StarTalk Radio - Mind Control & Neuroprosthetics with Cindy Chestek and Parag Patil

Episode Date: October 7, 2022

Are we nearing the singularity? Neil deGrasse Tyson and co-hosts Chuck Nice and Gary O’Reilly learn about brain machine interfaces and prosthetics you can control with your mind, with biomedical engineer Dr. Cindy Chestek and neurosurgeon Dr. Parag PatilNOTE: StarTalk+ Patrons can watch or listen to this entire episode commercial-free.Photo Credit: Pixabay, CC0, via Wikimedia Common Subscribe to SiriusXM Podcasts+ on Apple Podcasts to listen to new episodes ad-free and a whole week early.

Welcome to StarTalk, your place in the universe where science and pop culture collide. StarTalk begins right now. This is StarTalk Sports Edition. We're doing an entire episode on neuroprosthetics. And I don't even know what that means because my co-host, Gary O'Reilly, came up with this and created a show around it. Gary, how are you doing, man? I'm good, Neil. Thank you. Excellent. Excellent. Excellent. And you come to us from the UK. Mm-hmm.

Accent didn't otherwise give that away. Former soccer pro and current professional soccer commentator. So we love that about you. And every third show, I want to remind people that we see your sexy legs in a photo on your wiki page. You have your own wiki page. Gary O'Reilly, soccer pro. So we've got you there. Chuck doesn't have a wiki page yet, but we're working on that. Chuck, nice.

How you doing, man? What's up, Neil? How's it going, Gary? All right, man. All right. We're all good. We're all good. We're all good. So, Gary, just give us the overview before we introduce the guests. All right, here we go.

What prompted this and where's it going? All right. Spoiler. Not a lot of sport in this show, and I guess our audience will be cool with that once they realize where it is we're going to go with this. We're going to take them somewhere I hope they really want to go to. But in the now, in 2022, people with limb paralysis or amputations can restore some of their movement with the aid of prosthetics, you know, something like a running blade or some more sophisticated electrical engineering, robotic tech. But think about this. What if you could hook up a human brain interface and then found a way to record neural signals and decode them to learn our very own neural language?

It's been done. It was called the $6 million man. Ah, this would enable Chuck, a prosthetic user, to control ever more complex movements just by thinking. And our brain-machine interfaces, or if you prefer, neuroprosthetics, are taking the world of biomedical engineering to a new place. And that place is exactly where we are headed in this show. Okay, none of us have expertise in any of this.

I know, but that's why we have guests. Other than the fact that we, I presume, have brains. That's about, that's where it ends. Speak for yourself, Dustin. All right, first off, we have Dr. Cindy Chestek, Associate Chair for Research in Biomedical Engineering,

Associate Professor of Biomedical Engineering at the University of Michigan in Ann Arbor, where she runs the Cortical Neural Prosthetics Lab, which focuses on brain and nerve control and fingers. So that side of this subject is covered. On the other side is Dr. Parag Patil, Associate Professor of Neurosurgery, Neurology, Anesthesiology, and Biomedical Engineering, again at the University of Michigan. His focus is neurosurgery with emphasis on neuromodulation therapies for movement disorders. So, we are really going to get into this deeply, and Neil, these are our guests. Okay, that is beautiful. Welcome, guys, to StarTalk. Thank you. Thanks for having us.

Yeah, and you work in the same institution, so you guys work together, I presume. Yes. Absolutely. I do not do surgery. That's for all. So I do the engineering side. And you have strongly overlapping Venn diagrams in the mission statements. And I think that's where all this matters. So I want to know just each of you, just briefly tell me how you, let's start with Parag. Parag, how did you get interested in this? Well, so it actually, when I was 12 years old, I saw Star Wars, The Empire Strikes Back. And there's a scene where Luke Skywalker gets his hand cut off and he's just laying. Yes, exactly. You didn't know. First I find out you're my father, then you take my hand. Okay.

Okay. Sorry, I had to do it. I just, you cannot bring up that scene. That's in case no one remembers that scene. Chuck just reenacted it for us. There you go. Right. Thanks, Chuck.

A couple days later, there was an article in our paper about a computer-controlled knee, and that was being done at MIT. And basically, the 12-year-old me said, I want to go there, and I want to do that. Well, that was in the 80s, and we weren't quite ready. And it took me, I guess, 40 years to get to the point where we're actually developing those kinds of devices. You said it was a computer-controlled knee? At that time, yes.

There was a professor named Robert Mann who was working on a computer-assisted knee. Now, this is back when, you know, computers, the laptop hadn't been developed yet. So, it was quite an accomplishment. Wow. Yeah. So, a knee, that was like big-time stuff. I'm just thinking. Yeah.

If all you bring to this show is a knee, I'm thinking we got the wrong guy. Right? So, I'm just thinking if all you bring to this show is a knee, I'm thinking we got the wrong guy. I'm assuming the field has moved well beyond the knee since the 1980s, I presume. What's that, Chuck? I was going to say a knee is like kindergarten compared to

a hand. Right, right. Am I wrong when I say that the hand is the most complex moving part of the body? Is there anything else that has more complexity to its movement than the hand? Yeah, I mean, depending on how you count, there might be 20 degrees of freedom in here, right? Like how many of them you actually use when, you know, you're going about your life is debatable. There's very few things I do that

require, you know, all of them. But yeah, no, it's extremely complex and no way to make a robot right now, right? Like for anything that requires the human hand, Parag doing surgery, you know, assembling lots of things, we still have the hand and that's the most advanced motor controller we have. So just so people, just so we're on the same page, tell us what you mean by a degree of freedom. Yeah, okay. So a degree... That's a very engineering thing, right?

And so we want to know what that means. Okay, so it's sort of moving along one direction of movement. And I need to have full control along that dimension. I need to be able to stop anywhere. Like I need true freedom along that line, right? And that's my degree of freedom. And so, you know, your thumb can explore probably a three-dimensional space.

So I can call that three degrees of freedom. Sometimes we control the thumb as a two degree of freedom, but it's a consistent direction of movement and I have to have full control along that degree of movement. So it'd be up, down, left, right, forward, back. And I think twisting is also a degree of freedom, right? Yes.

Rotation? Yeah. So for example, yeah. So like, you know, the orientation of a point. So technically to control a whole point, you'd need six numbers or six degrees of freedom. Rotation and the X, Y, Z. Right.

So while we're on the explanation cycle here, what is a brain interface system, a BMI? What is that about? How does that work? Yeah, so yeah, that definitely requires some unpacking. So a brain machine interface is where we record from the brain, we apply algorithms to interpret that brain activity, and then we use it to control some kind of external device. For example, like a prosthetic hand, or we also stimulate,

if you have like a paralyzed muscle, you can take brain signals and, you know, turn them into stimulation commands for a paralyzed hand. So that's a brain machine interface. That's pretty well. This is Dr. Frankenstein. You bring something back to life that had no movement. Yeah. I mean, it's more like fixing broken wires. I love that Cindy was like, well, I wouldn't quite characterize it that way.

Thanks, Neil. Patients are much better looking. It's more like fixing a broken wire, right? Like, you know, if you have a spinal cord injury or an amputation, you know, the signals are perfectly good in the brain. You just have to somehow get them where they're going. And so, you know, it's replacing that descending pathway with machines. That's what the neck electrode nodules were for in Frankenstein.

In Frankenstein. Was that what they were for? I didn't know that. So, yeah, yeah. Wait, that's what... Are you making this up? Go into his neck. No, I...

Listen, dude. I think that's the stimulation. No, serious. That's where... No, that's where the... When the lightning charged the machine, the charge was sent

into the neck so it could stimulate the brain. Yeah. Did not know that. Okay, so... Cindy, I thought not know that. Okay, so... Cindy, I thought everybody knew that. No, I'm with Cindy. I just look at them and I had no idea what they were for.

I just thought it was a cosmetic thing. You hang your jewelry out of your glasses. No, Cindy is like this. Listen, I just work on brains in real life. Thank you. So, what is the difference between a brain-machine interface system and a neuroprosthetic? Is there any difference?

Is it the same thing or are they completely different here? I mean, I think prosthetic is used very widely for replacing any function. If you're replacing a function with technology, people will call that a prosthesis. A brain-machine interface, I think, is when you're connecting it

to something outside the body. You also A brain-machine interface, I think, is when you're connecting it to something outside the body. You also hear brain-computer interface. So, if you're just controlling a computer mouse or things like that, then it's, you know, not necessarily a BMI. But, you know, in my lab, we mostly control physical objects. Yep. Parag, I got a question for you. So, I'm going to be last in line to start hooking up machines to my brain, just so you know. Last. So when I think of what's going on in the brain, things are happening on a chemical molecular scale. We don't have machines that are that little to interact in the way my brain normally functions.

that are that little to interact in the way my brain normally functions. So what is the mismatch that you're trying, or let me not pre-bias the answer. Exactly. How are you taking machines and aligning them with the neuroelectrochemistry of our brains? That's a great question, Neil. So the brain actually is two different systems that are operating together. So you have all of these cells, little microscopic cells that have all

these chemicals in them and proteins and kind of the biology of them. And they're producing electrical signals. And so you can look at the brain in two different ways. You can look at it as the chemistry and the chemicals that are moving around, and you can look at it as the electrical signals. What Cindy and I work on accesses those electrical signals. Oh, so you don't need the biochemistry if the biochemistry's sole purpose is to make an electrical signal, because you can just put in the electrical signal, because we got that. We kind of can. The thing is that— That means no.

That's a polite no, I think. The trick is you have to put the right signal in the right place at the right time or read the right signal at the right time in the right place. It's a two-way street. And you have to know what you're reading. If it's written in a foreign language, it doesn't do you any good. And we don't know what the language of the brain is yet. So how do you create that baseline and do you find variation from brain to brain? Because brains tend to be very different from one person to the next.

So how do you create that baseline and then how do you account for any variations? That's easy. Yeah, so the short answer to that is it's a machine learning problem, right? Like, you know, you'd think we would need to understand the brain at a very low level to be able to do this, but we can just use correlations, right? So if we have somebody coming into the lab and we say, okay, move your thumb 10 times, move your index finger 10 times, you know, and then we learn that relationship. And we're only recording from a very small number of the neurons. Like, there's many, many neurons in the brain.

We get to look at a very small number of them, and we have to deduce that relationship. Think of it like your iPhone. Your iPhone recognizes your face, but does it really know your face? Does it really know it's you? No, it has a computer program in it that says, ah, that's the face that I've been taught to recognize and turn on, and that's kind of what we're doing.

So at the moment, there's no Rosetta Stone, no code breaker for people in your field to be able to just go, this is what this means. Because that would mean you'd have to standardize it. You would have to localize it to every individual brain. Oh, absolutely. Every brain would have to go through some brain interface standardization for that person. Yep. Yeah. No, and you're trying to get that down to the smallest amount of data. But basically, if you, you know, you're looking for the same kind of pattern over and over again, and then you relate that to the different movements. But yes, every single person that gets a brain implant, there's a whole calibration routine and they think about all these different movements and we learn what that is. So are the motor cortices of the brain centrally located or are they like other parts of workings

of the brain where they're in different spots? Yeah. By the way, first, Parag, how many brains have you poked at? Let's get that up. Yeah. So, I guess I wouldn't quite use the term poking, but, you know, I would say on average about maybe a hundred a year over the past 17 years. What? And then a bunch of spines as well. The technical term of that is a hunch. And how many of them have you turned into your own personal zombies?

I currently have no personal zombies, I'm afraid. So I kind of joke that I'm the doctor that you never want to have, right? Who wants to have a pediatrician? You want to have a primary care doctor, but you don't want to have a neurosurgeon. No, you don't. Right on. Okay, so you've 100 brains a year for 17 years. And Gary, what did you call that?

A lot. A lot. Technically speaking, the pure description of that is a lot. Okay, let's— Mathematically, that's a lot. You didn't get a chance to answer the question, though, about the locality, motor cortices. So it's related to the question we were talking about before. So the brain, in some sense, motor is distributed.

So we have different parts of the brain that contribute to the motor plan. And then signals converge on something called the motor cortex, which kind of sits about here. And those signals go directly down your spinal cord to your muscles. And so the trick is not only operating the muscles, but figuring out what you want the muscles to do. And that's what the rest of the brain is doing. But you said here, some people will only be listening to this via podcast. So what part of your head were you pointing at? I was pointing kind of

just above the ears. If you drew a line from the kind of halfway up your head over your ear to your eyeballs, that's roughly where the motor cortex sits. Got it. Okay.

All right. So, Dr. Chestek, how are you using assistive exoskeletons to help recapture the use in a person's hand who might have damage, might be an amputee? And how is that process linked into the work that you do? So, actually, Gary, we've got to take a break. Okay. When we come back, we will lead off with that question. Dr. Cindy Chestek and her collaborator, Dr. Parag Patil, who we have just learned has been inside the brains of thousands of people

when StarTalk Sports Edition returns. We're back. StarTalk Sports Edition. We're talking about neuroprosthetics with Chestek and Parag Patil. And Cindy's an engineer and Parag is a medical surgeon. And they're both focused in the same way with the same mission statement. And we're trying to learn more about what the present and future of this brain-machine interface is. And we left off with a question.

Gary, why don't you give us that question one more time? Okay. I came across the term assistive exoskeleton, right? And I was thinking, well, how is that going to work? How is that set within your program? Because it's looking to restore and recapture the movement in a hand of someone who's damaged or an amputee. So if you could explain how that fits in.

Yeah, so exoskeleton is one of the ways of giving the movement back. So if we're talking about an exoskeleton, we're probably talking about spinal cord injury and paralysis and trying to help somebody that can't move right now, as opposed to, for example, a prosthesis or a prosthetic hand. But yeah, we mostly in the lab are trying to stimulate paralyzed muscles, but you can just actuate the hand. And we do also, you know, sometimes just, you know, use a servo motor to move the fingers. And that's another way of trying to restore that movement.

Yeah, but if you do it with servo motors, then you're not doing it neurochemically within the nerves of the hand itself, correct? No, no. And I mean, let me maybe take you through the system a little bit. You know, so we record from the brain. We get maybe like 100 tiny needles worth of information. You know, we process that.

Needles that are inserted in the brain live. Yes, and you were asking earlier about... Sign me up. Yeah, you were asking earlier about detecting the signals. It's really, it's not as hard as you think. It's more like milliseconds and microns is the time scale and the space scale that we need. We got that.

Yeah, so it's not that bad. Milliseconds and microns. If we get our signal within like 50 milliseconds, that's going to feel pretty instantaneous if you're trying to control something.

But then, yeah. So a millisecond is not a millionth of a second. It's a thousandth of a second. Yeah, that's about the time scale of neurons. How long are these needles? How long are these needles?

Who's counting? In the skull, into the brain. Are they attached permanently or you just take them out? It's like acupuncture. I mean, it's about a millimeter, right, that goes into the brain and that gets you into, you know, the really great signals

that you want. But are you keeping those in on a permanent basis? Yes. So how do you overcome going through the skull into a brain that is floating freely? Do these people have like skulls when this is happening? So I'm guessing they are. I'm sorry, I'm sorry for the ignorance. But it was just a fact that if you had any slight impact

on that patient while those needles are in there, you've got all sorts of issues to overcome. Gary, they're not playing football while this is happening. What do you think these people are? Some people bump into something. It just happens. We're human.

Will you let the woman answer the question? This is actually a Parag question of why you can have tiny needles in your brain safely. So we've had over a century of being able to get into the surface of the brain safely and for patients to heal up well and be able to even go home from the hospital the same day or the next day. These electrodes really rest on the surface of the brain, but they go in just enough to be able to record the activity of the cells. So don't think about it as something going deep into the brain, even though that's something I do with deep brain stimulation, help sense the name. But this is something that's resting right on the surface. And then you're

right, the brain is soft. And so there's a very flexible wire that comes up to the surface where we can then attach it to something that can take it to the computer. And this is through a hole in your skull that you have drilled with like a drill bit. Yep. I was hoping you were going to say, no, we have more inventive ways. That's just a blunt, yep, yep, get out our Black & Decker drill and go in. Welcome to this old brain. We don't remodest. This one needs new siding. So wait a minute.

So I guess for either one of you, when it comes to the exoskeletons that are used for like lifting things and, you know, because they're already in existence. How are they communicating? Because you're telling that what to do. When you move your arm, that exoskeleton goes with your arm. So how is that machine receiving a signal? Yeah. I mean, you know, short answer is circuits, right? And every time we use a different output device, we got to, you know, short answer is circuits, right? And every time we use

a different output device, we got to, you know, build a little circuit to control it. I can tell you that some of the biggest problems we run into is if that thing you're controlling is not totally following your thoughts

or what you want it to do, you lose that sense of embodiment, right? So for example, prosthetic hands, they're just not that fast yet, right? And ex example, prosthetic hands, they're just not that fast yet, right? And exoskeletons, they're nowhere near the performance of the real arm. And so you're trying to maintain that sense of embodiment by doing a good job on the prosthesis. So the idea is to take that sensation of a phantom limb and make it the actual limb. Yeah, I think that's right.

That is the win condition. Sounds like it. Yes. Okay. And if you did that right, they can use it the very first time they pick it up. And it's going to feel like their real hand. But I'm not saying we're there yet, but that's the goal. Parag, please tell everyone about phantom limbs because Chuck just slipped that in there,

but that's an actual thing. Yeah. So phantom limbs basically are when you feel a limb that isn't there. And there are a couple of ways that that can happen. If, for example, you lose your arm, you might have the sense that your hand is clenching, even though you no longer have a hand.

And so that's phantom limb. What Cindy is talking about when we talk about embodiment is similar to that, but importantly different. So when we, let's say we're using a pencil, you're writing something with a pencil back in the days when people used pencils, and it almost felt like an extension of your body. So you didn't have to think about it as an independent object. What we hope to do is reanimate the hand and the arm so that the person… The paralyzed hand and arm. The paralyzed.

Got it. So that it feels like their own again. Right. Eric, let me ask you this question. I mean, from my own experience of surgery being done to me, you get scar tissue. You can't help it. Now, if you're

implanted... Wait, wait, Gary, have you been operating on your brain? No, that was a waste of time. The new pointless operation. So you're talking about in normal other surgeries, there's scar tissue. So even if you're inserting, and please tell me how large or small, if you like, these microelectrodes really are.

If you're going through the skin and going into spinal columns, if you're going into the cord, if you're going into the brain, you are going to encounter scarring. So how is that affecting how these electrodes operate? So this gives me a great excuse to talk about what I would say is the most science fiction thing my lab does. Right now, the electrodes that we're putting in are on the size scale of like 50 microns, right? And that's big for something that goes into the brain. You're absolutely going to develop a scar around that. And that is limiting the performance. So, you know, you wouldn't do this. And again, a micron, that would be 50 thousandths

of a meter. Yeah. Yeah. So 50 microns, you really need neurons to be within 30 microns. And actually, I'd love to talk about the physics of this. If you have a neuron within about 30 microns, you can see the little check marks every time it fires a spike on your signal. Scar tissue pushes that away, right? If I grow a scar around that electrode, those neurons get farther away. That's the hardest thing about making these systems. What, you know, my lab and others are doing to try to get around that problem is we're making wires that are smaller than neurons, right? So my lab uses carbon fiber. Carbon fiber is one of the strongest materials we have access to as engineers. And we can make something that is less than 10

microns, but can still go into the brain. And then that there's, you know, yes, there's still some scarring at that point, but it's dramatically less. Wow. If you're smaller than the stuff that's going on in your body, your body won't even know to make scar tissue, it sounds like. Right. That's the hope. I mean, eight microns is still vastly bigger than the kind of wires we have in computer, you know, microchips and things like that. So, like, it's not, it's a solvable problem. There's a design space where you can make something small enough to go in, but big enough to get the signal out. Okay, so you're overcoming the scarring issue. Fantastic. You can't just put any material into a human body. So you've got to find biocompatible materials.

So where have you- Unless you don't like the patient. There is always that. Chuck. Why, Chuck is not a medical doctor. That's fair enough. But you see, so where have you gone? Where have you gone for biocompatibility? What materials are you finding the most effective?

Yeah, so I mean, there's wonderful material scientists working on this, but we have a really good menu to work with. So, you know, we work with carbon. There's all kinds of like organic polymers that are in medical devices. Your body actually really likes things like titanium and platinum, right? So there's a lot of medical devices made of those materials. And so that's sort of our menu. And it's good enough.

There's a lot of like, you know, good conductive materials, good insulating materials. People are making small devices out of glass packaging. That does a good job of protecting electronics. And so, yeah, I'd expect to see a lot more devices that go in the body, you know, over the next 10 years. Wow. All right, so Parag, if you've got a damaged spinal column,

right, and there are no signals coming out of that, how can you place electrodes in that area to regenerate? Just gap the break. Or do you bridge it, as Neil was suggesting? Yeah. Yeah, so this is a really exciting time because, as you can hear, technology is advancing, computational speeds are advancing, and so our options are growing. And so one option is to get signals from the brain and to operate something like a

electrical stimulation of the muscle. Another option might be to create a jumper cable in the spinal cord. And people are actively working on all of these because in the end, we don't know what the right answer is going to be and what the one is going to be that's easiest to solve. People are trying to use new materials to bridge the gap, to let your own cells grow and reattach. There are lots of different approaches. And so I would say that the approach that Cindy and I are taking is an excellent possibility, but it's certainly not the only one. excellent possibility, but it's certainly not the only one. So when you're, I'll call it harvesting these signals from a patient's brain, how are you interpreting what check marks present themselves for you to be able to go, right, that's going to mean in ones and zeros, someone picking up a pen, someone holding a cup of coffee, someone waving.

So how are we interpreting these things? So there's a lot of steps along that pipeline. So, you know, we start with the voltage. The voltage has little deflections in it that mean when the neurons are firing. We then have to turn that into some sort of summary signal. So, you know, you can just count up the spikes like for some amount of time and look at firing rate of neurons. Then that's when you need your training data. So you need to say, okay, I know this set of spikes is for this movement.

This set of spikes is for this movement. And, you know, usually we're running something like a regression, right? Which is where, you know, you're basically trying to get these, you know, one set of signals to predict another set of signals. Or we're learning some kind of probabilistic model, like, you know, it's like, what is the odds that I saw these 10 spikes, you know, when you're moving this direction? And, you know, and then, well, then there's even more circuits, right?

Because then we have to turn it into something that goes in the prosthetic hand. So it's, I don't know, it's a fun engineering problem. And there's at least five or six steps down that pathway. And that's all trying to reproduce, you know, what was normally in the spinal cord. So an engineering and statistical problem. Yeah. Right. Because it's the distribution of cause and effect with signal and phenomenon has to be modeled. Yes.

Right. And that this is your job and your people. And the better our model is, the more likely it is that, you know, when you, when you go to grab the doorknob instead of the cup, it's still going to work. So, and that's challenging. So if I could take away one challenge that you have to overcome to get to perfect restoration of movement in a hand with all of the, what I would call, typical finger movements and thumb movements, what would that be?

What would the biggest challenge that I would need to remove right now for you? Yeah, so I would say the answer is probably not intuitive. I need the body to not destroy my devices. Right? Like anything you put in there, I can't make, like we know how to make tiny electronics.

We don't know how to make tiny electronics that can live in the brain and not get attacked, right? It's like a hot, salty water kind of environment. And if I could just make things survive forever, we could make things very small and get them into the brain. So what you need is a cloaking device. Yes.

You need to be able to couch the technology inside of something that the body thinks is itself. If we could also teleport, that'd be great. And then we wouldn't need Parag anymore to open the skull. And then the warp drives, yeah. Just go down the list, you know. Okay, so Cindy, give me, I mean, okay,

so the body has a self-defense mechanism and, you know, phases aren't set to stun, etc., etc., to carry on the Star Trek analogy. Just how long does this stuff last in the human body? Are we talking days, minutes, weeks, months, years? The good news is that the lifespan is measured in years. I just saw an article in Wired. They were talking about, you know, the new record being set at like a seven-year implant.

That's still functioning pretty well. So, you know, it's pretty good. But of course, if you're going to get a brain surgery, which you definitely only want to do once, you're going to want to last for a very long time. Yeah, seven years is not a good run. You don't want to be seeing a neurosurgeon every five, seven years. I mean, that in itself is traumatic to have to go through that amount of surgery,

particularly in that part of the body. Okay, so there's still something I don't understand. So, Cindy, it sounded like you gave this one, like that's the only thing left to resolve. But the two of you working together, are you saying you have succeeded in exactly what you intended? Did you recreate Luke's hand? Did you, you know, what are the successes? Maybe that's really what I'm asking here.

So I think Cindy nicely talked about the major engineering challenge, which is the brain-machine interface. And I would say from the biology side, we still don't understand the language of the brain. We don't know when we are recording all these signals, what are the signals telling us? And so that's a whole, we need the Rosetta Stone. You know, as Gary said, we need the Rosetta Stone for the brain to understand what these signals mean. So is that basically the hard yards of research and going through patient after patient after patient and just making AI work a 28-hour day. Yeah. I mean, I would say my perspective is slightly different on that. Where like, I totally agree.

We do not understand the language of the brain. There's so many wonderful neuroscientists that are working on that. But quantity has a quality all its own. So if we could solve the engineering problems and we could get, you know. Beautiful quote. A thousand. Oh my gosh. It's not mine.

I think it's Napoleon. Say that quote again. Quantity has a quality all its own. So if you were to give me a thousand stealth microelectrodes that the brain, you know, was not reacting to, I'm pretty sure I could get a lot of these degrees of freedom moving even without necessarily solving the language just because of the power of machine learning. Yeah, but Parag, we've been focusing this conversation on movement, sort of the

kinetics of the brain-human intentions. But what about thoughts? And how is a thought, a pure thought, different from a thought that creates movement? So that's a great question, and we don't have the answer. So people are also working on these kinds of interfaces to reproduce language. For example, you put the electrodes in the language area of the brain, and you try to interpret what a patient who is unable to speak is trying to say.

Ooh, right. And also those thoughts do have, I don't know how to put it. It's not a tangible quality, but a measurable quality. Like I read this study about how they put these people in an MRI and they did things to them just thought-wise and their brain lit up like it was actually happening. So there is that element of it too. Yeah. So that comes in where we're trying to think about where to put these electrodes. So imagine someone is paralyzed, so they can't move. How do we know where in that particular patient's brain the hand area is located? What we do is we have them think about moving their hand, and their brain activates in their individual hand area, which gives us some sense of where we need to go. which gives us some sense of where we need to go.

Even though that hand area has lost its electrical connection, the hand area of the brain has lost its electrical connection to the actual hand. Correct. Wow. Trippy. This completely is trippy. I have to say that if you lose the electrical connection, then the representation in the brain changes. It might, just like when someone has a stroke, other areas take over those functions. It might rewire, yeah. And so some of

that is also happening. So then how do you account for that kind of neuroplasticity that, you know, it seems like the brain is always changing, which, I mean, so you're always trying to hit a moving target. That's kind of, okay, Chuck, wait, wait. Chuck, put a pin in that. We're going to take a quick break. And when we come back, we're going to find out the answer to Chuck's question. And I want to start bringing up sort of the ethical issues here. Poking around in people's brains to change what they're doing. And maybe even one day what they're thinking.

That's part of the neuroplasticity. Stick any word after neuro and it will apply in this moment. Okay, when we come back, the StarTalk Sports Edition. We're back. StarTalk Sports Edition. We're talking about the brain with two brain experts, and I love it. We got one coming from the biology side, one coming from the engineering side, and these are the folks we need to get their heads together and figure out what the future of the human brain interface,

the brain machine interface, will become. And so, Chuck, we left off with a quick question of yours. I'll paraphrase it. Was it that we know the brain rewires itself from stroke victims and through other sort of brain can be opportunistic so that it doesn't completely lose abilities from an injury versus what it can recover from. So how does that plasticity interfere with your data? Yeah. So if it was a brand new brain every day, this would be really, really hard. So luckily, we don't have to deal with that. I can count on things being very, very similar from day to day, even though we do have to recalibrate all the time because

lots of mundane things change from day to day. So there's some base truths, base truths. Yeah, like you're more or less seeing the same signals from day to day. But of course, yeah, of course the brain can change. And there's really like two schools of thought in how to do a brain machine interface. I'm like the machine learning club, right? We're just going to use lots of algorithms, you know, same things they use for autonomous vehicles. There's also like the plasticity club, right? And they're going to try to figure out how to unlock the brain's natural plasticity.

And, you know, surely I control my own brain. And if you just give me my own enough practice with it, that I'm going to be able to control anything with it. Then we just go around the problem, and then the problem's not a problem anymore. And it doesn't matter where you put the electrodes. It's the brain, right? Yeah, you just rewire. Yeah.

You just wire around the break. Yeah. Okay. Wow. All right. That's pretty cool. Let me ask you, how long before these neural interfaces, once we get our Rosetta Stone, once we crack and code our signaling,

how long before this is common medical practice? And if it is, will it be limited to just your motor skills or will it have applications for neural damage in the brain, senses, or hearing, or sight? Or even more importantly augmentation oh yeah of course

how long before we're bionic there you go yeah yeah yeah I mean this this is one of those wicked problems that

you know if you asked me when I was headed up to college if I would have something like this that you know

has all this information I can reach all my friends I would have let the record show he held up a smartphone a smartphone yeah something like this that, you know, has all this information, I can reach all my friends. I would have no idea. Let the record show he held up a smartphone. A smartphone, yeah. I wouldn't know. And so the answer is we're making steady progress.

And I think that devices that are brain machine interfaces, you know, you could see them rolling out in the next 10 to 15 years. But at what point will we have something that we want to control our wheelchair in traffic? That's the question. So there's a question of reliability as well. So we're making steady progress, but we don't know. And there may be jumps in technology that we haven't even imagined yet that are going to be fundamentally important to what we're trying to do. But, right, so the future of this is sky's the limit, right? Because you're here working on motor skills, motor ability at all, but you're still poking the brain.

I'm sorry, you didn't like the word poke. You're, what's the word? Exploring. Probing. Exploring. Probing the brain. Exploring an electrode. Okay.

So when I think of all of the reasons why people are institutionalized for mental health reasons. If once you, Cindy, once you know the brain map, you ought to be able to go in and nip, tuck, cut, paste to change whatever might be a completely regressive societal behavior into someone who's a perfectly law-abiding citizen otherwise, or perfectly behaving in ways that we consider normal. But then again, who defines what's normal? Yeah.

So, yeah, I agree. In a sense, what we're developing is technologies for interfacing with the nervous system. And the nervous system does all kinds of things and is related, you know, we can help people in so many ways just by, you know, controlling the organs or trying to correct things in the brain. But yeah, I mean, I guess I hope that the future has a, you know, robust ethical framework around this.

I mean, I think those decisions have to be made, you know, by individuals, by in consulting with their doctors and that, you know, yeah, I don't have an answer to that because that's difficult. But I can say that this is coming and it's probably coming within the next couple of decades. And so now is the time to be talking about it. So, okay, this is so far off, but I still got to ask.

Okay, this is so far off, but I still got to ask. When you look at the neurochemical interactions of the brain and you look at what you're doing in terms of recording it, codifying it, and then manipulating it, wouldn't it be possible one day to take that and kind of change the way people think? Because like, for instance, something like racism, which is an irrational response to seeing someone different, you would be able to kind of dial that down

so that you might be able to eliminate certain human maladies with this type of technology? Or am I just too sci-fi and too bleeding hard? So that's a great question. And so if we- And by the way, there's a deep morality dimension of that too, because that means you're controlling people's thoughts and you have a nice progressive mission statement

that we all, you know, kumbaya, but in the hands of nefarious counterparts to Cynthia and Perlhock, you guys could start a revolution. So, yeah. What's up with that? I think it's like so many things. There's probably a potential to do bad, and there's also a great potential to do good. So for example, if someone is addicted to drugs, can we put a signal in that helps them get better? I was involved in a trial where we put electrodes in to try to help people with otherwise incurable depression and help them access things like therapy. So all of these things have the potential for good or bad. Frankly, I'm hoping that elementary education gets rid of racism in

this country rather than having to put electrodes in everyone's brain. Seems a less risky approach. I got to tell you, the way things are going, I'm betting on you. approach. I got to tell you, the way things are going, I'm betting on you. Cynthia, can you comment on the ethics of this? Yeah. So I want to say, you know, right now, we're starting out with little demonstrations. We're just getting started with all of this, but it is not too early to emphasize the ethics. And we need to put a framework around this. Like, you know, what you said, like changing people's thoughts by stimulating, that's possible. We're doing it.

We are exploring treatments. And so we need to put, you know, the legal framework, the ethical framework, and we need to do that now. And I will say like right now, we already have like, I love the FDA. I enjoy all of our interactions, right? You know, we need regulatory agencies like that.

But we need to start exploring it now. Okay. I love the FDA, but. Well, but if we didn't have them, like, this stuff should have an ethical framework around it. And I don't think we've had enough of a discussion as a society about what that should look like. it. And I don't think we've had enough of a discussion as a society about what that should look like. Let me just add that it's so important also to involve the people that you're trying to help and get their ideas and perspectives, making sure that what they want or what we think will help them will actually be of use to them. And so that's a whole other dimension.

Well, that's a good point. Good. Yeah. That's a whole other dimension. Right. That's a good point. Good, yeah. You are both in the field of biomedical electrical engineering, right? A, is there a group of chemical engineers studying the same thing? And is it going to be something like those joining forces that brings this thing to a speedier conclusion

and gives you what ultimately you're aiming for. Absolutely. So one of the most exciting things is that Cindy and I don't operate in a vacuum. We have a whole team at the University of Michigan. There are probably 40 investigators who are interested in what I call restorative neuroengineering. And there are material scientists, there are chemists, there are electrical engineers, there's data scientists. It's going to take the combined efforts of all these different fields

to lick this problem. So generally that's enabled when there's a journal that comes out that has syllables extracted from all those fields stapled together into a new word, like biochemical engineering, right? I mean, this has been all the rage in the last several decades. It's how you get astrophysics, you get biophysics, you get geophysics, you get geochemistry.

So it sounds like you guys need a little bit of everybody to tackle the brain. This universe probably won't need an astrophysicist. Physics comes in here a lot, actually. Totally. I'm all in with the physics. I'm all in with the astro part. Maybe a little less. I have another question because you've got my minute brain spinning here.

How are you powering this? I mean, are we getting to the point where you can solar power? Is it close enough to the surface of the skin? And the connectivity, are you at a Bluetooth? Is this Wi-Fi enabled or is this just basic? Is this cable at the moment? How far along are we on those two fronts? Yeah. So right now, every device that's inside of a person today, it's probably got a battery in it or it's got like an RF coil. So it's, you know, radio frequency power. There's a little bit of like optical powering. Where the powering gets really exciting

is we're trying to make these implants tiny, right? So for example, we're trying to do infrared, which passes through tissue really well. And we're trying to use that to take our, you know, bed of needles and make it a distributed group of like lots of little nodes that can record. And those would get infrared power. But people have looked at all kinds of things for this. Ultrasound, like that's another way of getting power into the body. So, but yeah, lots of items on the menu. If you have too much power around each electrode, do you not then run the risk of damaging cells and causing other problems?

Yes, and that's why you can't just, you know, power everything from the outside. Like you can always get power in there, but the question is did you deliver too much power to the brain tissue in order to power up your device, right? Because you have to say… We always want more power. No, no, the brain does not want more power. Things go very wrong if you deliver too much power.

More power. No, no, no, you can't put a V8 in there, right? Oh, yeah. Wait, just to be clear, V8 is an old-fashioned internal combustion engine for a car that has eight piston V4s. Yeah, you guys remember those. Yes.

Dinosaurs used to drive them. Wait, wait, wait. Wait, Gary, I just got to ask, just slip in here. So Chuck just hinted at this earlier, but I want to make sure we address this because it especially affects performance, either intellectual or athletic performance. What's to stop people from making themselves better than they ever would have been,

holding aside whether they were injured or missing limbs or had spinal cord injury? Right. Just give me that. I'm a, quote, average, normal, healthy human being. Right. Still give it to me so my reflexes are faster. Right. So I can jump higher so that I won't psych myself out in the big game. We already use, you know, chemicals to increase and augment certain brain functions.

Yes. Why not have— And as Yogi Berra said, 90% of the game is half mental. There you go. Okay. So, like, I want an implant that just singularizes my focus. Focus, yeah. I singularize my focus so that when you

throw a pitch at me, all I see is like the baseball is like big as a volleyball. And the pitches are just turning slowly. I can actually read Rawlings across as it comes across. So you know enhancing human performance is on the doorstep. So where do you see that taking place? Yeah, so I want to say for specifically the stuff that we do that involves doing surgery, we are not remotely talking about augmentation. It's nowhere near where we're at. We're trying to get bits per second out of the brain, right? Which is amazing if you're trying to control a prosthetic hand,

but not very helpful if you're an athlete, you know, trying to augment your performance. I can say that there's every possible kind of non-invasive way to do this is being explored right now, right? There's little headphones, there's little things you can wear on your forehead, there's earphones that, you know, theoretically stimulate your brain. There's where we have some of the physics problems, right? Like you're just, you know, you're trying to get a signal in or out that, you know, maybe you can do something and it's worth a try,

but like, you know, you're probably not going to see that much augmentation in the near future. Wow. But deep brain stimulation, either by electrodes that go deep in the brain or by some, like you're saying, Cindy, some external device that focuses energy.

Super madness. Yeah, for example. Probably not, but yeah, for example. So I always say that augmentation is nowhere near where we're at, you know, like 50 years, like, you know, not going to happen. However, this whole field didn't exist 20 years ago, right? And 20 years is an eye blink.

So by the time you're fast-forwarding this technology, I mean, like, I don't think that, you know, there's going to be brain implants anytime soon, but a thousand years is a long time, right? And there's maybe going to be other ways of doing this that are more non-invasive and more safe. And so you can never say never in engineering. So, Parag, if we can actually make permanent positive changes to the brain,

then we stitch you back up and Cindy's out of a job because your brain is doing it on its own, neurochemically, rather than with external help. Is that where you might land one day? Yeah, well, so that's another great question. So one of the things we do deep brain stimulation for is for Parkinson's disease and to help patients who have Parkinson's disease. And people ask me, well, what will you do if we find a cure to Parkinson's disease?

Who will you operate on? And the answer is, I will celebrate. Similarly, if we find a way to cure spinal cord injury, I will celebrate. We will all celebrate. There are so many. I'll start working on the problem of chronic pain. So with every technology,

it creates many more opportunities than it does limit our options. I didn't mean to express sympathy for you guys being out of a job. That was not my intent here. It was just, I'm just imagining if you have a cure for something, then so much of what Cindy is describing is no longer necessary. That's all. Because the body now handles it itself.

So that's a foreseeable future as well. Yeah. In a sense, I mean, you know, where there's people that are trying to regrow the spinal cord and there's people that are trying to bridge around it. And it's a, it's a friendly competition, but it's great if one puts the other one out of business. But if it's just electricity and, and, and chemistry, why can't you just solder a new connection across the broken spinal cord? Why is that so hard? Get some physicists in there. We do this all the time.

Okay? Some electronics people. Just give me a soldering iron. Yeah, so unfortunately, it's much easier to do with biological components. So, like, I'm a fan of the wires, right? Like, I'm trying to get wires in. But probably it's better if it's, like, neurons and axons, and they can, you know, reform the connections, make new synapses. But on the other hand, you still have the patterning problem, right? Like they have to know where to go. And so

people are using electricity to try to guide like, you know, regrowth of the spinal cord, for example. Neil, we're right on the doorstep. We just need three things. We need the wire, we need the soldering iron, and we need the wiring diagram. Once you give us those three things, we're there. Oh. Ooh.

Okay. And it sounds like the wiring diagram is a tough one at this point. That's hard to get back once you've lost it. Whoa. By the way, Parag, that is everything. So just pointing that out.

You said, we're right there, we only need three things. I'm like, yeah, that's everything? No, no, no. But it's deep that you can even list three things. Yeah. No, no, it is. And you can abstract those categories to many other walks of life. All right?

You need the tool. You need the materials, the tools, and the diagram. I mean, I think that applies to everything. But the cool thing is that every advancement you make touches upon other areas of our lives. It's an advancement everywhere. Any advancement where. It's an advancement everywhere. Any advancement where you are is an advancement everywhere.

Well, and it's a science and engineering. That's the hallmark of what that is. Yeah. Well, it's the rising tide, isn't it? This is rising tide that raises all ships. So it's just the benefit of this work. So Cindy and Parag, how do we find you on social media or your lab?

Does it have a home on the internet? Yeah, I have a great last name for science. So, it's a recent immigrant misspelling. So, if you search Chestic, you get me and my brother. And so… And that's it? Yeah. And nobody else?

So, Google, I'm very Google-able. And, you know, I run… Chestic, C-H-E-S-T-E-K. Yeah, so my lab and, you know, all the wonderful students. You know, I don't do any of this work. They do all of it. And I'm always looking for new trainees and new students. That's what students are for.

Are these the students who you put the electrodes in? No. Coding the programming. No electrodes are implanted in students in my lab. Okay. And Parag, where do we find, are you, how do we find your work? Well, similar to Cindy, P-A-T-I-L, if you Google that and Michigan Neurosurgery, my email will pop up. Okay.

Guys, this has been a delightful interview. I love learning about where we are and where we're going in not only engineering, but in all the sciences. And you know, Gary, Chuck, this is going to show up in sports very soon. Yeah. And hopefully it will show up in not just sports, but across the whole spectrum of society in such a bountiful class for so many people.

All right, guys. This has been great. Chuck, always good to have you. Always a pleasure. Pleasure, Neil. Cindy and Parag, we'll track your work. And if you get any new breakthroughs, give us a call. We'll put you right back on the air.

Oh, yeah. Thanks so much for having us. We'll learn about what that is. This has been StarTalk Sports Edition. Neil deGrasse Tyson here, your personal astrophysicist. Keep looking up.