By Dr. Lise Johnson (CSNE Education Manager)
Brains need blood, and plenty of it. Brain cells use up a lot of calories and oxygen, and the only place they can get them from is the blood. If the blood supply to the brain is disrupted, the brain tissue starts to die, and that is a bad thing. We call it a stroke. A stroke happens when a blood vessel in the brain either bursts or gets clogged; either way, the result is dead brain cells. How this affects the person who had the stroke depends on where in the brain stroke occurred and how much of the brain was damaged as a consequence. There is no such thing as a good stroke, but some strokes are worse than others. A brainstem stroke is particularly nasty because almost all of the signals going into and coming out of the brain have to go through the brainstem. Thus, damage to the brainstem can impair or completely prevent communication between the brain and the rest of the body. If this happens, the brain has no way to tell the body when and how to move and the body has no way to tell the brain what is happening to it. The result is complete paralysis.
That is exactly what happened to Cathy more than 16 years ago1. Since her stroke she has not been able to move her arms or legs, or even to speak. But last year, Cathy got to do something she hadn’t done in a long time. She was able to bring a bottle of coffee to her lips and drink it. She accomplished this with the help of a robotic arm – and not just your run of the mill, highly sophisticated robot arm, but a robotic arm that she controlled with her brain. Cathy was a participant in a research study at Massachusetts General Hospital for which she underwent surgery to have an array of electrodes implanted in her brain. The implant was placed precisely in the part of the brain that controlled the movement of her hand before she suffered her stroke. This part of the brain is called the primary hand-motor cortex and the neurons there are active when you move or try to move your hand. That part of Cathy’s brain is still intact, so when Cathy thinks about or tries to move her hand the electrodes implanted there detect the electrical activity of the neurons. The scientists and engineers working on the project were able to decode those electrical signals, and turn them into commands for the robotic arm. Thus, Cathy was able to move the robotic arm just by thinking about moving her own hand.
I think, and I hope you will agree with me, that that is amazing. If you don’t think it’s amazing, think about it again. This is a technology with the potential to really change someone’s life, to restore independence and the ability to interact with the world. It’s the sort of news that everyone likes to hear. It was widely reported in the popular media, so you may have heard about it already. However, you may not have heard the part of the story that makes it especially remarkable: the electrodes were implanted more than five years before the experiment. You may be wondering why that is such a remarkable fact, and the answer is, unfortunately, not the sort of thing that anyone likes to hear. The truth is that these kinds of implants don’t usually work for that long. Five years is actually an exceptionally good run. The reasons for this are complicated (as these things tend to be) and have to do with the way that brains and electrodes get along.
We know that neurons communicate using electrical signals, and electrodes are built to record electricity, so if you put an electrode into brain tissue (assuming you know what you’re doing and you do it properly) you can record the conversations that neurons are having close to that electrode. If you can make sense of that conversation (or decode it) then you can know what it is that that group of neurons is trying to accomplish at any given time. This is exactly how Cathy’s device works. I’m making it sound much easier than it really is, but that’s the gist of it. Of course, using terms like “the neural conversation” is vague and not very scientific so let me make it clear that in this context we are talking about recording the electrical outputs produced by individual neurons. You can’t record the activity of all the neurons in the brain with a single electrode any more than you can record all of the voices of the people in the world with a single microphone. You might not want to hear what everyone is saying anyway; it might be completely irrelevant to what you want to know. You only want to eavesdrop on the neurons that are saying something interesting for your problem (in this case, moving a robotic arm), and you want to record as much of what they are saying as possible. So, when scientists and doctors implant electrodes they are very selective about where they put them, and they put a bunch of them in at once (Cathy’s implant has 96 electrodes). If all goes according to plan, when you turn on the device after surgery, you can hear lots of neurons talking, and you can get a pretty good idea of what they’re collectively trying to say. That’s exactly what you want, so everyone is happy. The problem is that over time more and more of those neurons will go silent, and when that happens it gets more and more difficult to figure out what’s happening in the conversation. The performance of the device gets progressively worse, until eventually it doesn’t work at all.
Why do neurons go silent? I said it has to do with the way that brains and electrodes get along, and the real problem is that they don’t get along at all. Your brain does not like to find non-brain objects in its midst. That shouldn’t be surprising; if you think about it, it makes very good sense. How often does a piece of metal lodged in your brain mean good things are about to happen? I will tell you, for all of human history, the answer has been never. Something stuck in your brain is never a good thing. This is the assumption that your brain is operating under, and if it finds something has been stuck into it, it does the only logical thing – it tries to kill it. Of course, electrodes can’t be killed. This is partly because they aren’t alive to begin with, but also because the electrodes are enormous compared to the size of a cell, and that makes them difficult to dispose of. So the brain does the next best thing – it wraps the electrodes up in scar tissue so that they are at least contained. Unfortunately for our purposes, they are also electrically insulated. When electrical signals can’t get through, we can’t hear what the neurons are saying anymore.
That’s why I say it’s remarkable that an implant of this type would still be functional after five years. But, just because an implant is still functional, that doesn’t mean it is still functioning at its peak. In this case, even though the electrodes recorded the activity of enough neurons for the device to work, after five years they were recording fewer neurons than they did at first. This begs the question, what will happen in another five years? How long will that device be useful? More to the point, what can we do to extend the life of the recording?
Neural engineer Bill Shain thinks he can solve this problem, and that is what I’ll discuss in my next entry.
For more information about stroke, visit http://faculty.washington.edu/chudler/vessel.html
For more information about the research project Cathy is participating in and to see a video of her controlling a robotic arm visit: http://www.nature.com/news/mind-controlled-robot-arms-show-promise-1.10652
1Hochberg, L. R. et al. Reach and grasp by people with tetraplegia using a neurally controlled robotic arm. Nature 485, 372–375 (2012).