Presenter: Professor Tom Oxley, Founding Chief Executive Officer, Synchron
Category
🤖
TechTranscript
00:00Good afternoon everyone. I'm very excited to be here to talk about what is an
00:05upcoming space and what promises to be potentially very ground-shifting for
00:11humanity in ways that are hard to appreciate right now. My name is Tom
00:15Oxley. I'm the CEO of Synchron. I'm an Australian neurologist. I moved to the US
00:21about six years ago, but about 12 years ago I started this journey with funding
00:27from US Defense Agency DARPA and what's now been a critical infrastructure
00:31investment in this now emerging space. We are five years into human clinical
00:37trials at this point with ten patients that have now been
00:40implanted with our technology and the initial application for this technology
00:45is in the need for people who have paralysis. We anticipate this is going to
00:51be a very large area of medical device which has not yet been addressed by
00:57any currently available technology. Paralysis is a wide-ranging medical
01:01condition from a range of afflictions of the body and as just mentioned there's a
01:07prediction that this first-generation healthcare TAM is going to be in the
01:11range of four hundred billion dollars. As a neurologist I've been looking after
01:17patients who have suffered from injury or disease and one way to think about
01:22how the problem is going to be solved is how does the brain engage with the world
01:26around us. You may not appreciate this but there is a huge amount of
01:30information coming into the brain, limited information going out. At the
01:35bottom of that diagram there is the brain stem which carries all the
01:38information out of the brain. So we already engage in a digital world
01:43every day, we all use our devices to engage with our day-to-day activities. So
01:48one way to think about this is that we can expand our capacity to engage in the
01:53world if we have multiple avenues of outputs from the brain, in other words a
01:58virtual nervous system that can engage with the brain. So with people who have
02:02paralysis or disease the brain is often intact and it's how the brain engages
02:07with the outside world that needs to be corrected. For people who are suffering
02:10from that it's a horrible way to live when you're locked inside your own skull.
02:14So the world is now ready and you might have heard about BCI, it's getting a lot
02:19of attention in mainstream media now, but this is happening off the back of
02:23decades of neuroscience research, now infrastructure that enables edge
02:28computing, very active global leadership from the FDA, especially David McMullin
02:34and now mainstream awareness. Many of you have probably heard of Elon Musk's
02:37Neuralink and I would certainly say that over the course of fundraising over the
02:41last 12 years there's been a big flip in how venture capitalists are thinking
02:45about the potential for this space. But there's been one piece missing which
02:50relates to how this technology can scale because most people think about open
02:54brain surgery for brain computer interfaces and so what at Synchron we've
02:59done is we've looked back to take cues from the medical device field with what
03:04enables technology to scale out into market. And so the problem with open brain
03:09surgery right now is that none have scaled up to this point.
03:13Cochlear devices for hearing, deep brain simulation for Parkinson's disease,
03:17and then epilepsy are in the order of around 10,000 cases per year per annum.
03:32These are tiny numbers relative to what we've seen in medicine before and that's
03:36not going to be the solution for brain computer interface to go to treat the
03:40potential problem at hand. Instead transcatheter is what unlocked cardiac
03:46treatment over a period of decades. Going from open heart surgery to now
03:50transcatheter therapies we've seen a change from the 70s to now the
03:58year 2024. Low hundreds of thousands to now millions of procedures being
04:03performed every year. And this happened over a period of decades from balloons,
04:07stents, pacemakers, and valves there's now 1 million cardiac stents being
04:11delivered in the USA. The key really has been to get out of the operating room
04:15and really no other area in medicine has grown like this over the past several
04:18decades. So Synchron asked the question how can we take that revolution into
04:23neurotechnology and deliver technology at scale. So this is an example of how we
04:29deliver our first technology. We come up the venous system through the jugular
04:33vein using a catheter to come up to the top of the brain and then deliver
04:37electronics into the inside of the blood vessels that can interface with the
04:41brain from the inside. We utilize a lot of the materials and knowledge that have
04:45been the basis for solutions in cardiac over a number of decades and we've
04:50repurposed that to sense brain activity and turn that into outputs that are
04:55useful initially for people with paralysis. This is our first generation
05:00system. We're calling it SyncOne and it hopefully will become the first in
05:05class implantable brain computer interface. We've conducted two clinical
05:08trials over five years with ten patients implanted. The best way to think about
05:12this is a brain controlled system that enables you to navigate through your
05:17personal devices without having to use your hands. You may not need this in the
05:21audience today but for paralysis patients this can be life-changing with
05:26there's currently no available therapy. So here's one of our users. The system
05:31goes in, it switches on, you start trying to move your arms and legs. We train the
05:35system how to move very simple things around the screen. This is not yet at the
05:39level of control of what you can do with a mouse and a keyboard but it does
05:42enable basic use of iOS. So we've been using control of Apple products to
05:48engage in navigate and select as a minimal viable product. This is where the
05:53technology will be starting with our first generation system. We're now
05:59engaging in the development of a novel BCI HID Bluetooth profile and the best
06:05way to explain this is that the example of Steve Jobs. Steve Jobs didn't like the
06:10idea of the large iPhone because your hand or your fingers could never reach
06:14all the way around the screen. He thought all phones should be reachable by any
06:17finger at any point in time. So that speaks to the idea that our devices are
06:23currently constrained to the built form factor of our human bodies. BCI is going
06:29to be constrained not by the physical form factor of the human body but by the
06:33constraints of the brain and I think we all know that there's more going on in
06:36your brain than what you can express. Maybe fighting with your partner or
06:40trying to express your emotions. There's far more things happening in your brain
06:44than your body can get out. So coming back to the virtual nervous system I
06:50just want to explain how we're thinking about this problem. Probably the most
06:59common question I get asked by investors is how is this field going to progress?
07:03And semiconductor chips, that was the mechanism of Moore's law in
07:09computers. So how do we think about progress over time with BCI? The
07:14prevailing theory or thinking has been that the number of channels or electrodes
07:20inside the brain, think of them as sensors, is the limiting factor. So more
07:25electrodes into the brain, more horsepower, more bandwidth. But we don't
07:29believe that to be true. We actually believe that optimizing brain diversity
07:33from a data perspective is what's going to drive compute. So compute versus
07:37bandwidth and I want to explain that concept to you today. This is the motor
07:42cortex. This is where BCI is beginning. It's the most relevant area to begin
07:46because it's the most constraining in our body. This is the command center of
07:50your brain. It takes up about a third of the real estate in your brain. About the
07:54size of an A4 piece of paper if you unwrap it. Our first device only covers
08:01about 0.04% coverage of the whole motor cortex. It's a tiny amount and we've done
08:05that with 16 electrodes that we deliver into the brain. You might have heard of
08:09the and so as I mentioned minimal viable product is navigate and select and we're
08:14doing that on Apple iOS products. The neural link device has now been
08:19implanted in two subjects has a lot much larger number of electrodes with the
08:23idea had been that if you put in a much higher number of electrodes you can get
08:26much more data out of the brain. Navigate and select is also the same
08:32control mechanism but something we noticed something interesting. In that
08:37first neural link device 900 channels were lost in the first month and that's
08:41been publicly reported and then the system came back online with about a
08:45hundred channels about 10% of the number of channels and the function of the
08:49device didn't change and it got us asking the question okay so if you had a
08:54thousand down to 100 and there was no performance change why is that? Why were
08:58there was there redundancy in the system? And so we've had a theory for some time
09:03that there is a optimized amount of information and the reason for that is
09:07that brain function is preserved in packets of neurons around the brain so
09:13neurons are next to each other do the same thing. The best analogy I can give
09:17is listening to a symphony orchestra you put a thousand microphones on one
09:21violin and you're trying to hear the whole orchestra you lose 900 of them
09:25you're not losing that much information about what you're hearing in the room. So
09:28the question became if you're trying to optimize a system for commercial
09:32manufacturing commercial scale how do we think about redundancy and optimization?
09:37And we have a belief that at about the level of one millimeter you have
09:41information redundancy in the brain. So it's not just horsepower of electrode
09:45count it's have you spaced your electrodes in the brain at the right
09:48amount. So what we're saying is that at the microscopic scale there's
09:52redundancy the mesoscopic scale is where we think there's the right
09:55optimization of naturalistic brain function. So with that in mind we are
10:01building a system with multiple generations and I want to explain how I
10:05think this field is going to progress over time. Firstly, maximizing brain
10:09coverage diversity. Secondly, people like their skulls so leaving the normal
10:15physical architecture intact as much as possible and finally delivery at scale
10:20which I've covered already from the cardiac analogy. So at Synchron our
10:24second generation system is utilizing the blood vessels and the blood vessels
10:28are like branches of a tree they go down in order of size and they reach further
10:32and further. So when we're minimizing and miniaturizing our electronics into
10:38the brain a second generation system at a maximized at a maximized spacing is
10:44going to cover about 1% of coverage of the brain. We predict that will lead to
10:473D control of robotics and then our third generation system we're going to
10:52be targeting around 15% coverage in the brain. We don't exactly know what that's
10:56going to lead to but we estimate that at that point we're getting to full
11:00humanoid avatar control capabilities and it's got us thinking a lot about where
11:04this field is headed in the next couple of decades and what it means from a data
11:10perspective and it hasn't been obvious because so the the recent the recent AI
11:17revolution has shown us that a massive data source can label you to do these
11:21incredible things with natural language processing but from a BCI perspective
11:26it's we don't have the advantage of having huge amounts of brain data. We
11:30actually have a time series brain data set and what we've realized is the way
11:34the brain reacts is only relevant in the environment that it's reacting to. So you
11:39need to know how is the brain adapting to the environment, what is the context
11:43of how the brain is reacting in that moment and can we build recursive AI
11:49platform that is going to enable this virtual nervous system to flourish. So
11:54here's our vision of the future. Axes on one hand model model fidelity over time
11:59and on the other hand diverse brain region coverage. Our previous couple of
12:04clinical trials with our prototype has got us beyond click 1d click to 2d and
12:08click which with our first commercial system as we move towards a miniaturized
12:12second generation system getting towards 3d control and robotics and then finally
12:17moving towards an area that we don't fully appreciate yet but what I can say
12:22is that we will pass the precipice of improving that which the human body can
12:26do and so you know the 400 billion Morgan Stanley number comes around 12
12:31million user market potential. Beyond that it starts to get up to an economy
12:37that looks a lot like the automobile industry to me and it got me thinking
12:42about how do we think about where this is going to go and what it could mean. So
12:50I've been reflecting on on the ethics and the morals of where this is going to
12:54take us and it's it this might not happen for a period of decades but it
12:57does seem obvious to me that once we have enabled the ability for the brain
13:01to engage at the level that it can naturally process and not be limited by
13:05the muscles in our bodies we are going to go into the precipice of being able
13:10to achieve a level of human productivity that hasn't been possible and I think
13:14we'll look back as we all are now and think how slow we were and so what does
13:21that make you think of or what it made me think of was something we've seen
13:25before. Improved exponentially improved human productivity functioning beyond
13:30the physical limit of the body and coming with a minor risk that requires a
13:34procedure. Now the procedure risk is probably less than what we take risks in
13:41our daily lives and it made me think about the fact that when automobiles
13:45first appeared there was huge blowback from the community. Why would you
13:52get in a motorized car and risk your life? Then we had flying and the same
13:56question were asked about why would we take those risks and take that on. I
14:00think what's going to drive BCI out into the future is the need and demand for
14:04human productivity and I think that's where BCI is going to take us in the
14:08future. Thank you.