June 7, 2004Link
Wireless technology lets us talk on cell phones with people thousands of miles away, surf the Web without a cable and control our stereos, DVD players and televisions. But none of this technology works without pushing buttons or giving voice commands. Imagine what it would be like if we could turn our brains into remote controls, sending wireless commands to computers, robots and other machines.
It's not so farfetched. Like a computer, the brain is made up of many little units wired together to process information digitally. Where computers use zeros and ones, neurons encode our thoughts in all-or-nothing electrical impulses. And if computers and brains speak the same language, it should be possible for the two to speak to each other.
Researchers hope ultimately to eavesdrop on the brain's digital crackle with electrodes, transmit those signals to a computer that can read the brain's code and then use those signals to control a machine. Imagine a quadriplegic person able to operate a robotic arm mounted on a wheelchair with merely a thought. Imagine a digital stream flowing from a microphone into a deaf person's auditory cortex, where it could become the perception of sound.
These dreams have an official name: brain-machine interfaces. A decade ago they seemed little more than fantasy, but now their emergence seems like just a matter of time. At the Center for Neuroengineering at Duke University, monkeys with electrodes surgically implanted in their brains move robotic arms with their minds alone. The electrodes pick up signals from neurons that normally would produce hand movements, and a computer translates those instructions into commands that drive the robot. The translation happens almost instantaneously, and is sophisticated enough to allow the monkey to do more than move the arm. It can also squeeze the gripper at the end of the robotic arm as hard or as lightly as it pleases.
The Duke neuroengineers are now moving from monkeys to people. In the July 2004 issue of the Journal of Neurosurgery, they report their success at temporarily implanting their electrodes into the brains of volunteers. (The subjects were undergoing surgery for Parkinson's disease and other tremor disorders.) The patients then played videogames while the electrodes recorded the brain signals. The scientists trained a computer to recognize the brain activity corresponding to the different movements of the joystick‚€”the first step toward translating brain commands into computer ones. Now the Duke researchers want to do long-term research on electrodes implanted in quadriplegics.
In its current form, the Duke brain-machine interface isn't pretty. Cables run out of the test subject's skull, Borg style. The design is not just ugly, but unhealthy‚€”the opening for the wiring could let in infection. The Duke neuroengineers are hoping to make their brain-machine interface wireless: electrodes buried in the brain would relay signals to a transmitter embedded in the skull, which in turn would send them as radio waves to a receiver attached to the scalp. The receiver would then pass the signals to a miniature computer a person might wear on his or her belt. The device would wirelessly send commands to a robotic arm or some other machine.
If you don't need a cable to transmit signals from your brain, then you aren't limited by a cable's reach, either. You could send those signals through the Internet to a machine thousands of miles away. You could uplink them through a satellite to a rover prowling around on Mars. Consider the possibility of electrodes implanted in the language centers of the brain, wirelessly transmitting your inner voice thousands of miles away. You might choose instead to send them to someone standing nearby with electrodes implanted in his or her hearing centers. Telepathy, anyone? Or, if you take a bleaker view of the future, mind control?
Huge hurdles remain between today's state-of-the-art and these possibilities. The fact that scientists can decode hand-moving brain signals is no guarantee that other signals‚€”incoming touch or outgoing speech, for example‚€”will be as easy to master. Basic hardware challenges, such as getting more power to the internal transmitters, have yet to be solved. And making brains Wi-Fi will still involve surgery, unless someone can figure out how to monitor neurons from the outside. On the other hand, few would have imagined a decade ago that monkeys would now be running robots with their minds. When it comes to Wi-Fi, it may not be wise to bet against the future.
Copyright 2004 Carl Zimmer