One Question Quiz
(Photo: Getty Images)
(Photo: Getty Images)

ScienceApril 19, 2021

The New Zealand scientists trying to make brain-controlled robots a thing

(Photo: Getty Images)
(Photo: Getty Images)

Brain-controlled devices could give people with disabilities or severe injuries new access to the world. But it could also be used to enhance humans, create super soldiers or even transcend the human body entirely. Mirjam Guesgen looks at how far we are willing to go and New Zealand’s role in it all.

Brandon Prestwood gets choked up when retelling what it felt like getting sensation back in his fingers after years as an amputee. The moment his wife reached out and he could feel her touching his prosthetic hand was when he was whole again, he said

Prestwood sports a robotic prosthetic limb that’s wired directly to nerves in his arm, linking his robotic hand to his brain. He controls the hand with his mind. When he thinks about reaching out to greet someone, his hand forms the shape of a handshake. 

He’s the posterboy for a bioengineering lab in Ohio in the US, whose goal is to merge people and technology through so-called brain-machine interfaces.

The sky’s the limit when it comes to what brain-machine interfaces can, and will be able, to do. Giving amputees like Prestwood fully functional limbs, dispatching mind-powered robots to dispose of bombs from afar, and sending thoughts to loved ones like we do emails, are all possible through this pioneering technology.

But it’s not just labs in the US trying to meld mind and machine. A small group of New Zealand researchers and entrepreneurs are also making brain-powered devices. 

A force for therapeutic good

Dmitry Selitskiy runs Thought-Wired, an Auckland and Melbourne based startup that ultimately wants to “let anyone control anything with their mind” as he puts it. 

Thought-Wired’s first piece of tech is nousBlink, a system that picks up on brain waves and miniscule muscle movements through a headband on a person’s forehead and turns the signals into a click on a computer. 

It allows users, people who can’t move their limbs or speak for example, to use a computer to talk, access the web or control devices around their home. 

Selitskiy started looking into brain-machine interfaces after seeing a video of one online. He was excited by the prospect of what they could do for people like his younger cousin, who’s paralysed and doesn’t speak. “Coming across a technology that would allow him to do stuff was very exciting to me,” he says. 

A team of University of Otago computer scientists also hope to develop brain-machine interfaces that could be used for therapeutic purposes, specifically to help treat mental illness. They have investigated how to detect the brain signals when someone’s feeling anxious or depressed. Those signals could later be linked to a therapeutic device.

The device could, for example, give the person an alert to make them more aware of their mental state as a kind of mindfulness training. It’s a high-tech version of a technique called neuro-feedback, which trains people to regulate their own brains by showing them their brain activity. 

These neuro-feedback brain-machine interfaces could help people suffering from a host of mental health issues including ADHD, depression and anxiety.

So far, the team has developed a proof-of-concept headset that reads brainwaves and a computer algorithm that can translate those brainwaves into making a drone fly.

The technological challenges

Getting to the point of a working device, let alone a commercially-viable one, is no walk in the park. 

“I very quickly realised it was very complicated,” says Selitskiy. He says there are few bits of technology already out there that are robust enough to handle differences in people’s brain activity, don’t require wires to be implanted directly into their brains, comfortable and user-friendly enough so that someone actually wants to wear them, and cheap enough to be affordable. 

The last point was perhaps the hardest for Selitskiy to grapple with – finding a way to get the tech out to the people who need it. “The users that you build solutions for in most cases struggle to purchase those solutions,” Selitskiy explained. “If there’s funding or there are insurance schemes in the country where they are, oftentimes they’re either difficult to access or it takes a very long time.” 

The solution has been a “staircase of capability” approach. First, build something more simple that you can get out to people and that investors can latch onto, then build up to your dream tech. 

For Thought-Wired, that meant first focussing its efforts on analysing the muscles that move in the forehead when someone blinks, and developing an artificial intelligence algorithm that learns how to tell a person’s subconscious eye blinks from purposeful ones. Those blinks can then drive a device.

It’s also possible to simplify the system. Instead of trying to analyse many different types of brainwaves, an Otago-based research team has focused on a single channel. “We have external funding from a few Chinese companies who initially wanted video game controllers using EEG [electroencephalographic] technology, reading brainwaves and translating those into commands then sent to the game via Bluetooth,” Associate Professor Zhiyi Huang told University of Otago magazine.

One research team based at the Auckland University of Technology is tackling the artificial intelligence side of the challenge. To get around some of the difficulties in translating complex brainwaves into instructions for a machine, they’ve created a computer algorithm that mimics the human brain.

The artificial intelligence, called NeuCube, maps the brain’s structures and functions and could be used to control devices in the same way our brains control our real limbs.

“We want to make this system more brain-like, so we can better interpret the signals this system has learned,” said team leader Nikola Kasabov. “We want a better understanding of what the brain is doing.”

The ethical questions

Right now, in New Zealand and worldwide, only a rare few have access to this life-altering technology either through small start-ups like Thought-Wired, research projects at universities or through clinical trials. 

Who gets access ultimately comes down to the group developing the tech. Historically people of colour have been excluded or overlooked from medical trials or AI development, leading to issues with how well devices work. 

There are concerns, too, if people have to give the tech back at the end of a trial. “It’s like you’re injuring them all over again,” said Dustin Tyler, the director of the lab that created Brandon Prestwood’s bionic arm, in a presentation in 2019.

Then there’s the question of how far to take the tech. 

An extreme version of the neurofeedback systems is closed loop deep brain stimulation. Wires are implanted directly into a person’s brain and not only pick up on their signals but send signals back

For example, a system for treating extreme epilepsy predicts when a seizure is about to come on and then fires competing electrical signals back into the brain to quash it. 

A closed loop stimulation application raises the question of whether the patient really has control over if, and when, they get their treatment, since the computer algorithm is deciding for them. 

There’s also a fine line between healing and enhancement. The same tech that’s aimed at helping people recover from traumatic brain injuries by restoring memories can also be used to improve healthy people’s memory. 

Some have gone so far as to say that brain-machine interfaces are the beginning of transhumanism, where minds and machines become one.

In fact some of the biggest players in brain-machine interfaces worldwide, including Elon Musk’s Neuralink and the US Defense Advanced Research Projects Agency (DARPA), have openly said they want to enhance human capability. That might seem inconsequential if it helps people learn new information. But it becomes potentially problematic if it’s intended to create super-soldiers with bionic arms. The latter is something DARPA is actively pursuing.

Although New Zealand is currently a small player in the brain-machine interface game, it’s not beyond the influence of companies or agencies with bigger goals that could raise ethical questions. 

Funding for one project in 2015 came in part from Chinese companies Shenzhen Hampoo Science & Technology Co Ltd and its spinoff EEGSmart. Given how China uses another technology, facial recognition software, for surveillance of its people and assigning social credits based on behaviour, should this raise alarm bells? And how much onus then falls on the researcher to address these ethical concerns? Is it possible to create advances in healing, medicine and communication without also creating enhancements and weapons?

Selitskiy takes the stance that they as a company refuse to pursue particular applications, including military ones. “A single defence contract could fund the company for a substantial amount of time. But the humanitarian cost is too great if you go down that path.”

Government regulation to control what gets developed is a possibility, but is often too slow, he says. “It may work but it always lags behind… Technology, especially when it’s developed by private companies, moves at a much faster pace. The time lag is too great to prevent harm.”

Keep going!