LMPSeizNet: A Lightweight Multiscale Pyramid Convolutional Neural Network for Epileptic Seizure Detection on EEG Brain Signals DOI Creative Commons

Arwa Alsaadan,

Mai Alzamel, Muhammad Hussain

и другие.

Mathematics, Год журнала: 2024, Номер 12(23), С. 3648 - 3648

Опубликована: Ноя. 21, 2024

Epilepsy is a chronic disease and one of the most common neurological disorders worldwide. Electroencephalogram (EEG) signals are widely used to detect epileptic seizures, which provide specialists with essential information about brain’s functioning. However, manual screening EEG laborious, time-consuming, subjective. The rapid detection epilepsy seizures important reduce risk seizure-related implications. existing automatic machine learning techniques based on deep characterized by extraction selection features, leading better performance increasing robustness systems. These methods do not consider multiscale nature signals, eventually resulting in poor sensitivity. In addition, complexity models relatively high, overfitting issues. To overcome these problems, we proposed an efficient lightweight convolutional neural network model (LMPSeizNet), performs temporal spatial analysis trial learn discriminative features relevant seizure detection. evaluate method, employed 10-fold cross-validation three evaluation metrics: accuracy, sensitivity, specificity. method achieved accuracy 97.42%, sensitivity 99.33%, specificity 96.51% for inter-ictal ictal classes outperforming state-of-the-art methods. decision-making shows that it learns clearly discriminate two classes. It will serve as useful tool helping neurologists patients.

Язык: Английский

Cognitive load detection through EEG lead wise feature optimization and ensemble classification DOI Creative Commons

Jammisetty Yedukondalu,

Kalyani Sunkara,

Vankayalapati Radhika

и другие.

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Янв. 4, 2025

Cognitive load stimulates neural activity, essential for understanding the brain's response to stress-inducing stimuli or mental strain. This study examines feasibility of evaluating cognitive by extracting, selection, and classifying features from electroencephalogram (EEG) signals. We employed robust local mean decomposition (R-LMD) decompose EEG data each channel, recorded over a four-second period, into five modes. The binary arithmetic optimization (BAO) algorithm reduce feature space extract multi-domain modes, thereby optimizing classification performance. Using six optimized machine learning (ML) classifiers, we conducted an exhaustive that encompassed both lead-wise overall classification. improved our method combining R-LMD-based with BAO ensemble (OEL) classifiers. It was 97.4% accuracy (AC) at finding in MAT (mental task) dataset 96.1% AC it STEW (simultaneous workload) dataset. In same vein, this work introduces detection, which offers temporal spatial information regarding brain activity during tasks. analyzed 19 14 leads STEW, respectively. F3 lead notably noteworthy its ability analyze variety tasks, obtaining maximum 94.5% 94%, Our approach (R-LMD+BAO+OEL) outperformed existing state-of-the-art techniques detection.

Язык: Английский

Процитировано

7
EEG detection and recognition model for epilepsy based on dual attention mechanism DOI Creative Commons
Zhentao Huang,

Yuyao Yang,

Yahong Ma

и другие.

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Март 19, 2025

In the field of clinical neurology, automated detection epileptic seizures based on electroencephalogram (EEG) signals has potential to significantly accelerate diagnosis epilepsy. This rapid and accurate enables doctors provide timely effective treatment for patients, reducing frequency future risk related complications, which is crucial safeguarding patients' long-term health quality life. Presently, deep learning techniques, particularly Convolutional Neural Networks (CNNs) Long Short-Term Memory networks (LSTMs), have demonstrated remarkable accuracy improvements across various domains. Consequently, researchers utilized these methodologies in studies focused recognizing through EEG analysis. However, current models CNN LSTM still heavily rely data preprocessing feature extraction steps. Additionally, CNNs exhibit limitations perceiving global dependencies, while LSTMs encounter challenges such as gradient vanishing long sequences. paper introduced an innovative recognition model, that Spatio-temporal fusion epilepsy model with dual attention mechanism (STFFDA). STFFDA comprised a multi-channel framework directly interprets states from raw signals, thereby eliminating need extensive extraction. Notably, our method demonstrates impressive results, achieving 95.18% 77.65% single-validation tests conducted datasets CHB-MIT Bonn University, respectively. 10-fold cross-validation tests, their rates were 92.42% 67.24%, summary, it seizure STFFD significant accelerating improving patient prognosis, especially since can achieve high without or

Язык: Английский

Процитировано

2
Extremely multi-stable grid-scroll memristive chaotic system with omni-directional extended attractors and application of weak signal detection DOI
Yupeng Shen, Yaan Li, Weijia Li

и другие.

Chaos Solitons & Fractals, Год журнала: 2024, Номер 190, С. 115791 - 115791

Опубликована: Ноя. 29, 2024

Язык: Английский

Процитировано

8
Application of deconvolutional networks for feature interpretability in epilepsy detection DOI Creative Commons

Sihao Shao,

Yu Zhou, Ruiheng Wu

и другие.

Frontiers in Neuroscience, Год журнала: 2025, Номер 18

Опубликована: Янв. 24, 2025

Introduction Scalp electroencephalography (EEG) is commonly used to assist in epilepsy detection. Even automated detection algorithms are already available clinicians reviewing EEG data, many for seizure fail account the contributions of different channels. The Fully Convolutional Network (FCN) can provide model’s interpretability but has not been applied Methods To address these challenges, a novel convolutional neural network (CNN) model, combining SE (Squeeze-and-Excitation) modules, was proposed on top FCN. performance patient-independent evaluated CHB-MIT dataset. Then, module removed from model and integrated with Inception, ResNet, CBAM modules separately. Results method showed superior advancement, stability, reliability compared other three methods. demonstrated G-Mean 82.7% sensitivity (SEN) specificity (SPE) In addition, each channel task have also quantified, which led us find that FZ, CZ, PZ, FT9, FT10, T8 brain regions more pronounced impact epileptic seizures. Discussion This article presents algorithm accurately identifies seizures patients enhances interpretability.

Язык: Английский

Процитировано

1
Dual-Modality Transformer with Time Series Imaging for Robust Epileptic Seizure Prediction DOI Creative Commons
Jiahao Qin, Z G Liu,

Jihong Zhuang

и другие.

Applied Sciences, Год журнала: 2025, Номер 15(3), С. 1538 - 1538

Опубликована: Фев. 3, 2025

Automated EEG classification algorithms for seizures can facilitate the clinical diagnosis of epilepsy, enabling more expedient and precise classification. However, existing signal preprocessing methods oriented towards artifact removal enhancement have demonstrated suboptimal accuracy robustness. In response to this challenge, we propose an Adaptive Dual-Modality Learning Model (ADML) epileptic seizure prediction by combining time series imaging with Transformer-based architecture. Our approach effectively captures both temporal dependencies spatial relationships in signals through a specialized attention mechanism. Evaluated on CHB-MIT Bonn datasets, our method achieves 98.7% 99.2% accuracy, respectively, significantly outperforming approaches. The model demonstrates strong generalization capability across datasets while maintaining computational efficiency. Cross-dataset validation confirms robustness approach, consistent performance above 96% accuracy. These results suggest that dual-modality provides reliable practical solution prediction.

Язык: Английский

Процитировано

0
Long Short-Term Memory and Kolmogorov Arnold Network Theorem for epileptic seizure prediction DOI
Mohsin Hasan, Xufeng Zhao, Wenjuan Wu

и другие.

Engineering Applications of Artificial Intelligence, Год журнала: 2025, Номер 154, С. 110757 - 110757

Опубликована: Апрель 29, 2025

Язык: Английский

Процитировано

0
PAFWF-EEGC Net: parallel adaptive feature weight fusion based on EEG-dynamic characteristics using channels neural network for driver drowsiness detection DOI
Ali H. Abdulwahhab, Indrıt Myderrizi,

Muhammet Mustafa Yurdakul

и другие.

Signal Image and Video Processing, Год журнала: 2025, Номер 19(7)

Опубликована: Май 6, 2025

Язык: Английский

Процитировано

0
Interpretable classification of epileptic EEG signals using ALIF decomposition and attention-augmented cascaded deep neural networks DOI
Wei Zeng, Minglin Zhang,

Liangmin Shan

и другие.

Applied Soft Computing, Год журнала: 2025, Номер unknown, С. 113211 - 113211

Опубликована: Июнь 1, 2025

Язык: Английский

Процитировано

0
An efficient epileptic seizure detection framework based on optimized deep residual network DOI
B. V. N. Silpa, Malaya Kumar Hota

Neural Computing and Applications, Год журнала: 2025, Номер unknown

Опубликована: Июнь 5, 2025

Язык: Английский

Процитировано

0
ARNN: Attentive recurrent neural network for multi-channel EEG signals to identify epileptic seizures DOI
Salim Rukhsar, Anil Kumar Tiwari

Neurocomputing, Год журнала: 2024, Номер unknown, С. 129203 - 129203

Опубликована: Дек. 1, 2024

Язык: Английский

Процитировано

1