I did get a request to provide an extensive opinion about the Open Bionics Hero Arm.
What is a normal prosthetic arm?
Summary
A normal prosthetic arm is a balancing tool (i.e., asymmetry reducing tool and overuse prevention orthopedic tool) first, and a thing that may provide better appearance only second.
As most work that causes damage on basis of asymmetry and overuse falls under “hazardous work”, i.e., highly repetitive, very heavy or both [link], a prosthesis that manages to provide that, to be that type of arm, is sufficiently comfortable, reliable, robust, cost-efficient and fast/easy to repair.
There are eight simple questions to ask if you need to know whether a particular prosthetic arm will work for everyday use in real work situations [link]. All of that is not exactly rocket science or neurosurgery though, but really fairly simple [link].
Details
As I am a below elbow amputee and what you would call – from perspective of what is sold as prosthetic arm elsewhere – “extreme” user far out and beyond (the wording is used to denote that prosthetic industries do not cover this usually: I normally clean the house including kitchen, bathrooms and toilets, I normally do garden work, I normally ride a normal bicycle, and I am a normal forensic pathologist and specialist for clinical forensic medicine), my activity exposure appears to reside in the far upper end of the range of physical loading inasmuch as the possibilities of existing prosthetic components is concerned. As that, I have a considerable experience that you can also read up in our free : ) scientific article about just that: https://jneuroengrehab.
Now, nothing of what I do there is “inspirational” – it is just really heavy loads, activities that generate sufficient sweat to put myoelectrics out of order and so on and so forth. I was just last week on some location, examining a ~190 kg body of an obese person – undressing the body, turning the body, examining the whole skin surface for injury and signs, and so on. The funeral directors were just two men to carry the coffin – which goes to show that this type of weight handling is not all too extraordinary, but quite within the range of what we usually do. My prosthetic arm, as also described in our paper, does contain necessary parts to make such activities happen, service-free, for over almost a year, under full work exposure. There is no other prosthetic arm design or build, currently that delivers that type of force and reliability, and full comfort also under layers of protective gear, also when you sweat like a horse.
Review of the Hero arm technology and video
By and large, they build prosthetic arms that are definitely not for what I call real use. They build temporary plastic things that are far from sustainable in terms of rehabilitation. If anything, they look great. The looks can be easily customized by 3D- printing some other shape or spray coating it with different colors. But then, I spray coated my real body-powered arms too, so having fun with looks is not restricted to Hero Arms, they maybe just want to make it look that way.
Is its control reliable?
The choice of Open Bionics to build a myoelectric rather than body powered arm sadly is not good. Myo arms are ladden with inherent and very serious control problems so unfortunately they are not really fit for a truly reliable everyday use even when leading a rather sedentary life. See the article [link].
Myoelectric control remained an empty promise, in terms of improved reliability, ever since its invention in the fourties and fifties.The idea as such is nice: use skin electrodes and read the activity of underlying muscles, to control a prosthesis [see post about Russian Arm and history of myoelectric arm control].
In fact, the last four decades of myoelectric control research seemed to only have produced a stalled level of unacceptably high myoelectric control rate errors (link). There, unfortunately, not even a victory of tenacity over intelligence resulted, as signal wise, to put it bluntly, you get only “garbage out” if all you get is “garbage in”. That was as true fifty years ago, which was when some clever researchers voiced that for some of the first time. So control wise, you will not get a lot of reliable control out of a myoelectric arm unless you can keep the skin dry, and avoid any limb position effect [link].
Is its socket good?
The socket style the Hero arm has is a Munster type socket, where the material fixes to the arm by slightly squeezing the arm above the elbow.
It unfortunately only repeatedly fits on the arm if the stump has the same volume always, whereas a “fit” means that the shape of the socket exactly matches the shape of the arm.
I once had such a socket for my myoelectric arm and I was not lucky: my stump either was too small, arm a bit shrunken, on days where I was a bit dehydrated or cold maybe, and sometimes a bit large, arm a bit difficult to get in, on other days.
If the stump is too small, shrunken, the electrodes may not get contact to the skin at all, and I had days when the myoelectric arm just would not work, like, at all. With a swollen arm, electrodes do not only make contact but quite possibly, they may press into the skin with added pressure and possibly cause a bit of friction, imprinting, abrasion, there, depending on specifics. The prosthetic socket caused painful, heavy bruising on my elbow when I took it shopping, to carry grocery baskets. And when I wore the iLimb to the office for light office work for 1 day, my arm was a bit abraded or so [link].
So the socket style of the Hero arm may be a considerable problem, that, as the socket style itself, seems not particularly new.
Any myoelectric prosthesis’ socket has the difficult task to keep the electrodes firmly on the skin always, which is why it is important to have a socket that always truly fits the stump, i.e., it is neither too large nor too narrow. In addition, with a conical stump and a socket as shown in the Hero arm video (or in any other video with that type of socket) the force required to pull the arm just a tad bit off the elbow or stump seems so little, that it will in all likelihood lose electrode skin contact all the time. There are moments in the video where one can see the socket coming off the arm at the slightest touch. You may not be able to even take a full regular trash bag (35 l and a few kilograms of weight) out with such an arm. Mine slipped off my arm far too often. That is one of the reasons why I stopped using my iLimb socket, and we moved to a totally different setup (liner, magnet electrodes). We are still troubleshooting that though.
The Open Bionics Hero arm comes off and back on so easily that it is clear that one cannot carry even a light shopping basket with it – it will lose electric contact to the skin first and then fall off entirely shortly afterwards. This type of socket from my experience is almost absolutely useless for real work and real life. But a utilitarian perspective may be lost on this beauty of 3D printed cosmetic cover design. Just for 5000 GBP plus covers maybe, it is still a very expensive toy.
(C) Copyright Open Bionics – advertising material
Are the 3D-printed parts a mature design?
The hand appears to have very fragile fingers, where one wonders why exactly they could not be more robust.
We build our own 3D objects and devices – totally different scale and application – some of which we may use on our deployments or for case investigations. We develop and test and improve the designs until they are ready for prototype testing or even for real use. I also make my own prosthetic 3D thingies, like so [link]. And from here, the joints of these fingers look like not even a Beta but early Alpha type development. Just wondering why the development had to stop there.
Does the hand offer a reliable predictable grip geometry?
The most important aspect of everyday prosthetic gripping in my view is that the hand or hook has a very predictable grip geometry.
An iLimb does not have that, to name a commercial example – the swiveling thumb swivels automatically or inadvertently, and that has caused significant problems. Check these links (1, 2) for some Cybathlon analyses, and the grip issues of multi-articulated hands are explained in all detail. They did not fix that until today.
The main issue is that a prosthesis needs to bring what I do not have when not wearing one: as an arm amputee with below elbow amputation on the right side, I have a very useful “power grip” or “cylinder grip” that I approximate with the stump/elbow. I do not need a prosthetic hand for that.
Where I really depend on the device is for the precision grip: where the thumb meets the index fingertip. The best device to deliver that, as little elegance as it has, is the prosthetic hook. It always places the tips together, regardless of how many times I open and close it and regardless of what I touch, push against or do. Not so the multi-articulated hands – they have swiveling thumbs as some type of feature for reasons I have not yet fully understood. But ultimately, the rounded fingertips make it almost impossible to reliably plan a precision grip, and a swiveling thumb can seriously worsen that. So the degree of coordination and the grip surface geometries are absolutely crucial here; check an in-depth analysis of grip angles of various devices if you are interested in this [link]. Before considering an Open Hero arm and spending any money on one, that needs to seriously be considered.
What about the wrist connector?
The wrist is usually an extremely critical point in that it needs to be both mechanically very stable and electrically functional. Otto Bock wrists usually die once one starts to halfway using the prosthesis. So the stability would actually need to be examined physically. I could not tell from the videos.
What about a glove?
The 3D design part is something I know well and regard differently. From my view and experience, we use 3D design and printing to only develop a particular shape or model, part or piece. Once we have that, the question is always what we use as material for the final product. I currently am building a special belt buckle, with a 3D extruded logo, a great design, but once the plastic 3D printed version fits the belt and works, I will get it done in metal. And the same goes for other parts we build and use. Some 3D-designed parts will have to be made from high definition acrylic, others are cool with cheap 3D printed PLA.
A hand that has plastic fingers will not have a great grip. Hard plastic really sucks unless all you do is swap bed sheets. So you want to wear a glove with ample padding. Soft gloves with some padding greatly enhance the grip of a prosthetic hand, even a hook – I wear silicone tubing on my hook fingers if I need to handle a lot of hard objects. For my Becker Mechanical hand, I used various hardware store-derived materials – car cleaning leather, rubbery sheet stuff, tapes – to pad the hand before putting on the glove and without surprise: insane grip, great.
Also, a glove is a really sensible step to using a prosthetic hand in real life – I would suggest performing a CT scan of an Open Bionics hand, then take it to (manual) pizza-eating, then wipe it off or wash it (it does not mind immersion and tap water, we surely will find out then) and then perform another CT scan to see where we have remnants of pizza.
My Becker hand would accumulate considerable amounts of Pizza Margarita leftovers, stuck in various finger joint mechanism parts, clearly illustrating the real need for a protective glove. Also, gloves get really dirty over time, accumulate dirt and damage, need to be swapped. And there, the only way to not make users really mad with having the prosthetic hand require a glove is to design the hand so a normal, human-type glove fits. The kind of glove we get around the corner. The stuff we can buy ourselves for little money. The only hand that I know of that has that is the Becker Mechanical hand: that one works with any type of ~size 8 to 9 gardening glove, package handling glove or other types of work glove you can get for a few bucks and that is absolutely great. All myoelectric hands seem to require special shapes for gloves and as these have to be custom made they are insanely expensive. A glove for an iLimb costs some 200 (thin, cheap, dies within minutes) to 700 USD, and it stays intact for not more than 10 minutes of holding the water hose while car washing, I tried, I know. So glove, cover, padding, that is a really relevant subject. Now, a prosthetic hand that is very light as the Hero Arm apparently is usually has a relatively weak motor. Gloves however require extra force just to move the glove, and the thicker the glove or under the material, the better the padding and the better the form closure and grip, so you do not want a really weak motor once you want to wear the hand to reliably hold hard objects such as a wine glass for a reception, a hammer for building something. I have no idea what they really use these Hero Arms for but the plastic finger/glove aspect is relevant.
What about bike riding?
They show a bike riding sequence in their video. That, I would never do. The socket is definitely not safe for pull forces. The myoelectric control stops working once you sweat. We ride bikes seriously, and extensively, we have both road and mountain bikes at home, spent a few weeks doing alpine bike tours this year, and to advertise bike riding in a video for that type of prosthesis is a bit like claiming you can hop on the moon wearing just a bikini.
For safe and comfortable riding, a shift of sitting position while riding is relevant. For that, you need a totally different arm. Check my approach for modifying a Colnago road bike for real use here [link].
(C) Copyright Open Bionics – advertising material
Ultimately these are anything, but from my estimation, they are not rehabilitation devices and not what I call functional devices.
You cannot, or maybe should not, really use it to ride a bike, or reliably vacuum the house, or carry a grocery basket. Functionally, this is neither new, nor good, nor groundbreaking in any way.
What to really use such a device for then?
So while these devices appear to be not actually too functional, that may not at all be a problem: some amputees have never gone past a level of relative manual inactivity, and then, such an arm may be a great gift. I like their designs, they look great. The photo and video poses that we have been shown are absolutely fabulous. They are more of enactment than they are a manual proof of physical real work though, and as I feel over the years that it is the hard and heavy, and/or repetitive work where I need the prosthesis most, there is no room for enactment or role-play type prostheses here. Not from my very experienced applied rehabilitative viewpoint.
One may have to consider also that the users will mostly be kids or adolescents, that may wear this for a day or two, and then only wear it if television, the uncle that bought it, or Santa Claus come to visit.
Even myoelectric “bionic” arms are often worn that way: store them, only pull them out three times a year.
Given that myoelectric control sucks, the arms break too soon anyway, that (rather than my work and private life-related heavy) usage pattern is what the makers have in mind.
However, I then wonder why they charge 5000 GBP (around 6500 USD) rather than 50 for the lot of one arm.
I would go out of my way and pay 200 CHF for an arm I will wear just for show and tell – but for 6500 USD, you can get an actually useful body-powered arm, and I do wonder why one would not really want to do that. That one can go bicycle riding, real working, real gardening, really carrying grocery baskets, for real.
How to really build a prosthesis
Check here [link].
