The U.S. military’s scientific and technological wing, the Defense Advanced Research Projects Agency (DARPA), is tasked with predicting and constructing the future. Robotic soldiers, artificial intelligence (AI), and “vampire drones” that disappear in sunlight have all been in development for some time. Now, DARPA has announced that it is seeking to develop an electrical interface between the brain and a computer system, allowing the two to “talk” to each other.

This isn’t DARPA’s first foray into brain implants. Only just last year, a project designed to use electrodes within the brain to stimulate memory formation in those suffering from neurological damage was green-lit. This new program, Neural Engineering System Design (NESD), aims to turn a science fiction concept into reality.

The brain operates using electrical signals, which are initiated by the transmission of specific chemicals called neurotransmitters between brain cells. These electrical signals are not too dissimilar to those used by computers, with the key difference being that a computer’s communication language conventionally uses binary signals, which represents information as a series of ones and zeros.

The brain, on the other hand, is far more complex, with billions of electrochemical transmissions per second translating into thoughts and actions. DARPA recognizes that these electrical signals could be isolated and translated into information that an artificial interface may be able to understand. After all, if they can develop thought-controlled prosthetics, where an artificial limb is in direct communication with the brain, why can’t a computer be directly connected to the brain in the same way?

^{DARPA’s Revolutionizing Prosthetics program has managed to develop fairly precise thought-controlled limbs. DARPA}

These types of neural interfaces already exist, produced as part of DARPA’s Reliable Neural-Interface Technology (RE-NET) program, but they are fairly primitive. Up to 100 implants or “channels,” each connected to tens of thousands of neurons, are able to record and encode information that a computer can recognize as representing specific neurological activity. However, this data is full of “noise,” and is frequently inaccurate.

The new project hopes to drastically up the ante: DARPA is hoping to be able to translate and encode information from more than one million neurons simultaneously. Not only that, but they hope to develop a feedback system, wherein the computer interface is able to send electrical signals back to the brain, and stimulate at least 100,000 neurons – particularly those associated with audio, visual and somatosensory (touch, pain, pressure, movement) functions.

As with many DARPA projects, the specific aims, goals and technologies involved in NESD are not made explicit. It does admit, however, that enormous advances in many scientific fields are required, including those in synthetic biology, electronics and neuroscience. The issue of making the implants safe to use in humans, and finding willing subjects, also looms large over the project.

In addition, although connecting a computer and allowing it to communicate with up to a million neurons does sound impressive, it’s worth noting that the average adult brain contains around 86 billion of them – so there’s a lot of ground left to cover.

This announcement, then, is really about letting the world know that NESD is up and running. Regardless, the ambition is clear to see: DARPA hope to have a working demonstration of the system within the next four years.