They can’t extract secret terrorist plots yet, but Utah bioengineers have implanted a brain chip in human test subjects that enables researchers to download brain data onto hard drives. The team working with the chip is hoping to make immediate improvements in the lives of people with epilepsy, paralysis or blindness, but say the chips may one day enable brain-native Internet browsing or most any other function currently possible with a computer.
The Utah Electrode Array’s purpose is analogous to a modem: It relays data from the brain to a computer, and vice-versa. It may soon enable thought control of bionic limbs like Luke Skywalker’s in Star Wars and, further in the future, may help the blind to see.
Neural Engineering Lab supervisor and University of Utah assistant professor Bradley Greger describes the chip as “a platform technology that is going to enable many, many new things.” With a grant from the National Institutes of Health, Greger and dozens of other scientists are pioneering brain-computing technology.
Greger, who described himself and his colleagues as “neuro-geeks and science-fiction fans,” says it’s fun to talk about sci-fi fantasies like brainnative Internet, but they’re focused on life-changing advancements that can help patients now. The chip’s most impressive feat to date is demonstrated by quadriplegic patients “thought typing” and manipulating a computer cursor with brain waves only.
But, Greger says, “If you can get a link to a computer, then you’ve got the whole world.” The Internet’s vast knowledge and tools could be just a thought away. Foreign language translation, photographic memory and brain-data backup could become a reality with use of more advanced chips.
Greger’s not confident that he will live to see brain-native Internet access, but he hopes at least one seemingly sci-fi chip advancement will be accomplished within his lifetime: creating or restoring rudimentary vision to the blind.
The chip is smaller than a dime and implanted in only about 15 people worldwide—some only temporarily. It was invented in 1997 by U of U bioengi neering professor Richard A. Normann and manufactured by Salt Lake City’s Blackrock Microsystems, a U of U startup. Though the chip is a decade old, the device only recently gained U.S. Food and Drug Administration approval for human trials.
Brown University researchers have implanted the Utah array in seven paralyzed patients who are using the chip to “thought type.”
U of U researchers have implanted the chip in four others to study the causes of epilepsy and to gather data for the brain-control of prosthetic limbs. Greger says another four or five epilepsy patients have received the brain chip at Columbia University—and that’s it. “This is the only human-approved device of this type,” Greger says. “These are the very earliest human trials.”
The current chip is “wired.” That is, patients’ brains must be physically “plugged in” to a computer to harness its powers. Blackrock has a wireless prototype planned for submission to the FDA in the coming years, however. Removing the wires is important because wired devices leave an exposure of the brain to the outside world, creating a risk of infection.
But wireless communications present significant engineering problems. The brain produces so much data even for simple tasks—like controlling one arm—that transmitting wirelessly is not possible with current wireless modem technology of any size.
Improving wireless transmission, then shrinking that capacity into a tiny chip, may take decades, Greger says. The wireless modem also needs a power source, but Greger says transmitting power to the chip may be possible wirelessly in the future as well.
Medical ethicists contacted by City Weekly declined to comment on the Utah Electrode Array, stating they are unfamiliar with its capabilities. Greger admits ethical issues will need to be addressed as the technology advances. What becomes of privacy in the brain-computing future? And should anyone who wants a brain chip be eligible for one, or must a patient have a medical need? Greger’s not worried about those issues yet because advancements thus far have come so slowly. “I’m not worried about anyone invading neural privacy yet,” Greger says.
The Utah patients are volunteers already scheduled to receive brain surgery for treatment of epilepsy. The U of U neurosurgeons implant the chip on the section of the brain known to cause seizures. Over the course of several days, they collect data during seizures. Then the chip comes out. At Brown, the chip is left in, patients go home, and they have someone—usually a research assistant—plug in their brains when they want to thought-type an e-mail or play Solitaire.
They know basically how the chip reads the brain data—electrons travel over platinum wires—but they can’t say precisely how or why the brain “learns” to use it.
“After a length of implantation the brain starts to use this and the control gets better with time,” says Kyle Thomson, a graduate student on Greger’s team, “not because the electrode got better or the decodes got better, but simply because the brain got better and says, ‘I want to use this. We’re going to use it, no matter what.’”
Once implanted, the brain chip uploads a terabyte of brain data every 46 hours, the equivalent of about 1,500 fulllength movies. That’s only a tiny fraction of the brain’s total data, but it’s as much at the Utah array’s 100 electrodes can collect. A better chip with more electrodes would provide more data.
Downloading the brain data is the easy part, Greger says, but no one understands very well the brain’s language of “voltage signals in a funny hybrid of analog and digital signals.” They can tell from the brain data whether the patient says “yes” versus “no” or moved a right arm versus a left, but that’s almost the extent of their understanding. A mathematical algorithm that makes sense of the brain’s language, Greger says, probably will earn someone a Nobel Prize when it is discovered.