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.

Nocturnal paroxysmal dystonia to sleep-related hypermotor epilepsy: A critical review.

Sleep-related paroxysmal motor episodes (SPMEs) have been described by various names, including nocturnal paroxysmal dystonia, nocturnal frontal lobe epilepsy (NFLE), and sleep-related hypermotor epilepsy. The underlying pathophysiology has been debated over the years, with these episodes assumed to be a form of paroxysmal dystonia or parasomnia versus a form of epilepsy. In most studies published on SPMEs and their variants (paroxysmal arousals, nocturnal paroxysmal dystonia, and episodic nocturnal wanderings) in the early 1990s, the authors speculated on the pathophysiology but did not commit to one idea. It was not until the mid-1990s that epilepsy became the leading prospect. We performed a narrative review of the major articles that have described this syndrome in a chronological fashion. We identified three eras, 1972-1993, 1994-1998, and 1999 to the present, each era marked by a landmark study. Our critical review of these early studies shows that the neurophysiological data supporting epilepsy as the sole basis for all SPME cases is very weak. In 1994-1995, a familial pattern of this syndrome was described and the term autosomal dominant NFLE was coined, with the authors claiming that all their patients had a form of frontal lobe epilepsy. With the exception of a few reference cases, the neurophysiological evidence that all patients had frontal lobe epilepsy was very weak. Compared to articles published on surgical series of frontal lobe epilepsy, the percentage of SPME cases with positive interictal/ictal electroencephalograms remained very low, seriously questioning the epileptic basis of the syndrome. Our critical review and analysis of the published literature shows that the evidence presented in favor of SPMEs being a homogenous focal epilepsy syndrome is very weak. Neurologists must recognize that SPMEs could be a form of movement disorder, parasomnia, or epilepsy. We recommend a pragmatic semiology-based classification of these episodes using the four-dimensional classification system.

County-Level Socio-Environmental Factors Associated With Stroke Mortality in the United States: A Cross-Sectional Study.

We used machine learning methods to explore sociodemographic and environmental determinants of health (SEDH) associated with county-level stroke mortality in the USA. We conducted a cross-sectional analysis of individuals aged ≥15 years who died from all stroke subtypes between 2016 and 2020. We analyzed 54 county-level SEDH possibly associated with age-adjusted stroke mortality rates/100,000 people. Classification and Regression Tree (CART) was used to identify specific county-level clusters associated with stroke mortality. Variable importance was assessed using Random Forest analysis. A total of 501,391 decedents from 2397 counties were included. CART identified 10 clusters, with 77.5% relative increase in stroke mortality rates across the spectrum (28.5 vs 50.7 per 100,000 persons). CART identified 8 SEDH to guide the classification of the county clusters. Including, annual Median Household Income ($), live births with Low Birthweight (%), current adult Smokers (%), adults reporting Severe Housing Problems (%), adequate Access to Exercise (%), adults reporting Physical Inactivity (%), adults with diagnosed Diabetes (%), and adults reporting Excessive Drinking (%). In conclusion, SEDH exposures have a complex relationship with stroke. Machine learning approaches can help deconstruct this relationship and demonstrate associations that allow improved understanding of the socio-environmental drivers of stroke and development of targeted interventions.

Semiology and Neurophysiology of Clonic Seizures: A Report of 39 Patients.

Background and Objectives: Clonic seizures are currently defined as repetitive and rhythmic myoclonic contractions of a specific body part, producing twitching movements at a frequency of 0.2-5 Hz. There are few studies in the literature that have reported a detailed analysis of the semiology, neurophysiology, and lateralizing value of clonic seizures. In this article, we aim to report our findings from a retrospective review of 39 patients. Methods: We identified 39 patients (48 seizures) from our center who had been admitted with clonic seizures between 2016 and 2022. We performed a retrospective review of their video-EEG recordings for semiology and ictal EEG findings. Seventeen patients also had simultaneous surface-EMG (sEMG) electrodes placed on affected body parts, which were analyzed as well. Results: The most common initial affected body parts were face, arm, and hand. In most of the cases, seizures propagated from lower face to upper face and distal hand to proximal arm. The most common seizure-onset zone was the perirolandic region, and the most common EEG seizure pattern was paroxysmal rhythmic monomorphic activity. The lateralizing value for EEG seizure onset to contralateral hemisphere in unilateral clonic seizures (n = 39) was 100%. All seizures recorded with sEMG electrodes demonstrated synchronous brief tetanic contractions of agonists and antagonists, alternating with synchronous silent periods. Arrhythmic clonic seizures were associated with periodic epileptiform discharges on the EEG, whereas rhythmic clonic seizures were associated with paroxysmal rhythmic monomorphic activity. Overall, the most common etiology was cerebrovascular injuries, followed by tumors. Discussion: Clonic seizures are characterized by synchronized brief tetanic contractions of agonist and antagonistic muscles alternating with synchronized silent periods, giving rise to the visible twitching. The most common seizure onset zone is in the perirolandic region, which is consistent with the symptomatogenic zone being in the primary motor area. The lateralizing value of unilateral clonic seizures for seizure onset in the contralateral hemisphere is 100%.

Rejuvenating classical brain electrophysiology source localization methods with spatial graph Fourier filters for source extents estimation.

EEG/MEG source imaging (ESI) aims to find the underlying brain sources to explain the observed EEG or MEG measurement. Multiple classical approaches have been proposed to solve the ESI problem based on different neurophysiological assumptions. To support clinical decision-making, it is important to estimate not only the exact location of the source signal but also the extended source activation regions. Existing methods may render over-diffuse or sparse solutions, which limit the source extent estimation accuracy. In this work, we leverage the graph structures defined in the 3D mesh of the brain and the spatial graph Fourier transform (GFT) to decompose the spatial graph structure into sub-spaces of low-, medium-, and high-frequency basis. We propose to use the low-frequency basis of spatial graph filters to approximate the extended areas of brain activation and embed the GFT into the classical ESI methods. We validated the classical source localization methods with the corresponding improved version using GFT in both synthetic data and real data. We found the proposed method can effectively reconstruct focal source patterns and significantly improve the performance compared to the classical algorithms.

EEG based oculographic analysis of epileptic nystagmus.

Epileptic nystagmus (EN) is a subtle seizure semiology, most commonly seen in seizures originating in the posterior cortical regions. EN is broadly categorized into type I and type II. Type I EN consists of contralateral repetitive saccadic eye movements alternating with post-saccadic slow drifts with an overall contralateral deviation. Type II EN is characterized by ipsilateral slow drift alternating with contralateral corrective saccades. In this article, we report a method to perform oculographic analysis of eye movements using EEG only. We used this method to classify the type of EN in three patients with parieto-occipital seizures. In all three patients, the ictal EEG demonstrated repetitive saccadic eye movements, directed contralateral to the seizure onset zone. With prolonged time constant, we were able to identify this eye movement pattern as EN with distinct slow and fast phases. We were able to further characterize the type of EN as type I and type II. In all three patients, the direction of EN (direction of fast phase or saccades) was contralateral to the seizure onset zone. EN can be easily missed on video-electroencephalography (vEEG) recordings because of various reasons. Our study demonstrates a systematic method of eye movement analysis on EEG, which can be used to not only identify EN as seizure semiology but also classify it, without requiring additional electrodes.