Epilepsy Detection Based on Graph Convolutional Neural Network and Transformer DOI Creative Commons

Shibo Nie

BIO Web of Conferences, Journal Year: 2024, Volume and Issue: 111, P. 03017 - 03017

Published: Jan. 1, 2024

Epilepsy detection is a critical medical task, but traditional methods face challenges in accuracy and reliability due to the difficulty of EEG data acquisition limitation number sample seizures. To overcome these challenges, this paper proposes new model for epilepsy that combines Graph Convolutional Neural Network (Graph Network, GCN) Transformer, aiming significantly improve sensitivity detection. The core adopts GCN, which utilizes its powerful inter-node relationship capturing capability graph feature learning mechanism. However, GCN integrating global features, incorporates Transformer structure enhance aggregation reduce irrelevant interactions. After multiple rounds testing GHB-MIT dataset, demonstrated excellent performance, with an average 92.97%, specificity 94.60%, 94.59%, was better than method. Further comparison latest literature also confirms advantages present In summary, we developed based on convolutional neural network not only shows significant improvement sensitivity, provides more accurate reliable technical support diagnosis, valuable reference research related fields.

Language: Английский

Graph convolution network-based eeg signal analysis: a review DOI
Hui Xiong, Yan Yan, Yimei Chen

et al.

Medical & Biological Engineering & Computing, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 30, 2025

Language: Английский

Citations

0
Deep multi-view information-powered vessel traffic flow prediction for intelligent transportation management DOI Creative Commons
Huanhuan Li, Yu Zhang, Yan Li

et al.

Transportation Research Part E Logistics and Transportation Review, Journal Year: 2025, Volume and Issue: 197, P. 104072 - 104072

Published: March 21, 2025

Language: Английский

Citations

0
A Systematic Review of Artificial Intelligence Techniques Based on Electroencephalography Analysis in the Diagnosis of Epilepsy Disorders: A Clinical Perspective DOI
Seyyed Ali Zendehbad, Athena Sharifi‐Razavi, Nasim Tabrizi

et al.

Epilepsy Research, Journal Year: 2025, Volume and Issue: 215, P. 107582 - 107582

Published: May 16, 2025

Language: Английский

Citations

0
EEG-Based Patient Independent Epileptic Seizure Detection Using GCN-BRF DOI

Raghad Alqirshi,

Samir Brahim Belhaouari

Communications in computer and information science, Journal Year: 2024, Volume and Issue: unknown, P. 351 - 366

Published: Jan. 1, 2024

Language: Английский

Citations

1
GBMPhos: A Gating Mechanism and Bi-GRU-Based Method for Identifying Phosphorylation Sites of SARS-CoV-2 Infection DOI Creative Commons
Guohua Huang,

Runjuan Xiao,

Weihong Chen

et al.

Biology, Journal Year: 2024, Volume and Issue: 13(10), P. 798 - 798

Published: Oct. 6, 2024

Phosphorylation, a reversible and widespread post-translational modification of proteins, is essential for numerous cellular processes. However, due to technical limitations, large-scale detection phosphorylation sites, especially those infected by SARS-CoV-2, remains challenging task. To address this gap, we propose method called GBMPhos, novel that combines convolutional neural networks (CNNs) extracting local features, gating mechanisms selectively focus on relevant information, bi-directional gated recurrent unit (Bi-GRU) capture long-range dependencies within protein sequences. GBMPhos leverages comprehensive set including sequence encoding, physicochemical properties, structural provide an in-depth analysis sites. We conducted extensive comparison with traditional machine learning algorithms state-of-the-art methods. Experimental results demonstrate the superiority over existing The visualization further highlights its effectiveness efficiency. Additionally, have established free web server platform help researchers explore in SARS-CoV-2 infections. source code publicly available GitHub.

Language: Английский

Citations

1
GMAEEG: A Self-Supervised Graph Masked Autoencoder for EEG Representation Learning DOI
Zanhao Fu, Huaiyu Zhu, Yisheng Zhao

et al.

IEEE Journal of Biomedical and Health Informatics, Journal Year: 2024, Volume and Issue: 28(11), P. 6486 - 6497

Published: Aug. 15, 2024

Annotated electroencephalogram (EEG) data is the prerequisite for artificial intelligence-driven EEG autoanalysis. However, scarcity of annotated due to its high-cost and resulted insufficient training limits development Generative self-supervised learning, represented by masked autoencoder, offers potential but struggles with non-Euclidean structures. To alleviate these challenges, this work proposes a graph autoencoder representation named GMAEEG. Concretely, pretrained model enriched temporal spatial representations through signal reconstruction pretext task. A learnable dynamic adjacency matrix, initialized prior knowledge, adapts brain characteristics. Downstream tasks are achieved finetuning parameters, matrix transferred based on task functional similarity. Experimental results demonstrate that emotion recognition as task, GMAEEG reaches superior performance various downstream tasks, including emotion, major depressive disorder, Parkinson's disease, pain recognition. This study first tailor specifically learning considering Further, connection analysis may provide insights future clinical studies.

Language: Английский

Citations

0
Enhancing Epileptic Seizure Detection with Random Input Selection in Graph-Wave Networks DOI
Yonglin Wu,

Jionghui Liu,

Yangyang Yuan

et al.

Published: July 15, 2024

Graph neural networks show strong capability of learning spatial relationships between channels. In recent studies, they greatly advanced automatic epileptic seizures detection via multi-channels scalp electroencephalography (EEG). this work, we used WaveNet to extract and temporal dependencies seizures. However, EEG signals often contain noise, leading unsatisfactory model performance. This study compared effects four input preprocessing strategies on robustness. The fast Fourier transform (FFT) features, the network, were preprocessed by intact, hard, learnable, random selection. Results that with selection (30% dropout) FFT features outperforms other benchmarks an AUROC 88.57% detect Random effectively mitigates over-fitting noise promotes identification task-related frequencies through global exploration. strategy proves be a simple yet effective method improve robustness, without prior knowledge additional computational expense.

Language: Английский

Citations

0
Epilepsy Detection Based on Graph Convolutional Neural Network and Transformer DOI Creative Commons

Shibo Nie

BIO Web of Conferences, Journal Year: 2024, Volume and Issue: 111, P. 03017 - 03017

Published: Jan. 1, 2024

Epilepsy detection is a critical medical task, but traditional methods face challenges in accuracy and reliability due to the difficulty of EEG data acquisition limitation number sample seizures. To overcome these challenges, this paper proposes new model for epilepsy that combines Graph Convolutional Neural Network (Graph Network, GCN) Transformer, aiming significantly improve sensitivity detection. The core adopts GCN, which utilizes its powerful inter-node relationship capturing capability graph feature learning mechanism. However, GCN integrating global features, incorporates Transformer structure enhance aggregation reduce irrelevant interactions. After multiple rounds testing GHB-MIT dataset, demonstrated excellent performance, with an average 92.97%, specificity 94.60%, 94.59%, was better than method. Further comparison latest literature also confirms advantages present In summary, we developed based on convolutional neural network not only shows significant improvement sensitivity, provides more accurate reliable technical support diagnosis, valuable reference research related fields.

Language: Английский

Citations

0