What would it take to define the future of biology and medicine?|WKF 2020

what this means is that in biology and medicine we test hypothesis does sleeping more help your brain feel better does taking a certain vitamin may make you healthy or not these are yes no questions that you can answer but this does not provide a platform for you to engineer an outcome and so currently in biology and medicine a lot of the effort to try to develop a new drug for example is based on a treasure hunting model you based on various cues you try and find clues to come up with something that might work but there is no systematic way to engineer it and so the question of what would it take to define the future and transform the biology and medicine comes down to can we build an engineering platform for biology and medicine so to just to give you an example the top 10 pharma companies current market cap even though we talk about a lot of these pharma companies that have a really large profit make a lot of the drugs that we currently use none of the companies reach a trillion dollar valuation like we see in the i.t space this is largely because of the fact that the drug development even though it's a very profitable business and it makes they have exclusive rights to a lot of the drugs because of the fact that these are based on a treasure hunting model the fact that you have very valuable drugs right now doesn't guarantee that you will be able to make a new drug for the future because you have to be in the game of treasure hunting altogether but if someone draws a clock like number 10 does that person have alzheimer's disease if they draw a clock like number one do they have alzheimer's disease you can see that there is something wrong with the person once you start to go down the scale but this is definitely not uh good enough to understand what the state of your brain is and so despite the fact that this is such an important problem we don't know what it is that we need to solve and therefore we have no solutions and so so far how many drugs do we have in alzheimer's disease space to be able to modify the disease any guesses among the few of you that are here we have exactly zero and uh the latest failure was uh in 2019 last year by biogen where many of the large pharma companies have left the space because of the repeated failures and large investments while biogen last year failed the clinical trial and this wiped out 18 billion from their market value and this drug was not going to cure alzheimer's disease even if it succeeded and yet just the expectation of something that might be able to work uh had a market value of 18 billion dollars and even the most exciting new treatments that are going through clinical trials have uh the effect of this is the group that received the drug and this is a group that didn't and there is a little bit of difference in the cognitive score that is what we are looking to achieve when you say there is an exciting new treatment in my opinion this is because of the fact that we are completely ignoring the fact that the brain is a circuit in the effort to both diagnose and find treatment for the disease the brain in no doubt is a circuit and this is not a debated fact this is a known fact and this means that the brain has integrated circuit components integrated together to control our behavior and so in order to understand what is going on we need to know what the circuit function is and the main objective of treating brain disorders is to restore the brain function your goal of treating neurological disease is not to remove plaques if you removed all the plaques that were in the brain and the brain doesn't function where you still can't remember anything or you still have a tremor that's not a cure the cure by definition for brain disease is that you have to restore the function to its normal state and so in a mathematical term it means your objective function is the normalization of brain circuit function while your different variables might be to reduce plaques or reduce inflammation or many other things that are under investigation but we are currently not using any of it to diagnose or treat neurological disease because of the fact that this is very difficult to measure so given this fact what do we need to do then to measure this and utilize this for neurological disease control there are two things from the perspective of an engineer to be able to try and address this problem one is in order to take control of the system it you need one thing that is the brain interface you need an input output system that can communicate with the system so that you can start to manipulate it you can think of this as a keyboard or a usb that you plug into your computer another very important aspect is that just because you have an interface doesn't mean you can fix your computer just like that what you need to do you need to know in order to take control of our brain is to have algorithmic access to the brain function where you have to be able to understand the system architecture underneath it to be able to read and write codes algorithms associated with the brain and interestingly there was an interesting announcement regarding building improved io system recently by neuralink in 2020 august 28 where elon musk that all of you know has announced an input output device that is commercially made along with a surgical robot to implant this device and they've demonstrated that in a pig they can put these i o device and record from various neurons there has been many efforts to develop these input output device and that are being used in various labs but this was a first commercial attempt to make something that can be easily implanted into the brain but as i said oh and also the input output device currently exists in clinics where it's being used it's not as many channels as the one that elon musk showed but for example this is a clinical input output device that uses four different channels to pulse signals into the brain it's called the deep brain stimulator where these deep brain stimulators are currently clinically used where you implant the device into the brain and pulse current to restore our brain function and it's most prominently used in parkinson's disease and also in epilepsy where what you do is through this device you pulse currents to normalize your brain function however these things are currently done in an empirical fashion we're not putting in this i o device because we know exactly what to do it's kind of like having a tv and like hitting it or trying to shock it to see if it will start working it actually sometimes works and sometimes you can manage but without knowing exactly how your system works this can often be a very difficult task where many of the companies that pursue new uh electroceuticals this is what a lot of the time it's called pursue these electroceuticals often have to go through hundreds of millions of dollars trial which leads to minimal improvements or fails at different times and so ideally like i said io is not enough we need to be able to understand the algorithm to precisely engineer what we want to achieve and for the first time in 2017 my lab has been able to reconstruct a cell type specific whole brain circuit code algorithm where in this case the circuits called the d1 medium spinning neuron and d2 medium spinning neuron related circuits these are critical circuits for example in parkinson's disease and you can see that now we can reconstruct how these specific cell types interact with the whole brain in order to create the behaviors of interest and once we have a large-scale interaction map of how things are working together we could then also model things at a single cell level in the brain there's large-scale communication that is happening but also on a single neuronal level they are giving codes in zero one just like in our computers although the architecture is very significantly different and now that we have the overall scale of algorithm we can now also model things at a single neuronal level which allows us to understand what algorithm is executed in controlling our behavior also very importantly in our recent study we also saw the code our goal is to restore the code to normal but we also wanted to understand okay our goal is to restore the code but is there any relationship between the code and the hardware defects that you see in the brain because one of the biggest things that in neurodegenerative disease that people look at and try to do is to figure out why you have these different plaques these hardware defects that accumulate in the brain and how to clear it very interestingly what we have found is that the code change also changes the pathology once we restore the function to normal it allows you to also clear up some of the hardware defects that is happening and so with this it also allows us to understand that the software restoration is not only a goal but it also cleans the hardware and interestingly what i want to note is that we are now starting to decode the brain circuit algorithms which means that we can start to build an engineering platform for the brain so bio healthcare needs to move from treasure hunting to systems engineering approach in order to define the future of where this needs to go and interestingly when i was in korea last time very recently in july of 2020 sk biofarm ipo'd at a close to around 20 billion valuation and this is an amazing success that many celebrated but one thing that's not necessarily discussed is the fact that something like this took 20 years of treasure hunting to get this drug approved however once we have an engineering platform for the brain that can enable system engineering for discovering neurologic and psychiatric disease treatment we may no longer need to do this treasure hunting and move away from this model where the treasure hunting is not necessarily done even without a compass or maybe a hint of a map and so i'm going to show you a brief video of what we're able to do as an example you can take things like brain signals and find important events through ai technology that are clinically relevant and once you find the important events you can directly understand what the network activity underlying the disease is and once you can directly visualize the network mechanisms associated with it you can predict what sort of therapy will have what outcome and also design new therapy against what you are able to visualize directly and we are starting this developing these solutions for epilepsy but we are also expanding it to many other diseases at this point and our goal is to bring digital innovation to neural health everybody wants a quick solution you want a magic vial of drug that says okay here you're going to cure alzheimer's disease for example the world knowledge forum

You May Also Like