The implant in each of these schemes is powered by an electromagnetic induction coil outside the skull. A cap or headband containing a coil no more than a few centimeters wide could send power to the device. An external power source for neural chips makes more sense than batteries, Harrison says, because the chips consume a hundred times as much current as pacemakers do. This means their batteries would need to be replaced much more often than a pacemaker battery, whose typical life span is seven years.
Wireless neural implants open up the possibility of embedding multiple chips in the brain, enabling them to read more and different types of neurons and allowing more complicated thoughts to be converted into action. Thus, for example, a person with a paralyzed arm might be able to play sports. "When we hit a tennis ball or kick a soccer ball, we plan things first...and then execute an action based on input," Nurmikko says. "The brain is furiously calculating what it's going to do with this thing that's coming at you."
Eventually, you would want to listen in on hundreds or even thousands of neurons. But then infrared or RF transmission bandwidth would be a constraint, observes Babak Ziaie, an electrical and computer engineering professor at Purdue University, in West Lafayette, Indiana. At the EMBC meeting, Ziaie presented an optical approach: An LED array, possibly attached to the skull, could convert the electrodes' signals into light pulses that are captured by a high-speed camera chip and reconverted into electrical signals. He plans to test the scheme on animals this fall.
