Phantom Neuro is developing a muscle-machine interface that allows individuals – particularly those with limb loss or motor impairments – to intuitively control advanced robotic systems like prosthetics and exoskeletons using the natural electrical signals generated by their muscles.
Connor Glass, founder and CEO of Phantom Neuro, spoke with MobiHealthNews about the company’s technology for amputees and the potential for robotic limbs to one day exceed the capabilities of the human body.
MobiHealthNews: What is Phantom Neuro’s technology?
Connor Glass: Phantom Neuro, at the end of the day, if I had to describe it, is a control company, both metaphorically and literally. So, we are trying to make people’s lives better by giving them better control over their lives, and in particular, helping people that have lost some semblance of control through some kind of an injury or disability regain that control and then maybe even perhaps surpass where they were ahead of that.
The way that we’re doing that is by literally enabling control through this emerging paradigm of robotic systems, primarily.
So, there’s incredible robotic systems that other people are building – humanoid robots, prosthetic limbs, wheelchairs, exoskeletons, surgical robots, all of these things. But the problem that exists is that patients and individuals are not able to meaningfully control those systems in order to capture their true intended utility.
So even though a robot may be capable of moving each individual finger and rotating its wrist and doing really complicated things, a user does not actually have a way to control that functionality, and so what could be an incredible robot is relegated to a very simplistic device that does not justify a really high price tag.
Phantom is about bridging that gap between robotic systems that are emerging and the future of those systems, and a human’s ability to capture that functionality within the robot and, therefore, enable control of it so they can regain their functionality.
And then there are, of course, applications for this technology beyond just an injured state and giving somebody control over a robotic limb or an exoskeleton to recreate someone’s function, but also by allowing these robotic systems to function more like humans, either semi-autonomously or autonomously, where you can take the same data that drives direct control of these different robotic prosthetic systems, for example, and use that as a training layer to be able to actually train the robots to move like humans.
So, you take human movement data and put it into a robot to tell a robot how to move like a human. The best way to train a robot to do something like a human would do it is by taking data associated with a human doing that action and then teaching the robot to do that. And so, we’re really trying to allow complicated robots to function like humans.
MHN: What you are offering is a muscle-machine interface, right? How does a muscle-machine interface differ from a brain-computer interface?
Glass: There are two key facets to that. The first is, how is it technically different? So, it’s technically different in that we are recording electrical activity from the muscle instead of recording that electrical activity from neural tissues like the brain or the spinal cord or nerves.
Most people are not aware that the muscles actually generate electricity when they do what they are supposed to do, just like the brain or spinal cord or nerves. And you can find patterns in that electricity the same way you can in a brain to do the same thing.
So, at the end of the day, whether it is a muscle machine, nerve machine, brain-machine interface, you are finding electrical data from the body, you are finding patterns in that data and using those patterns to basically do interesting things.
We have chosen to focus on muscle as our source of electrical activity, specifically for this purpose of control because how do humans actually control the world around them? It’s with their limbs. Their limbs are driven by muscles. We actually do not use the Force to just directly control the world around us. We use muscles that flex and extend to move our limbs, and that is how we are naturally set up to interact with our environment, which means that using electricity from muscles in order to have people control things is the most natural way that a person can wirelessly control robotic systems, because that’s how our body is designed to function.
There are also technical differences between muscle and the brain. So the muscle, for example, the signals are much bigger than the signals coming out of the brain. So one of the challenges in the brain is, how do you get these really teeny, tiny signals in a reliable way to be able to utilize them?
So, in the brain, they are teeny, tiny, and there is lots of noise in there and it is very chaotic, which makes it a very hard problem. Muscle signals are close to 1,000 times larger in amplitude. So much of how all of these systems work is determined by how big and clean the signals you are detecting are. The bigger and cleaner they are, which translates to signal-to-noise ratio, the easier they are to utilize in a given scenario.
With our signals, they are quite large. There are far fewer of them, and they are easier to detect. And so it is still a very challenging problem, but it is simpler than trying to do the same task directly from the brain. So that’s the first part.
The second part is everything else apart from the technology. One thing that a lot of people, in my opinion, do not think about is everything else that goes along with the technology in order to make it successful. So, you can have the world’s greatest technology but if you need a specialized surgeon, general anesthesia, a prolonged hospital stay, hundreds of thousands of dollars to be able to pay for it, continuity of care that is specialized, a functional MRI ahead of everything – that is a lot of infrastructure required for each patient to receive something that can totally negate that technology’s ability to scale regardless of how good the technology is.
For us, because our technology just gets implanted right under the skin and a residual limb, that is an outpatient procedure. You do not need a specialized surgeon. There are 70,000-plus surgeons capable of doing it compared to neurosurgeons, where there are a handful of thousands that are capable of doing it. You don’t need general anesthesia.
It can, therefore, be done at a price point that is much more accessible to patients … not hundreds of thousands of dollars, but more like tens of thousands of dollars, which is much more palatable from a payer perspective.
MHN: Do you think robotic prosthetics could be superior to a human arm?
Glass: Absolutely. I think that. As they stand today, they are not because they are heavier. They require a lot of battery power, which has a lot of problems associated with it. Some of them, theoretically or in practice, have all the different movements available to them with all these motors, but there are all these problems with that.
But we are in the early infancy of these robotic systems being developed. Now, in order for those robotic systems, for there to be an incentive for these robotic systems to approach true human function or surpass it, there has to be a financial incentive in place, which means they have to work really, really well for people and generate revenue for people for people to be willing to put the time, money and effort into making them better and stronger.
And our system is an example of one that can allow them to have the financial incentive to make the robots better, faster and stronger.
Now, the natural future of it, and the not so distant future in my opinion, is that these robotic limbs are better than human limbs and they are replaceable. So, if something goes wrong, you can replace it or if your needs change over time. Let’s say you start off as a Special Forces individual and you want a robotic leg that still allows you to go and operate in the field and do all of that, but then as you age, maybe once you are an elderly person, you want one that is much better at stabilizing you and doing all these different things that are more just quality of life things.
So you can change what the appendages actually do over time. And then you can also imagine a wacky world where you can have different attachments for these appendages.
Let’s say that you are a handyman and you have lost your upper limb in some workplace accident in a machine or whatever it might be. Eventually you will be able to have a robotic limb in the very near term, where you can have different attachments that allow you to do your job better than you otherwise were able to do when you were limited by just this form factor. It’s a wild thing to think about, but all of that tech is there and exists. It is just how do you control it and make it useful.
MHN: The device has been tested on a pig, correct?
Glass: We have tested the device on two pigs at this point, which has been really interesting. So, the first pig that we did, we implanted the very first way back when versions of the system in the intact leg, on top of the muscles in the leg, and then we trained the pig to walk on a treadmill. And we had the pig leg control a virtual reality robot leg that basically mimicked what the intact leg was doing based purely off the electricity coming from the muscles. And that is the same way that a person controls a prosthesis or an exoskeleton. And so that was our initial proof of concept.
And then we recently did another animal where we implanted the final system that we will soon be implanting in humans to do kind of like a final systems check – to make sure that everything works in a living thing and that all the software systems and the wireless communications and the robustness of the device and all of that is according to plan. It has all gone extremely well.
MHN: Have you started human clinical trials yet?
Glass: We have not started implantable clinical trials yet. We have done some surface recording clinical trials technically under IRB approval, but we are implanting humans for the first time this year in Australia.
So, we will be implanting a cohort of upper limb amputees out in Australia towards the end of this year to control multiple different commercially available robotic prosthetic limbs.
MHN: You said a lot of other devices are very heavy. What is your device made of?
Glass: Our device is extremely lightweight. We designed this system to be as safe as humanly possible and to have as many knowns as humanly possible.
So, there are people that are working on all these novel materials and really thin film arrays and things like that. We elected to not go down that path because we did not need to go down that path, and because that introduces a lot of questions from the FDA about, “Is this safe? Will it last a long time?”
We wanted to use technologies that have been used in humans for a very long time successfully. So, we have an electrode array that has platinum electrodes, which are the standard in the field, that is silicon-based.
So, just like a normal silicone array, similar to spinal cord panel arrays, similar to breast implants, similar to every implant essentially that has a silicon derivative. We have an electronics housing that is made out of titanium, which is the standard within implantable medical devices. We have some gold in there, lots of precious metals, but it is all standard materials that have been used in medical implants for literally decades with a lot of success that are known to be very safe.
