And it’s more than a little ironic that one of the world leaders in brain-machine interfaces (BMIs) continues to push the technology to its limits, all the while thinking predictions that we will one day be able to recreate the human brain are a “bunch of hot air”.
Miguel Nicolelis, originally from Brazil but now at Duke University, made use of his ties by coupling the rats all the way from Brazil to North Carolina. The techniques he used are nothing new: scientists routinely use electrodes to record or stimulate brain activity in rodents. Nicolelis’ lab itself has pushed BMI technology in the past. A 2011 experiment monkeys controlled a virtual arm, and in 2008 they controlled a robot walking on a treadmill – the monkey was at Duke, the robot in Japan. What the current experiment achieves is a combination of recording and stimulation to produce, according to the researchers, the first brain-to-brain interface.With recording electrodes implanted in its brain, the rat in Brazil was presented two levers and taught to press the lever above which a light turned on. If it picked the correct lever it received a drink reward. After some training the rat was able to press the correct lever 95 percent of the time. Each time it performed the task correctly, its recorded brain signals were sent over the Internet to Duke where they were decoded to a pattern of electrical stimulation in the brain of the second rat. This decoder rat was presented the same choice of levers but did not receive any light cues – it relied solely on the signals sent from the encoder rat to make its choice. Even without cues, it chose the correct lever 70 percent of the time.
In a second experiment the encoder rat had to use its whiskers to discern the size of an opening in its cage wall. If the opening was narrow it was supposed to poke its nose through a hole on the left, if the opening was wide it would poke its nose through a hole on the right. Again, its “correct choice” brain signals allows the decoder rat halfway across the world to also make the correct choice even though it was given no wall opening to discriminate. It chose the correct hole 65 percent of the time, which statistical analysis showed to be better than chance.
The brain signals sent from the encoder rat weren’t so explicit as “Hey, it’s the one on the right!” Rather, the encoder’s motor cortex, a part of the brain that controls movement, became more active when it pushed the correct lever or poked its nose through the right hole. That increase in neuronal firing was sent to the decoder rat’s motor cortex where neurons there were likewise stimulated. The researchers then trained the decoder rat to push the correct lever in response to this stimulation.
And communication didn’t only flow in one direction. In an attempt to train the encoder rat to send clearer signals to the decoder rat, a reward was given to the encoder rat if the decoder rat got the task right. The result was loud-and-clear telepathy as the signal-to-noise ratio of the encoder’s signal improved.
The study was published recently in Nature. Here’s short demonstration of what cross-world rat telepathy looks like.
[Source: BeyondBoundariesBook via YouTube]
Aside from probably the coolest parlor trick you’ve ever seen, what use could this brain-to-brain interface (BTBI) serve? The authors write, “…by coupling the animals’ brains…BTBIs can…enable networks of animal’s brains to exchange, process, and store information and, hence, serve as the basis for studies of novel types of social interaction and for biological computing devices.”
That’s right, Nicolelis and colleagues are attempting nothing short of creating a biological computer. But while quick to acknowledge the group’s numerous and triumphant contributions to BMI field, some fellow neuroscientists are more conservative about the study’s implications.
Lee Miller, a neuroscientist at Northwestern University in Evanston, Illinois, told The Scientist, “I’m afraid one would have to be forgiven, on reading the current account, with its references to, a ‘grid of multiple reciprocally interconnected brains…solving heuristic problems deemed non-computable by a general Turing-machine,’ for mistaking it for a poor Hollywood science fiction script.”
But who knows? If we were to be naively optimistic about technology – as Nicolelis himself apparently thinks Singularitarians are – we could envision a day when a thousand brains coupled through BTBIs could solve a task through the cumulative progressions spurred by millions of minuscule hunches – a kind of instantaneous, subconscious crowdsourcing. And yet here he is being accused of exaggerating the implications of this own work.
Yes, brain-to-brain computers do sound like something out of Hollywood. But Miller and other critics may want to recall the words of Michael Faraday, one of the founding fathers of electromagnetic theory. In 1816, following a lecture, a woman asked, “But, Professor Faraday, even if the effect you explained was obtained, what is the use of it?” Faraday’s reply, so it’s told, was, “Madam, will you tell me the use of a newborn baby?”