Don't give up

Why are bionic prostheses not becoming more accessible?

Evgeny Zhvansky, Forbes, 13.06.2017

The main goal of prosthetic developers is to create a ready–made and user-friendly device that you can buy, put on and use without a complicated learning process. How far is it to this?

From the Middle Ages to the present day, mankind has been striving to create prostheses that are most similar to the lost limb both externally and functionally. The future belongs to bionic prostheses, which are mechanically closest to the functionality of a healthy person's body, but the problem of quality control of such devices still does not have a ready–made solution today.

 Over the past 5 years, there have been many companies engaged in the development of bionic prostheses. The focus is mainly on cheap bionic devices made of plastic parts, including those made using 3D printing technologies. There are already ready-made products, for example, from OpenBionics, which are now at the stage of approval by the FDA. The mechanical part of such market players as OttoBock or iLimb is also developing, but this development is not aimed at reducing the cost of prostheses, but rather at the mechanics of movements (smoothness, naturalness, accuracy). With this approach, the functional part of the prosthesis develops, but the controllability remains the same.

From hook to bionics

The history of prostheses begins in ancient times – the most ancient is considered to be an eye prosthesis, which dates back to the III millennium BC. In the Middle Ages, well-known "pirate" wooden supports began to appear instead of lost legs or hooks instead of brushes. Such prostheses performed a limited number of functions that a particular person needed, based on his type of activity. A similar approach can be found in prosthetics today.

When it comes to rehabilitation after amputation of the arm, the simplest solution is a cosmetic prosthesis. In addition to the aesthetic purpose, such prostheses perform practically no functions and have no advantages compared to medieval hook prostheses.

Another solution is traction prostheses. Their hands can already contract and unclench due to, for example, movements of the wrist or elbow joint of the remaining part of the arm. These movements guide the mechanical tension of the threads that drive the "fingers" into action. Such a brush "knows" only how to clench and unclench his fist. It is characterized by high performance and good reliability. Traction prostheses are developed by domestic innovative companies, they can also be made independently according to the instructions (which is also practiced in third world countries).

The third class is mechanical prostheses controlled by muscle activity. Such devices, as a rule, are made of metal, have great strength, but have only two degrees of freedom – compression and decompression. It is not very convenient to operate a mechanical prosthesis: in order to unclench the fist, you need to strain the outer side of the forearm, and in order to squeeze – on the contrary, strain the inner side of the forearm. This is the so–called trigger control method: either there is muscle activity - then the movement is activated, or there is no muscle activity. Unfortunately, such a control system can lead to false positives. Mechanical prostheses have the "appearance" of cosmetic ones and the functionality of traction ones, they are powered by a battery that is placed on the prosthesis. The metal frame and the motor that drives the brush make it possible to call the design reliable: for example, if you need to hold an object, a mechanical hand will be able to squeeze it hard and for a long time, and this will require practically no effort on the part of a person.  Inconvenient control and limited functionality are the main disadvantages of mechanical prostheses.

The last, fourth class – bionic prostheses, in which each finger is controlled by a separate motor – this gives a greater advantage in terms of manipulating objects. The control system of the bionic brush is the same as that of the mechanical one, based on compression and decompression – therefore, these prostheses are difficult to use. To facilitate use, add any external switches - levers on the prosthesis or applications on the smartphone.

High cost and low functionality

"Bionicity" implies, in addition to replenishing the mechanical functions of the lost hand, the naturalness of its use. The developers are focused on optimizing the structure of prostheses – we need the most durable, ergonomic, functional solutions from the point of view of mechanics. However, the task of providing maximum management functionality does not have a ready-made solution on the market. And uncomfortable and limited functioning prostheses cost from $30,000 to $70,000.

All today's R&D projects are focused on two directions: making the prosthesis cheaper and improving the management system. If there are more or less suitable solutions for the first problem, then everything is just beginning in the field of control system development.

Ideally, a person using a prosthesis should not notice the control system. That is, the interface between a person and a prosthesis uses natural control mechanisms that a person learned in childhood. Thus, the question is acute, which interface of interaction between a person and a prosthesis should be used and how to adjust this interaction to the individual characteristics of each?

Perfect interaction with a person

3D printing technologies are used to reduce the cost of production. The cost of such prostheses is low due to the use of plastic parts, and there are quite a lot of companies that are engaged in 3D printing of prostheses around the world, including in Russia. Foreign companies create models of bionic prostheses and make them publicly available, contributing to the development and accessibility of prosthetics. Other development companies optimize and refine the design and mechanics of freely available 3D models.

But it is much more difficult to solve the problem of improving human interaction with a prosthesis. The most "natural" approach is a full–fledged hand transplant. Muscles and nerves at the same time work exactly like in a healthy hand, but the procedure is very expensive, requiring donor material, additional therapy and risks of rejection. Of course, such a method, in one form or another, has a future that will come only after revolutions in related fields – in 100 years. So far, it is important to create rehabilitation devices that sufficiently replenish the functions of the lost brush and allow you to control yourself in a natural way.

There are four main types of human interaction with a prosthesis:

The first, the most radical, is various kinds of implants in the motor and sensory areas of the cerebral cortex. Such an interface has the same disadvantages as a transplanted arm. Implants in the brain are especially appropriate in the case when, for some reason, the connection between the brain and the hand is disrupted. In other cases, it is worth additionally evaluating the benefits/risks of using such an interface.

The second control method is the use of electroencephalography (EEG). The EEG method is based on the registration of bioelectric activity of the brain resulting from the propagation of the action potential through neurons. The method is considered promising, but has a number of technical difficulties that prevent the appearance of an interface based on it. Firstly, due to the peculiarities of registering the brain activity map, the system needs to be "trained" anew when moving the electrodes. And secondly, the signal itself is very unstable to various kinds of electrical interference and interference.

Third: implantation of electrodes to peripheral neurons in the remaining part of the arm. This method has all the same problems as transplantation and brain implants, besides it requires long and individual work of doctors.

And the last type of interface is electromyography (EMG). Its simplest implementation – trigger – is used in mechanical prostheses, directing the compression or unclenching of the hand. Exactly the same control system is implemented in bionic prostheses. But, as already mentioned, EMG in them is used only for two degrees of freedom – flexion and extension of the fingers. Also, a third degree of freedom can be added to them – simultaneous tension of both muscles on which EMG activity is measured.

Electromyography is a method of analyzing muscle activity based on measuring the potential difference at two points between which an action potential spreads under the skin along the membranes of muscle fibers (this potential is the propagation of a wave of muscle activity from the zone where the action potential of the motor neuron enters, forcing our muscles to "work"). This method allows you to record the signal of muscle activity with a minimum noise level. Most of the movement of the fingers and hand is closely related to the muscles of the forearm. It is easy to check this by placing one hand on the forearm (just below the elbow) and moving the fingers of the other hand – you can feel how the various muscles of the forearm contract. Using a control system that is individually tuned to the patterns of hand movements of a particular person brings us closer to creating a natural interface between a person and a prosthesis. On the one hand, it is non–invasive and has great functionality, on the other hand, it is quickly configured and resistant to external influences. The problem may be atrophy of the remaining muscles, but the method allows you to extract the maximum of the preserved natural patterns of muscle activity.

Current status of developments in the world

Prosthetic control systems are also being developed, but there are significantly fewer companies focused on this task. Basically, developers use ready-made electromyographic amplifiers and, having received a signal, process it primitively (one way or another, everything comes down to a "trigger" system, the only question is the number of thresholds and the number of EMG recording channels). In some cases, cluster analysis is resorted to, but this is mainly found in scientific articles, which also claim that such methods are not adapted for use in real life due to the variability of muscle activity. Trigger systems use smartphones or other devices that switch the modes of contractions, by analogy with existing prostheses. Nevertheless, in combination with the cheapness of 3D printing and a similar control system for "expensive" prostheses, these companies will take their market share. There is another approach to solving the problem of controllability – more detailed processing of the EMG signal and the selection of patterns of specific movements, in order to subsequently reproduce them on the prosthesis after training using machine learning. That is, it is necessary to train the control system for each individual movement for a specific patient, which will be reproduced with repeated tension of the muscles corresponding to a specific movement. This training of the control system can take place within 1-2 minutes, while the accuracy of motion recognition will depend on the quality of EMG processing algorithms and machine learning algorithms and will be at least 99%, depending on the variety of recognized movements. Such a control system can be integrated into almost any bionic prosthesis, which will distinguish it in the market among competitors. There are not so many companies developing in this area all over the world. In our country, a number of companies are also engaged in this (the company "Myonics", which the author represents, is one of them – Forbes)

Feedback systems are also being developed – from vibrational tactile feedback to artificial skin integrated with the human nervous system. This is a separate layer of development, which is certainly necessary for subtle manipulations with complex objects, for example, fragile or soft. Without feedback, a prosthesis, as a rehabilitation device, will not be a full-fledged replacement for a lost limb. It is noteworthy that, as a rule, the development of feedback does not intersect with the development of improved prosthetic mechanics and, moreover, the control system of bionic prostheses.

The direction of bionic prostheses is developing all over the world. The main goal of this development is to create a ready–made, easy-to-manage prosthesis that you can buy, put on and use without a complicated learning process. Unfortunately, at the moment such a product has not been created, and the demand for it is growing every year. We believe that in the near future we will be able to see a cheap prosthesis with a convenient, simple and personalized control system and feedback. Such control systems will also give impetus to the development of exoskeletons controlled by small muscular efforts.

Portal "Eternal youth" http://vechnayamolodost.ru  14.06.2017