 All right, so this talk began with asking two questions. First, why have we failed so spectacularly at finding treatments in medicine for the most common diseases that we see? And the second question is why do we continue to repeat that mistake over and over and over despite that failure? I'll give you just another second to appreciate this. All right, so one of the reasons that I directed my, decided to direct my interest in the neurosciences into a career in neurology was because I thought that we were on the verge of some major breakthroughs in a lot of the most common diseases that we faced, as were many other people, so I was led to believe that as well. And that was a bit of a mistake. So if you look at this timeline here, these are the sort of major breakthroughs, pharmaceutical breakthroughs that have been made in some of the major categories of neurological disease and the, so the mark is essentially for the last significant breakthrough that was not perhaps clinically considered a breakthrough, but at least a stepwise change from what we'd had previously. So if we look to this, the last one would have been, was migraine, 1991, with the drug Sumitriptan, a drug taken for migraine relief. And it's debatable, even with that one, whether it's been a net benefit or done more harm than good. So if we were to throw that one out, we have to go all the way back to 1961 with Levodopa for Parkinson's disease. So to say that this lack of progress is surprising is an understatement. So there we have the therapeutic winter between 1961 and 2019 to illustrate a lack of progress. And notable things absent from here, things like Alzheimer's, multiple sclerosis, diabetic neuropathy, other common things that we neurologists see. And this isn't limited to neurology, so these are some of the major categories of disease in other areas. And if we look here, you know, we'll see 1970 Prozac for Depression being sort of the last landmark. That's not to say there aren't new drugs being developed, but nothing that's been a stepwise change from anything that preceded it. So Santa Claus has gotten in on the fun and the therapeutic winter. So the, as I said, one of the reasons that I decided to pursue a career in neurology was because I thought we were going to have some major breakthroughs within that period of time. And I still vividly remember as a senior in medical school asking one of the prominent Alzheimer's researchers at my institution when he thought we might have a cure for Alzheimer's disease and his estimation was 10 years. And that was the year 2000. And we have made zero progress on the pharmaceutical front. And some of you may have seen headlines like these that the pharmaceutical industry is giving up a search for an Alzheimer's cure because it's so futile and has spent billions of dollars on that project. So if we might have to revise our projection of when we might find that breakthrough from, if we consider that it was first reported in 1906, Alzheimer's disease by Louise Alzheimer's. And maybe you could argue that the current drugs are a minor progress. We've made a tiny little step forward, but maybe our realistic projection is in the next 2,000 years or so. So I think anyone taking a sober look at this has to ask themselves the question, what the flock is going on, which brings me to Angry Birds. So the parable of the Angry Birds is a thought experiment to help try to answer our first question. And so imagine that an iPhone lands on an alien planet and these aliens, they had never seen iPhones. They don't have any computers. They don't play any games where you launch birds at pigs, forts. But they are a curious and competitive species of aliens. So they're intrigued and they decide they're going to have a competition. And so they're going to break into two teams, meet back in a month and to see which team is better and crown one team the victor. So one team goes off and does what most of us would probably do, which is to play the game over and over again and try to get really good at it. And so that's their approach. The other team, they have a few scientists in their midst, so they start taking the game apart and they realize that the game itself is just an illusion and that underneath it, sort of specifying everything about the game is this programming language and underneath that is machine language and really it's all just transistors in their momentary state and all just at the root ones and zeros. So they decide that they're super excited, by the way. They think that they've cracked the code so that other teams can have no chance. And their goal is going to be to manipulate the source code as their approach to winning the game. So once again, team game level, they're going to figure out how the game works, learn the controls, learn the game mechanics, get really good at that part. Team source code is going to try to manipulate the source code in real time and win the game that way. So game day arrives and the team game level crushes it, right? They post a super high score and team source code's turn comes up and it's a massacre. They can't even get the game to run once without it crashing. And that's because even the person or people who coded the game could not do such a thing. There's no computer scientist alive who could look at the machine language of any piece of software and have any idea what it did. So they are doomed to fail. So we are here because we think that an evolutionary perspective is essential to understanding the foundations of help and we also believe that mismatches between our ancestral and modern environment drive the pathogenesis of chronic diseases. And what this parable hopefully illustrates is that our failure to translate immense progress in science and technology into any therapeutic progress comes from the absence of an evolutionary perspective in our approach to therapeutics. So modern medicine has almost exclusively taken team source code's approach to finding new therapies. And another way of thinking about source code interventions is that they are evolutionarily novel and so mismatched as well to our physiology and so by adopting evolutionarily novel types of treatments we're essentially introducing all the same hazards and risks that led to the conditions that we're trying to treat in the first place and I would argue that that approach of favoring source code interventions over game level interventions explains our therapeutic winter. On the other hand, the ancestral health therapeutic paradigm is to take team game levels approach. So this is how we would integrate an evolutionary perspective into how we evaluate and develop therapies. So unlike source code interventions, game level interventions are evolutionarily familiar and so matched to our evolutionary history. And so I think if we can fully embrace this paradigm then we have the opportunity to leverage all of the advances that we have made in science and technology so that medical therapeutics can catch up with all of the other areas of human progress. And so because it's, I think, a very useful frame for understanding what an ancestral health perspective on therapy looks like. Let's explore this distinction between game level and source code interventions a little bit more. So as I mentioned, one way to think about a game level intervention is that it's evolutionarily familiar. So these are the evolutionary forces which we've been adapting to throughout the history of our species and beyond and really some total of actions that we might take in the game of life whereas source code interventions are, by their nature, evolutionarily novel. So from a safety perspective, game level interventions are safer because they engage evolved regulatory mechanisms whereas source code interventions are going to be inherently riskier because they bypass those evolved regulatory mechanisms and so are much more likely to crash the system. Game level interventions are also more powerful because they act upstream in the physiologic cascade so impacting many pathways whereas source code interventions are inherently weaker because they act downstream and with a narrower scope of influence. Game level interventions are also better suited towards intervening in a complex adaptive system where the outcome of a source code intervention can't be predicted. And the source code interventions are most appropriate for complicated systems or scenarios where we know a condition is caused by a single factor and where we need a targeted intervention and it's no coincidence that pretty much all of the most significant breakthroughs in medicine come from that sort of condition where we have a single target like meningococcus causing meningitis for example where all the major breakthroughs have been made. So just to review, game level interventions are evolutionarily familiar. They engage evolved regulatory mechanisms and so are inherently safer. They have an upstream locus of action so they're stronger, not stronger and they're best suited for intervening in a complex of adaptive systems whereas source code interventions are evolutionarily novel. They bypass evolved regulatory mechanisms and so are inherently riskier. They have a downstream locus of action and as such are inherently weaker and are best suited for intervening in complicated systems where the impact of the intervention can be predicted. So some of the implications that fall out from this. Randomized controlled trials are the appropriate way to acquire knowledge about source code interventions. So where you're going to be manipulating a single variable and you have a static intervention that's not going to change. Now it's reasonable to think given sort of the complexity of biological systems if we should ever expect a breakthrough when that is our primary therapeutic paradigm impacting a single variable with a static intervention. So you could argue that we've sabotaged ourselves from the outset by requiring that as the only means of validating new therapeutic knowledge and by the same token randomized controlled trials are the wrong way to acquire knowledge about game level interventions where you're going to be manipulating multiple variables and that manipulation is going to be dynamic. So if you think about a baseball player trying to learn how to throw a curve ball it would be a bit ludicrous to consider manipulating one variable at a time especially considering that you change one thing that's almost certain that you're going to need to change another to accommodate it. So it would be needlessly tedious and time-consuming and unlikely to ever arrive at the optimal solution. But one of the most common objections that we would hear to those of us who advocate an ancestral health paradigm or ancestral approach to therapies is that we need more trials and I would argue that it's not, and by that they mean more randomized controlled clinical trials. I would argue that's not what we need. We do need research but we need to develop an entirely new research ecosystem that's designed towards developing game level knowledge. So this is how we acquire game level knowledge. We play the game, we make some action, we assess what happens, we get feedback, we refine our approach and then we try again. So it is an iterative trial and error process and most of the knowledge that we all have in our heads was acquired in this manner including a lot of the knowledge that we know about human health. So we have no randomized controlled trial telling us that humans need water to stay alive or that they need to breathe or that decapitation leads to a permanent loss of consciousness. Yet we all consider these things as incontrovertible truths and if someone would argue that all we have is anecdotal evidence we would think that's absurd. Another important implication here is that mechanistic understanding isn't required to win the game. So again, if I'm a baseball player trying to learn to throw a curveball I don't need to know anything about the underlying physics to do that well. I only need to know about what I need to do at the game level and furthermore, somewhat counter-intuitively even knowing the underlying physics doesn't typically help me throw that curveball any better. By the same token we knew that air and water were essential to life and decapitation was not long before we knew the mechanisms involved at why we need water or why we need to breathe long before we knew about oxygen and the electron transport chain. And yet in spite of that if we can't provide mechanistic explanations for something that we do it's considered undermining that intervention and another common objection to someone who would advocate for a game level intervention. And of course that particular objection comes from the paradigm itself which says that we should be looking at the mechanisms first to guide our pursuit of a therapeutic intervention which is the very paradigm that has led to our therapeutic winter. So we have kind of two forms of knowledge here that we can develop. One being mechanistic understanding and one being simply research at how to play the game. And these are both valuable forms of knowledge. They just have different ways of application. And if we want to get better at helping our patients developing new and transformative therapies and we should be redistributing this allocation of resources significantly imagine if we had been pouring the same amount of effort into developing research on how to play the game as we have in advancing mechanistic knowledge. So let's now revisit the two questions that I started with. The first was why have we failed so spectacularly in finding new treatments? And the second being why have we continued to repeat that same mistake or that same approach in spite of overwhelming evidence that is not working? And I think it's important or I think one of the potential explanations here one explanation is that we all find reductionism inherently seductive. So because it's a product of science and reason and because it reveals to us levels of explanation that are initially hidden it feels as if we've uncovered some deeper truth and it feels like we have found the man behind the curtain and sort of everything else is just an illusion. So and I think we're all biased towards thinking that's a more privileged form of knowledge and hopefully I've illustrated that that's not the case that the different kinds of knowledge have different domains of application but I think we're all again biased in this direction and into thinking that reductionist knowledge is inherently superior and again I'm not complaining not claiming that it doesn't have a role even in therapeutics but I would say that by not factoring this into account we're missing out on tremendous opportunity to usher in a revolution in therapeutics and because we are all prone to this bias or most of us are prone to this bias I know I am it's helpful to have safeguards in place to kind of keep us from being lured by the seduction of reductionism and so I think it's helpful to kind of formalize what an ancestral approach to therapeutics looks like what incorporating an evolutionary perspective looks like so that we can also have a common language for talking about it and a compass to help orient ourselves and keep us focused on what we consider to be most important so to that end I'll now present a four quadrant model that kind of formalizes how I think about an ancestral approach to interventions and my own personal way of safeguarding myself against the reductionist trap so on the vertical axis we have the level of the intervention so again we have game level interventions and we have source code interventions so game level interventions once again are ones that act at the level of evolutionary forces kind of the top most level of biology whereas source code interventions generally act at the level of cells so things that impact enzymes, neurotransmitters, receptors and so on and then on the horizontal axis we would have things that are supportive meaning we're trying to support whatever the body is trying to do at any given moment and generally speaking from an evolutionary standpoint what we're trying to do there typically is to minimize mismatch between our ancestral and modern environment and then the other category in terms of the goal of our intervention would be things that are disruptive or exploitative so where we are essentially taking the physiologic status quo and overriding it in some way redirecting it because we think that doing so has some sort of benefit so why would you do such a thing probably the main case would be when you have a regulatory system that you think is broken beyond repair so an example being giving insulin to someone with type 1 diabetes so they can't manufacture it anymore that's their physiologic status quo so you administer it because the system is broken exploitative interventions would be where we are taking our understanding of physiology and doing something that exploits it in some way so for example we can take our understanding of what happens to the body in temperature extremes whether it's heat or cold and use that to improve health in some way or that would be a game level intervention whereas a source code might be to give a vaccine so that we can help our immune system fight off a pathogen in the future so again exploiting some knowledge of biology and you see those so here are just a few more examples of what those would be it's not meant to be an exhaustive list but just to kind of help illustrate what I'm talking about here so all the things in that top left category are the things that we talk about most at these times of conferences and then some of the exploitative type things like I mentioned we'd have heat and cold exposures breathing methods fasting, high intensity training, mindfulness, psychotherapy again all things that take our understanding of what we know about biology and are intervening at the game level to confer some benefit and then source code interventions that would be supportive mainly the main thing I think about here are where we think that there's a nutrient or that's insufficient and the body needs more of to do its jobs so we're trying to support its ability to do what it needs to do and then source code interventions that are disruptive would essentially be all pharmaceuticals things like no tropics and then hyper supplementation so where you're supplementing with something past the point of correcting a deficiency and so the decision algorithm from an ancestral point of view would be to prioritize category one which would essentially be the foundations of health over intervention two and then three and four unless you had to have evidence that would indicate otherwise that you should prioritize one over another so which there that can exist so another reason why I think that we've continued to repeat this same mistake over and over again is that we need solutions that scale and I mean not diet but drugs and supplements scale really well and another problem is that the way we currently or our current regulatory system for evaluating new drugs allows companies to make or to create a blockbuster drug even if it's no better than its predecessors and that's been the prevailing strategy for the last few decades I think we're starting to squeeze the last drops out of that particular strategy but it has prevented market forces from forcing their hand into looking for something that's truly a breakthrough truly an improvement upon what currently exists and I feel quite certain that there are people who are in the drug research industry who've recognized this and know that we that true breakthroughs will require a paradigm shift but realize that the only route to profitability is to create these me too drugs which is all I've seen in my own career which is almost 20 years so that's to say that to combat this issue we do need to have credible models for developing new therapies that can be better scalable and I think that's possible so we'll revisit the model for how we acquire game level knowledge so I think it's possible to create a model that can scale and that leverages advances in science and technology at every step along the way so at the level of gameplay most of the things or a lot of the things that we would be thinking about require behavior change so there's an entire evolving science of behavior change that can be applied there along with technological tools that are becoming increasingly sophisticated for helping us to do so and then we have other types of technologies that we would use perhaps to mitigate the impact of being indoors all day so lighting solutions, architectural solutions you have sleep technologies from wearing sleep masks to chili pads to audio to entrain certain EEG rhythms so we have all manner of potential technological solutions that can help solve this problem and help to provide game level interventions and then at the assessment piece here we're talking about things where we're trying to figure out how we're doing with whatever action we've taken in the game and this is where source code knowledge is best applied so disease markers can help us get an idea of how well we're doing so labs imaging, biometrics subjective data like surveys or just talking to people digital phenotyping all of these things can be brought into the assessment phase and then with that data we can apply all of the emerging tools of data science including machine learning to then figure out how to what variables we should manipulate again and take our next action in the game and continue so creating a virtuous cycle that's self-amplifying and where we're able to leverage the tools of science technology at every step so just to give a couple of examples to help make this more concrete so one example would be the ketogenic diet so nutritional therapies are a game level intervention and at that level obviously you can give your change as a big part of what you're trying to accomplish then you have the assessments that you'd make you can have process based assessments so how well someone is adhering to the intervention which can be all manner of things including measuring things like whether they're in ketosis or not and then you have primary metrics so the very thing that you're trying to treat if you're trying to treat diabetes for example then you'd be measuring hemoglobin A1C or fasting glucose as your primary metrics or of potential secondary metrics that you might want to care about such as impact on blood pressure, body fat CRP and so on and then in the refinement phase there are all manner of variables that you can manipulate the protocol itself the diet the nature and frequency of their support and so on and another example so there are all sorts of variables within sleep that you can manipulate one being trying to improve slow wave sleep which the research would suggest can have potentially broad ranging effects and could be a force multiplier on many other things in medicine and so there's all sorts of possible technologies we might develop for enhancing slow wave sleep including things already in development but here you would whatever your intervention is you then assess based on metrics like the time in slow wave sleep and we already have tools for doing that and tools that will likely improve in terms of their fidelity and then you have secondary sleep metrics so anything you might want to monitor that we believe might be connected to sleep or things that we don't even know yet are and then of course taking that data making a manipulation to the variable and then starting to loop over again and of course there are multiple domains within just sleep that you might care about and then the next frontier would be for doing this at scale seeing how all these different areas interact so using the tools of systems analysis to figure out what things we're doing at the momentary what things might be at odds what things might be multiplicative or subtractive it also may help us to resolve some paradoxical observations so things like we know that there's pretty robust data that people who drink an alcoholic beverage once or twice a day tend to live longer yet we know that also has sleep and we know that sleep quality is also associated with longevity so how do we reconcile that we currently don't have any real mechanisms for understanding the contextual dependencies of all of our interventions so how does doing something in one context differ from doing it into another which is incredibly important information so just to summarize our myopic focus on evolutionary novel source code interventions explains our inability to translate advances in knowledge and technology into better treatments so when you prioritize source code interventions over game level interventions you have a spectacular failure that you repeat over and over again and number two that an ancestral health paradigm for chronic disease prioritizes evolutionarily familiar with game level interventions and since we are all prone to the lure of reductionism it's useful to have safeguards and a way of kind of formalizing our approach so that was the goal of the four quadrant model proposed third the advancing knowledge about game level interventions requires different research tools and methods than source code interventions randomized controlled trials are not the appropriate tool for that we need an entirely new research ecosystem for learning how to play the game and doing so if we create that ecosystem we can finally bring about the revolution in therapeutics that I was promised 20 years ago and we can in my cheesy success graphic we can have exponential growth and perhaps most importantly the last point is this amazingly doesn't get old or can I no I wish we have like two or three minutes for questions if anybody has pull them up here Josh you spoke I don't know when it was but on migraines years ago and I remembered 14 I remember that presentation really well you're a great speaker and this is a phenomenal presentation and concept and I think everybody needs to hear it so couple comments I did work with Dean Ornish in 2000 interesting thing we did was look at a variety of different complementary mechanisms to see if we could get greater outcomes than the individual parts alone what you lose by doing that is certainty about what worked if it did but you could engineer better outcomes overall and I think if one thing that we might do with lifestyle based stuff is instead of trying to have certainty engineering for certainty gamify okay who can get the biggest effect and then that all create a different type of thinking into we'll go in here you mentioned the funny I was writing down silver bullet and I looked up and it said silver bullet on the screen I think one of the reasons why we're looking to it's like if we can find the silver bullet we can manipulate it and keep living the way that we do we're subconsciously searching for license to just keep on doing what we're doing and then shoot I had one more thought shoot I'll forget but I'll mention it to you afterwards but yeah thanks again it was great yeah and I think one of the first point you're saying that once we dissociate this we're trying to acquire mechanistic knowledge from trying to understand how to play the game you can pursue all these other types of research we're currently constrained by feeling like we have to prove we have to link in the mechanism every time we do research and so you can't it limits us in many ways including what you talked about I remember my point challenging your idea play assess refine why not just play yeah is the assess and refine simply behavioral if we get some validation then we can get people to commit more to the play so there are certain save for a diabetic or some of the hypertension you make a game level intervention you have to figure you have to have some assessment to understand whether the outcome if it's opaque to the patient or the client or whatever you still need measures for assessment sometimes the feedback is right there and you can just play the game and so much of what we do so much of the day in and day out we have our own feedback systems but there are times when our own endogenous feedback systems aren't enough that's where we can deploy all of our source code knowledge to get more granular detail about how well we're playing the game it's just interesting to me that source code knowledge at that point might simply be behavioral it's the game base of it I mean sometimes yes but oftentimes it's just getting people to hear more to circadian rhythm only if you just play the game we don't even need putting that in there I think there's value in that but understanding that you can create a scalable concept through a kind of virtuous cycle awesome thanks and that's time thank you so much sure