Investigating Upper Limb Movement Classification on Users with Tetraplegia as a Possible Neuroprosthesis Interface

Area of Science:

• Neuroscience
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Spinal cord injury (SCI) and stroke can cause paralysis, limiting independence.
• Existing assistive technologies like robotics and neuroprosthetics are challenging for individuals with tetraplegia to operate.
• Effective control interfaces are crucial for neurorehabilitation and restoring movement.

Purpose of the Study:

• To develop and evaluate a novel movement classification system for individuals with tetraplegia.
• To compare the accuracy of different calibration algorithms for controlling assistive devices.
• To enhance the usability of robotic and neuroprosthetic technologies for upper limb movement restoration.

Main Methods:

• Developed a system using a single inertial measurement unit (IMU) on the upper limb.
• Applied Principal Component Analysis (PCA) for movement classification.
• Analyzed three calibration algorithms: unsupervised, supervised, and adaptive learning.
• Tested with eight participants with tetraplegia (C4-C7) piloting a robotic hand through three postures.

Main Results:

• The supervised learning algorithm achieved 89% accuracy in classifying performed movements.
• Offline simulations showed 76% accuracy for unsupervised learning and 88% for adaptive learning.
• The IMU-based system demonstrated high potential for controlling assistive devices.

Conclusions:

• A single IMU system with supervised learning offers a highly accurate method for movement classification in tetraplegia.
• This technology can significantly improve the control of assistive devices for neurorehabilitation.
• Further development could enhance functional independence for individuals with severe paralysis.