Artificial Intelligence Enables Real-Time and Intuitive Control of Prostheses via Nerve Interface

Area of Science:

• Neuroscience
• Biomedical Engineering
• Artificial Intelligence

Background:

• Developing advanced prosthetic limbs requires seamless integration between human neural signals and machine control.
• Current prosthetic technologies often lack the dexterity and intuitive control of natural limbs.

Purpose of the Study:

• To present a neuroprosthetic system utilizing an artificial intelligence (AI) agent for translating amputees' movement intentions.
• To demonstrate real-time, intuitive control of a prosthetic hand through a peripheral nerve interface.

Main Methods:

• An AI agent, based on a recurrent neural network (RNN), was developed to decode six degrees-of-freedom (DOF) from multichannel nerve data in real-time.
• The system's performance was evaluated in motor decoding experiments with three human amputees, assessing accuracy, reaction time, and information throughput.

Main Results:

• The AI agent enabled amputees to intuitively control prosthetic hand and wrist movements with 97-98% accuracy.
• Real-time performance was validated through gesture matching tasks, and the decoder demonstrated robust predictive performance over 16 months.

Conclusions:

• This study highlights the potential of AI-enabled nerve technology for next-generation prosthetic hands.
• The developed system offers a pathway towards more dexterous and intuitive prosthetic limb control for amputees.