A framework for seizure detection using effective connectivity, graph theory, and multi-level modular network

Akbarian, Behnaz; Erfanian, Abbas

doi:10.1016/j.bspc.2020.101878

Cited by 41 publications

(17 citation statements)

References 68 publications

(111 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The use of FC in epilepsy studies is a growing trend, even for seizure detection, as shown in (6,(17)(18)(19). The advantage of applying FC to seizure detection is that a multitude of implicit information about the spatial configuration and propagation of epileptiform events can be discovered.…”

Section: Introductionmentioning

confidence: 99%

RETRACTED: EEG Signal Connectivity for Characterizing Interictal Activity in Patients With Mesial Temporal Lobe Epilepsy

et al. 2021

View full text Add to dashboard Cite

Over the last decade, several methods for analysis of epileptiform signals in electroencephalography (EEG) have been proposed. These methods mainly use EEG signal features in either the time or the frequency domain to separate regular, interictal, and ictal brain activity. The aim of this work was to evaluate the feasibility of using functional connectivity (FC) based feature extraction methods for the analysis of epileptiform discharges in EEG signals. These signals were obtained from EEG-fMRI sessions of 10 patients with mesial temporal lobe epilepsy (MTLE) with unilateral hippocampal atrophy. The connectivity functions investigated were motif synchronization, imaginary coherence, and magnitude squared coherence in the alpha, beta, and gamma bands of the EEG. EEG signals were sectioned into 1-s epochs and classified according to (using neurologist markers): activity far from interictal epileptiform discharges (IED), activity immediately before an IED and, finally, mid-IED activity. Connectivity matrices for each epoch for each FC function were built, and graph theory was used to obtain the following metrics: strength, cluster coefficient, betweenness centrality, eigenvector centrality (both local and global), and global efficiency. The statistical distributions of these metrics were compared among the three classes, using ANOVA, for each FC function. We found significant differences in all global (p < 0.001) and local (p < 0.00002) graph metrics of the far class compared with before and mid for motif synchronization on the beta band; local betweenness centrality also pointed to a degree of lateralization on the frontotemporal structures. This analysis demonstrates the potential of FC measures, computed using motif synchronization, for the characterization of epileptiform activity of MTLE patients. This methodology may be helpful in the analysis of EEG-fMRI data applied to epileptic foci localization. Nonetheless, the methods must be tested with a larger sample and with other epileptic phenotypes.

show abstract

Section: Introductionmentioning

confidence: 99%

RETRACTED: EEG Signal Connectivity for Characterizing Interictal Activity in Patients With Mesial Temporal Lobe Epilepsy

et al. 2021

View full text Add to dashboard Cite

show abstract

“…In our work, on the other hand, we chose a constant threshold for each patient, and we are confident that with our approach, if we use a more elaborate technique for selecting a threshold or if we extract more features from the inferred synchronization network and combined with a good machine learning technique, we can improve the performance of the seizure detector. Recently, a paper has shown that using multiple features with a machine learning model (e.g., a neural network) improves the accuracy of the seizure detector 50 .…”

Section: Discussionmentioning

confidence: 99%

“…Furthermore, visual inspection was used to identify the seizure onsets. Various more elaborate methods have been proposed in previous studies, which could be adopted instead of using a simple threshold and visual inpsection, e.g., the method that determines the threshold based on the moving average of previous data points 49 and neural networks 50 .…”

Section: Inferring Connections Of a Brain Network (Icon Method)mentioning

confidence: 99%

Real-time Inference and Detection of Disruptive EEG Networks for Epileptic Seizures

Bomela

Wang

Chou

et al. 2020

Sci Rep

View full text Add to dashboard Cite

Recent studies in brain science and neurological medicine paid a particular attention to develop machine learning-based techniques for the detection and prediction of epileptic seizures with electroencephalogram (eeG). As a noninvasive monitoring method to record brain electrical activities, eeG has been widely used for capturing the underlying dynamics of disruptive neuronal responses across the brain in real-time to provide clinical guidance in support of epileptic seizure treatments in practice. In this study, we introduce a novel dynamic learning method that first infers a time-varying network constituted by multivariate eeG signals, which represents the overall dynamics of the brain network, and subsequently quantifies its topological property using graph theory. We demonstrate the efficacy of our learning method to detect relatively strong synchronization (characterized by the algebraic connectivity metric) caused by abnormal neuronal firing during a seizure onset. The computational results for a realistic scalp EEG database show a detection rate of 93.6% and a false positive rate of 0.16 per hour (FP/h); furthermore, our method observes potential pre-seizure phenomena in some cases. The human brain is undoubtedly one of the most complex dynamic systems known to mankind, which consists of billions of interconnected neurons 1. The connectivity patterns across different brain regions could convey evidences for advancing and supporting the understanding of cognitive and behavioral functions 2-5. In the past decades, many research studies have been devoted to discovering the process and mechanisms that govern the communication between neurons; in particular, how and why neuronal networks are formed for various neurological disorders. Epilepsy is one of the most common central nervous system disorders, following Alzheimer's disease and stroke. According to recent statistics 6 , nearly 4% of people across different ages are diagnosed with epilepsy and suffer from epileptic seizure occurrence and recurrence during their lifetimes. Epileptic seizures develop with a sudden abnormal surge of electrical activities of pathological, synchronous neuronal firing in all brain regions or some parts. Seizures can be controlled with medications and/or invasive surgeries in 70% of diagnosed patients 7. However, prompt detection or even prediction of seizure onsets remains a critical and challenging task to date because of its high variation patterns of occurrence and manifestation from one individual to another. To this end, a lot of efforts have been given to developing subject-specific, data-driven methodologies aiming at ultimately improving individualized seizure onset detection and prediction 8. Electroencephalography (EEG) recordings in high temporal resolution, regardless of arguable evidence of poor spatial resolution and low signal-to-noise ratio (SNR), are useful amongst neuroimaging techniques allowing clinicians to collect and monitor global neuronal activity information in the brain of epileptic patients. Epilepti...

show abstract

“…But threshold method has issue with threshold selection. A higher threshold may cause the problem of not being able to construct a brain network and a lower threshold may cause the problem of no meaningful connectivity measures [30]. Empirically, selecting threshold as 0.2 produced the best results.…”

Section: Threshold Methodsmentioning

confidence: 98%

“…In this study, we used to construct the effective connectivity in four frequency band which delta (0.5-4 Hz), theta (4-8 Hz), alpha (8-13 Hz) and beta (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) respectively between ADHD and HC. An average effective matrix for an ADHD subject and a HC subject of delta band was shown in Fig.…”

Section: Effective Connectivity Analysismentioning

confidence: 99%

Graph theoretical analysis based on EEG effective connectivity in ADHD children

Shen

Wen

et al. 2021

Preprint

View full text Add to dashboard Cite

This paper reports a new method to identify the ADHD children using EEG signals and effective connectivity techniques. In this study, the original EEG data is pre-filtered and divided into Delta, Theta, Alpha and Beta bands. And then, the effective connectivity graphs are constructed by applying independent component analysis, multivariate regression model and phase slope index. The measures of clustering coefficient, nodal efficiency and degree centrality in graph theory are used to extract features from these graphs. Statistical analysis based on the standard error of the mean are employed to evaluate the graph theory measures in each frequency band. The results show a decreased average clustering coefficient in delta band for ADHD subjects. Also, in delta band, the ADHD subjects have increased nodal efficiency and degree centrality in left forehead part and decreased in forehead middle.

show abstract

A framework for seizure detection using effective connectivity, graph theory, and multi-level modular network

Cited by 41 publications

References 68 publications

RETRACTED: EEG Signal Connectivity for Characterizing Interictal Activity in Patients With Mesial Temporal Lobe Epilepsy

RETRACTED: EEG Signal Connectivity for Characterizing Interictal Activity in Patients With Mesial Temporal Lobe Epilepsy

Real-time Inference and Detection of Disruptive EEG Networks for Epileptic Seizures

Graph theoretical analysis based on EEG effective connectivity in ADHD children

Contact Info

Product

Resources

About