One day, James Johnson wants to drive a car again. He will do it if he does.
In March of last year, Johnson broke his neck in a go-carting accident, leaving him almost completely paralyzed below the shoulders. He understood his new reality better than most people. He had been caring for people with paralysis for decades.
He was invited to join a clinical trial of a brain interface by researchers from the California Institute of Technology in Pasadena. He would have to have his brain implanted with two grids of electrodes. The researchers would use the data from the electrodes to decode his thoughts and intentions. The system would use Johnson's brain activity to operate computer applications. It would take years and require hundreds of training sessions.
The first time he used his BCI, Johnson moved a cursor around a computer screen.
Johnson has used the BCI to control a robotic arm, play video games, and drive a car through a virtual environment.
35 people have had a BCI implanted in their brain, and Johnson is one of them. The number of laboratories conducting such research is growing. In the past five years, the range of skills these devices can restore has expanded. Last year, scientists described a study participant using a robotic arm that could send sensory feedback directly to his brain, a person who was unable to speak due to a stroke, and a person who was able to communicate by imagining himself.
The majority of implants for recording long-term from individual neurons have been made by a single company. Commercial interest in BCIs has increased in the past seven years. Neuralink was launched in San Francisco, California, by Musk, with the goal of connecting humans and computers. The company has raised a lot of money. Several newer BCI companies received major financial backing last year.
Bringing a BCI to market will require transforming a custom technology road-tested in a small number of people into a product that can be manufactured, implanted and used at scale. Large trials will need to show that BCIs can work in non- research settings and that the market can support them. For thousands of years, we have been looking for a way to heal people who have paralysis.
In June 2004, researchers put a grid of electrodes into the motor cortex of a man who had been stabbed. He was the first person to have a long-term implant. He was like most people who have received BCIs. He had lost the neural pathways between his motor cortex and his muscles. After decades of work in many labs in monkeys, researchers were able to decode their movements from real-time recordings of activity in the motor cortex. They wanted to know if a person's brain activity was in the same region.
A landmark paper was written in 2006 about how a man learned to use a computer, a television and robotic arms. The study was led by a neuroscience professor at Brown University in Providence, Rhode Island and Massachusetts General Hospital in Boston. It is the first of a multicentre suite of trials called BrainGate.
Hochberg says it was a very simple demonstration. It was shown that it could be possible to record the cortex of a person who was unable to move and to allow that person to control an external device.
Scientific American newsletters are free to sign up for.
BCI users have access to a wider range of skills. Researchers began to implant multiple BCIs in different brain areas of the user and devised new ways to identify useful signals. Hochberg says machine learning has improved the ability to decode neural activity. Machine learning links activity patterns to a user's intentions, rather than trying to understand what activity patterns mean.
We know what the person who is generating the neural data is trying to do, and we are asking the algorithms to create a map between the two.
People with paralysis are the most likely to say what they want from the technology. This usually means restoring movement for people who can't move their limbs.
One way to control the muscles of a person's own limbs is by implanting electrodes that directly stimulates the muscles.
A participant in the study drank a cup of coffee and fed himself while using the system.
The command signals that his system can decode are for grip force. Several labs are trying to give BCI users a sense of touch.
In 2015, a team led by Robert Gaunt at the University of Pittsburgh reported implanting an array in the hand region of a person. The person felt something akin to being touched when they used the electrodes.
Gaunt and Collinger collaborated to advance the control of robotic arms. They created a robotic arm with pressure sensors embedded in its fingertips, which fed into electrodes in the somatosensory cortex to evoke a synthetic sense of touch. Sometimes it felt like pressure or being prodded, other times it was more like a buzzing. The time it took to pick up an object was halved because of the feedback.
Brain regions that have different roles can be added to with implants. Richard Andersen, who is leading the trial at Caltech in which Johnson is participating, is trying to decode users' more abstract goals by tapping into the posterior parietal cortex. The motor cortex directs the hand to the coffee, whereas the thought is that I want a drink.
The dual input aids BCI performance by contrasting use of the two cortical regions alone or together. Tyson Aflalo, a senior researcher in Andersen's laboratory, says that un published results show that Johnson's intentions can be deduced more quickly in the PPC. He says that motor-cortex activity lasts throughout the movement, making the trajectory less jittery.
Johnson and others are being helped by this new type of neural input. Johnson uses a driving simulation and another can use her BCI to play a virtual piano.
Edward Chang, a neurosurgeon and neuroscientist at the University of California, San Francisco, says that the loss of ability to communicate is one of the most devastating outcomes of brain injuries. In early BCI work, participants could imagine grasping letters and moving their hand around a computer screen in order to communicate. Chang and others have made rapid progress by targeting movements that people naturally use to express themselves.
A team led by Krishna Shenoy, a neuroscientist at the University of California, set a benchmark for the amount of time it takes for a message to be communicated.
This group reported last year that an approach that enabled study participant Dennis Degray, who is paralysed from the neck down, to double the pace.
Frank Willett suggested that Degray imagine handwriting while they recorded from his motor cortex. The system sometimes struggled to figure out signals relating to letters that are handwritten in a similar way, such as r, n and h, but generally it could distinguish the letters. When statistical language models are used similar to predictive text in phones, the decoding algorithm jumps to 99%.
You can do that in 90 characters per minute.
Degray has had a functional BCI in his brain for nearly 6 years, and is a veteran of 18 studies by Shenoy's group. It is remarkable how easy tasks become. He compares the process of learning to swim to the process of learning to drive.
Chang uses a similar principle in his approach to restoring communication. Speech is similar to writing in that it is formed of separate units called phonemes. There are around 50 phonemes in English and each is created by a stereotyped movement of the vocal tract, tongue and lips.
The part of the brain that produces phonemes and speech is called the dorsal laryngeal cortex. The researchers applied these insights to create a speech-decoding system that displayed the user's intended speech as text on a screen. They reported last year that this device enabled a person with a brainstem stroke to communicate using a pre-selected vocabulary of 50 words and at a rate of 15 words per minute.
Chang didn't record from single cells. Instead, he used a device that measured the activity of the brain's cortex. The approach is less intrusive than the one from the cortex.
In a locked-in state, people are unable to speak or move, which is the most profound loss of communication. A group of people including a neuroscientist, reported in March that they were starting to communicate with a man with amyotrophic Lateral Sclerosis. The man used to use eye movements to communicate, but eventually lost his ability to do so.
The man's family consented to the team of researchers implanting a BCI and trying to get him to imagine movements to use his brain activity to choose letters on a screen. They tried playing a sound that mimicked the man's brain activity and taught him to change his neural activity to signal more activity. He was able to pick out a letter every minute.
The method is different from the one used in the paper published in 2017: it uses a non-invasive technique to read brain activity. Questions were raised about the work and the paper was taken out.
There has been a noticeable increase in both the number of clinical studies and the number of case studies.
Although such achievements have attracted attention from the media and investors, the field remains a long way from improving day-to-day life for people who have lost the ability to move or speak. Study participants operate BCIs in brief, intensive sessions; nearly all must be physically wired to a bank of computers and supervised by a team of scientists working constantly to hone and recalibrate the decoders and associated software.
Leading academics are collaborating with companies to develop new products. A not-for-profit company co-founded by him, is called "ALS Voice", and it is focused on developing neurotechnologies for people locked in a locked-in state.
Blackrock's existing devices have been used in clinical research for 18 years, and the company wants to market a BCI system within a year. The FDA put the company's products onto a fast-track review process to facilitate developing them commercially.
Solzbacher wants to show how people's lives can be improved by using four implanted arrays and wires.
Blackrock is working on a fully implantable wireless BCI that will be easier to use and remove the need to have a port in the user's cranium. These features have been a goal of Neuralink and Paradromics from the beginning.
The two companies are trying to increase signal bandwidth by increasing the number of recorded neurons. Currently being tested in sheep, the interface has 1,600 channels divided into 4 modules.
Neuralink's system uses very fine, flexible electrodes that are designed to bend with the brain and reduce immune reactions, according to Shenoy, who is a consultant and adviser to the company. The goal is to make the device more stable. Neuralink has not published any peer-reviewed papers, but a post in 2021 reported the successful implant of threads in a monkey's brain to record over one thousand sites. Neuralink has been testing its system in animals, but academics would like to see the technology published for full scrutiny. If what they claim is true, it will be a game-changer.
Blackrock Neurotech is one of only a few companies that has implanted a BCI long-term in humans. Synchron in New York City has developed a set of 16 electrodes. In an outpatient setting, the device is threaded through the jugular vein to a vein on top of the motor cortex. The technology was put on a fast-track review path by the FDA a year after it was implanted in a person.
The stentrode lacks the resolution of other implants so can't be used to control complex prosthetics. It allows people who can't move or speak to use a computer to surf the internet and control connected technologies.
The company is submitting the results of a four-person feasibility trial for publication, in which participants used the wireless device at home whenever they chose. Oxley says that it is always working. He says that the next step is a larger-scale trial to assess whether the device meaningfully improves functioning and quality of life.
The challenges before researchers are realistic.
Commercial devices will have to work without expert oversight for months or years, and that they need to function equally well in every user. She thinks that machine learning will address the first issue by giving users recalibration steps. It might be more difficult to achieve consistent performance across users.
I don't think we know what the scope of the problem is because of variability from person to person. In non-human primate, even small variations in positioning can affect which circuits are tapped. She thinks that there are important quirks in how different people think and learn and how their brains are affected by their conditions.
There is widespread acknowledgement that ethical oversight must keep pace with technology. Privacy is one of the concerns presented by BCIs. Users must retain full control of their devices. Current technologies can't decode people's private thoughts, but developers will have vital data about their brain health. There is a new type of cybersecurity risk presented by BCIs.
There is a risk that the companies that make the devices will fold, or that they won't be supported forever. Users were let down when their devices were not supported.
Degray wants to see BCIs reach more people. He says he would like to be able to scratch his eyebrow. In the middle of the night, a spider walks across your face. That is the bad stuff.
It is about human connection and a hug from a loved one.
The article was first published on April 20 2022.
All Rights Reserved © 2024
