A brain implant that lets people who are unable to speak because of paralysis communicate by thought could go on sale for the first time in 2022.
The technology, known as a brain-computer interface, involves a small electrode that is put permanently into the brain to detect neural signals, allowing people to control a computer. It has seen success in research studies, including in recognising people’s wishes to move a computer cursor and imagining writing letters by hand.
There are still question marks over the technology, because electrodes put in the brain tend to get slowly covered with immune cells called microglia, which makes a device less sensitive. But in November, biotech firm said it plans to put such a device on sale in the US in 2022 – if it is approved by the US Food and Drug Administration.
Advertisement
żěè¶ĚĘÓƵ spoke to the firm’s president, to find out more.
How many people have had this type of implant in a research setting so far?
Right now, we’re at about 34 or so worldwide, and about 31 of those are ours.
How do we know they are safe? Could they cause infections, for instance?
The safety of patients is always paramount. Any percutaneous [inserted through the skin] implant such as a wire or connector creates a potential site of infection that needs to be carefully managed with regular cleaning protocols and check-ups. The experience of the past 10 years of working with human patients has, however, shown that this can be very well managed. We do have a wireless version of the implant in preparation that is currently going through testing.
Why the focus on asking people to imagine handwriting?
Traditionally, people had a sort of keyboard, and then you would imagine where a cursor went. You needed to click every character. That is doable, and if you don’t have anything else, it is valuable. But if I now take a piece of paper and handwrite, I’m definitely faster.
How does the implant distinguish between the tiny hand movements needed for different letters?
It doesn’t have to be the entire process of controlling your hand – all that we’re looking for is characteristic patterns. It’s a bit like looking through a spyglass, and seeing part of the pattern. As long as that is distinct enough that it can be reliably correlated with: “I’ve just moved my finger the following way”, or “I turned my wrist by 15 degrees”, then you’re good.
Is there a training process before it can work properly?
When it first started, you’d have to have a number of [scientists] standing around somebody, and it would take a couple of weeks until something was working. We’ve been able to make a lot of the implant technology much more reliable and robust, make certain components of the electronics smaller, and surgical procedures have improved. And, of course, the algorithms and available computer power have improved significantly. Today, the patient recovers from anaesthesia and if you plug them into the system, you immediately have good signals. Of course, over time, the patient learns, and the AI that analyses the signals becomes better.
How long does the training process take?
The more complex the task, the longer it may take to perfect the control. But that is not different from an able-bodied person: some things we do intuitively well from the onset; some we have to train more. The basic skills to control a cursor on a computer screen, to control your wheelchair or to turn thoughts into text tend to intuitively work quite well. We have patients that generate art with their interface and some that practise computer games. To do this really well does require some training and with every week they get better.
Do immune cells impair performance after several years?
The reality is that everything you put inside the body does cause a response. The question is, is it manageable and how long can you get signals and have a function? There’s a certain amount of redundancy, because you don’t necessarily need quite that many electrodes to get all the information that we need.
How long will the implants last?
Our longest patients have had implants running for around seven years. They haven’t stopped working. The practical functionality that patients can achieve appears to not have suffered. Granted, the numbers are still small. In non-human primates, the longest ones have been over nine years. I believe we can offer a product that can run for 10 years, based on the data that we now have. We probably will initially [market the device] with a timeframe of five years or 10 years.
Interestingly, some of the patients have said they would do the surgery, even if [the device lasted] only two years. But that really depends on the severity of their cases.
How much will an implant cost?
We have not said how much it will be, but I want to make sure that as many people as possible get access to it. And so, we will position it similarly to where cardiac pacemakers and cochlear implants are today – in the $8000 to $25,000 range. If somebody is tetraplegic and they buy a vehicle with a wheelchair ramp, that will cost you over $70,000. The cost for a tetraplegic person, over their lifetime, is around $5 million for the family and insurance, so this is really at the low end. We want to get it out there and have as many people benefit from it as possible.
Sign up to our free Health Check newsletter for a round-up of all the health and fitness news you need to know, every Saturday
