MOSCOW, 07 Jan 2022, RUSSTRAT Institute.
Brain chips went into implementation: a paralysed Australian wrote the first ever tweet using a neurocomputer interface.
62-year-old Australian Philip O’Keefe, a partially paralysed disabled person with amyotrophic lateral sclerosis, became the first person in the world to write a message on Twitter with just the power of thought transmitted through a neuroprosthesis in the brain. “hello, world! Short tweet. Monumental progress” was the first of seven remarks O’Keefe left on December 23, 2021 during a 36-minute experiment.
It was conducted by the American company Synchron, which implanted its brain computer interface (BCI) under the name Stentrode in April 2020 in O’Keefe. Philip is not the only owner of it: another similar prosthesis was implanted in 76-year-old Graham Felstead, also an Australian. According to the developer company, both paralysed patients can now use a computer, including sending mail and shopping.
Despite the beautiful presentations and enthusiastic reviews in social networks, Synchron has yet to prove the revolutionary nature of its development: is it really so practical and safe? In any case, the US Food and Drug Administration (FDA), although it issued Synchron in 2020 the first ever BCI-implants approval for testing Stentrode as a “breakthrough device”, has not yet certified it.
Anyway, two circumstances force us to take a closer look at this company. Firstly, it is headed not just by another “visionary in a turtleneck”, but by Thomas Oxley, head of the Vascular Bionics Laboratory at the University of Melbourne, who has hundreds of scientific papers in refereed journals. And secondly: Synchron received the initial grant for its research back in 2016 from the notorious DARPA – the Defence Advanced Research Projects Agency of the US Department of Defence.
Neuro-prosthesis in the brain. Without surgery
What is the essence of Oxley’s invention? His device takes impulses from the cerebral cortex, processes them and transmits them to a computer, allowing to control the latter. It does this “from inside” the skull, but it does not require open surgery on the brain, since it is injected endovascularly – through blood vessels. And this is the most important advantage of Stentrode, which combines the strengths of invasive (with penetration into the human body) and non-invasive BCIs, and without the disadvantages of both – high health risks and low accuracy of “data collection”.
Imagine a wire made of a flexible alloy of titanium and nickel, woven in the form of a grid 40 mm long and 8 mm wide, on which there are 16 sensors reading brain signals. This is the neuro-prosthesis from Thomas Oxley. The device is inserted using a catheter through the jugular vein into the upper sagittal sinus (venous canal running between the hemispheres of the brain) – almost under the very crown, in the area of the motor cortex.
This part of the cerebral cortex is responsible for human movements. As a rule, patients with amyotrophic lateral sclerosis have nerve cells that transmit impulses from there to the extremities, but the motor cortex itself remains intact. So, Stentrode is just designed to replace out-of-order neurones.
The prosthesis registers signals from the cortex and transmits them via a flexible 50-centimetre wire running inside the jugular vein to a small receiving device in the chest. It is implanted under the skin and powered by induction, without wires. From this “receiver” the wireless signal goes further, to an external transmitter attached to the human body, and through it – already in processed digital form – to a tablet or computer with a special program.
Tablet management looks like this. The long-known computer technology of gaze tracking is used to move the cursor across the screen – this can be observed both in the explanatory picture and in the Synchron demo video. And the neuro-prosthesis itself transmits the command to click the cursor directly from the brain to the computer at the right moment. That is, strictly speaking, the process of writing a tweet by poor O’Keefe consisted not in freely inventing text that would immediately appear on the screen, but in concentrated “clicking” in the right places on the screen through mental commands.
Perhaps the main disadvantage of this technology, rarely mentioned in presentations, is connected with this – it needs a careful setup and long training. Both Australians spent several months preparing, training and testing until their accuracy in selecting computer clicks exceeded 90%.
But, unlike the more common neurocomputer interfaces created on the basis of electroencephalography and attached to the scalp, the Stentrode prosthesis “understands” not one, but at least three different commands for the cursor: “no click”, “short click” and “long click”. And in this sense, we are facing another breakthrough, in addition to the non-surgical insertion of a neuro-prosthesis.
From Vidal to Musk
BCI interfaces – more precisely, attempts to create workable models of them – have long been known to science. They have been studied since the 1970s and are divided into a fair number of subspecies – both implanted in humans and external equipment. However, all such work is still at best limited to the level of a “walking stick for the disabled”.
Nevertheless, already half a century ago, experiments on controlling mechanisms using signals from the human brain seriously interested the American military. It is no coincidence that the author of the term “brain-computer interface” Jacques Vidal from the University of California at Los Angeles received a contract from DARPA. The Pentagon’s interest was understandable: it could actually be about creating combat cyborgs – half-humans-half-machines. Or “at least” about the control of weapons by mental commands.
The first neuro-interface successfully implanted into the human brain appeared back in 1998 at Emory University (Atlanta, USA). Even then, the Vietnam War veteran Johnny Ray, who suffered a stroke, learned to use such a prosthesis to control the cursor on a computer screen.
In 2004, scientists from Brown University (Providence, USA) implanted a matrix with BrainGate microelectrodes into the area of the motor cortex of ex-football player Matt Nagle, who was paralysed as a result of a knife wound. With its help, he was also able to control a cursor, as well as control a robotic arm. The 96-electrode BrainGate blocks have become a kind of standard in the field of invasive BCIs.
Experiments with quadrocopters and artificial hands controlled mentally continue to this day, but more recently, all of them have been eclipsed by the well-known entrepreneur Elon Musk. In the spring of 2017, the head of SpaceX and Tesla announced the Neuralink startup to create its own neurocomputer interfaces, and ideally – not for sick, but for healthy customers.
According to Musk, in comparison with his future implants in the form of polymer threads, the vaunted BrainGate looks like a “medieval instrument of torture”. So far, however, the progress of Neuralink is limited to the implantation of a “chip” into the brain of a pig and a somewhat more impressive game of a monkey in “mental ping-pong” on a computer screen.
One should not, however, think that work on neural interface chips has reached a dead end, turning into the lot of lone freaks. Dozens of American universities, hundreds of technology companies and research centres are involved in research in this area right now. Not least because they are generously paid from the US state budget. More precisely, from their military budget.
DARPA prepares for the future
According to legend, Australian researcher Thomas Oxley almost from the street called people at DARPA to ask them for money for his development, and they immediately gave him $1 million. Maybe that’s how it really was. What is known for sure is that the Management has introduced the activities of Synchron into the framework of its program to create the so-called reliable neural interface technology (RE-NET).
According to the US military, RE-NET is an exclusively peaceful technology designed to efficiently extract information from the brain to control prosthetic limbs in seriously wounded servicemen. But it is clear that such programs always have a dual purpose: it is only necessary to teach the brain to cope with a robotic leg, as it already fires a machine gun with the power of thought.
However, RE-NET is just one of several DARPA programs within the framework of a whole bunch of research called the BRAIN Initiative. The initiative, announced by the White House in April 2013, received initial funding of $100 million. In addition to similar dual-use programs that can be easily converted to weapon control, there is mentioned, in particular, the so-called Next-Generation Nonsurgical Neurotechnology (N3).
We are talking about a portable interface chip, which, it seems, will be temporarily inserted directly into a person’s head in order to read and write data to several points of the brain at once – moreover, in healthy people. And here the Pentagon’s interest in Oxley’s similar development becomes clear.
In DARPA’s veiled terminology about the N3 technology, it says this: “The Battelle team <...> aims to develop a minimally invasive interface system connecting an external transceiver with electromagnetic nanotransformers that are non-surgically delivered to the neurons of interest.” Very similar to Stentrode, only there’s no need to implant the BCI itself.
Battelle’s own website describes this more specifically: “wearable brain-machine interfaces that could ultimately enable diverse national security applications such as control of active cyber defence systems and swarms of unmanned aerial vehicles, or teaming with computer systems to multitask during complex missions.”
There is also a recognition of the curator of the N3 program, Dr. Al Emondi, who puts the dots on the “I”:
“DARPA is preparing for a future in which a combination of unmanned systems, artificial intelligence, and cyber operations may cause conflicts to play out on timelines that are too short for humans to effectively manage with current technology alone. By creating a more accessible brain-machine interface that doesn’t require surgery to use, DARPA could deliver tools that allow mission commanders to remain meaningfully involved in dynamic operations that unfold at rapid speed.”
In the future, military personnel will be able to put on a headset with a neural interface, use the technology as they need, and put the instrument aside at the end of the mission, Dr. Emondi continues. Moreover, such a headset can be built directly into the helmet.
It remains to add that in Russia, too, work is underway to create various brain-computer neural interfaces, some of which are described in the open press. And although such devices are positioned as purely civilian, the profile of the development companies, as well as their parent organisations, eloquently testifies: our professors are also preparing for battle.