I wonder about the reliability and speed of myoelectric prostheses - an issue that very much stays with us also with some new bionic arms such as the DEKA arm, the Michelangelo hand or the iLimb. Trigger for that question were a number of what could be termed possible urban myths. One guy gut stuck in a bus holding on to a bar - and his prosthesis did not open any more. A nurse was giving a patient an urgently needed injection when shortly before the injection her prosthesis gave up. Other people apparently reported the prosthesis going haywire in proximity to electrical appliances. Furthermore, videos one can round up covering bionic and myoelectric prostheses seem to show a significant lag of action - a real pause, when a prosthesis is supposed to grab some object or perform otherwise. Most of the videos available for myoelectric prostheses seem to be considerably slower than - compared to a myoelectric prosthesis - my own experience with my body powered arm.
This - if true - would be a problem to myoelectric prosthetics that is not easy to dismiss. After all, one is already humiliated enough by being disabled - the humiliation of a suddenly and unpredictably dysfunctional prosthesis is a truly unnecessary and cumbersome burden I'd want to avoid by all means. I'd wear an expensive technologically advanced prosthesis to show-case the stereotype of 'high tech hero' - not the stereotype of the 'person that bought the expensive nightmare'. Let us not forget that the dream of "electric switches" also is not new and so one has to be careful to not adopt dreams that may be outdated as too promising. Let us not forget that after all we are living in a world that shows us very elegant mechanical prosthetics such as the X Finger:
So good looks and design are a necessary requirement, but the really unavoidable values in analogy to real estate ('location, location, location') for the field of upper extremity prosthetics are 'performance, performance, performance'. Unless you know in your heart how it feels to be publicly exposed to failure of your manipulations due to disability you will never understand why clumsy or non-functional artificial arms may not just be an acceptable inconvenience but something closer to sheer horror.
If myoelectric systems contain systematic constraints that limit their speed and reliability, one would be better advised to investigate and consider these constraints regardless of manufacturers' promises. Then, cable controlled or body powered prostheses - that are both lighting fast and relatively reliable - would be the technology to push, the prosthetic development to invest.
Because oddly enough, a major part of current prosthetic hand developments seem to focus entirely on myoelectric prostheses. There is no really modern, prosthetic high-tech hand that mimics a normal hand and is purely mechanical even though that would definitely be the next killer product. There is the X-Finger. But manufacturers of prosthetic hands seem to strongly favor myoelectric hands.
What seems to be behind these assumptions or myths of reliability and speed issues?
Watch this video featuring an older model of the "iLimb" hand, that features a delay each time an action is ready to be shown, abd that somewhat seems to get out of control due to reasons pertaining to myoelectric control - not the terminal device "iLimb":
The following video of the Otto Bock Sensorhand Speed shows a similar time lag involved in handling items using a myoelectric setup:
This footage shows the new Michelangelo hand by Otto Bock and while its so-called bio-mimicry (see the guy wave his hand and how it appears 'relaxed') is certainly impressive, the time lag is as substantial as it is for other myoelectric prostheses:
The delay seen in the videos is likely to not get much faster as it seems to be intrinsic to myoelectric prosthetic technology: cited from Todd A. Kuiken, Guanglin Li, Blair A. Lock, et al. (2009) Targeted Muscle Reinnervation for Real-time Myoelectric Control of Multifunction Artificial Arms. JAMA. 2009;301(6):619-628:
Faster is clearly better, but what is good enough? Farrell et al found that participants did not appreciate a time delay of less than 100 ms and others have advocated that a delay of up to 300 to 400 ms is acceptable [2,3,4]. Thus the motion selection times of TMR patients for arm function are quite good (!220 ms), and the motion selection time for hand grasps is perhaps marginal at 380ms. It should be noted that this delay is not due to computational processing,which takes only a few milliseconds. The motion selection time delay is intrinsic to EMG control, because it represents the need for sufficient data to accumulate before an accurate decision can be made.
The issue of intrinsic design and control problems (that cannot be overcome by design) related to both myoelectric and body-powered prosthetic arm usage / design / build is discussed in this freely available article in depth:
Myoelectric control error rates
In about four decades, academically published myoelectric arm control error rates remained (a) unacceptably high and (b) became slightly worse [very detailed analysis here].
 Farrel TR.Multifunctional Prosthesis Control: The Effect of Targeting Surface vs. Intramuscular Electrodes on Classification Accuracy, and Effect of Controller Delay on Prothesis Performance [dissertation]. Evanston, IL: Northwestern University; 2007
 Englehart K, Hudgins B. A robust, real-time control scheme for multifunction myoelectric control. IEEE Trans Biomed Eng. 2003;50(7):848-854.
 Englehart K,Hudgins B,ChanAD. Continuous multifunction myoelectric control using pattern recognition.
Technol Disabil. 2003;15(2):95-103.
 Hefftner G, Zucchini W, Jaros GG. The electromyogram (EMG) as a control signal for functional neuromuscular stimulation, I: autoregressive modeling as a means of EMG signature discrimination. IEEE Trans Biomed Eng. 1988;35(4):230-237.