EMS-Net: A Deep Learning Method for Autodetecting Epileptic Magnetoencephalography Spikes

Deep learning based automatic detection and dipole estimation of epileptic discharges in MEG: a multi-center study

Magnetoencephalography (MEG) provides crucial information in diagnosing focal epilepsy. However, dipole estimation, a commonly used analysis method for MEG, can be time-consuming since it necessitates neurophysiologists to manually identify epileptic spikes. To reduce this burden, we developed the automatic detection of spikes using deep learning in single center. In this study, we performed a multi-center study using six MEG centers to improve the performance of the automated detection of neuromagnetically recorded epileptic spikes, which we previously developed using deep learning. Data from four centers were used for training and evaluation (internal data), and the remaining two centers were used for evaluation only (external data). We used a five-fold subject-wise cross-validation technique to train and evaluate the models. A comparison showed that the multi-center model outperformed the single-center model in terms of performance. The multi-center model achieved an average ROC-AUC of 0.9929 and 0.9426 for the internal and external data, respectively. The median distance between the neurophysiologist-analyzed and automatically analyzed dipoles was 4.36 and 7.23 mm for the multi-center model for internal and external data, respectively, indicating accurate detection of epileptic spikes. By training data from multiple centers, automated analysis can improve spike detection and reduce the analysis workload for neurophysiologists. This study suggests that the multi-center model has the potential to detect spikes within 1 cm of a neurophysiologist's analysis. This multi-center model can significantly reduce the number of hours required by neurophysiologists to detect spikes.

Lightweight and wearable magnetoencephalography system based on spatially-grid constrained coils and compact magnetically shielded room

Magnetoencephalography based on optically pumped magnetometers can passively detect the ultra-weak brain magnetic field signals, which has significant clinical application prospects for the diagnosis and treatment of cerebral disorders. This paper proposes a brain magnetic signal measurement method on the basis of the active-passive coupling magnetic shielding strategy and helmet-mounted detection array, which has lower cost and comparable performance over the existing ones. We first utilized the spatially-grid constrained coils and biplanar coils with proportion-integration-differentiation controller with tracking differentiator to ensure a near-zero and stable magnetic field environment with large uniform region. Subsequently, we implemented the brain magnetic signal measurement with the subject randomly moving fingers through tapping a keyboard and with the condition of opening and closing the eyes. Effectively induced brain magnetic signals were detected at the motor functional area and occipital lobe area in the two experiments, respectively. The proposed method will contribute to the development of functional brain imaging.

Magnetoencephalography (MEG) Data Processing in Epilepsy Patients with Implanted Responsive Neurostimulation (RNS) Devices

Drug-resistant epilepsy (DRE) is often treated with surgery or neuromodulation. Specifically, responsive neurostimulation (RNS) is a widely used therapy that is programmed to detect abnormal brain activity and intervene with tailored stimulation. Despite the success of RNS, some patients require further interventions. However, having an RNS device in situ is a hindrance to the performance of neuroimaging techniques. Magnetoencephalography (MEG), a non-invasive neurophysiologic and functional imaging technique, aids epilepsy assessment and surgery planning. MEG performed post-RNS is complicated by signal distortions. This study proposes an independent component analysis (ICA)-based approach to enhance MEG signal quality, facilitating improved assessment for epilepsy patients with implanted RNS devices. Three epilepsy patients, two with RNS implants and one without, underwent MEG scans. Preprocessing included temporal signal space separation (tSSS) and an automated ICA-based approach with MNE-Python. Power spectral density (PSD) and signal-to-noise ratio (SNR) were analyzed, and MEG dipole analysis was conducted using single equivalent current dipole (SECD) modeling. The ICA-based noise removal preprocessing method substantially improved the signal-to-noise ratio (SNR) for MEG data from epilepsy patients with implanted RNS devices. Qualitative assessment confirmed enhanced signal readability and improved MEG dipole analysis. ICA-based processing markedly enhanced MEG data quality in RNS patients, emphasizing its clinical relevance.

Optical Co-Registration Method of Triaxial OPM-MEG and MRI

The advent of optically pumped magnetometers (OPMs) facilitates the development of on-scalp magnetoencephalography (MEG). In particular, the triaxial OPM emerged recently, making simultaneous measurements of all three orthogonal components of vector fields possible. The detection of triaxial magnetic fields improves the interference suppression capability and achieves higher source localization accuracy using fewer sensors. The source localization accuracy of MEG is based on the accurate co-registration of MEG and MRI. In this study, we proposed a triaxial co-registration method according to combined principal component analysis and iterative closest point algorithms for use of a flexible cap. A reference phantom with known sensor positions and orientations was designed and constructed to evaluate the accuracy of the proposed method. Experiments showed that the average co-registered position errors of all sensors were approximately 1 mm and average orientation errors were less than 2.5° in the X -and Y orientations and less than 1.6° in the Z orientation. Furthermore, we assessed the influence of co-registration errors on the source localization using simulations. The average source localization error of approximately 1 mm reflects the effectiveness of the co-registration method. The proposed co-registration method facilitates future applications of triaxial sensors on flexible caps.

Deep learning in neuroimaging of epilepsy

In recent years, artificial intelligence, particularly deep learning (DL), has demonstrated utility in diverse areas of medicine. DL uses neural networks to automatically learn features from the raw data while this is not possible with conventional machine learning. It is helpful for the assessment of patients with epilepsy and whilst most published studies have been aimed at the automatic detection and prediction of seizures from electroencephalographic records, there is a growing number of investigations that use neuroimaging modalities (structural and functional magnetic resonance imaging, diffusion-weighted imaging and positron emission tomography) as input data. We review the application of DL to neuroimaging (sMRI, fMRI, DWI and PET) of focal epilepsy, specifically presurgical evaluation of drug-refractory epilepsy. First, a brief theoretical overview of artificial neural networks and deep learning is presented. Next, we review applications of deep learning to neuroimaging of epilepsy: diagnosis and lateralization, automated detection of lesion, presurgical evaluation and prediction of postsurgical outcome. Finally, the limitations, challenges and possible future directions in the application of these methods in the study of epilepsies are discussed. This approach could become an essential tool in clinical practice, particularly in the evaluation of images considered negative by visual inspection, in individualized treatments, and in the approach to epilepsy as a network disorder. However, greater multicenter collaboration is required to achieve the collection of sufficient data with the required quality together with the open access availability of the developed codes and tools.

An artificial intelligence-based pipeline for automated detection and localisation of epileptic sources from magnetoencephalography

Objective.Magnetoencephalography (MEG) is a powerful non-invasive diagnostic modality for presurgical epilepsy evaluation. However, the clinical utility of MEG mapping for localising epileptic foci is limited by its low efficiency, high labour requirements, and considerable interoperator variability. To address these obstacles, we proposed a novel artificial intelligence-based automated magnetic source imaging (AMSI) pipeline for automated detection and localisation of epileptic sources from MEG data.Approach.To expedite the analysis of clinical MEG data from patients with epilepsy and reduce human bias, we developed an autolabelling method, a deep-learning model based on convolutional neural networks and a hierarchical clustering method based on a perceptual hash algorithm, to enable the coregistration of MEG and magnetic resonance imaging, the detection and clustering of epileptic activity, and the localisation of epileptic sources in a highly automated manner. We tested the capability of the AMSI pipeline by assessing MEG data from 48 epilepsy patients.Main results.The AMSI pipeline was able to rapidly detect interictal epileptiform discharges with 93.31% ± 3.87% precision based on a 35-patient dataset (with sevenfold patientwise cross-validation) and robustly rendered accurate localisation of epileptic activity with a lobar concordance of 87.18% against interictal and ictal stereo-electroencephalography findings in a 13-patient dataset. We also showed that the AMSI pipeline accomplishes the necessary processes and delivers objective results within a much shorter time frame (∼12 min) than traditional manual processes (∼4 h).Significance.The AMSI pipeline promises to facilitate increased utilisation of MEG data in the clinical analysis of patients with epilepsy.

Magnetoencephalography-based approaches to epilepsy classification

Epilepsy is a chronic central nervous system disorder characterized by recurrent seizures. Not only does epilepsy severely affect the daily life of the patient, but the risk of premature death in patients with epilepsy is three times higher than that of the normal population. Magnetoencephalography (MEG) is a non-invasive, high temporal and spatial resolution electrophysiological data that provides a valid basis for epilepsy diagnosis, and used in clinical practice to locate epileptic foci in patients with epilepsy. It has been shown that MEG helps to identify MRI-negative epilepsy, contributes to clinical decision-making in recurrent seizures after previous epilepsy surgery, that interictal MEG can provide additional localization information than scalp EEG, and complete excision of the stimulation area defined by the MEG has prognostic significance for postoperative seizure control. However, due to the complexity of the MEG signal, it is often difficult to identify subtle but critical changes in MEG through visual inspection, opening up an important area of research for biomedical engineers to investigate and implement intelligent algorithms for epilepsy recognition. At the same time, the use of manual markers requires significant time and labor costs, necessitating the development and use of computer-aided diagnosis (CAD) systems that use classifiers to automatically identify abnormal activity. In this review, we discuss in detail the results of applying various different feature extraction methods on MEG signals with different classifiers for epilepsy detection, subtype determination, and laterality classification. Finally, we also briefly look at the prospects of using MEG for epilepsy-assisted localization (spike detection, high-frequency oscillation detection) due to the unique advantages of MEG for functional area localization in epilepsy, and discuss the limitation of current research status and suggestions for future research. Overall, it is hoped that our review will facilitate the reader to quickly gain a general understanding of the problem of MEG-based epilepsy classification and provide ideas and directions for subsequent research.

Data-driven approach for the delineation of the irritative zone in epilepsy in MEG

The reliable identification of the irritative zone (IZ) is a prerequisite for the correct clinical evaluation of medically refractory patients affected by epilepsy. Given the complexity of MEG data, visual analysis of epileptiform neurophysiological activity is highly time consuming and might leave clinically relevant information undetected. We recorded and analyzed the interictal activity from seven patients affected by epilepsy (Vectorview Neuromag), who successfully underwent epilepsy surgery (Engel > = II). We visually marked and localized characteristic epileptiform activity (VIS). We implemented a two-stage pipeline for the detection of interictal spikes and the delineation of the IZ. First, we detected candidate events from peaky ICA components, and then clustered events around spatio-temporal patterns identified by convolutional sparse coding. We used the average of clustered events to create IZ maps computed at the amplitude peak (PEAK), and at the 50% of the peak ascending slope (SLOPE). We validated our approach by computing the distance of the estimated IZ (VIS, SLOPE and PEAK) from the border of the surgically resected area (RA). We identified 25 spatiotemporal patterns mimicking the underlying interictal activity (3.6 clusters/patient). Each cluster was populated on average by 22.1 [15.0–31.0] spikes. The predicted IZ maps had an average distance from the resection margin of 8.4 ± 9.3 mm for visual analysis, 12.0 ± 16.5 mm for SLOPE and 22.7 ±. 16.4 mm for PEAK. The consideration of the source spread at the ascending slope provided an IZ closer to RA and resembled the analysis of an expert observer. We validated here the performance of a data-driven approach for the automated detection of interictal spikes and delineation of the IZ. This computational framework provides the basis for reproducible and bias-free analysis of MEG recordings in epilepsy.

Fully-Automated Spike Detection and Dipole Analysis of Epileptic MEG Using Deep Learning

Magnetoencephalography (MEG) is a useful tool for clinically evaluating the localization of interictal spikes. Neurophysiologists visually identify spikes from the MEG waveforms and estimate the equivalent current dipoles (ECD). However, presently, these analyses are manually performed by neurophysiologists and are time-consuming. Another problem is that spike identification from MEG waveforms largely depends on neurophysiologists' skills and experiences. These problems cause poor cost-effectiveness in clinical MEG examination. To overcome these problems, we fully automated spike identification and ECD estimation using a deep learning approach fully automated AI-based MEG interictal epileptiform discharge identification and ECD estimation (FAMED). We applied a semantic segmentation method, which is an image processing technique, to identify the appropriate times between spike onset and peak and to select appropriate sensors for ECD estimation. FAMED was trained and evaluated using clinical MEG data acquired from 375 patients. FAMED training was performed in two stages: in the first stage, a classification network was learned, and in the second stage, a segmentation network that extended the classification network was learned. The classification network had a mean AUC of 0.9868 (10-fold patient-wise cross-validation); the sensitivity and specificity were 0.7952 and 0.9971, respectively. The median distance between the ECDs estimated by the neurophysiologists and those using FAMED was 0.63 cm. Thus, the performance of FAMED is comparable to that of neurophysiologists, and it can contribute to the efficiency and consistency of MEG ECD analysis.

Multilevel Feature Learning Method for Accurate Interictal Epileptiform Spike Detection

Interictal epileptiform spike (referred to as spike) detected from electroencephalograms lasting only 20- to 200-ms can provide a reliable evidence-based indicator for clinical seizure type diagnosis. Recent feature representation approaches focus either on the concrete-level or on abstract-level information mining of the spike, thus demonstrating suboptimal detection performance. Additionally, existing abstract-level information mining methods of the spike based deep learning networks have not realized the effective feature representation of long-term dependent distinguished information within similar waveform cycles caused by morphological heterogeneity, which affects detection performance. Thus, a multilevel feature learning method for accurate spike detection was proposed in this study. Specifically, the spatio-temporal-frequency multidomain information in concrete-level first are inferred the common mimetic properties of the spike using the multidomain feature extractors. Then, the effective feature representation of long-term dependent distinguished information within similar waveform cycles caused by morphological heterogeneity is suitably captured using the temporal convolutional network. Finally, the spatio-temporal-frequency multidomain long-term dependent feature representation of spike is calculated using the element-wise manner to fuse the feature representation in concrete- and abstract-levels. The experimental results indicate that the proposed method can achieve an accuracy of 90.62±1.38%, sensitivity of 90.38±1.52%, specificity of 91.00±1.60%, precision of 90.33±4.71%, and the false detection rate per minute is 0.148±0.020m^-1, which are higher than when using the feature representation in the concrete- or abstract-level alone. Additionally, the detection results indicate that the proposed method avoids the subjectivity and inefficiency of visual inspection, and it enables a highly accurate detection of the spike.

Informed MEG/EEG source imaging reveals the locations of interictal spikes missed by SEEG

Determining the accurate locations of interictal spikes has been fundamental in the presurgical evaluation of epilepsy surgery. Stereo-electroencephalography (SEEG) is able to directly record cortical activity and localize interictal spikes. However, the main caveat of SEEG techniques is that they have limited spatial sampling (covering <5% of the whole brain), which may lead to missed spikes originating from brain regions that were not covered by SEEG. To address this problem, we propose a SEEG-informed minimum-norm estimates (SIMNE) method by combining SEEG with magnetoencephalography (MEG) or EEG. Specifically, the spike locations determined by SEEG offer as a priori information to guide MEG source reconstruction. Both computer simulations and experiments using data from five epilepsy patients were conducted to evaluate the performance of SIMNE. Our results demonstrate that SIMNE generates more accurate source estimation than a traditional minimum-norm estimates method and reveals the locations of spikes missed by SEEG, which would improve presurgical evaluation of the epileptogenic zone.

Sensitivity of a 29-Channel MEG Source Montage

In this paper, we study the performance of a source montage corresponding to 29 brain regions reconstructed from whole-head magnetoencephalographic (MEG) recordings, with the aim of facilitating the review of MEG data containing epileptiform discharges. Test data were obtained by superposing simulated signals from 100-nAm dipolar sources to a resting state MEG recording from a healthy subject. Simulated sources were placed systematically to different cortical locations for defining the optimal regularization for the source montage reconstruction and for assessing the detectability of the source activity from the 29-channel MEG source montage. The signal-to-noise ratio (SNR), computed for each source from the sensor-level and source-montage signals, was used as the evaluation parameter. Without regularization, the SNR from the simulated sources was larger in the sensor-level signals than in the source montage reconstructions. Setting the regularization to 2% increased the source montage SNR to the same level as the sensor-level SNR, improving the detectability of the simulated events from the source montage reconstruction. Sources producing a SNR of at least 15 dB were visually detectable from the source-montage signals. Such sources are located closer than about 75 mm from the MEG sensors, in practice covering all areas in the grey matter. The 29-channel source montage creates more focal signals compared to the sensor space and can significantly shorten the detection time of epileptiform MEG discharges for focus localization.

Categorizing objects from MEG signals using EEGNet

Magnetoencephalography (MEG) signals have demonstrated their practical application to reading human minds. Current neural decoding studies have made great progress to build subject-wise decoding models to extract and discriminate the temporal/spatial features in neural signals. In this paper, we used a compact convolutional neural network-EEGNet-to build a common decoder across subjects, which deciphered the categories of objects (faces, tools, animals, and scenes) from MEG data. This study investigated the influence of the spatiotemporal structure of MEG on EEGNet's classification performance. Furthermore, the EEGNet replaced its convolution layers with two sets of parallel convolution structures to extract the spatial and temporal features simultaneously. Our results showed that the organization of MEG data fed into the EEGNet has an effect on EEGNet classification accuracy, and the parallel convolution structures in EEGNet are beneficial to extracting and fusing spatial and temporal MEG features. The classification accuracy demonstrated that the EEGNet succeeds in building the common decoder model across subjects, and outperforms several state-of-the-art feature fusing methods.

A deep learning framework with multi-perspective fusion for interictal epileptiform discharges detection in scalp electroencephalogram

Objective.Interictal epileptiform discharges (IEDs) are an important and widely accepted biomarker used in the diagnosis of epilepsy based on scalp electroencephalography (EEG). Because the visual detection of IEDs has various limitations, including high time consumption and high subjectivity, a faster, more robust, and automated IED detector is strongly in demand.Approach.Based on deep learning, we proposed an end-to-end framework with multi-scale morphologic features in the time domain and correlation in sensor space to recognize IEDs from raw scalp EEG.Main Results.Based on a balanced dataset of 30 patients with epilepsy, the results of the five-fold (leave-6-patients-out) cross-validation shows that our model achieved state-of-the-art detection performance (accuracy: 0.951, precision: 0.973, sensitivity: 0.938, specificity: 0.968, F1 score: 0.954, AUC: 0.973). Furthermore, our model maintained excellent IED detection rates in an independent test on three datasets.Significance.The proposed model could be used to assist neurologists in clinical EEG interpretation of patients with epilepsy. Additionally, this approach combines multi-level output and correlation among EEG sensors and provides new ideas for epileptic biomarker detection in scalp EEG.

Enhanced Fast-VESTAL for Magnetoencephalography Source Imaging: From Theory to Clinical Application in Epilepsy

A novel magnetoencephalography source imaging approach called Fast Vector-based Spatio-Temporal Analysis (Fast-VESTAL) has been successfully applied in creating source images from evoked and resting-state data from both healthy subjects and individuals with neurological and/or psychiatric disorders, but its reconstructed source images may show false-positive activations, especially under low signal-to-noise ratio conditions. Here, to effectively reduce false-positive artifacts, we introduced an enhanced Fast-VESTAL (eFast-VESTAL) approach that adopts generalized second-order cone programming. We compared the spatiotemporal characteristics of the eFast-VESTAL approach to those of the popular distributed source approaches (e.g., the minimum L2-norm/ mixed-norm methods) using computer simulations and auditory experiments. More importantly, we applied eFast-VESTAL to the presurgical evaluation of epilepsy. Our results demonstrated that eFast-VESTAL exhibited a lower dipole localization error and/or a higher correlation coefficient (CC) between the estimated source time series and ground truth under various conditions of source waveforms. Experimentally, eFast-VESTAL displayed more focal activation maps and a higher CC between the raw and predicted sensor data in response to auditory stimulation. Notably, eFast-VESTAL was the most accurate method for noninvasively detecting the epileptic zones determined using more invasive stereo-electroencephalography in the comparison.