A bioelectric neural interface towards intuitive prosthetic control for amputees.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Artificial Intelligence

Background:

• Current prosthetic hands offer limited grasp patterns, hindering daily activity improvements for amputees.
• Existing human-machine interfaces (HMI) fail to fully leverage the capabilities of advanced prosthetic devices.

Purpose of the Study:

• To develop a technology platform enabling intuitive, high-dexterity control of prosthetic hands.
• To bridge the gap between advanced prosthetics and amputees' daily needs through neural decoding.

Main Methods:

• Utilized a bioelectronic neural interface with implantable intrafascicular microelectrodes.
• Employed an ultra-low-noise neural recording system for electroneurography (ENG) signal acquisition.
• Applied deep learning artificial intelligence (AI), specifically recurrent neural networks (RNN), to decode motor intentions from peripheral nerve signals.

Main Results:

• Demonstrated high-accuracy control of a prosthetic hand with up to 15 degrees of freedom (DOF) in a pilot study with a transradial amputee.
• The developed interface provides an intuitive control system, directly mapping complex prosthesis movements to the user's intended actions.
• Established a sufficient information channel for advanced prosthetic control.

Conclusions:

• The study lays the foundation for robust and dexterous neuroprosthetic control strategies approaching natural hand function.
• This technology offers an intuitive conduit for mind-machine connection via peripheral neural pathways.
• Paves the way for more useful active prostheses, significantly improving amputees' quality of life.