Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Glutathione-responsive PROTAC for targeted degradation of ERα in breast cancer cells.

Bioorganic & medicinal chemistry·2023
Same author

Tomato leaf curl New Delhi virus: an emerging plant begomovirus threatening cucurbit production.

aBIOTECH·2023
Same author

Telomere-to-telomere genome assembly of melon (<i>Cucumis melo</i> L. var. <i>inodorus</i>) provides a high-quality reference for meta-QTL analysis of important traits.

Horticulture research·2023
Same author

Association between radiotherapy and prognosis in patients with small cell carcinoma of the bladder undergoing bladder-sparing surgery.

Frontiers in oncology·2023
Same author

Accurately predicting the tool influence function to achieve high-precision magnetorheological finishing using robots.

Optics express·2023
Same author

Aponermin or placebo in combination with thalidomide and dexamethasone in the treatment of relapsed or refractory multiple myeloma (CPT-MM301): a randomised, double-blinded, placebo-controlled, phase 3 trial.

BMC cancer·2023

Related Experiment Video

Updated: Jun 20, 2026

STFEEG-Tool: A Spatial-Temporal-Frequency EEG Analysis Tool for Motor Imagery Brain-Computer Interfaces
05:36

STFEEG-Tool: A Spatial-Temporal-Frequency EEG Analysis Tool for Motor Imagery Brain-Computer Interfaces

Published on: March 10, 2026

A weighted multi-scale attention-enhanced temporal convolutional network for motor imagery EEG decoding in

Zhongchen Song1, Xuejun Zhang1,2

  • 1School of Electronic and Optical Engineering & Flexible Electronics (Future Technology), Nanjing University of Posts and Telecommunications, Nanjing, China.

Frontiers in Bioengineering and Biotechnology
|June 19, 2026
PubMed
Summary

This study introduces a novel Weighted Multi-scale Attention-enhanced Temporal Convolutional Network (WMA-TCNet) for decoding motor imagery electroencephalogram signals. The WMA-TCNet model significantly improves brain-computer interface performance in neurorehabilitation applications.

Keywords:
attention mechanismbrain-computer interfacechannel-preserving prior pathelectroencephalogrammotor imagerytemporal convolutional networkweighted feature fusion

More Related Videos

Simultaneous Scalp Electroencephalography (EEG), Electromyography (EMG), and Whole-body Segmental Inertial Recording for Multi-modal Neural Decoding
11:25

Simultaneous Scalp Electroencephalography (EEG), Electromyography (EMG), and Whole-body Segmental Inertial Recording for Multi-modal Neural Decoding

Published on: July 26, 2013

Motor Imagery Brain-Computer Interface in Rehabilitation of Upper Limb Motor Dysfunction After Stroke
09:42

Motor Imagery Brain-Computer Interface in Rehabilitation of Upper Limb Motor Dysfunction After Stroke

Published on: September 1, 2023

Related Experiment Videos

Last Updated: Jun 20, 2026

STFEEG-Tool: A Spatial-Temporal-Frequency EEG Analysis Tool for Motor Imagery Brain-Computer Interfaces
05:36

STFEEG-Tool: A Spatial-Temporal-Frequency EEG Analysis Tool for Motor Imagery Brain-Computer Interfaces

Published on: March 10, 2026

Simultaneous Scalp Electroencephalography (EEG), Electromyography (EMG), and Whole-body Segmental Inertial Recording for Multi-modal Neural Decoding
11:25

Simultaneous Scalp Electroencephalography (EEG), Electromyography (EMG), and Whole-body Segmental Inertial Recording for Multi-modal Neural Decoding

Published on: July 26, 2013

Motor Imagery Brain-Computer Interface in Rehabilitation of Upper Limb Motor Dysfunction After Stroke
09:42

Motor Imagery Brain-Computer Interface in Rehabilitation of Upper Limb Motor Dysfunction After Stroke

Published on: September 1, 2023

Area of Science:

  • Neuroscience
  • Biomedical Engineering
  • Machine Learning

Background:

  • Accurate decoding of electroencephalogram (EEG) signals is crucial for brain-computer interfaces (BCIs).
  • Existing multi-scale temporal methods for EEG analysis often fail to capture both transient and long-term neural dynamics effectively.
  • There is a need for advanced models that can handle scale-specific importance in neural signals for improved BCI performance.

Purpose of the Study:

  • To propose a novel Weighted Multi-scale Attention-enhanced Temporal Convolutional Network (WMA-TCNet) for enhanced decoding of motor imagery EEG signals.
  • To address the limitations of existing methods in capturing multi-scale temporal dynamics and scale-specific importance.
  • To improve the performance and robustness of BCIs for neurorehabilitation and assistive technologies.

Main Methods:

  • Developed WMA-TCNet, employing parallel multi-scale temporal convolutions to capture distinct EEG rhythm patterns.
  • Integrated a global-aware scale attention mechanism to adaptively weight temporal information based on task relevance.
  • Introduced a weighted Channel-Preserving Prior Path to maintain channel-wise dependencies and enhance spatial modeling.
  • Utilized a temporal attention-guided TCN for joint capture of local and long-range temporal dependencies.

Main Results:

  • WMA-TCNet achieved high accuracies: 85.8% and 90.0% in subject-dependent settings on BCI Competition IV 2a and 2b datasets.
  • The model demonstrated robust performance in cross-subject scenarios, reaching 68.6% and 79.5% accuracy.
  • The proposed method showed improved decoding performance and robustness compared to existing approaches.

Conclusions:

  • WMA-TCNet offers a biologically meaningful framework for modeling multi-scale neural dynamics in EEG signals.
  • The model significantly enhances decoding performance for motor imagery, benefiting BCIs and neurorehabilitation.
  • The attention-enhanced multi-scale approach provides a robust solution for complex neural signal processing.