Morning Coffee Science

Mind-controlled prosthetics that can sense touch


Limb loss is a widespread and costly health issue. In the United States alone, there are nearly 2 million people living with limb loss. Each year that is more than 1 million annual limb amputations globally - one limb amputation every 30 seconds. Most people think that the most common cause of limb amputations are severe injuries incurred in accidents on the road or in the workplace. However, the most common reason for amputation is the peripheral arterial disease which means that blood and nutrients can not reach the limb because of damaged or narrowed arteries (atherosclerosis).

While prevention and education about the dangers of peripheral arterial disease should be more widespread, a post-factum solution to limb loss is also important.

The current alternatives for amputees are subpar. Hand prostheses after transhumeral (upper arm bone) amputations are usually attached to the humerus with a socket that compresses the amputee's stump and is controlled by the biceps and triceps muscles. The crucial thing here is that these prostheses do not have somatosensory feedback - amputees do not feel what their artificial hand is feeling. Patients have to rely on visual or auditory cues of the movement of the prosthesis in order to gauge what kind of muscle contractions result in what movement. Current prostheses are unintuitive and hard to learn to use for the patient. In addition to that, activities that require delicate touch by a hand become nearly impossible to learn (e.g. picking up an egg).

A study published in the New England Journal of Medicine reports three Swedish patients who have lived with e-OPRA, a new, "feeling", neuromuscular interface, for 3 to 7 years without medical supervision. The technology that underpins e-OPRA is called osseointegration (osseo = bone). The latter technology has been discovered in Sweden in the 50s and has been applied to dental implants in the 60s.

Now, scientists have created new generation prostheses that would directly connect the implant system of the prosthesis to the skeleton of the stump of the amputated limb. This process is what is called osseointegration. More seamless control of the artificial arm is ensured because electrical muscle and nerve signals are captured by electrodes on the muscle of the amputation stump. These signals are passed down to the implant through the skin that connects to the prosthesis. The signals are then decoded by sophisticated artificial intelligence algorithms which result in a controlled signal for the prosthetic hand's movement.

But how does the patient feel the touch of his artificial limb? Here is how it works. The force sensors located in the thumb of the prosthesis measure contact and pressure applied to an object while grasping. This information then is turned into an electric potential and sent to the nerves in the amputated stump that used to be connected to the biological hand before the amputation. The nerves then lead to "touch" centers in the brain and it can then perceive the extent of pressure and force applied by the hand.

It is interesting how scientists had to develop this technology. They firstly had to "calibrate" the A.I. algorithm so that the electric activity in the amputated stump could be decoded to the desired movement by the prosthesis. The first steps where understanding simple commands like what neural signals lead to the opening and closing of the artificial limb palm. Then came more complex commands: turning the wrist, moving individual fingers, exerting a specific amount of pressure on an object to hold it. All of these are classification problems that require a lot of trial runs by scientists.

A similar calibration protocol had to be followed for the "perceiving" part of the prosthesis. The scientists had to classify what is the perceived "big" pressure/force for the patient and adjust the algorithms that would decipher the force recorded by the thumb of the prosthesis to electrical signals sent to the brain.

This is not the first time that an artificial limb that can feel touch is created. The novelty of this report is that patients have been using these interfaces in their daily for 3 to 7 years. The three patients reported noted that their daily lives have been overwhelmingly positively influenced. 2 of them said that they think that there is almost no difference between a real and an artificial limb. Moreover, the unsupervised use of this interface shows that it is safe and stable to use in the long term. Patients have been using them in their everyday activities and the prostheses were effective not only in restricted and controlled lab environments.

Dr Max Ortiz Catalan, one of the leaders of this project, says this about the future of the technology: "We expect this system to become available outside Sweden within a couple of years, and we are also making considerable progress with a similar technology for leg prostheses, which we plan to implant in a first patient later this year.”

Article was prepared by Matas Vitkauskas on behalf of INA

Watch a video by the Charlmers University of Technology on osseointegration.