Giving paralysed patients the gift of movement

Before Elon Musk, computer brain interfaces have to this point been solely a medical pursuit. It dates back to the late 90’s, when pioneers like John Donoghue of Brown University were first beginning to understand how they might restore abilities to people with paralysis. I caught up with the man himself to chart the early evolution of this technology...

John - We had a pretty reasonable understanding of how the part of the brain that controls movement was structuring its output. It was putting out patterns and we could interpret what those patterns meant. And about the same time we had these electrode arrays that could pick up enough neurons so you could read out enough of that pattern and you could make sense out of it. So we said at that time, well, this is really something that can help people. People who are paralysed from a disease called ALS where the connection between the brain and the muscles degenerates, also, a spinal cord injury will cut off the messages from the brain to the body. Even at that time in the late 1990s, I was saying that the real goal would be to use the advances in electronics to reconnect the brain back to the body. In 2017 we did just that with Bob Kirsch and Hunter Peckham at Case Western University, a person who was completely paralysed from a spinal cord injury was able to pick up and drink and eat and move his own arm using this kind of device.

James - And throughout all of these developments, it's crucial to mention that they've all involved implanting something within the head, a surgical procedure to implant, as you mentioned, that electrode array which has on it the strands which are able to sit in our head and read the neuronal activity that's going on.

John - Exactly. The resolution, that is, the high fidelity resolution that we need to read out these patterns in the brain, requires that something be put in the brain. This is about the size of an aspirin, five millimetres or so, that is put into the brain surgically in a small opening and then closed up. There are still ongoing attempts to get signals from just reading on the surface of the head, but they're very crude and you can't get much more than a 'yes, no' signal.

James - What were the efficiency gains needed from helping paralysed people to move cursors on a screen to later being able to move prosthetic limbs to pick things up just with the power of the mind?

John - There are biological and technological challenges here. The electrode array that was around in the 1990s - until the slightly different technology that Elon Musk's Neuralink is using - basically all of the few dozen people that have been implanted have all used that same technology. The only difference would be is we put maybe two of them in so instead of 100 samples, we get 200 samples. The biological question is, is more samples good? It seems to be that the more information we read, the more we can get a better estimate of what you want to do. The second one is technology. To be honest, the reason we didn't control a robot is we couldn't afford one. Then, the more sophisticated work that showed really elaborate robotic arm control, that was more about what we learned of the complexity of the signals. It's like having a QR decoder for your camera to scan a product and it doesn't work very well. Over the years, we learned how to make it work better, and that's called the decoding problem, reading out the brain and understanding the signals. The next step though is that all of the people who have the implants through the research setting have a plug on their head, and that plug on their head is something that's not desirable. The step that Neuralink has made is they can put the electronics under the scalp so everything is inside the body.

James - This field was started purely as a medical pursuit. People are speculating that, in the not too distant future, brain computer interfaces could be looking at expanding our functions as people, expanding our memories or potentially uploading our consciousness to the cloud, perhaps. Are we in a position where, implanting neural devices into healthy individuals, the probability and severity of side effects is getting towards that critical point where regulatory bodies are going to be satisfied for this to be happening more regularly?

John - Brain surgery, I think, is something that puts everyone aback or most people aback, but I'd say brain surgery is actually quite safe. I'm not a clinician, so I don't have exact experience. I've worked with many neurosurgeons and I do know that there are complications in any surgical procedure - I think the overall rate is around 5% - but those range from minor skin infections to maybe a postoperative fever, which is not uncommon. I do think it's not a worry for a person who is paralysed. That small risk of a surgical procedure is worth it with a device that has been approved by regulatory bodies. Now, for able bodied people, is a 5% risk worth it? I don't think clinicians would ordinarily say, I'm going to do a surgical procedure on a person that has no clinical issues, no health problems, and that is an ethical dilemma. Where will we go with that kind of thing? Personally, I think it's not a good idea because it's not zero risk, but it's not a big risk.

James - Transitioning the focus away from helping people with paralysis to potentially expanding our memory, for example, how do you feel about that?

John - Well, my first reaction is that we have very good devices that expand our memories and do all these functions without having to have surgery or have any device implanted in our brain. And that is, for example, our smartphones. We've already expanded our memory! I'm sure those devices will only get better and better. I don't see the remarkable advantage. The fundamental question is, how many neurons over how much space do you have to record in your brain to know what you're thinking? And one side of that argument, which I sort of lean to, is if I wanted to know exactly what's going on in your brain, I would need to record from every neuron or pretty close to all of them. We're not going to do that. Sticking that many wires in your brain is not reasonable.