Neuroprosthetic Limbs: The Future Of Prosthetics

Neuroprosthetic technology that allows for articulate prosthetic limbs controlled by the brain is advancing rapidly. Every day, researchers at universities and private companies such as Elon Musk’s Neuralink announce new breakthroughs in mind-controlled technology. The technology is enabled by a brain-computer interface and it allows for input and/or output signals to travel between the brain and a machine. Current myoelectric prosthetic arms already interpret signals from the brain through the muscles in a residual limb in order to perform the desired movement.

Advances in neuroprosthetics are already allowing for devices in which wearers can control individual fingers in their prosthetic and/or experience tactile feedback. They are able to access the nervous system via electrodes attached to the wearer. Brain activity is mapped so the prosthetic will move based on the same signals that would control a biological limb. For tactile feedback, computing technology converts the tactile information from the prosthetic into signals for the brain. Electrodes can be attached to the scalp, to peripheral nerves on the body, on the surface of the brain, or even embedded in the brain itself. As you can imagine, each of these methods offers different benefits and come with different risks.

Scalp electrodes are the least invasive and use Electroencephalography (EEG) to read electrical activity in the brain. However, due to the barrier of the skin and skull, the brain waves interpreted by EEG only allow for simple control options.

Electrodes implanted directly in the brain, also known as intraparenchymal electrodes, have allowed test subjects to control a prosthetic arm to drink from cups and pick up eggs. A major challenge to be solved in both cases is that the body’s immune system eventually causes degradation and loss of functionality of the electrodes.
While still invasive, electrodes on the surface of the cortex represent the middle road. They use a type of neural recording called electrocorticography (ECoG) to measure brain activity. A test subject at the University of Pittsburgh used this technology to control a 3D robotic arm.

How do we get a prosthetic to learn to move based on the brain’s intention? There needs to be a training and calibration period, just like with current myoelectric prosthetics. With new AI technology, however, this machine learning process will become much quicker. Current myoelectric calibration requires extensive work with a prosthetist and lots of trial and error. Future neuroprosthetics will use technologies like error-related potential (ErrP) to be able to assess via brainwaves whether the intended action was carried out. As a result, it will learn the wearer’s intention much quicker.

Hardware has already caught up to the point where we can create truly amazing prosthetics. Hugh Herr is the head of MIT Media’s Biomechatronics group and a double leg amputee. He has created myoelectric bionic leg prostheses that move naturally based on environmental stimuli and signals from the residual leg. The team at MIT is now experimenting with growing nerves, transected nerves, through channels, or micro-channel arrays. These technologies would allow prosthetics to provide sensory feedback that feels like a natural limb.

It won’t be long before we see examples of a neuroprosthetic limb that is almost indistinguishable to the wearer from a biological one. Although the technology still has some hurdles to cross to become viable to the public, we are seeing rapid advancement. The brain is a complex organism and learning how to effectively establish two-way communication between it and a machine is not a simple task., but advancements in AI and technology has gotten us closer than ever before.