IM this article to a friend!

September 12, 2005

Medical Implants Edge Toward 'Bionic Man'

From: InformationWeek, NY - Sep 12, 2005

Electronic components can already shore up faltering human capabilities, from artificial limbs to implantable heart devices. Can chips in the brain be too far off?

By Chappell Brown, EE Times
Sept. 12, 2005

Cell phones have been a big hit because they untether communications technology from the desk, giving anyone that capability anywhere. The logic of this trend would suggest that the next step might be to implant a tiny cell phone-on-a-chip directly into the cerebral cortex. With dense electrode connections to the speech center of the brain and RF communications to the nearest Wi-Fi hotspot, it would be the next best thing to being psychic.

Sound extreme? Maybe. But this is a pivotal moment in the history of electronics, when technology is morphing from something that extends the body to something that merges with it. Electronic components can shore up faltering human capabilities, in the form of artificial limbs that use real-time embedded control, implantable defibrillators that correct the heart's rhythms, and artificial ear and eye systems that help the deaf and the blind.

"There are a variety of people in the investment community, and I share their view, who believe that this century is clearly the century of the 'bionic man' — the human being whose life style has been improved substantially through the use of these electronic enhancements," said

Nicholas Colella, senior vice president in the Product Miniaturization Division at Tessera Inc. (San Jose, Calif.). Tessera has been developing miniaturized chip-scale packaging over the past 10 years, a critical technology in this area.

Yet, for all their promise, the developments in implant electronics also carry profound implications for medicine and ethics. On a more mundane level, they present major design challenges. Connecting inorganic silicon circuits to the delicate nerve nets of the body poses a host of problems in itself. And then comes the hurdle of building safe, highly compact, extremely low-power components that can exist independently in the body for decades.

The market, meanwhile, has barely begun. Cochlear implants — hearing aids that link directly to the auditory nerves in the ear, while the rest of the system is worn externally — are now being marketed as products, but vision research, which is inherently more difficult, is still in the R&D stage. Compact vision systems that can be worn on glasses and connect directly to the optic nerve have been demonstrated. Elsewhere, implantable control systems that monitor and correct heart function have become common. Further down the road are electrode implants that would cure paralysis resulting from nerve damage.

The quest to heal such ills has the blessing of large institutions like the National Institutes of Health, hospital research centers and charities. It also has a feel-good component that makes it possible to accept — or at least contemplate — the idea of rather bizarre intrusions into the body. But when the usual dynamic of commercialization takes hold, ethics and conventional morality tend to work against the notion of the bionic man.

Consider, for example, a small step toward such futuristic visions of im-plantable electronics: A couple of years ago, a number of individuals volunteered to have radio frequency ID tags implanted under their skin. This is a medically benign operation, but the social implications are a little scary. The demo project touched off a heated debate over issues such as personal privacy and whether the body should be declared off-limits to invasive technology.

If a device is surgically implanted, it needs to be highly reliable and ultralow in power consumption in order to remain viable for periods lasting decades. "Medical devices have to have very good reliability, requiring long-term tests on your material," said Mike Warner, chief engineer at Tessera. Then, too, "everything has to be sterilized, and you sometimes have to use specialized materials that are compatible with the body."

Indeed, a major problem for the field of implantable electronics is the lack of long-term studies on circuits and materials. That data is crucial for the product design stage of development, and explains why the market will be very slow to take off.

Another need is ultrasmall packages. "If you think about the traditional electronic system such as a radio or television, if you used [Tessera's] miniaturized 3-D packaging approach that is now available to the medical market, you could reduce that to a 10th of the volume and weight," Warner said. "This technology is being used in commercial products in very high-volume PDAs, cell phones, things like that, so it is a technology that is mature from the manufacturing and cost standpoint."

As for what will be going into the chip-scale packages that Tessera has developed, it most likely will not be state-of-the art digital processors. The power of Pentium-class CPUs would certainly be welcomed for the demanding information-processing algorithms needed in medical implants — but not the power consumption and waste heat.

It is here that a new design frontier using CMOS as a low-power analog technology is emerging. Analog systems can attain the computational levels required for interfacing with the nervous system at very low power, said Rahul Sharpeshkar, director of MIT's Analog VLSI and Biological Systems Group. Sharpeshkar is developing a fully implantable replacement for the cochlea, the sound-processing system that resides in the inner ear.

"The constraints with a fully implanted system are that it has to operate on a battery without [additional] surgery for about 30 years. If that is the case, the electronics have to be very, very low power," he said. "I have just designed a processor that can run on a 100-milliampere battery for 30 years. There is a company that hopefully will commercialize it soon."

Sharpeshkar and his colleagues had to come up with a bandpass filter that could operate in the 100- to 200-Hz range with a dynamic range of 66 dB while consuming only a few tenths of a microwatt. To achieve those kinds of figures, they used a 1.5-micron BiCMOS technology and ran it at very low voltages.

"Our low power is achieved by running the transistor where everyone thinks it's off, and it has very low leakage currents," Sharpeshkar said. "We exploit its action in that region."

The implantable cochlea makes a good test system for this new design frontier. The circuit has to do signal processing on 16 channels with a sampling rate between 500 Hz and 2 kHz. Even at those specs, the actual sound-processing capability is fairly crude. A basic constraint has been the inability to make very many connections to the auditory-nerve network in the ear. The cochlea has hundreds of millions of tiny hairs that respond to sound vibrations and stimulate nerves. The result is a rich experience of detailed sound.

Artificial cochleas must reduce the design objective to something achievable and vital, such as speech recognition. That requires only a rather crude reception of a series of amplitudes, whereas the full experience of a symphony orchestra, say, would need both amplitude and phase information. With the scaled-down requirement, it is possible to get by with only 16 connections to the auditory nerves.

Such constraints are common in implantable designs. "Whether they are pacemakers or defibrillators — and now people are working on almost fully implanted systems for paralysis — they have to be small, they have to be low power and they have to be wireless," Sharpeshkar said.

While the field is only in its infancy, the prospect of merging information technology intimately with the brain has some startling implications. For example, work on implanting control systems in the brain at Miguel Nicolelis' lab at Duke University Medical Center has led to the speculation that it would be possible to not only control remote robotic systems mentally, but actually perceive them as part of the body. The same type of technology would make it possible to implant databases in the brain that would allow someone to recognize people and know their detailed history, without ever having met them.

And then there is the implantable cell phone — an RF device that could link minds directly. Enhanced vision systems, meanwhile, could detect infrared, ultraviolet light or RF radiation.

It can be done. The technological hurdles are large, but not impossible. Social, ethical and personal issues are, of course, exceedingly complex. When the technology moves beyond therapy into the realm of creating a superior being — a bionic man — there will be much more discussion. Indeed, merging electronics with the body may ultimately change not only how we conceive of ourselves, but what we are.

Copyright © 2005 CMP Media LLC