Can we, by reading electrical signals in the brain, control a robotic arm? That is the question that has driven cutting-edge research into advanced prosthetics by Michael Boninger, MD.
Dr. Boninger is vice chair for research, physical medicine and rehabilitation at the University of Pittsburgh and senior medical director post acute care UPMC. He is the author of five U.S. patents, and is also recognized for extensive research on spinal cord injury, assistive technology and overuse injuries, particularly those associated with manually propelling a wheelchair. Dr Boninger sat down recently with MossRehab for a wide-ranging conversation about his many areas of research.
Here is a transcript of our conversation with Michael Boninger, MD:
Welcome to a MossRehab conversation with Michael Boninger, MD, vice chair for research, physical medicine and rehabilitation, at the University of Pittsburgh, and senior medical director, post-acute care, UPMC. Thank you for joining us, Dr. Boninger. Reading some of your achievements here, you are the author of five U.S. patents, and recognized for extensive research on spinal cord injury, assistive technology, and overuse injuries, particularly those associated with manually propelling a wheelchair, which we will come back to in just a bit. But I'd like to start with your work with robotics controlled via brain implants, allowing patients to manipulate arms and hands to rotate, grasp and perform other movements. Please explain the brain implants. How does that work?
Dr. Boninger: The brain implants conceptually can be relatively easy to think about. Basically we put small chips into the brain that are capable of recording the electrical activity of the brain. The cells in the brain, just like the cells elsewhere in the body, have electrical charges. When a cell fires, which is how cells communicate with each other, that electrical charge can be measured. What we're trying to do with a brain-computer interface is look at the electrical potentials in the brain, look at these firing, and say, "Can we determine what someone wants to do based on how these cells fire?" In a normal person, you would think about, "I want to move my hand." It's almost not even a thought and your hand moves.
The question is, can we, by reading these electrical signals in the brain, determine that someone wants to move a hand? If we can, can we then interpret those electrical signals, which can then control a robotic arm? That's essentially exactly what we've done. We have implanted these chips into individuals with paralysis who, when they think they want to move their hand, the signal doesn't go through because they have a spinal cord injury or another disorder. Then we've been able to develop techniques to be able to understand what those electrical signals in the brain are so that a person who has a spinal cord injury can think, "Hey, I want to move a hand," and have a robotic hand move exactly as they think they would like it to move.
What drew you to this work? Why this particular area of research?
Dr. Boninger: I would say that probably the thing that drew me to this area the most was an opportunity. I'd like to say that one of the things that I've done in my career is be an opportunist and looked at things that presented to me and how they came about, and jumped at things that were exciting and I thought were potentially life-changing. There is a brilliant basic scientist who works at the University of Pittsburgh, named Andy Schwartz. He and another member of my department, named Doug Weber, and I were sitting down for a beer. We talked about the potential of these types of neural interfaces, which is another name for a brain-computer interface. Andy's contention at the time was that if he could get these implants into a human, that he could get them to play the piano. He had done work in monkeys that was really groundbreaking. I then brought a person named Jennifer Collinger into the picture, who works in my department. Jennifer really ran with those ideas, but the origin of it was this beer. My question for Andy, "Well, we can certainly this into a person. It had been done before. Let's do that at the University of Pittsburgh." That really started me down this pathway.
And yet, as promising as this work is, it's still not available outside of the lab?
Dr. Boninger: No, that's right. I mean, this technology is really highly sophisticated, highly complex, and there is many things that we still need to learn. You can kind of break it down into a couple of different buckets. One bucket is that we need to make sure that this technology is really transparent, that people don't see it. We need to go wireless. We need better electrodes to record the signals. We need electrodes that are stable over long periods of time, so maybe you have surgery if you have a spinal cord injury at one point, and then that lasts for years so that you're not having re-surgery. Then we need to get better at not only controlling the limb, but one of the things our group did recently is actually being able to get sensation back from a robotic limb.
In the case of someone with a spinal cord injury, I think a whole nother thing would be most people with spinal cord injury would not want to control the robotic limb as much as they'd want to be able to control their own arm. Figuring out how to interface this brain-computer interface with their own arm to enable independence is really another goal. Many of these things, technological hurdles, basic science understanding of how the brain works hurdles, clinical hurdles, and then eventually we'll hit the hurdle of paying for these devices. Those are all things that have to happen before a patient can go to a doctor's office and just get a prescription.
How many patients have you done this work with so far?
Dr. Boninger: Well, there's different technologies that you can implant in the brain. The one that we've gotten the most control from is a technology where there's actually small electrodes that actually penetrate a very short distance into the brain. That's called microelectrode technology. We've had two subjects that we've implanted with that. There have been probably a dozen subjects, maybe more planted worldwide, but we've had the most success in terms of getting great control over a hand. Being able to control multiple aspects of hand and arm movement is really where we've had some tremendous success. We have done implants of another technology where the electrode just sits on the surface of the brain. We've done that in three people with disabilities. That would be five total at the University of Pittsburgh, and again, those also have been done outside of the University of Pittsburgh as well.
Any predictions on when this will get to general availability?
Dr. Boninger: You know, that's that million dollar question that everybody wants to know the answer to, and the one that I'm always hesitant to answer. The reality of it is is that I think that we are learning things today that already can inform care related to how the brain changes, how the brain is plastic, so how with training you can do things that you couldn't otherwise do before. And we're learning the basic science of the brain, which is all very important to our understanding this technology and how we can make this technology fully functional. What I would say is, we're learning things that can inform care today, and that different forms of this technology I'm hopeful will be available in my lifetime, and in maybe a decade, but I think we're still a little ways away from this being something that a doctor can prescribe for a patient.
Among practical advances that can help people today, I mentioned earlier you've put some focus on the wheelchair. Tell us about that.
Dr. Boninger: My original work was really ... Just by way of background, I have an engineering degree, and the medical degree. All of my work has been fairly engineering-focused, and it's enabled me to understand and contribute to the science involved in both the brain-computer interface work and in the wheelchair work. My wheelchair work has been looking at many different aspects of wheelchair. The initial work was related to people who push a manual wheelchair day in and day out. Imagine you have a spinal cord injury, you're no longer able to walk, you learn to use a wheelchair, and that's how you get freedom. I've talked to people on how important that freedom is.
That freedom is in danger if you push the wheelchair the wrong way or get the wrong wheelchair, because you can actually develop arm pain over time. You can get disorders like carpal tunnel syndrome or rotator cuff tears, that I think everybody's familiar with that terminology, which can lead to pain. That pain can be very disabling. We did work where we looked at how someone propels a chair, what the chair looks like, and came up with some very specific guidelines that say, "Hey, this is what you should do differently. This is how you should propel a chair. This is the type of chair you should get," that I believe have really led to a decrease to the amount of pain and increased function for people who use these chairs day in and day out.
Dr. Boninger, give us a quick rundown of some of your other past work.
Dr. Boninger: I think a lot of the work has been in, again, assistive technology. I can tell you that the other area that we've done some work in recently ... One example is just looking at how often wheelchairs fail. There's unique aspects of the US healthcare system that tend to pay for the least expensive chair. What we found in our research is that wheelchairs fail at an unbelievably high rate. In a six month period of time, if you look at a population of people with spinal cord injury, over half of them will have a wheelchair failure in that period of time. If you think of the last time your car broke down on you, you'll realize how unacceptable that level is. That's kind of one area where we branched off.
I think another on, and this is kind of an interesting area to look at, especially given my brain-computer interface work, is we've been recently looking at walking in people with spinal cord injury and gait training. Again, if you picture a person who's been injured, the spinal cord injury causes a paralysis and they can't move their legs like they used to, the question is, will they walk again, and what should we be teaching them when they're in the hospital and in rehab about walking? What we have found is that many people who will likely never walk again get walking training, get ambulation training during rehab, and that that actually can cause detrimental effects. I think that the source of this is that everybody who's hurt wants to walk again. Walking is seen as healthy, and there's no part of me that blames them, but what we found is that that can cause problems, that can cause unrealistic expectations. People don't learn how to use a wheelchair correctly.
A line of research that we have right now, in addition to looking at this issue further, is can we come up with better predictors of who's going to be a successful ambulator? The way this all ties in with the brain-computer interface is that I'm a firm believer in doing groundbreaking research that will impact patients in the future that can be transformative. That's really where the brain-computer interface work lies. But we always have to temper that with one, saying, "Hey, it's not here yet, and you need to do the hard work of rehab." We have to tell our patients that. I also think I'm very lucky to be able to be involved in research that is also helping people right now, today, in terms of propelling a chair better, or directing initial therapies better so that people really learn what they need to when they're in a rehab setting.
Aside from moderating expectations, what are some of the other practical applications of your work for clinicians?
Dr. Boninger: The most practical applications I have are in the wheelchair realm. An example of a practical advice to give to a clinician is to really get people very lightweight chairs that fit them well, and then to spend time training someone how to use a wheelchair and how to propel it. It's really simple stuff, so you just tell someone when they're pushing on the wheelchair, to make sure that their stroke is a long, smooth stroke, and that to go a distance, what they're trying to focus on is using as few strokes as possible to get someplace. Pushing on the wheelchair as few times as possible while they're pushing from point A to point B, that gets people to take these long strokes that have less impact, less trauma on the arm over time. If they get a chair that is fit to them, that is lightweight, and they're appropriately propelling it, you're probably doing a lot to prevent injury. I think that in rehab, what we really have to say is, "This is the skillset you need to learn. You need to learn how to get in and out of this chair. You need to learn how to wheel from this chair. By the way, if you can walk, we can do that after you leave the hospital." Then we can explore whether ambulation might be a goal for them after they've already had the skillset in place.
We talked earlier about your frustration over the unavailability to the public of brainwave-controlled prosthetics. Aside from technology, what other significant challenges have you run into?
Dr. Boninger: The huge challenge is really in the basic science of the brain. The reality of it is that right now, what we're able to do is we're able to get an individual to move a robotic arm by thinking. They think, "I want to pick up a ball," and the robotic arm moves there to pick up the ball, just like your or my arm would. What we found is, once you pick up the ball, everything changes. You have this movement that you want, but then you're interacting with the ball. To make this a more realistic example, let's picture it as a styrofoam cup of coffee. If you squeeze it too hard, you're going to crush the cup, and it's going to spill all over your hand. If you don't squeeze it hard enough, it's going to slip out of your hand and fall. What that requires is sensory input, so you can feel the cup, and that tells you how hard you have to squeeze.
What we have to do is now move out of a realm where we're just controlling movement, to moving into a realm where we're controlling force, and that what you need to control force is sensation, so how do we restore sensation? We're working on restoring sensation, but what we need to do is we need to get to a point where we can restore sensation and look at this interaction between movement and force and dexterity. How do you dexterously manipulate objects? Which is the whole reason we move our hand to a position is so that we can type, or pick up a cup of coffee, or shake hands with someone. That's a whole nother realm of basic science that we're working very hard to understand right now.
Well, my next question was going to be, is there a current project you're especially excited about? That sounds like that would be it.
Dr. Boninger: That's exactly right. The current projects that I'm most excited about are our work to integrate this sensory feedback, this motor control, and come up with new algorithms that are based not just on the speed of movement, but also on the force that's applied in moderating that force so that we can get dexterous control of a hand. That's area one. The second one is informing clinicians about who are the people who are most likely to walk, how do we predict who's most likely to walk, and then how do we direct their inpatient therapy so that they're learning what they need to do in rehab? I have this futuristic work that I get to be involved with, and then I have this very practical work that I get to be involved with.
Dr. Boninger, thank you so much for joining us on this MossRehab conversation. We have been speaking with Michael Boninger, MD, vice chair for research, physical medicine and rehabilitation, at the University of Pittsburgh, and senior medical director, post-acute care, for UPMC. Look for more conversations to come on our website at mossrehab.com/conversations. I'm Bill Fantini. Thanks for listening.
Each issue, we provide links to the three articles from the PM&R literature that you'll want to know about.
Make an appointment or schedule a tour online:
Consistently ranked among the best in the U.S. and top-ranked in Pennsylvania.
Sign up for updates