Dispatch April 2009

The Real Arms Race

In the foreseeable future, artificial limbs may be able to perform as well as or better than natural ones.
More

Regenerative medicine, which seeks to replace organs, tissues, and cells damaged by disease, injured by trauma, or worn by time (see my earlier posting on this subject, "Stem Cells and Beyond," March 9), holds much promise. But at present it is not possible to regenerate entire limbs lost to accident or war. Until the field of regenerative medicine advances further, artificial limbs offer the most hope for those who have lost an extremity. And thanks to remarkable progress in microelectronics, computer science, micro-engineering, and neurosurgery, we are nearing a time when artificial arms may move as spontaneously and naturally as the original.

A recent issue of the Journal of the American Medical Association* describes one notable advance by a team of neurosurgeons, engineers and rehabilitation specialists working in Chicago and Canada, funded by both the Defense Advanced Research Projects Agency and the National Institutes of Health. Their goal is to create an artificial arm that naturally responds to signals from the brain.

The first step of the process is to surgically relocate nerves that signal absent muscles. For example, the ends of severed nerves that would connect to the biceps or forearm are stitched to separate muscles of the chest. After the body heals, a signal from the brain that normally contracts a muscle in the arm now tightens one in the chest. An array of sensors placed over the chest to detect the slightest muscle tension relays signals to a computer, which in turn sends signals to a set of tiny motors that move the new limb.

The computer must then learn how to interpret the signals, which it does by connecting the output signals from the computer to a television model of an arm. Patients direct the virtual arm to perform specific tasks such as raising and lowering the forearm, flexing the wrist, opening and closing the hand, and pinching the thumb and forefinger. The patient learns by watching how the virtual arm responds to the intent as signaled by the brain. The computer learns as data from the multiple sensors is sorted into reproducible patterns that will provide input to the individual motors of the mechanical arm.

The researchers compared the abilities of the amputees and people with normal limbs to direct actions of the virtual arm. For the control group, the sensors were placed over muscles normally used to move the hand and arm. Both groups learned simple arm motions, such as raising and lowering the arm, equally well. Those with normal limb function learned fine hand movements more rapidly than the amputees, although some of the amputees learned such movements almost as fast.

Patients were then fitted with mechanical arms, most of which had been developed by DEKA Integrated Solutions Corporation and were capable of ten separate movements. (An arm developed by the Johns Hopkins Applied Physics Laboratory and collaborators was also used.) All the patients were able to perform basic operations using mechanical arms on the very first day of testing. Over the next two weeks, all gained proficiency. Usually each motion was performed separately, although occasionally two separate motions would occur simultaneously, particularly as people reached for objects to pick up. The speed of movements was nearly normal. Eventually, some patients became expert at complex tasks such as picking up checkers pieces rolling across a table, stirring a spoon in a cup, and moving small blocks (see www.jama.com for a demonstration).

This is a remarkable advance. The movement of the mechanical arm, directed by the transplanted nerve, is as intuitive as the motion of a natural arm. People learn to control the positions of the elbow, wrist, and hand in a sequential, ordered, and useful way.

Of course, there are several caveats. The work was done with a group of only five patients. And although during the training period all patients learned to move the arm, some never learned fine movement control. The biggest drawback is that the mechanical arm must be controlled visually. The arm cannot send sensory input to the brain. Designing limbs that transmit sensory signals from the arm to the brain is the next frontier. In the foreseeable future, it's very likely that mechanical arms will be able to perform as well as or better than the original.


* "Targeted Muscle Reinnervation for Real-time Myoelectric Control of Multifunction Artificial Arms" Kuniken et. Al. Journal of the American Medical Association. Volume 301, pages 619-628, February 2009.

William Haseltine is a scientist, businessman and philanthropist. For much of the '70s, '80s and '90s he was a professor at Harvard Medical School, where he researched cancer and HIV/AIDS. He is also the founder of several companies, including Human Genome Sciences, where he served as Chairman and CEO. He is President of the William A. Haseltine Foundation for medical Sciences and the Arts.
Jump to comments
Presented by

William Haseltine is a former professor at Harvard Medical School, where he researched cancer and HIV/AIDS. He is the founder of Human Genome Sciences, where he served as chairman and CEO, and the president of the William A Haseltine Foundation for Medical Sciences and the Arts. He lives in Washington, D.C.

Get Today's Top Stories in Your Inbox (preview)

Sad Desk Lunch: Is This How You Want to Die?

How to avoid working through lunch, and diseases related to social isolation.


Elsewhere on the web

Join the Discussion

After you comment, click Post. If you’re not already logged in you will be asked to log in or register. blog comments powered by Disqus

Video

Where Time Comes From

The clocks that coordinate your cellphone, GPS, and more

Video

Computer Vision Syndrome and You

Save your eyes. Take breaks.

Video

What Happens in 60 Seconds

Quantifying human activity around the world

Writers

Up
Down

More in Technology

More back issues, Sept 1995 to present.

Just In