. How to fuse the EEG and MEG in a more principled manner under the deep learning paradigm remains a challenging question. This study develops a Multi-Modal Deep Fusion (MMDF) framework using Attention Neural Networks (ANN) to fully leverage the complementary information between EEG and MEG for solving the ESI inverse problem, which is termed as MMDF-ANN. Specifically, our proposed brain source imaging approach consists of four phases, including feature extraction, weight generation, deep feature fusion, and source mapping. Our experimental results on both synthetic dataset and real dataset demonstrated that using a fusion of EEG and MEG can significantly improve the source localization accuracy compared to using a single-modality of EEG or MEG. Compared to the benchmark algorithms, MMDF-ANN demonstrated good stability when reconstructing sources with extended activation areas and situations of EEG/MEG measurements with a low signal-to-noise ratio.

Functional Neurophysiological Biomarkers of Early-Stage Alzheimer's Disease: A Perspective of Network Hyperexcitability in Disease Progression

Network hyperexcitability (NH) has recently been suggested as a potential neurophysiological indicator of Alzheimer’s disease (AD), as new, more accurate biomarkers of AD are sought. NH has generated interest as a potential indicator of certain stages in the disease trajectory and even as a disease mechanism by which network dysfunction could be modulated. NH has been demonstrated in several animal models of AD pathology and multiple lines of evidence point to the existence of NH in patients with AD, strongly supporting the physiological and clinical relevance of this readout. Several hypotheses have been put forward to explain the prevalence of NH in animal models through neurophysiological, biochemical, and imaging techniques. However, some of these hypotheses have been built on animal models with limitations and caveats that may have derived NH through other mechanisms or mechanisms without translational validity to sporadic AD patients, potentially leading to an erroneous conclusion of the underlying cause of NH occurring in patients with AD. In this review, we discuss the substantiation for NH in animal models of AD pathology and in human patients, as well as some of the hypotheses considering recently developed animal models that challenge existing hypotheses and mechanisms of NH. In addition, we provide a preclinical perspective on how the development of animal models incorporating AD-specific NH could provide physiologically relevant translational experimental data that may potentially aid the discovery and development of novel therapies for AD.

Reproducibility of EEG-MEG fusion source analysis of interictal spikes: Relevance in presurgical evaluation of epilepsy

Fusion of electroencephalography (EEG) and magnetoencephalography (MEG) data using maximum entropy on the mean method (MEM-fusion) takes advantage of the complementarities between EEG and MEG to improve localization accuracy. Simulation studies demonstrated MEM-fusion to be robust especially in noisy conditions such as single spike source localizations (SSSL). Our objective was to assess the reliability of SSSL using MEM-fusion on clinical data. We proposed to cluster SSSL results to find the most reliable and consistent source map from the reconstructed sources, the so-called consensus map. Thirty-four types of interictal epileptic discharges (IEDs) were analyzed from 26 patients with well-defined epileptogenic focus. SSSLs were performed on EEG, MEG, and fusion data and consensus maps were estimated using hierarchical clustering. Qualitative (spike-to-spike reproducibility rate, SSR) and quantitative (localization error and spatial dispersion) assessments were performed using the epileptogenic focus as clinical reference. Fusion SSSL provided significantly better results than EEG or MEG alone. Fusion found at least one cluster concordant with the clinical reference in all cases. This concordant cluster was always the one involving the highest number of spikes. Fusion yielded highest reproducibility (SSR EEG = 55%, MEG = 71%, fusion = 90%) and lowest localization error. Also, using only few channels from either modality (21EEG + 272MEG or 54EEG + 25MEG) was sufficient to reach accurate fusion. MEM-fusion with consensus map approach provides an objective way of finding the most reliable and concordant generators of IEDs. We, therefore, suggest the pertinence of SSSL using MEM-fusion as a valuable clinical tool for presurgical evaluation of epilepsy.

Code-modulated visual evoked potentials using fast stimulus presentation and spatiotemporal beamformer decoding

When encoding visual targets using various lagged versions of a pseudorandom binary sequence of luminance changes, the EEG signal recorded over the viewer’s occipital pole exhibits so-called code-modulated visual evoked potentials (cVEPs), the phase lags of which can be tied to these targets. The cVEP paradigm has enjoyed interest in the brain-computer interfacing (BCI) community for the reported high information transfer rates (ITR, in bits/min). In this study, we introduce a novel decoding algorithm based on spatiotemporal beamforming, and show that this algorithm is able to accurately identify the gazed target. Especially for a small number of repetitions of the coding sequence, our beamforming approach significantly outperforms an optimised support vector machine (SVM)-based classifier, which is considered state-of-the-art in cVEP-based BCI. In addition to the traditional 60 Hz stimulus presentation rate for the coding sequence, we also explore the 120 Hz rate, and show that the latter enables faster communication, with a maximal median ITR of 172.87 bits/min. Finally, we also report on a transition effect in the EEG signal following the onset of the stimulus sequence, and recommend to exclude the first 150 ms of the trials from decoding when relying on a single presentation of the stimulus sequence.

MEG-EEG Information Fusion and Electromagnetic Source Imaging: From Theory to Clinical Application in Epilepsy

The purpose of this study is to develop and quantitatively assess whether fusion of EEG and MEG (MEEG) data within the maximum entropy on the mean (MEM) framework increases the spatial accuracy of source localization, by yielding better recovery of the spatial extent and propagation pathway of the underlying generators of inter-ictal epileptic discharges (IEDs). The key element in this study is the integration of the complementary information from EEG and MEG data within the MEM framework. MEEG was compared with EEG and MEG when localizing single transient IEDs. The fusion approach was evaluated using realistic simulation models involving one or two spatially extended sources mimicking propagation patterns of IEDs. We also assessed the impact of the number of EEG electrodes required for an efficient EEG–MEG fusion. MEM was compared with minimum norm estimate, dynamic statistical parametric mapping, and standardized low-resolution electromagnetic tomography. The fusion approach was finally assessed on real epileptic data recorded from two patients showing IEDs simultaneously in EEG and MEG. Overall the localization of MEEG data using MEM provided better recovery of the source spatial extent, more sensitivity to the source depth and more accurate detection of the onset and propagation of IEDs than EEG or MEG alone. MEM was more accurate than the other methods. MEEG proved more robust than EEG and MEG for single IED localization in low signal-to-noise ratio conditions. We also showed that only few EEG electrodes are required to bring additional relevant information to MEG during MEM fusion.

Three-dimensional reconstruction of a cardiac outline by magnetocardiography

A 3-D cardiac visualization is significantly helpful toward clinical applications of magnetocardiography (MCG), but the cardiac reconstruction requires a segmentation process using additional image modalities. This paper proposes a 3-D cardiac outline reconstruction method using only MCG measurement data without further imaging techniques. The cardiac outline was reconstructed by a combination of both spatial filtering and coherence mapping method. The strength of cardiac activities was first estimated by the array-gain constraint minimum-norm spatial filter with recursively updated gram matrix (AGMN-RUG). Then, waveforms were reconstructed at whole source grids, and the maximum source points of an atrium and ventricle were selected as a reference, respectively. Next, the coherence between each maximum source point and whole source points was compared by the coherence mapping method. A reconstructed cardiac outline was validated by comparing with an overlapped volume ratio when the reconstructed volume was identically matched with the original volume. The results obtained by the AGMN-RUG were compared to the results by other spatial filters. The accuracy of numerical simulation and phantom experiment by the AGMN-RUG was superior 10% and 8%, respectively, than the accuracy by the standardized low-resolution electromagnetic tomography. This accuracy demonstrated the efficacy of the proposed 3-D cardiac reconstruction method.

Gamma band activity associated with BCI performance: simultaneous MEG/EEG study

While brain computer interface (BCI) can be employed with patients and healthy subjects, there are problems that must be resolved before BCI can be useful to the public. In the most popular motor imagery (MI) BCI system, a significant number of target users (called “BCI-Illiterates”) cannot modulate their neuronal signals sufficiently to use the BCI system. This causes performance variability among subjects and even among sessions within a subject. The mechanism of such BCI-Illiteracy and possible solutions still remain to be determined. Gamma oscillation is known to be involved in various fundamental brain functions, and may play a role in MI. In this study, we investigated the association of gamma activity with MI performance among subjects. Ten simultaneous MEG/EEG experiments were conducted; MI performance for each was estimated by EEG data, and the gamma activity associated with BCI performance was investigated with MEG data. Our results showed that gamma activity had a high positive correlation with MI performance in the prefrontal area. This trend was also found across sessions within one subject. In conclusion, gamma rhythms generated in the prefrontal area appear to play a critical role in BCI performance.