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 Experiment Video

Updated: Jun 25, 2026

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

GaitNet: Transfer Learning-Enhanced CNN-GRU Architecture for Intention Detection in Healthy and Post-Stroke

S Hossein Sadat Hosseini, Rudri Purohit, Shuaijie Wang

    IEEE Transactions on Neural Systems and Rehabilitation Engineering : a Publication of the IEEE Engineering in Medicine and Biology Society
    |June 23, 2026
    PubMed
    Summary

    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

    Phenotyping population-level chronic condition prevalence: The importance of forcing factors from the ecological framework.

    Public health·2026
    Same author

    Change in Physical Activity Patterns Through a School-Based Intervention: The Altus Health and Wellness Academy Model.

    AJPM focus·2026
    Same author

    Genome-Wide Resequencing Reveals High Connectivity and Localized Adaptive Signals in Manila Clam (<i>Ruditapes philippinarum</i>) Populations Along the Southeastern Coast of China.

    Animals : an open access journal from MDPI·2026
    Same author

    Predicting the Complex, Multilevel Ecology of US County-Level Diabetes Prevalence: A Cross-Sectional, Artificial Intelligence Analysis.

    Diabetes/metabolism research and reviews·2026
    Same author

    Change in nutrition patterns through a school-based intervention-The Altus health and wellness academy model.

    Journal of education and health promotion·2026
    Same author

    The ecological framework of population health - Adding public trust as a forcing factor for U.S. life expectancy and COVID-19 mortality.

    Public health in practice (Oxford, England)·2026

    GaitNet, a new AI model, accurately detects human motion intention in real-time for stroke recovery robots. It significantly reduces data needs for post-stroke patients, improving rehabilitation robot control.

    Area of Science:

    • Robotics
    • Artificial Intelligence
    • Neurorehabilitation

    Background:

    • Accurate human motion intention detection is crucial for assistive and rehabilitation robots, especially for post-stroke recovery.
    • Existing methods often require extensive subject-specific data and are less effective for post-stroke individuals, limiting clinical use.
    • There is a need for data-efficient, real-time motion intention detection applicable to both healthy and post-stroke populations.

    Purpose of the Study:

    • To introduce GaitNet, a novel transfer learning-enhanced dual-branch CNN-GRU framework for robust and data-efficient motion intention detection.
    • To evaluate GaitNet's performance and data efficiency in healthy and post-stroke participants.
    • To demonstrate GaitNet's suitability for real-time robotic control in rehabilitation.

    Main Methods:

    More Related Videos

    Soft Pneumatic Robot Modulates Graph Theory Metrics of Brain Network for Hand Rehabilitation After Stroke
    05:30

    Soft Pneumatic Robot Modulates Graph Theory Metrics of Brain Network for Hand Rehabilitation After Stroke

    Published on: October 10, 2025

    Related Experiment Videos

    Last Updated: Jun 25, 2026

    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

    Soft Pneumatic Robot Modulates Graph Theory Metrics of Brain Network for Hand Rehabilitation After Stroke
    05:30

    Soft Pneumatic Robot Modulates Graph Theory Metrics of Brain Network for Hand Rehabilitation After Stroke

    Published on: October 10, 2025

    • Developed GaitNet, a dual-branch CNN-GRU framework incorporating convolutional layers for spatial features and GRUs for temporal modeling.
    • Utilized transfer learning, pre-training on healthy participants and fine-tuning on post-stroke data, including joint supervised pretraining (JSP).
    • Assessed accuracy, precision, recall, and F1-score, alongside inference time, and deployed in a closed-loop robotic control system.

    Main Results:

    • GaitNet achieved high accuracy (99.29% healthy, 98.96% post-stroke) using the full dataset.
    • With transfer learning (JSP), GaitNet maintained >98% performance metrics using only 45% of the training data for post-stroke participants, a 65% data reduction.
    • Inference time was 27.46 ms, suitable for real-time applications, and successful closed-loop control was demonstrated.

    Conclusions:

    • GaitNet offers a practical and scalable solution for highly accurate, data-efficient motion intention detection.
    • The framework significantly reduces data requirements for post-stroke individuals, enhancing clinical applicability.
    • GaitNet's real-time capabilities and successful deployment confirm its potential for advanced rehabilitation robotics.