Myoelectric prosthetic arms nowadays represent a rather stale technology that remains getting built mostly according to the schematics of the sixties.

It remains being sold at absolutely screaming prices, and we are continuously told that the very complex wiring and mind-boggling technology actually warrants prices of 45’000 to 120’000 for one prosthetic arm of that type, despite its clear and massive functional deficits. However, cars with loads of highly evolved technology sell for such prices, and cars are built and sold under free market assumptions (where build a car for, say, 1500 CHF and selling it for, say, 25000 CHF, would be absolutely in order).

But myoelectric arms never saw that type of technological breakthrough. As that and as of today, that technology should be avoided by users (with only a few exceptions). And not surprisingly, rejection rates remain rather high: in terms of function and overuse prevention, one might be better off not wearing a myoelectric arm. It is surprising to see how little the industry, how little researchers and how little prosthetic technicians overall seem to care by, say, addressing the issues in a manner that does, in fact, address the issues that this technology not only has but, always had.

After all, a colloquial slang expression saying “myo arms with hard sockets suck” does not quite nail it now, does it. Because more literally, they don’t. Hard suction sockets do not work well as they may slip off, lose electrode contact, restrict motion and can be painful to wear,  localized surface EMG is problematic as it tends to be unreliable due to a range of reasons including sweat, hard socket slip and others, and energy consumption as well as component stability lead to a poor overall performance. Read up on it elsewhere.

And yet, for over seven decades, mechanic,  robotic and engineering labs have been mesmerized with the myoelectric arm concept. Institutions of all kind realized that portraying a prosthetic hand on their annual report or logo would attract funding. So they found that taking funding that was actually dedicated to health and rehabilitation could be branched off into their own funds if they played that card right. To this day, however, not much came out of this. Nothing at all, really, if one is to examine the papers  that were written back in the days and then examine what is being sold and built today. And when the best thing I can buy still is a body powered arm with a hook [how it works][performance], clearly, the last 7 decades of prosthetic arm research point to one thing as a necessary step: there needs to be tight control over these projects.

Until that happens, mechanical engineers will continue to ride what has to be termed a ‘dead horse’ today: in about four decades, academically published myoelectric arm control  error rates remained (a) unacceptably high and (b) became slightly worse [very detailed analysis here].

Why might that be?

Maybe the clue lies in the origins of this concept.

Culturally, the Russians jammed a sword into the soft belly of the Western Bloc during a time that could not have been more massive. The Cold War was at its height when Kobrinsiki et al. presented their paper [1] titled “Problems of bioelectric control”. While the title was an absolutely massive understatement of what the paper actually reported, the development as such must have represented a significant and powerful investment in research and development that has not found any similarly useful and impacting precedent in prosthetic arms or hands to this day. Previous publications by Kobrinski [2, 3] were dug out and translated later for English readers, but then, the English speaking community – “the West” – was already confronted with the fact that in spite of thousands of disabled war veterans, the Russians had outplayed everyone else with this move. And this must have come across as a direct slap in the face that no one can wipe away in any polite or elegant way – even more so as a number of governments suddenly had to find words facing families with birth defects due to Thalidomide. Without explicit mention, the burden of hope placed on medical and engineering development must have been massive at that time.

The time for government funding of prosthetic developments – particularly in Germany – must have come to an end after a series of traumatic developments that involved Sauerbruch and the Carnes’ arm exploit [link, link].

After World War I, public funding had been used to get disabled veterans off the streets and to cut down on begging and what begging did to the public – so then, the societal involvement with the aftermath of disability and disfigurement was wholesome and comprehensive.

The amputees then were mostly men in their working age and they were used to going out, to going to work, and so the public was used to having male amputees wear prostheses that were made for work.

Skin electrodes for myography were already around after a publication of H. Piper in 1912 [4].

A self-funded German inventor, Reinhold Reiter, allegedly [5] patented his “Elektrokunsthand” in 1945 and presented it to the public in 1948, but with no particular success. So in other words – real users, industrial users, hard-working users for prosthetic arms, or insurances, were not to approve of these designs then.

It required a new kind of situation to allow for myoelectric prostheses to become more popular.

In that light, it also becomes apparent that myoelectric prostheses never exceeded that new type of application paradigm. They were never, ever, designed for, meant for or approved of hard-working blue collar job affiliated men or women.

But the scene changed.

The public must have started to suspect Thalidomide as a cause for birth defects in children at the latest, one would assume, by the end of the year 1960.

With Thalidomide, however, the victims were bound to stay at home, and as the families suffered societal discrimination rather than inclusion in any major way, the public (and the elected government officials) were likely not nearly as concerned as in the early 1900s.

Having the Russians come and point out this problem in such a way certainly must have come across as a strong message.

Secondly, the mandate to push for prosthetic arms must have been different with parents struck with massive disability of children rather than grown-up men (many with trained manual professions): parents can get away with a child that toads around heavy useless myoelectric arms.

They can tell the child to not be ungrateful, to train harder, to submit to some type of testing procedures, and mostly, to try to look pretty without having to accept a no for an answer. At the same time, the fix for the disabled child had to fulfill the requirement of being worthy, costly, modern and admirable in order to make up for the massive blow to the narcissistic ego of the parents, leaving alone the more serious health problems with costly upkeep of the growing child. In other words, myoelectric arms to a great extent must have served as a fix for the parents and for society – not primarily for the amputees.

So it thus makes sense that research left it at that.

That is what myoelectric arms are still best at: being paraded around as gadgets, being admired by society. In any amputee view, they are demonstrably and understandably, repeatedly and repetitively worse than not wearing a prosthetic arm.

Myoelectric arms were always designed as and hyped up as a solution to make parents of children – that were too intimidated, shy or otherwise incapacitated to rebel or refuse – look better in a society that looked down on, and rejected, visually different people, people with disfigured extremities.

Sherman (1964) [6] then went ahead, as others did at the time, to pick up on this, and to report the Russian invention to make it understandable to others.

Following the publicity in the medical and lay press concerning a revolutionary artificial arm designed and developed in Russia, an investigation team sponsored by the Research and Training Unit of the Rehabilitation Institute of Montreal visited the United Kingdom and the U.S.S.R. from July 11 to 23 to obtain further information about this device. The official arrangements for the visit were made by the Department of National Health and Welfare, acting in concert with the Ministry of External Affairs who consented to implement this request by action through the appropriate diplomatic channels. The Rehabilitation Institute of Montreal is indebted to the Canadian Government for its efforts which helped to ensure the success of this mission [6].

Imagine what it must have been like, at the time of the Cold War, to have to go and re-publish the developments of the Russians in order to play mere catch-up [6].

The externally powered active component of this device is the hand, which houses the drive motor in the metacarpal area and produces a single movement of simple pinch or grasp with a maximum force of 15 kg. available at the fingertips [6].

Here is a photo from  this paper [6]:

A detailed analysis of the motor could not be made but, according to the British group, it contains an integral gear reduction and a worm gear drive. The motor operates at a high speed with an acceptable noise level, and the small size of the battery which provides the power indicates that this type of motor is a very efficient energy converter. This development is probably a by-product of space navigational research. Control is effected by myoelectric signals picked up by twin electrodes placed in contact with muscles selected for this purpose. The signals are magnified by a miniature transistorized amplifier, which in turn drives an electronic gating unit or relay to control the motor. The motor then runs in the desired direction to open or close the hand, stopping when the signal is discontinued. The position of the hand is then mechanically held until a reverse signal is received by the amplifier. Should both control muscles issue commands simultaneously, the system automatically shuts itself off to conserve the battery. A simple contraction of the control muscles is sufficient to turn on the amplifier. Visual feedback though the eyes of the operator is used to follow the action of the hand. Normally, antagonistic muscles are used to provide “close” and open commands through individual electrode pairs and amplifier channels, operating through the same motor drive. The response of the system is rapid, without noticeable delay; the stopping and reversing is almost instantaneous and the time delay from signal input to motor action is negligible [6].

Without much evolution, a current myoelectric arm still is wired exactly like this [6]:

Really, the center of gravity at the distal end, increased weight, very limited grasp function, hard sockets not built for pulling action and electrodes that fail with sweat, myoelectric arms remain very limited in terms of functionality and consequentially, overuse reduction. Back in the days, first results were reported as follows:

The amputee using this prosthesis does not appear to become tired, as he uses only  isometric contractions of muscles of the stump to perform the movements of the artificial hand. With the ordinary prosthetic appliance, the amputee has to use gross movements of the shoulder girdle to activate the terminal devices, and he tires more rapidly. When the amputee uses the isometric contraction of one muscle in his stump to perform one movement, without resort to any other compensatory movement, he has the impression that he performs the pinch or grasp with his own hand. This is virtually realistic, as the pinch or grasp is performed by contraction of the flexor muscles and the fingers are extended by contraction of the extensor muscles. As there is direct contact of the stump with the interior wall of the socket, the stump sock is completely eliminated. With this new type of prosthesis, the fingers of the artificial hand can be flexed and extended at whatever position the elbow and the shoulder might be at the time. The utilization of this hand therefore considerably improves the performance and dexterity of the amputee [6].

 Leaving orthopedic shortcomings aside, no currently commmercially available prosthetic arms have gone beyond this limitation:

The prosthetic training of an amputee wearing a prosthesis with myoelectric control will be quite different from the training of the amputee wearing a conventional prosthesis or a prosthesis activated by carbon dioxide. Such training is divided into two phases. The initial training will take place in the research laboratory where the amputee, with the aid of oscillographs, will learn how to individualize isometric contractions of the selected muscles. It is essential that the amputee should be able to individualize contractions of two, four, or six muscles in his forearm. The second phase of the training will be carried out in the occupational therapy department [6].

Managing to individualize contractions really is a problem. It is a lot more intuitive to use body-powered control (really, study that technology, it has great advantages for all things related to everyday usage).

It is relevant to point out that the Western Bloc has, to this day, left it at that. It still is the amputee’s problem to try to wiggle their muscles to get the isolated control signals right.

So we read Sherman’s conclusions [6], wondering why it is 2013 and we still only have two motions for myoelectric arms – and mind you: Otto Bock’s Michelangelo, RSL Steeper’s BeBionic 3 and Touch Bionics’ iLimb all only have two motions, only two controls, only two electrodes, exact same principle as the Russian Bioelectric arm:

The Russian bioelectric prosthesis at present has only two motions: opening and closing of the hand. Powered supination and pronation, wrist flexion and extension, as well as elbow flexion for above-elbow amputees, would be highly desirable. (…) There is no doubt that the Russian development of the bioelectric artificial limb represents a major contribution in the field of prosthetics. The future of advanced prosthetic appliances for the upper extremity lies in research in the further application of electronic principles [6].

We might also consider, in detail, the premises or research questions of the last 70 years of research that was done into myoelectric arms. And discuss who wrote what, and what became of them.

But as long as we continue to see Russian Arm copies hailed as “thought controlled arms”, we can rest assured that in all likelihood, nothing moves very fast in that area of engineering as so far they managed to keep the whole thing frozen at what appears to be a late 50s Russian research project.

Obviously and already at the time, research was done and exchanged internationally.

There were prior works in myoelectric control (e.g., [7]; [PDF]), that, however, remained without the impact that the Russian Arm then had. If anything, the reported (well, read it then!) placement of an electrode near or over the brachioradialis muscle of 1955 still caused trouble in the Cybathlon 2016 prosthetic arm race (which is why when they stubbornly copy mistakes over 70 years, I cannot accept that they call myoelectric arms of such a build “modern”). So, if you wonder where prosthetic industries today got all their ideas from? Stale pioneering attempts that never got examined, understood or troubleshooted.

In terms of having a primordial impact, prior research would be only of academic interest, so to speak.

More constructively, we should consider a very tight governmental control over such research (as the history of the myoelectric arm definitely questions the funding and research approach of the Western Bloc in 1960) (if you for some reason landed here and still cannot see how fundamental this problem is, you probably did not read what this is about). If anything, the Russians at the time did have tight governmental control over their research, and it must have to be assumed that only dedication of that kind will yield results of such magnitude.

Instead, we have seen a steady, unacceptably high remaining error rate for myoelectric control over the last 40 years [link]. The fact that no research group anywhere – Russia, USA, etc. – managed to push their error rates down to orders of magnitude where the daily use of myoelectric arms for physically demanding work achieves acceptably low error rates such as body-powered arms (really study this in detail if you have not done it so far) clearly shows that the technique, as ground-breaking as it may be for higher-level arm amputees, contains serious intrinsic problems. More disturbingly, there seems to be an academic belief, a type of religious collective superstition, that any control success rate that reads something with a 9 in it – such as “successful grip control in 90%”, or “96%” – seems to be sufficiently high for researchers anywhere on this planet to relax and publish their article – notwithstanding the fact that even “98%” is not anywhere close enough for users in daily life as is – and notwithstanding the fact that these laboratory results of some 1-2% error on average still deteriorate to a real-life error rate in the 10-30% range. All of these, even the proposed 1-2% errors under ideal laboratory conditions, are far too high/many for a real-life prosthesis user. Particularly if you know well-built body-powered technology that achieves far, far, far lower control error rates.

If it were any different, I am sure we’d know that by now.

References

• McKenzie (1965) Russian Myoelectric Arm

• Sherman (1964) Russian Bioelectric Arm

@article{kobrinski1960problems,
  title={Problems of bioelectric control},
  author={Kobrinski, AY and Bolkovitin, SV and Voskoboinikova, LM and Ioffe,DM and Polyan,EP and Popov, BP and Slavutski,YL and Sysin,AY and Yakobson,YS},
  journal={{Automatic and Remote Control, Proc. 1st IFAC Int. Congress}},
  volume={2},
  pages={619},
  year={1960}
}

@article{kobrinski1959first,
  title={Utilization of biocurrents for control purposes},
  author={Kobrinski, AY},
  journal={{Report of the USSR Academy of Science, Department of Technical Sciences, Energetics and Automation (translation by P Barta, UNB, 1966)}},
  volume={3},
  year={1959}
}

@article{kobrinski1960second,
  title={Bioelectric control of prosthetic devices},
  author={Kobrinski, AY},
  journal={{Herald of the Academy of Sciences USSR, Moscow (translation by US Office of Technical Services, Washington DC, USA, 1960)}},
    volume={30},
  issue={7},
  pages={58--61},
  year={1960}
}

@book{piper1912elektrophysiologie,
  title={Elektrophysiologie menschlicher Muskeln},
  author={Piper, H.},
  year={1912},
  publisher={Springer-Verlag Berlin Heidelberg}
}

@incollection{mclean2004early,
  title={The early history of myoelectric control of prosthetic limbs (1945--1970)},
  author={McLean, L and Scott, RN},
  booktitle={Powered Upper Limb Prostheses},
  pages={1--15},
  year={2004},
  publisher={Springer}
}

@article{sherman1964russian,
  title={{A Russian bioelectric-controlled prosthesis: Report of a research team from the Rehabilitation Institute of Montreal}},
  author={Sherman, E David},
  journal={{Canadian Medical Association Journal}},
  volume={91},
  number={24},
  pages={1268},
  year={1964},
  publisher={Canadian Medical Association}
}

@article{battye1955use,
  title={The use of myo-electric currents in the operation of prostheses},
  author={Battye, CK and Nightingale, A and Whillis, J},
  journal={{Journal of Bone \& Joint Surgery, British Volume}},
  volume={37},
  number={3},
  pages={506--510},
  year={1955},
  publisher={{British Editorial Society of Bone and Joint Surgery}}
}