Bionic hand

A bionic hand is a powered artificial hand, usually controlled by the electrical signals of the wearer's own muscles, that restores grasping function to a person who has lost a hand or was born without one. In clinical terms it is a multi-articulating myoelectric prosthesis: each finger typically has its own motor, so the hand can form many distinct grips rather than a single pinch.^[1][2] "Bionic hand" is the popular name for these devices; prosthetists more often call them myoelectric or externally powered hands. They sit at the high-functionality end of a spectrum of upper-limb prostheses. At the simplest end are passive or cosmetic hands, which look like a hand but do not move, and body-powered prostheses, which use a shoulder harness and a steel cable to open and close a split hook or a mechanical hand.^[3][4] A bionic hand replaces that cable with battery-driven motors and muscle-sensing electrodes, trading mechanical simplicity for a wider repertoire of grips and a more lifelike appearance. This article covers bionic hands as prosthetic devices for people. For the robotics counterpart, the multi-fingered manipulators used on robots and research platforms, see dexterous hand; the two fields increasingly share technology, as described below.

How bionic hands work

The defining feature of a bionic hand is myoelectric control. When a muscle contracts it produces a small electrical signal, the electromyogram (EMG). Surface electrodes embedded in the socket rest against the skin of the residual limb and pick up these signals, which a controller amplifies and converts into motor commands.^[1][2] No implant or surgery is required for standard myoelectric control; a prosthetist identifies the strongest muscle sites and casts a custom socket around the residual limb.

The classic scheme is two-site direct control. Two electrodes read an antagonistic muscle pair, usually a wrist flexor and a wrist extensor. Contracting one muscle opens the hand, contracting the other closes it, and the strength of the contraction sets the speed or grip force through proportional control, so a gentle signal moves the hand slowly for delicate tasks and a firm signal moves it quickly.^[1][2] Switching between grip patterns is done with a special trigger such as a quick double contraction, a button, a smartphone app, or a physical gesture. This works but is not intuitive: the user must think about muscles rather than about the hand.

A more advanced approach is pattern recognition. Instead of two channels, an array of electrodes captures the pattern of activity across many muscles, and a machine-learning classifier trained on the individual user infers the intended movement. The user thinks about opening the hand or making a pinch, and the algorithm decodes it. Pattern-recognition control was commercialized by companies such as Coapt, whose Complete Control system is a machine-learning add-on compatible with hundreds of prosthesis configurations.^[5] Research continues into high-density EMG and deep learning to improve the number of movements a hand can reliably decode.^[6]

Mechanically, a modern bionic hand contains individually powered fingers driven by small DC motors, often through worm gears or tendon-driven linkages, packaged with a battery and control electronics inside a hand-sized shell of carbon fiber or plastic. The number of independently driven joints defines the hand's degrees of freedom; most commercial hands actively drive the fingers and thumb while some joints move passively to conform to an object. The finger positions combine into preset grip patterns, for example a power grasp, a precision pinch, a tripod grip, a lateral (key) grip, and a pointing index finger. Many hands add a powered or manually positioned thumb and a rotating wrist; some offer powered wrist flexion. A rechargeable lithium battery typically powers the hand for a day of use. Because comfort and signal quality both depend on the socket, fitting and suspension (how the socket stays attached to the limb) are as important to function as the hand itself.

History and milestones

Powered artificial hands grew out of a much older tradition of body-powered devices. The split hook patented by dentist David W. Dorrance in 1912 gave amputees a durable, cable-operated grasp using two spring-loaded steel tines and a shoulder harness; its descendants, such as the Hosmer hook, remained among the most used upper-limb terminal devices for more than a century.^[4] Body-powered hooks are cheap, rugged, and give a form of felt feedback through the harness cable, and many users still prefer them for heavy work.^[18]

Myoelectric control emerged during the Cold War. In 1960 the Soviet scientist Alexander Kobrinski demonstrated what is generally credited as the first clinically useful myoelectric hand, a transistorized, battery-portable device often called the "Russian Hand." A 1964 assessment judged it operationally satisfactory but noisy, with only two motions (open and close) and a single size.^[6] West German firm Otto Bock, founded in 1919 in Berlin to serve wounded veterans of the First World War, brought myoelectric arm components to market in 1965 under Max Naeder and became a dominant supplier.^[6][7] For the next four decades most powered hands were single-motor devices in which all the fingers opened and closed together in one fixed grip.

The modern bionic hand arrived in 2007, when the Scottish company Touch Bionics launched the i-Limb, developed by David Gow and colleagues out of the Princess Margaret Rose Hospital in Edinburgh. It was the first commercially available hand with five individually powered digits, allowing multiple grips and articulated fingers.^[8][9] TIME magazine named it one of the best inventions of 2008.^[10] Britain's RSL Steeper followed in 2010 with bebionic, and Ottobock introduced the Michelangelo hand around the same period.^[11] The industry then consolidated: Iceland's Ossur acquired Touch Bionics in April 2016 for a reported 27.5 million pounds, and Ottobock acquired RSL Steeper's bebionic business in February 2017.^[11][12] In parallel, a new wave of startups used 3D printing to cut costs by an order of magnitude, most visibly Open Bionics, whose Hero Arm brought a multi-grip myoelectric hand to market at a fraction of the price of established devices.^[23]

Modern landscape

The market spans large medical-device incumbents, specialist manufacturers, and a newer generation of lower-cost and startup makers. The table below summarizes representative bionic hands and their makers. Figures are manufacturer-stated unless noted and vary by model, size, and configuration.

Scroll sideways for more →

Coapt, listed above as a control provider, does not sell a hand; its Complete Control pattern-recognition module attaches to hands from other makers to give more intuitive, machine-learned control.^[5] Several of the newer makers, notably Ottobock's Michelangelo, PSYONIC's Ability Hand, and the COVVI and Vincent hands, add some form of vibrotactile feedback, the subject of the next section.

Sensory feedback and neural interfaces

A central limitation of most bionic hands is that control flows only one way. The hand receives commands from the user's muscles, but it sends almost nothing back: the wearer cannot directly feel how hard the fingers are gripping or whether an object is slipping, and instead relies on vision. This absence of touch makes delicate or eyes-free tasks difficult and contributes to the mental effort of using the device.

The first commercial response is non-invasive feedback built into the hand. PSYONIC's Ability Hand carries pressure sensors in the fingertips and conveys contact by vibrating against the skin, and PSYONIC markets it as the first commercial hand to provide touch feedback; the Vincent evolution and the COVVI Hand offer comparable vibrotactile cues, so the distinction is one of degree rather than a clear first.^[16][21][24] Such feedback is indirect (the user learns to associate a buzz with a grip force) but requires no surgery.

More ambitious approaches restore feedback and control through the nervous system. Targeted muscle reinnervation (TMR), developed by Todd Kuiken and Gregory Dumanian at Northwestern University and the Rehabilitation Institute of Chicago, surgically reroutes the residual arm nerves to spare chest or upper-arm muscles. Those muscles then produce strong EMG signals that map naturally onto hand and elbow movements, giving more intuitive, simultaneous control.^[30] A related technique, targeted sensory reinnervation, redirects sensory nerves to a patch of skin so that touching that skin feels like touching the missing hand, creating a channel for natural sensation.^[30]

Researchers have also implanted electrodes directly into nerves. In the LifeHand 2 study led by Silvestro Micera at EPFL and Scuola Superiore Sant'Anna, intraneural TIME electrodes stimulated the nerves of amputee Dennis Aabo Sorensen, letting him feel graded contact and judge the stiffness of objects in real time, and later work showed that such feedback can improve grip-force control and coordination.^[31] A further frontier combines a bone-anchored socket with implanted electrodes: osseointegration attaches the prosthesis directly to the skeleton (the e-OPRA system pioneered in Sweden), and the "osseointegrated human-machine gateway" routes permanently implanted muscle and nerve electrodes through the bone anchor to provide stable control signals and artificial sensation independent of a surface socket.^[32] These neural interfaces remain largely in clinical research and specialist centers rather than routine care.

Cost, access, and regulation

Advanced bionic hands are expensive. Multi-articulating myoelectric hands commonly cost tens of thousands of dollars, and a complete high-end arm system can approach or exceed 100,000 dollars; the DEKA LUKE arm, for example, was reported at around 100,000 dollars.^[20] Reported ranges for multi-grip hands run from roughly 20,000 to 100,000 dollars in the US and tens of thousands of pounds in the UK.^[10][20] Cost, alongside weight and maintenance, is a major reason many amputees choose simpler devices.

In the United States, prosthetic components are billed under Healthcare Common Procedure Coding System (HCPCS) "L-codes," such as L6880 for a powered multi-articulating hand, and pattern-recognition control has its own code (L6700).^[25][5] Medicare and private insurers cover prosthetic hands that replace a body part and restore function, though coverage is decided case by case; broadening Medicare coverage of devices like the Ability Hand has been cited as a way to expand access well beyond the historically small share of amputees whose costs were covered through veterans' or workers' compensation programs.^[6][21] In the United Kingdom, NHS England approved funding for multi-grip prosthetic hands in 2022, allowing eligible users to be assessed and fitted with devices such as the i-Limb or a COVVI hand rather than only single-grip hands.^[34][24] The affordability push, led by makers such as Open Bionics, Unlimited Tomorrow, and Alt Bionics, relies heavily on 3D printing and remote fitting to lower prices.^[23][25][29]

Bionic hands are regulated as medical devices. In the US they are generally FDA Class II devices cleared through the 510(k) premarket-notification pathway, which requires showing substantial equivalence to an existing device; the LUKE arm was cleared in a new integrated-prosthesis category in 2014.^[20] In Europe and the UK, hands carry the CE or UKCA mark under the EU Medical Device Regulation and equivalent British rules, and reputable makers hold ISO 13485 quality certification.^[24]

Crossover with robotics

Bionic hands and the robotic dexterous hand have converged. Both are compact, multi-fingered manipulators built from small motors, tendon-driven or geared transmissions, and increasingly tactile sensing in the fingertips, and both must grasp a wide range of everyday objects reliably. The main difference is the command source: a prosthetic hand takes its orders from a human's muscles, while a robot hand takes them from a controller running artificial intelligence or teleoperation software.

Because the hardware is so similar, several prosthetics firms now sell their hands as robotic end effectors. PSYONIC offers the Ability Hand for research and robotics with a programmable API and streamed tactile sensor data, and reports early robotics clients including NASA, Apptronik, Sanctuary AI, and Meta; the hand has been fitted to humanoid platforms for manipulation research.^[22] Alt Bionics likewise builds both a prosthetic (Genesis) and a robotic (Surge) hand and has reported selling its robotic hand to buyers including NVIDIA.^[25] This overlap means advances in one field, such as durable low-cost actuation or better touch sensing, tend to benefit the other. For the robot-focused side of this technology, see humanoid robot hands and robotics.

Challenges and future

Despite decades of progress, bionic hands face persistent hurdles. Control is still not fully intuitive: two-site myoelectric control forces users to think in terms of muscles, and even pattern recognition can be disrupted when a limb changes position or when sweat shifts the electrodes.^[6] The lack of natural touch feedback keeps the cognitive load high. Multiple motors add weight and reduce robustness compared with body-powered hooks, and the devices are costly to buy and maintain.^[3]

These factors show up in abandonment. A widely cited 2007 review by Biddiss and Chau found that roughly one in four adults abandoned electric-powered upper-limb prostheses, with rejection rates for some myoelectric devices reported even higher, and most who gave up did so within the first months of use.^[33] Studies attribute abandonment mainly to comfort, limited function, weight, durability, and control difficulty rather than to any single cause.^[33]

The response is incremental improvement on several fronts at once: high-density EMG and machine-learning decoders for more natural control; lighter, faster, more durable and waterproof mechanisms; better and cheaper touch feedback; and surgical interfaces (TMR, osseointegration, implanted electrodes) that promise closer integration with the body.^[6][30][32] The parallel boom in humanoid robotics is also pulling investment into the shared underlying hardware. Whether these advances meaningfully lower abandonment will depend as much on comfort, cost, and access as on raw capability.

In simple terms

A bionic hand is a robotic hand for a person who is missing one. Tiny sensors in the arm socket listen to the electrical "twitch" signals your muscles make when you try to move, and small motors in the fingers move accordingly, so you can open and close your hand and switch between grips like a pinch or a fist. Older prosthetic hands used a shoulder strap and cable to work a metal hook; bionic hands use batteries and muscle signals instead. They can do a lot, but they are expensive, do not usually let you feel what you touch, and can be heavy, which is why some people still prefer simpler hooks. The same kind of hand is now also being bolted onto robots.

See also

• Dexterous hand
• Humanoid robot hands
• Tactile sensing
• PSYONIC
• Open Bionics
• COVVI
• Alt Bionics
• Robotics

References

1. "Myoelectric control of prosthetic hands: state-of-the-art review." PMC. https://pmc.ncbi.nlm.nih.gov/articles/PMC4968852/ ↩
2. "How does a bionic arm work?" Open Bionics. https://openbionics.com/how-does-a-bionic-arm-work/ ↩
3. "What is a bionic hand." PSYONIC. https://www.psyonic.io/bionic-hand-definition ↩
4. "Hosmer: 100 Years of Prosthetic Development and Manufacturing." The O&P EDGE. https://opedge.com/Articles/ViewArticle/2012-09_08 ↩
5. "The Best in Myoelectric Technology." Coapt. https://coaptengineering.com/technology ↩
6. "The Cold War Arms Race Over Prosthetic Arms." IEEE Spectrum. https://spectrum.ieee.org/prosthetic-arm ↩
7. "100 years of Ottobock: the company's history." Ottobock. https://corporate.ottobock.com/en/company/history ↩
8. "I-LIMB Hand." Wikipedia. https://en.wikipedia.org/wiki/I-LIMB_Hand ↩
9. "iLimb: world's first fully articulating and commercially available bionic hand." New Atlas. https://newatlas.com/ilimb-worlds-first-fully-articulating-and-commercially-available-bionic-hand/7661/ ↩
10. "The Bionic Hand, Best Inventions of 2008." TIME. https://content.time.com/time/specials/packages/article/0,28804,1852747_1854195_1854132,00.html ↩
11. "Bebionic." Wikipedia. https://en.wikipedia.org/wiki/Bebionic ↩
12. "World-leading artificial limb maker, Touch Bionics, acquired by Ossur." News-Medical. https://www.news-medical.net/news/20160412/World-leading-artificial-limb-maker-Touch-Bionics-acquired-by-Ossur.aspx ↩
13. "Ottobock Michelangelo Hand." Bionics For Everyone. https://bionicsforeveryone.com/ottobock-michelangelo-hand/ ↩
14. "Bebionic Advanced Myoelectric Prosthetic Hand." MCOP Prosthetics. https://mcopro.com/prosthetics/technology/bebionic-hand/ ↩
15. "i-Limb Quantum Bionic Hand." Ossur. https://www.ossur.com/en-us/prosthetics/arms/i-limb-quantum ↩
16. "VINCENTevolution4." Vincent Systems. https://www.vincentsystems.de/en/vincent-evolution4 ↩
17. "Introduction to the TASKA and the TASKA CX." Arm Dynamics. https://www.armdynamics.com/upper-limb-library/introduction-to-the-taska-and-the-taska-cx ↩
18. "Perspectives on the comparative benefits of body-powered and myoelectric upper limb prostheses." Journal of NeuroEngineering and Rehabilitation. https://link.springer.com/article/10.1186/s12984-024-01436-4 ↩
19. "The LUKE/DEKA advanced prosthetic arm." U.S. Department of Veterans Affairs. https://www.research.va.gov/research_in_action/The-LUKE-DEKA-advanced-prosthetic-arm.cfm ↩
20. "DARPA-developed next-generation bionic arm hits the market." New Atlas. https://newatlas.com/darpa-deka-luke-prosthetic-arm-mobius/44267/ ↩
21. "World's first touch-sensing bionic hand with lightning-fast response." Fox News. https://www.foxnews.com/tech/worlds-first-touch-sensing-bionic-hand-lightning-fast-response ↩
22. "Robots." PSYONIC. https://www.psyonic.io/robots ↩
23. "Hero Arm." Open Bionics. https://openbionics.com/hero-arm/ ↩
24. "COVVI's Technical Specifications: The Key Details." COVVI. https://www.covvi.com/covvi-hand/product-overview/ ↩
25. "How a UTSA grad and Alt-Bionics are building a bionic hand." San Antonio Report. https://sanantonioreport.org/bionic-hand-startup-alt-bionics/ ↩
26. "BrainRobotics AI Prosthetic Hand: Best Inventions 2019." TIME. https://time.com/collections/best-inventions-2019/5733081/brainrobotics-ai-prosthetic-hand/ ↩
27. "Aether Biomedical, bionic hand startup, gets $5.8 million to scale." CNBC. https://www.cnbc.com/2023/11/13/aether-biomedical-bionic-hand-startup-gets-5point8-million-to-scale.html ↩
28. "Esper Hand: The 200 Best Inventions of 2022." TIME. https://time.com/collection/best-inventions-2022/6228818/esper-hand/ ↩
29. "TrueLimb." Unlimited Tomorrow. https://www.unlimitedtomorrow.com/truelimb/ ↩
30. "Targeted reinnervation." Wikipedia. https://en.wikipedia.org/wiki/Targeted_reinnervation ↩
31. "Intraneural sensory feedback restores grip force control and motor coordination while using a prosthetic hand." PubMed. https://pubmed.ncbi.nlm.nih.gov/30736030/ ↩
32. "Grip control and motor coordination with implanted and surface electrodes while grasping with an osseointegrated prosthetic hand." PMC. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6460734/ ↩
33. "Upper limb prosthesis use and abandonment: A survey of the last 25 years." Biddiss and Chau. https://journals.sagepub.com/doi/full/10.1080/03093640600994581 ↩
34. "NHS approved Funding for Multi-Grip Prosthetics Hands." Ossur UK. https://www.ossur.com/en-gb/prosthetics/nhs-funding-for-multi-grip-hands ↩