Spectral information of EEG signals with respect to epilepsy classification

Tsipouras, Markos G.

doi:10.1186/s13634-019-0606-8

Cited by 93 publications

(48 citation statements)

References 39 publications

(247 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…As per the summarised report, for the Z–O–N–F–S classification task, our proposed method exhibited higher accuracies than the methods reported in the table. The proposed work suggested in [4, 17, 42] achieved above 90% ACC for the five‐class problem using the Bonn dataset, whereas in [10, 28] below 90% ACC is achieved. The approaches presented in [8, 14, 15] exhibited higher accuracies than the above works and close to the ACC achieved by our proposed method.…”

Section: Classification Results and Discussionmentioning

confidence: 99%

“…Though the proposed technique has proved successful for EEG classification with two datasets, this will also be applied to different other human disease datasets for evaluating its efficiency. Table 6 Comparison with different state-of-the-art methods using Bonn dataset and Bern-Barcelona dataset Using Bonn University dataset (Z, O, N, F, S for different cases) Cases Classification task Authors Classifier ACC, % case-1 Z-O-N-F-S Tzallas et al [10] STFT features + ANN 87.45 Tsipouras et al [17] RF classifier + frequency sub-bands/energy, total energy, fractional energy, entropy 90.78…”

Section: Conclusion and Future Scopementioning

confidence: 99%

“…Liang et al [28] SVM + fast Fourier transform (FFT) features 87.13 Deriche et al [8] SVM-RBF kernel + time-frequency (TF) histogram features and SVD 97.15 Guler et al [14] WT features + ANFIS 97.00 Guler et al [15] WT, Lyapunov exponents + SVM 98.36 proposed approach 99.69 case-2 Z-N-S Acharya et al [3] ApE, samEn, PhaseEntr 1, phase ent 2 + fuzzy classifier 98.43 Naghsh-Nilchi et al [6] eigenvector spectral estimation and multi-layer perceptron 97.8 Naghsh-Nilchi et al [6] max, min, average, mobility feature 97.39 Ghosh-Dastidar et al [18] principal component analysis + cosine RBFNN 97.63 Ghosh-Dastidar et al [16] mixed-band wavelet chaos nine parameters + LMBPNN 98.67 proposed approach 100 case-3 Z-F-S Zhang et al [22] GST + SVD + RF 98.97 Zhang et al [19] WPD + fuzzy distributed entropy and Kruskal-Wallis variance + KNN classifier 98.48 Tzallas et al [10] STFT features + ANN 100 Tzallas et al [11] time frequency analysis + ANN 99.35 Liang et al [28] SVM + FFT features 98.85 proposed approach 100 case-4 O-N-S proposed approach 100 case-5 O-F-S proposed approach 100 case-6 ZO-NF-S Zhang et al [19] WPD + fuzzy distributed entropy and Kruskal-Wallis variance with KNN classifier 98. 48 Zhang et al [22] GST + SVD + RF 98.28 Li et al [13] DWT + neural network ensemble 95.01 Tzallas et al [11] time frequency analysis + ANN 95.90 Li et al [23,25] MODWT + RF classifier 98.0 Tsipouras et al [17] RF classifier + frequency sub-bands/energy, total energy, fractional energy, entropy 98.15…”

Section: Conclusion and Future Scopementioning

confidence: 99%

“…In [16], WT has been utilised to find features that are supplied to Lavenberg Marquardt algorithm‐based BPNN (LMBPNN) classifier to identify three classes of EEG signals. In [17], a radiofrequency (RF) classifier has been utilised with WT extracted frequency sub‐bands based on which energy, total energy, and fractional energy are obtained for solving five‐class problems. Further discrete WT (DWT) based features along with different neural network classifiers have been used for EEG epileptic seizure classification.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Multi‐kernel‐based random vector functional link network with decomposed features for epileptic EEG signal classification

Parija

Dash

Bisoi

2020

IET signal process.

View full text Add to dashboard Cite

This study proposes an improved hybrid model built with empirical mode decomposition (EMD) features combined with weighted multi-kernel random vector functional link network (WMKRVFLN) where the kernel parameters are optimised with an efficient optimisation algorithm known as water cycle algorithm (WCA) for diagnosis and classification of epileptic electroencephalogram (EEG) signals. The proposed model with optimisation is known as WCA-EMD-WMKRVFLN. The tanh and wavelet kernel functions are contributing together to the effectiveness of the proposed model. The features generated from EMD in terms of intrinsic mode functions (IMFs) are modulated to find important statistical and entropy based features and these features in a reduced form are employed as inputs to the model to classify epileptic EEG signals. The presented approach is evaluated in terms of percentage correct classification accuracy (ACC), specificity and sensitivity using two datasets and is compared with different classifiers and state-of-the-art techniques. The highest accuracies of 99.69% (five-class) and 100% (three-class) achieved using the Bonn-University dataset and 99.0% ACC (two-class) achieved using the Bern-Barcelona dataset. The achieved results report that the presented approach is a promising approach for EEG signal classification and is superior to several state-of-the-art techniques and is highly comparable to many such techniques.

show abstract

Section: Classification Results and Discussionmentioning

confidence: 99%

Section: Conclusion and Future Scopementioning

confidence: 99%

Section: Conclusion and Future Scopementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Multi‐kernel‐based random vector functional link network with decomposed features for epileptic EEG signal classification

Parija

Dash

Bisoi

2020

IET signal process.

View full text Add to dashboard Cite

show abstract

“…Some studies make use of the online-available EEG datasets [26] that include recordings for both healthy and epileptic subjects [27] in the context of classification and spike detection. The idea was further developed by [28] where the classifier was tested with varying sets of frequency sub-bands other than the four fundamental ones. Power spectral density analysis showed a significant power decrease in the θ-band for a group of patients with drug-resistant mesial temporal lobe epilepsy (MTLE group) as compared to patients with non-MTLE (NMTLE group) [29].…”

Section: Introductionmentioning

confidence: 99%

Penalized Functional Connectivity Maps for Patients With Focal Epilepsy

et al. 2021

View full text Add to dashboard Cite

Objective: This study demonstrates how functional connectivity (FC) patterns are affected in direct relation to the lobe that is mostly affected by seizures. Methods: The novel idea of penalized FC (pFC) maps is compared against standard FC maps in the four fundamental EEG frequency sub-bands. The FC measure between any two specific electrodes is scaled depending on the probability of true FC between them, and their power content with respect to the two electrodes of maximum power within each frequency sub-band. The algorithm is automated and introduces adaptive power penalization based on the power distribution of the different sub-bands. Results: The pFC maps were found to be more effective at suppressing the local connectivity in the lobes that are less affected by the interictal epileptiform discharges (IEDs). More precisely, given the least amount of power penalization, pFC maps of the theta sub-band reveal statistical significance in terms of increased local connectivity margin of the affected region as compared to the standard FC maps. However, they cannot be solely relied upon as other sub-bands could alternatively show high local connectivity across different patients within the region of interest. Conclusion: penalized functional connectivity maps intrinsically provide more information regarding the whole brain network in context to regions of interest where the active lobe is determined by the neurologists to contain the focal source. Significance: Findings suggest that (1) the significant sub-band varies from patient to patient while remaining relatively consistent within the IED segments of a same patient, and (2) the pFC maps have an advanced capability in terms of pinpointing to a region of interest of the active lobe, and as such can play a critical role in providing insight as to a region of interest where the 3D source might be located when solving the ill-posed inverse problem.INDEX TERMS Focal epilepsy, Interictal Epileptiform Discharges (IEDs), penalized Functional Connectivity (pFC), penalized Weighted Phase Lag Index (pWPLI).

show abstract

A spike is a spike: On the universality of spike features in four epilepsy models

Sargsyan,

Casillas‐Espinosa,

Melkonian

et al. 2024

Epilepsia Open

View full text Add to dashboard Cite

ObjectiveFrequency properties of the EEG characteristics of different seizure types including absence seizures have been described for various rodent models of epilepsy. However, little attention has been paid to the frequency properties of individual spike–wave complexes (SWCs), the constituting elements characterizing the different generalized seizure types. Knowledge of their properties is not only important for understanding the mechanisms underlying seizure generation but also for the identification of epileptiform activity in various seizure types. Here, we compared the frequency properties of SWCs in different epilepsy models.MethodsA software package was designed and used for the extraction and frequency analysis of SWCs from long‐term EEG of four spontaneously seizing, chronic epilepsy models: a post‐status epilepticus model of temporal lobe epilepsy, a lateral fluid percussion injury model of post‐traumatic epilepsy, and two genetic models of absence epilepsy—GAERS and rats of the WAG/Rij strain. The SWCs within the generalized seizures were separated into fast (three‐phasic spike) and slow (mostly containing the wave) components. Eight animals from each model were used (32 recordings, 104 510 SWCs in total). A limitation of our study is that the recordings were hardware‐filtered (high‐pass), which could affect the frequency composition of the EEG.ResultsWe found that the three‐phasic spike component was similar in all animal models both in time and frequency domains, their amplitude spectra showed a single expressed peak at 18–20 Hz. The slow component showed a much larger variability across the rat models.SignificanceDespite differences in the morphology of the epileptiform activity in different models, the frequency composition of the spike component of single SWCs is identical and does not depend on the particular epilepsy model. This fact may be used for the development of universal algorithms for seizure detection applicable to different rat models of epilepsy.Plain Language SummaryThere is a large variety between people with epilepsy regarding the clinical manifestations and the electroencephalographic (EEG) phenomena accompanying the epileptic seizures. Here, we show that one of the EEG signs of epilepsy, an epileptic spike, is universal, since it has the same shape and frequency characteristics in different animal models of generalized epilepsies, despite differences in recording sites and location.

show abstract

Spectral information of EEG signals with respect to epilepsy classification

Cited by 93 publications

References 39 publications

Multi‐kernel‐based random vector functional link network with decomposed features for epileptic EEG signal classification

Multi‐kernel‐based random vector functional link network with decomposed features for epileptic EEG signal classification

Penalized Functional Connectivity Maps for Patients With Focal Epilepsy

A spike is a spike: On the universality of spike features in four epilepsy models

Contact Info

Product

Resources

About