Sensory nerve recording for closed-loop control to restore motor functions

Area of Science:

• Biomedical Engineering
• Neuroscience
• Rehabilitation Technology

Background:

• Functional electrical stimulation (FES) aims to restore motor function.
• Controlling FES with neural recordings offers a promising avenue for improved prosthetic and therapeutic devices.
• Challenges remain in developing robust and adaptable control systems for FES.

Purpose of the Study:

• To develop and evaluate a neural recording-based control method for functional electrical stimulation (FES).
• To design a rule-based controller for generating rhythmic ankle movements during locomotion using neural signals.
• To compare simple threshold detection with adaptive neural networks for FES control.

Main Methods:

• Chronic cats were used to record neural signals from tibial and superficial peroneal nerves using cuff electrodes.
• Muscular signals from ankle flexors and extensors were simultaneously recorded.
• A low-noise amplifier with a blanking circuit was designed for real-time processing and artifact reduction.
• Threshold detection and adaptive neural networks were employed for rule-based control system design.

Main Results:

• Cuff electrodes proved effective for long-term, non-damaging peripheral nerve recording.
• Both threshold detection and adaptive neural networks demonstrated robustness in handling neural recording variability.
• Adaptive neural networks effectively mapped transfer functions, suitable for determining gait invariants for closed-loop FES control.

Conclusions:

• A neural control system for FES was successfully developed and tested in a feline model.
• Simple rule-based controllers are suitable for initial human FES applications.
• Adaptive neural networks show potential for more complex FES applications requiring advanced signal processing and control.