EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Research on epilepsy detection and recognition based on the combination of time frequency transform and deep learning model

Canhui Wang, Yan Li iD, Haoran Tang, Tianqi Xu, Zongfang Ren

PLOS One · 20.03.2026 · 21 (3) · e0336764

DOI: 10.1371/journal.pone.0336764 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, online erschienen am 20.03.2026. Verfasst von Canhui Wang und 4 weiteren Autor:innen. Beteiligte Einrichtung: School of Electrical and Information Technology, Yunnan Minzu University, Kunming, China. Hauptthemen laut MEDLINE: Epilepsy, Deep Learning, Electroencephalography. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 39 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Young Academic and Technical Leaders Program of Yunnan Province, Scientific Research Fund of the Yunnan Provincial Department of Education, the Scientific Research Fund of the Yunnan Provincial Department of Education.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

To improve the detection performance of epileptic electroencephalogram (EEG) signals and address their non-stationary characteristics,this paper compares the combined effects of continuous wavelet transform (CWT) and short-time Fourier transform (STFT) with three neural network models—EEGNet,AlexNet,and Shallow ConvNet—and incorporates targeted optimization designs. Specifically,Focal Loss,dynamic data augmentation,and an early stopping mechanism are introduced in the training phase to enhance model robustness. For EEGNet,optimizations are implemented by integrating a Squeeze-and-Excitation (SE) attention module,improving depthwise separable convolution,and dynamically adapting dimensions to reduce classification errors. For Shallow ConvNet,improvements include layered convolution for extracting “time-frequency” features and average pooling to adapt to long-duration data blocks. Experiments are conducted based on subject-independent validation,and the results show that the CWT-based feature extraction method outperforms STFT comprehensively. Among all combinations,the CWT+Shallow ConvNet pair exhibits the optimal overall performance,while the CWT+EEGNet combination follows closely with excellent precision. These findings verify the effectiveness of combining precise time-frequency features (extracted by CWT) with optimized neural network models,providing reliable technical support for clinical epileptic EEG signal detection.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

EpilepsyDeep LearningElectroencephalographyHumansFourier AnalysisConvolutional Neural NetworksWavelet AnalysisAlgorithmsNeural Networks, ComputerSignal Processing, Computer-Assisted

Autor:innen und Institutionen

  • School of Electrical and Information Technology, Yunnan Minzu University, Kunming, China.
  • Yunnan Key Laboratory of Unmanned Autonomous System, Kunming, China.
  • Key Laboratory of Cyber-Physical Power System of Yunnan Colleges and Universities, Kunming, China.
  • Department of Gastroenterological Surgery, the Second Affiliated Hospital of Kunming Medical University, Kunming, China.
  • Department of Critical Care Medicine, the Second Affiliated Hospital of Kunming Medical University, Kunming, China.

ORCID-Profile: Yan Li

Förderung

  • Young Academic and Technical Leaders Program of Yunnan Province
  • Scientific Research Fund of the Yunnan Provincial Department of Education
  • the Scientific Research Fund of the Yunnan Provincial Department of Education

Zitieren

APA

Wang, C., Li, Y., Tang, H., Xu, T., & Ren, Z. (2026). Research on epilepsy detection and recognition based on the combination of time frequency transform and deep learning model. PLOS One, 21(3), e0336764. https://doi.org/10.1371/journal.pone.0336764

Vancouver

Wang C, Li Y, Tang H, Xu T, Ren Z. Research on epilepsy detection and recognition based on the combination of time frequency transform and deep learning model. PLOS One. 2026;21(3):e0336764. doi:10.1371/journal.pone.0336764

BibTeX
@article{Wang2026,
  author = {Wang, Canhui and Li, Yan and Tang, Haoran and Xu, Tianqi and Ren, Zongfang},
  title = {Research on epilepsy detection and recognition based on the combination of time frequency transform and deep learning model},
  journal = {PLOS One},
  year = {2026},
  volume = {21},
  number = {3},
  pages = {e0336764},
  doi = {10.1371/journal.pone.0336764}
}
Kostenlos starten →