I sat in on the EUROHAND 2008 conference that the organizers were nice enough to allow me for a special rate. I am posting my (and not an official) summary here.
Ther are 3 major directions of research aimed at helping arm / forearm amputees:
2) Regrowing limb.
3) Prosthetic technology.
About 1): Immunsuppression is a very, very serious problem. The presenter said it was a very important consideration. I completely agree. I did my doctoral thesis on some immunosuppression side effects (Sandimmun / Cyclosporin versus some previous drug combinations) and I am not sure whether transplant extremity replacement (with its own surgical and recovery related problems) is just good to recommend. It is definitely NOT a treatment you'd "just get". You could be seriously limited otherwise, i.e., trading in one disability for another, probably more serious one. But then, losing a hand is at least as much an emotional issue as it is a practical one so it must have a role for some (few, select) individuals. Ultimately, the decisions are made between doctor, patient and maybe insurance - and I firmly believe that is how it should be.
About 2): Still science fiction, this is a very very cool direction of research. Maybe one day they can regrow our hands. Right now, this is not for application.
About 3): This direction was portrayed as having one major advantage: it is basically risk free for the wearer, and, prostheses can be repaired free of risk / pain for the wearer - at least for the most part, that is, if no parts are implanted.
One major direction of research is picking up nerve signals. The problem to pick up nerve signals through the skin is that they are very weak. One way is to tap the nerve directly by placing an implant, the other way is to amplify the nerve signals by giving each nerve a small pack of muscles to twitch.
One British cybertechnology researcher had exerimented with nerve implants,and lived with a nerve implant for a while.
The technique to surgically split nerves into bundles and to assign each bunlde a pack of muscle fibers by surgically splitting a muscle is now used in a number of arm amputees. This procedure is called targeted reinnervation. For that, surgery is performed that provides these split muscles / nerves. The result allows the patient to separately trigger some 5-10 single switches on a myoelectric prosthesis.
A Swedish group is going another route. They place up to 14 electrodes on a (surgically unmodified, regular) forearm stump, and at the same time, have the patient wear a 3D coordinate sensing data glove on the (healthy) hand. The patient then activates stump muscles "as if" performing a certain complex movement and does the same movement with the (glove wearing) hand. Thus, the system wants to learn different activation patterns of the electrodes and allows "natural thought" / "naturally stored information" (you don't have to re-learn or get operated on specifically for that) to trigger even complex hand movements. However, they are still very much working on prototypes.