Serendip is an independent site partnering with faculty at multiple colleges and universities around the world. Happy exploring!

Brain-Machine Interfaces: An Upgrade to Nature's Design?

SerendipUpdate's picture
Biology 103
2005 First Paper
On Serendip

Brain-Machine Interfaces: An Upgrade to Nature's Design?

Zach Withers

Brain-Machine Interfaces, or BMI, are an attempt to artificially replicate the electrical-to-mechanical translation done naturally in animal muscles. The technology, as it currently stands, consists of electrodes placed in the brain to pick up the signals traveling down individual motor neurons. It has been found that this can be done by "polling" the neurons – taking a sample of the neurons you want, and extrapolating the signals in the rest of them. Like any poll, there is a margin of error. Also like any poll, the population of neurons is simply too large to ask them all. The process works by analyzing the subject's neural activity – what firing neurons correlate with what moving limbs? When that relationship is well understood, the BMI can then extrapolate what movement the subject intends based on which neurons are firing, and replicate that intended motion with an outside mechanical object. (1)

The seminal study in the field of BMI was run in the mid 1990s at Hahneman University, when researchers wired a rat's brain with electrodes and taught it to control a lever with its mind. The rat was taught to push a bar for water; the bar was connected to a lever, which was connected to a button which dispensed the water. As it did this, its brain functions were recorded. After some time, the mechanism on the lever was changed. It was disconnected from the bar; rather, it was computer controlled, and the computer was connected to the rat. Now, rather than pushing the bar, it had to think of whatever it was thinking when it pushed it before. It worked. Despite initial frustration that the bar was no longer working, the rat eventually figured out how to get the water flowing again. Soon, it wasn't even touching the bar, simply looking at it from across the room, and presumably imagining touching it. (1)

As computer technology improved through the early 2000s, more sophisticated studies were done, first using smooth-brained spider monkeys, where computers were able to accurately predict the monkeys' arm movement from its neural signals and make two robotic arms move in synchrony with the monkey's, and later with wrinkle-brained macaque monkeys, which were first taught to hit targets on a screen with a joystick – and later, to hit them by thinking about using the joystick. (1)

At present, the studies have advanced to humans. One of the hoped for products of this technology would be the restoration of limb movement, or at least some way of interacting with the world, to paralyzed individuals. At least two studies are currently underway in this direction, one at the University of Freiberg, and one joint project between Brown University and Cyberkinetics Corp. (2) (5) The Cyberkinetics study appears to be making excellent progress – they report a 25 year old quadriplegic has successfully been able to switch on lights, adjust the volume on a TV, change channels and read e-mail using only his brain, and was able to do these tasks while carrying on a conversation and moving his head at the same time. (4) The original project for more realistic prosthesis continues apace as well, as a team at Pitt has designed a BMI-prosthesis assembly precise enough to allow a monkey to pick up and eat food using only the robotic arm. (3)

As with so much of new technology, the military is one of the greatest backers of BMI. Some military uses of BMI seem fairly logical, such as directly integrating a pilot with his aircraft. (4) Pilots have long said that after a while, their machines feel like extensions of their body. Why not make it literally true? However, the military seem to be going well beyond this scope. If Cyberkinetics seems inspired by Luke Skywalker's prosthetic hand, the Pentagon is more intrigued not by the Terminator, but by Neo's knowledge of kung fu. DARPA's Eric Eisenstadt described a few of the military's goals for BMI in a recent speech. The projects he describe are far from what I would expect. Rather than building armies of invincible machine men, Eisenstadt wants to use the technology to prize data out of the biological brain. He talks of implanting electrodes into bomb-sniffing dogs' olfactory lobes to determine exactly what it is they are smelling, of perfect lie detectors, of directly recording datastreams from soldiers' minds. Also of interest to the Pentagon – and a project they have already begun and shown results from – is reverse-BMI, the translation of artifical nerve stimulus into a real brain, rather than real nerve stimulus into an artificial limb. Dr. Eisenstadt showed his audience a movie of "roborat", a real rat putting itself into highly unlikely positions, due to instructions being fed to its motor neurons through BMI. (6)

BMI, in its medical form, raises few if any moral qualms for me. I simply fail to see the difference between the brain's controlling a meat limb and a robotic one. Even the Pentagon's more straightforward projects don't really phase me. A pilot mentally fusing with his aircraft may be kind of weird, but it's only a natural extension of what they already do – what anyone using a tool does. The projects deeply involving themselves with the internal workings of the mind, on the other hand, deeply disturb me. Unquestionably, knowledge is power, and the military needs all the power it can get. Nonetheless, there is some power that I would not grant it, and power over the mind is one of those things. Even more disturbing than the fact that they might use these things on the enemy is that that they might use them on our own soldiers. An army of automatons would unquestionably be more efficient than an army of men, as long as the processes involving decision-making were left untouched. Brainwashing has always been a military trademark, but there has always been limits to it. If given unlimited access to a soldier's thoughts, I worry what the army would turn them into. These projects, however, pale in contrast to the Roborat experiment. This is just monstrous. A simple baseline fact of the world has always been that you can't make anyone do anything. You can ask, tell, beg, plead, cajole, and threaten, but in the end, they have to do it. The implications of Roborat alter this fundamental principle.

One has to question where BMI is taking us. The promised medical benefits and the question of whether we can decode neural impulses are simply to great to ignore – this is not a line of research we simply can close down. It also is one with which we must proceed with great caution.

Works Cited

1. Controlling Robots with the Mind - Scientific American article by leader of Duke monkey experiments
2. Brain-Machine Interface - Project C5 - Synopsis of project to develop BMI for prosthetic limbs
3. Scientists demonstrate a mind-controlled future - Account of U Pittsburgh BMI study
4. 'Briangate' Brain-Machine-Interface Takes Shape - Description of Cyberkinetics Co, and their prototype BMI for quadriplegics
5. Cyberkinetics Inc. - Company referenced above
6. Remarks by Dr. Eric Eisenstadt - Speech by DARPA manager on military applications of BMI
7. The Brian-Machine Interface – Paper on BMI by U South Carolina
pre-med Nidhi Kuma
8. Monkeys Adapt Robot Arm as Their Own - Duke University Press Release



Comments made prior to 2007

Has anyone at MITRE corporation or at DARPA ever claimed that they have created a brain machine interface using nanochips? ... Kyle Ogden, 30 January 2006