By leveraging your personal health data with Artificial Intelligence, from blood samples to psychological surveys, it is possible to make accurate predictions on biological age and interventions to keep you at your best.
Fedor Galkin and Deepankar Nayak of DeepLongevity, a health data platform that looks to transform approaches to ageing and longevity, discuss the upcoming launch of their mental health support site ‘FuturSelf.ai‘, models of psychological wellbeing and the importance of data throughout healthcare systems and patients lives.
Image Credit: Lightspring/Shutterstock
Transcript: The following transcript is automatically generated.
Will Mountford: Hello, I’m Will. Welcome to research pod.
Testing your personal health from blood samples to in depth DNA profiles, even to psychological surveys, can determine how old you really are inside. And with that answer you can create a personalised wellness plan by leveraging health data with artificial intelligence, it is possible to make accurate predictions on biological age, and interventions to keep you at your best.
Today we’re speaking with Fedor Galkin and Deepankar Nayak of Deep Longevity – a health data platform that looks to transform approaches to ageing and longevity – about the upcoming launch of their mental health support site- Futurself.ai, their models of psychological wellbeing and the importance of data throughout healthcare systems and patients.
Hello Fedor, good afternoon, good morning, whichever the time zone is appropriate.
Fedor Galkin: Hello Will, yes pleased to meet you on this podcast. I’m really excited to talk to you about ageing and wellbeing and how these concepts are connected to each other.
Will Mountford: And Deepankar, thank you very much for joining us as well.
Deepankar Nayak: Well, uh, I’m, I’m completely excited with, with what we are building here and looking forward to this conversation with you.
Will Mountford: Fedor, if you could tell us, maybe a little bit about yourself – what has led you down the road of ageing and wellness and what brings us all here today.
Fedor Galkin: Yes, sure. Well. My name is Fedor Galkin. People call me Tau. I graduated from the Luminous State University with a major in bio informatics, seen 2018 and right after that I joined Insilico Medicine and back there was a department doing research on biomarkers, Beijing, I was part of that department and in 2020 these department kind of split into a spinoff company called Deep Longevity.
Ageing is one of the greatest problems we are facing in the world on a global scale, so I was just interested in solving something big, contributing to solving one of the oldest issues the humankind has faced since the very beginning. Not over species, but even of probably life, because as soon as there was life, there appeared ageing. In today’s world, more and more people are, are healthy enough to grow old, but even despite all the recent advances in medicine, in therapeutics and all our knowledge, there is still no cure to ageing. So, this is what drives me in my research and this is why I joined disciplinary tendons to help in the fight against ageing.
This is a shared interest and, and a shared passion. I have spent all my career helping life sciences companies adopt technologies. That is where I have realised the power and the benefit that enterprise technology can bring, and it can really help patients. It can help the various stakeholders who are working so hard to bring new cures to society. So for me this was a natural extension of the work that I have been doing with my pharma customers in the past to join such an innovative company and then take this science not only to life sciences customers, but enterprises across the board who can benefit from our ageing. There is a clear distinction which is emerging which is ageing needs to be considered as a disease right rather than as a natural consequence, and that is the, the signs that we are adding, right?
And our biomarkers are a fantastic diagnostic tool which is going to help all the research that’s happening in this domain.
Will Mountford: In terms of what we’ve mentioned so far about time as the enemy and ageing as a disease to try and treat, to manage to maybe even overcome kind of ties into a a movement, kind of a global awareness of wellness in oneself and how that wellness is a part of enrichment of the self.
How do those overarching concepts either influence the concepts of ageing that we’re dealing with specifically in your work, or fit into this movement of being as well as you can for as long as we can?
Deepankar Nayak: We know that we should try and get our eight hours of sleep and we should drink lots of water and we should relax and we should try and keep stress away and, and get the daily bit of exercise.
We know that they are beneficial. What becomes hard is to do that in a consistent manner and then to know that this is this is benefiting you, and that is where we come in. So we are looking and. Let’s, let’s talk about mental health as an example, right? So we are looking to create an ecosystem where we can go to employers, we can go to large companies, small companies and say we can create an ecosystem with you where we will have your employees engage with us through our platform, do certain tasks, do certain activities that would be recommended for each employee in a very personalised way. And if they do that, then we’ll be able to actually track how their mental wellbeing is improving. Right, so we are going to be the, the diagnostic we are going to be the marker which will make it very evident that there is progress and that when you know becomes an encouragement for you to actually keep doing and keep staying on that path of wellness.
Will Mountford: To kind of look at the evidence of ageing, the biomarker, quite you know, exact cellular definition of an ageing a new refreshed cell – hen are you old?
Fedor Galkin: Ageing is, so to speak, a blanket term, and it is always defined alongside some specific research topic. There are people who can make an argument that ageing starts as soon as you are born. There is no clearly defined border between someone who is young and who is old – it’s a constantly ongoing process and, the lines that we try to draw in between the old and the young – they are pretty conditional.
But as far as our research goes, we use this concept of biological age as contrasted to chronological age, right, chronological age is pretty simple. It’s your date of birth, right, and the current date, and this is your chronological age – the difference between the two. But biological age is somewhat more complex. Now we know a number of processes, molecular processes, that constitute ageing. They are called hallmarks of ageing, they include such stuff as genomic instability, mitochondrial dysfunction, damage accumulation, till attrition, all these processes are constantly going on in your body as you grow older, but in some people, these processes are more intense than in other people.
In this case, we can tell that some people have higher biological age than others. If this is, these processes are too intense, a person is faster that, or also we can say that their pace of ageing is increased. And in this case, whether you’re old or young, it stops being some kind of an absolute quality of a person, it’s more of a relative thing. Are you too old, too biologically old, for your current chronological age, or are you actually younger than your chronological age from a biological viewpoint.
Deepankar Nayak: We know for sure that the human lifespan is, is increasing, right? People are living longer, it is statistically all over the world. It is, it is a phenomena that is that is well understood and recognized and societies are living longer, the the focus is that, not only do we keep enabling that, but also we help people live healthier, right, and to your question therefore, and it can happen, sort of in three ways.
So the first goal therefore is that can we have the right interventions which will slow down the process of ageing so you age slower, right? And the next stage after that would be that, can we even look to their full stop it, right, you stop the clock, right, if you do the right things, your cells, the way they are degenerating, and then you do the right steps and they start to regenerate and therefore they are balancing themselves out, and that’s why the, the leaders and the, and the scientists who are now coming into this field, and they’re all very excited and this broad goal is: human beings should be living till an age of 120 and it is achievable, right? You do the right things, it is absolutely achievable.
Then I would say the, the final goal would be, once we look to achieve that, is can we even look to reverse it, right, can we think that, well, in my next birthday, the best gift I can give to myself is I do all the right things in this year and I actually gain a year uh, when I’m supposedly getting a year older chronologically, and that is the ultimate goal which this industry is going to.
Fedor Galkin: Just to wrap up, I would like to add that in our research we do not treat age, old age, young age as some absolute quality. We try to define what a person is right now, are they too biologically old for their current chronological age, or are they actually ageing really, really slow? In this case, it is a creating to be a slow ageing and we are trying to find ways that will help, help people get from their increased pace of ageing to a reduced pace of ageing and in these, being young means being biologically younger than you, your chronological age is right now. And as a rule of thumb, we usually measure it as a number of years. If a person is biologically at least five years younger than they really are, it’s a great thing, they receive a lot of health benefits, they have reduced mortality rate, they have lower probability of getting a non-transmissible disease, or they have lower susceptibility to infectious diseases as well. it’s just a great thing to be, uh, biologically younger. But sometimes, it is still an open debate whether you can actually turn back time and make somebody look or and be biologically a 20 year old person when they are actually 60. Lots of people are trying to find out the solutions for these rejuvenation techniques in humans. For example, elder Law Labs, backed by Jeff Bezos are, they’re trying to implement epigenetic reprogramming for these purposes and there are also various scenes that try to work out gene therapy for similar purposes. But our approach is more about everyday life and just regular people. We want to find the ways that can help everybody with fairly low investment barrier to find a way to slow down the old age and to prevent the damage that these ageing process may do to their body before it happens.
Will Mountford: And all of this leads onto kind of the other very modern concept, besides being young for as long as you can, the idea of personalised medicine. The idea that one’s entire lifespan can be spent adhering to your own personal biome, and there’s lots of different ways that you can gather lots of different information over a very long period of time. So that’s going to be generating a lot of data, and I mean now we can start getting into some of the hard technology behind what it takes to manage that amount of data, and how artificial intelligence or machine learning processes are carrying that weight.
Fedor Galkin: It sure is the age of data. The 21st century is started with the, the march of Internet and uh over the of the globe now Internet is only present and in in the twenty tens, the what are they called the teens right? In the teens, there was a huge boom of AI technology and machine learning, Everybody started trying out data-driven approaches, and trying to manage these huge volume of data that Is, is constantly generated and now there are even more ways to generate data, even if we just limit ourselves to biologically relevant data or health relevant data. Everybody nowadays is wearing these trackers – Apple Watch and other similar devices, are to keep track of their heartbeat of their activity levels. Most recently, there appeared constant sugar monitors, right, before that, diabetic people only could measure their blood sugar from time to time, but nowadays you can just attach a small clip to your body and it will constantly monitor your blood sugar levels. It’s really impressive technology and even more exotic data types are getting available and more and more so with the creation of huge, you know, biobanks. Many of them offer, they their collections publicly and the methods of AI and machine learning are absolutely necessary to process all this information and find the signals which can be used to create something useful, In our case, this is an ageing related signal, right, we want to find the signatures of ageing in various data types we use some exotic data types such as you mentioned, all make data. Only data types such as transcriptomes, microbials, even epigenetic data, and we also use much simpler data types such as data from your wearable trackers, even just psychological surveys.
But our most popular model of ageing involves data generated from clinical blood tests.
Blood tests are everywhere in every clinic. It’s a really neat data type because there is no data bias from samples produced in different clinics and there are lots and lots of blood samples everywhere across the globe. Most importantly, it is a very biologically relevant data since blood is its issue which flows throughout the whole body and it gathers signals from every organ. If you have a damaged liver, there will be certain biomarkers elevated in your bloodstream. If you have some other type of damage, this will also can also be detected in blood, and this is this property has been used in medicine for ages. We are trying to enhance these insights physicians and doctors can get from reading your blood samples to the max with the power of AI, we are at the point where we can definitely tell, so what is going wrong with you in terms of your ageing processes, in your body, from your blood samples. We can also, quite importantly, tell what you should be doing differently in your life to become biologically younger. Now we can measure the effect of, for example, running in the morning every day, how many years will that detract from your biological age? Or if you want to experiment with a new diet, we can tell right away, without you actually trying out this new diet, whether you will become younger, or if it is not the diet for you, whether you will become older from following this diet. It’s really interesting stuff and it’s, it’s impossible to gather the process this amount of data and reach similar insights just by human review. You definitely need the power of contemporary machine learning techniques to process it and find the signatures weijing and test your hypothesis about how to remove these signatures and help a person become young again.
Will Mountford: And all of this is kind of covered as a, a case study optimising future wellbeing with artificial intelligence self-organising maps. If we could look at that paper a little bit more closely as kind of putting some of this to the test, putting some of these models in motion, what is a self-organising map and what do we mean by emotional stability in the first place?
Fedor Galkin: Yes, I I started talking about blood and how we can detect ageing with blood samples and psychological ageing is a topic we started actively researching just a couple of years ago. Currently there are, there is a load of publications about how attitudes and your lifestyle, how they can affect your physical health. And we can clearly see that your happiness, especially in the elderly, it is directly correlated to different metrics of physical wellbeing. And people who are happier, who have a better network of support or more independent, or have purpose in life, these people suffer less from various diseases such as heart, heart arrest, such as stroke or cancer, and similar findings have been shown in meta-analysis of studies in which both these data dimensions are available, the psychological dimension and the medical history.
Deepankar Nayak: So we know data is available and we know that it is available in in volumes in terms of every person, so our approach from the very beginning is about personalisation. We know what works for one, it’s known in the field of medicine as well doesn’t work for everyone. A medicine is limited because it is very expensive, if you figure out how to make one pill for every person in the plan, right, but and and and and mental health is a, is a perfect example, right? But here I think we have the advantage that using technology we can really bring that personalisation in a very democratic way without breaking the bank.
And a simple example would be that, well, we would like, you’d like everyone to kind of get their heartbeat up really high three times a week, and how you do it is completely up to you, but if I give a personal example, I’m not a big fan of running – I, I like long walks, but I love playing badminton and if I’m playing badminton, then that is very high cardio that I am doing twice a week and that works for me.
And that’s that’s what we would like to do with with our data, right, we would like to understand what are the the preferences of a person, we’d like to give very personalised pathways. And then Fedor can explain how that works through our songs.
But this is also a cycle of feedback where we, where we understand that if these tips and recommendations are not being followed and if it is not something which is which a particular person is receptive to, we can look to have a different set of combination and different combination of actions that we would like that person to do in order to be on a higher plane of wellness. So, one thing which is there in our in our vision and we are sort of doing a lot of research and work towards that already, is to understand that the needs of different segments and different parts of the population is different.
The anxiety that you’d be feeling working from home, would be different to an anxiety which a teenager is having, to an anxiety which an expectant mother is having, to an anxiety an elderly person would be having, one with a chronic patient is going to have and these would need specific interventions these would need different models right for us to help these segments and and these patients and people who can benefit from our technology, so that’s clearly something that we have on our minds.
Will Mountford: And to come back to this paper as a specific example, how does one begin to engage with the breadth of emotional wellbeing and stability and then code those surveys those data points into a numerical model to track who is doing well and how that is going to change?
Fedor Galkin: Just as Deepankar pointed out at the very core of our approach to mental illness personalisation, we want to make sure that the recommendations we provide to people are relevant to them and they understand that it is something that they can carry out. And something that we can help them in particular rather than just a general truth about stand things that can help you get out of depression or get you over a tough day. We want the recommendations that speak to people directly, and to do that, we applied machine learning more particularly self-organising maps which is a method of unsupervised learning. Unsupervised learning means that you do not need to know or what group people belonging belong to to start defining the psycho types among them. In our particular case, we used publicly available data from medals. This is a really well structured questionnaire, colts meet life in the United States. It contains all sorts of questions, some of them are about your psychological attitudes, but also many of them are about your physical health, about your life situation, it’s a great data set to kind of, find connections between different modalities of, of the human life. And among these thousands and thousands of questions, we first of all selected only the questions that correlated with one’s future wellbeing. Since it is a longitudinal data set, all people in these data set that we used have two time points: one was in the first wave in the 1990s and the one was a follow-up wave in 2012, four and 2000 and this allowed us to find the properties of the variables, features that correlate not just with your current wellbeing, but with your future wellbeing, and so we used this future wellbeing as a target.
We wanted to improve people, so to speak, we, the task of defining your current wellbeing is quite trivial since you can just ask a person a couple of questions and then there you are, wou know what their current will be, but, and we wanted to make sure that not only your current wellbeing is prioritized, but your happiness potential is prioritized – how long you can stay happy. And we want to make so that people are not just happy in the moment, and in this process sacrifice quite important aspects of their life, but stay happier for longer.
OK, let’s get back to the technical side once we employed the methods of feature selection such as boruta, it’s a random forest feature selection method, we arrived at a questionnaire which consists of 30 something questions which could be used to quite accurately determine what your future wellbeing in ten years will be. This is not as trivial as measuring your current wellbeing. It’s kind of a glimpse into the future, which lets us assess whether a person will be happy in a decade.
To help people with recommendations, we created recommendation engine based on self-organising maps, which as I said is a method of unsupervised learning and basically what I saw is it is a two dimensional map of all the different site types. If you imagine a map, it has South and North West and east, and it is divided into small cells by latitude lines, and in our case in a real map, these cells contain mountains and seas or some other geographical features, but in our case each cell on this map contains a group of people. And we can define these groups of people by their current and future wealthy, and just as a map can be used to find your place on a geographical location, a self-organising map we created as home can be used to find your place among all the different psycho types, where you’re currently standing and based on your position and the information we have about the people who kind of form this location, this psychological location, this psycho type your map too. We can tell what kind of problems you are probably encountering in your life.
Deepankar Nayak: So let me see if this example helps. We are looking at let’s say a person and the person you know appears to be sad. Now we know from the study and we just we we have data of 10s of thousands of people that there are there are hundreds if not thousands of people who are sad in different ways, and the objective of this platform and the engagement that we are offering is is not to transform people into what they are not right. So a sad person is not going to become happy by by just doing what algorithms are suggesting, but the idea is that we all can become a better version of ourselves. We can all become slightly better and slightly better and therefore be on a path which which brings the best out of what we have to offer. And in this example, what this would mean is by having, having the person interact with us and answer some questions for us, we would sort of know in that sea of sadness or lack of a better phrase where that person can be right, in which sort of island can that person be right, is it like the absolutely dark and despair, or is it, well it is slightly off, right, and then where does the path go from there, right? So people on adjacent islands can have overlapping paths. And people on on one part of the island would be encouraged to slightly move towards a better, higher plane, right, so when Fedor says that this is a two dimensional map, it is a feeling it is an emotion, right? How do you take it, how do you hold it? And you put it onto a coordinate. It is, it is very hard.
But that’s actually what the artificial intelligence is trying to do it is trying to break it down into a set of coordinates and is putting you on a point on the map, which we think looking at the rest of the population is sort of where you belong, like like a club, right, and and from there how do we, together, right as club members do better things to do, fun things to do, exciting things to do, to become better version of ourselves.
Will Mountford: Yeah, I think it’s worth mentioning some of the maps that are kind of featured as figures in the paper. They look quite like topographical maps, Ordnance Survey maps where you have, I mean, even perhaps like a weather, a weather front. We have areas of high pressure or low pressure in different colors and the boundary lines between I guess that’s where those interventions come in.
Fedor Galkin: Yes, there probably is an important property of self-connecting maps I might have overlooked earlier.
Well, the beauty of this approach is that while human psychology is multifaceted and probably cannot be described by just two variables right in our case it is still something variables that we have determined are really important for your future long term wellbeing self cognised maps, they offer a way to visualise and navigate these multidimensional 30 dimensional space of human psychology as if it was just a 2 dimensional object, just like a graphical map is set.
When we met a person find their location in this map as if it is a topographical map, we can tell whether they are in a pitch like right in this analogy. Let’s say, that if it is a bad place to be in our research article, we have marked it as locations in to, towards which depressive people gravitate. They have high density of people with depression and overall low psychological wellbeing, both current and future. So, if you happen to be in such a pit, or maybe not such a deep pit, maybe into something that, well, that is slightly better than a full blown depression. You can clearly see on this map allocation we add or switch people with high future wellbeing gravitate, right?
And since you’re on the map, you can also draw an itinerary that can bring you from where you currently are to where you want to be. Our approach, the recommendation system that we have devised can find the shortest path from your current point to point big to finish line where you start feeling great every day and your current and future wellbeing are significantly improved the conventional approach that uses quite general recommendations, these approaches, they ignore your own personality. They try to draw a straight line across this map that we have devised, but data is not usually the best case to approach mental health problems and improve mental resilience, because if you for example stand in a pit in this map and you draw a straight line, then at some point, this straight line might actually be going through an even deeper pit or this straight line might try to, might suggest that moving across these line you need to introduce changes that you are, under no circumstances would like to introduce the change, right, For example, one parameter of wellbeing that we studied, it’s called positive relations with others, right, for example, how well you interact with your relatives, and for many people, families that are really important. In some cases, taking a person to a region of high future wellbeing requires them to sacrifice their personal relationships with others. For some people, that is acceptable. For many others, it is not acceptable under any circumstances, and using now both zoom and the method to draw itineraries on it, we can imagine a roundabout way which might take longer for you to get there, but it does not involve you making such such heavy sacrifices right, and this is, I believe, why the service we are trying to build is something special. It respects who you are, what you are willing to do with your life, what your priorities are. And respecting these priorities, it tries to find the best way for you to improve yourself.
Will Mountford: To think then how that comes back to ageing, looking forward to wellbeing ten years in the future. There’s a lot of stuff that can happen over those ten years. And how are we kind of centering the emotional valuation within these surveys as the heart of that ageing you know, ten years on, can we figure out how these people are doing?
Fedor Galkin: Well, definitely if you take a psychological test which maps you to a certain location on the Somme is just a snapshot. Definitely repeated measurements are required to visualise your progress, but it’s not prohibited to take the same psychological test, once in a while, you can take the test or see where you currently are, then receive a number of recommendations and see how you have improved and maps, in this case, they offer a great visual way to observe your progress since you always have your initial point A, and your point B where you would like to end up eventually. You can see your psychological journey. Surely the change is all internal, but with this map you can make it slightly more physical. You can see whether you’re, for example about 10% done towards your final goal or you’re halfway there already and this is something that is important for someone who has just started their path towards self-improvement, right?
The first steps are the most difficult, especially because you do not see the progress right away, it takes until you can see the benefit of changing your ways and sometimes you just want to give up, but this map is something that you can use to cheer yourself up, even if you do not see that the positive change is entering your life quite as fast as you originally was hoping, you can still rest assured that the progress is being made, and that even if you just follow another day with your new adopted lifestyle on the path towards self-improvement. But then you can hope that you have become slightly better once again, and you can put it into concrete specific quantified terms that your mental resilience has become 5% better than it was, 5% is hard to notice, but our approach can notice this kind of progress and reassure you that you’re on the right way, and that even if things look tough for you right now due to various circumstances, maybe this just sends some rough patch you’re going through in your life, even if things still seem rough, you can see that, well, your resilience is slowly improving. It’s like building muscle, for example, these quantified approach, which both keeps you curious about how far you can go, what your gains will be in the next month, and it also keeps you motivated by creating a physical evidence of your progress, and we want to kind of recreate the same experience for people who are taking the self-improvement part right, it’s the same body mass measurements as in as in a gym, but for your mind and for your psychology, and it allows you to track how your wellbeing potential is improving, and even if you can’t see the progress, and feel it right now, the progress is there and if you just keep on the same track, you will definitely arrive where you want to be
Will Mountford: These models, the SOM models, have been trained now in this large data set. How broadly applicable are they to other kinds of data outside of psychological evaluations outside of surveys? Taking the tool that you have now and moving forwards with how that will either fit in with different data types, different types of analysis and then looking towards, you know, fitting it into health systems into one’s daily life into one’s lifetime even.
Fedor Galkin: The self-managed map approach the sum approach, it is data agnostic. It does not really care what data is used to train it and what data you throw into it. So the, this approach we applied to a mental wellbeing can be easily transferred to any other area of research, and this is also part of our reasoning behind this work we often use psychological data dimension to try new ideas, new approaches and experiment and try new experiments and machine learning techniques. It is in an easy data type to work with – the iterations are quick, the data is readily available from multiple sources, and if you want to run a pilot study and collect even more data from real people and see how your novel hypothesis holds true in the real world, not some academically collected data, data set that has been there for years.
Then you can easily do that and use the findings you have made, and also the experience of interacting with these algorithmic approaches and these new methods of artificial intelligence.
You can re-apply them to something more biological, right to medical history and find similar ways, for example to to to to treat our real physical diseases and you can also re-apply the same self-organising map approach to clinical blood tests and similarly group people not by their psycho types but by their blood profiles and draw conclusions based on that.
Some regions will probably be linked to specific diseases, physical disease. And Gary, and if a person is located in, for example, a region on such a blood derived self-organising map, you can tell that they have an increased risk of developing a kidney condition or a heart condition using the same recommendation engine with different content. But the mechanics, the workings, the inner workings of this recommendation engine, they remain the same. You can also help people go away from this region with increased risk of developing a kidney condition to help bring them to a region or where they will stay longer in good health. And just as I said, any data type should work with this approach, not only clinical blood data, but epigenetic data should also work as well, and epigenetics of ageing is quite an actively developing field, lots of studies, there are lots of insights and all the different applications these general approach entails.
Will Mountford:From the inside of a health system that might want to incorporate these tools, are they you know, accessible? Are they something that you think a health system on a kind of a regionally, maybe a national level would be able to incorporate into their care platform.
Fedor Galkin: Oh, it will be a huge step for the global healthcare system to adapt more data driven approaches.
If we speak about our health care systems, currently one of the bottlenecks there is the capacity of human professionals, right, a physician can can accept only these many patients in a day, right and at the same time stay attentive and not tired and make diagnosis but artificial intelligence, it has almost unlimited capacity and it is one of the greatest use cases for artificial intelligence is the initial screening.
It takes the load from human professionals and can process thousands, millions of people in an instant, provide some insights, raise some red flags for certain people and then based on these red flags, direct them towards a physician, at a general practitioner, or maybe some specialised doctor already, if the red flag is specific enough.
And this way we can not only greatly increase the capacity of any healthcare system in the world with national or local, but also find help people better understand their health and whether they’re in risk of something or what particular organ systems. They need to pay more attention to this is, I’d say it would be a great implementation of artificial intelligence and it will greatly improve the lives of everyone on this planet. Important to understand that artificial intelligence is not silver bullet, definitely it has its drawbacks and so screening done massively, will definitely raise a couple of false positives, right, if for an AI system tells that something is wrong, but it’s just a statistical mistake, this artificial intelligence system made, there is a risk of making people too nervous, too neurotic about their state of health while while there was actually no no reason to be so. It is one of the challenges of implementing such systems in the real world.
It’s currently a matter of balance, and so, for example, one such example would be the atrial fibrillation detector that is implemented in Apple Watch. It’s quite important that if you are using such, such a watch, you need to understand that there is such a thing as a false positive, and if an artificial intelligence built into this watch is telling you that you might suffer from AF, you should probably be not so worried until you’ve seen a real doctor to verify it.
Will Mountford: Still room in life for actual human doctors.
Deepankar Nayak: The recommendations that are being made by the by the AI have all been put together by physicians and professionals and psychologists, right, so the source of these recommendations and these actions, and these interventions and and what the the platform would encourage, right, its users to undertake is all coming from professionals and is is sound vetted professional advice right?
So it’s not just being pulled out randomly in in any sort of way so these are absolutely defined decision trees, looking at a number of combinations which is then being analyzed by the technology analyzed by algorithms and then proposed to the users.
Will Mountford: Are there any other current avenues of research at the longevity you can sing these songs for any other topics, or are there any other projects in the work that it might be useful to raise for listeners now?
Fedor Galkin: We have many different projects going on at Deep Longevity, as far as self-managed maps go, we are still experimenting with them and trying to build our first applications with them, currently focusing on the the application of songs to mental health, wellbeing and increasing your long term happiness. These application will be publicly available on the website futurself.AI in future without the E on the end self.ao, but definitely we will be looking for the ways to integrate these technology into more clinically relevant approach. But currently the haematological agent block is something we are paying something we are focusing on. It is a technology that, as I mentioned, probably several times on this podcast, it is a technology that allows you to receive insights about your base of ageing and very basic, very molecular ageing processes based on your clinical blood tests, really simple, but it’s nothing specific, no hormones, very few minerals, just a complete blood count and your comprehensive metabolic profile stuff like lipids, cholesterol, liver enzymes, stuff like that. We have already developed a system that can issue recommendations, find similar path towards physical longevity are based on clinical blood tests.
But in the works we also have of projects that involve epigenetic ageing, right, DNA methylation of ageing clocks, and how methylation state status of particular genes affects your pace of ageing.
Hopefully we will soon create a method that allows you to manipulate the epigenetic properties of your organism to stay younger for longer. Another project is involving microflora, it’s a really great topic.
Probably it deserves another episode to talk just about microflora, since so much research is being done in this area.
Deepankar Nayak: We’d also like to mention that it’s an ongoing process, right, so we are collaborating with universities in different parts of the world, with hospitals with clinics with researchers, the field of longevity and ageing is gathering momentum. There are, there’s a lot of interest there are every day we are getting to know there are new longevity and ageing centers being established inside universities and hospitals, so we’d also like to take advantage of this medium to reach out to those researchers and mention that.
If if they are looking to do something and and they think that an ageing clock would be a perfect tool for for their research, then we’d love to hear from them and and establish even more collaborations.
Fedor Galkin: AI technology is holds great potential and it is applied wisely to the field of health care on any level.
If you’re just a physician and you would like to get a fresh perspective on things or find a new way to interact with your patients, AI can help you with, and if you are policymaker, or you manage a large clinical organization in this case, I can not just improve the services that you are providing, but also greatly, greatly improve your capacity of healthcare system that you are managing or greatly reduce the costs in certain aspects.
And help you reach out to many more people and help them as well live longer, healthier lives. Generally speaking, the problem of global ageing – people on the planet Earth are getting on average older and older each year and I see that AI can definitely if if not provide, the whole solution, it is, and it is indispensable part of a solution that we, as a species, come up with to solve this problem.
Deepankar Nayak: When I thought about joining Deep Longevity, what got me excited was all the innovation that we are doing, but also this is the perfect platform to kind of stay immersed and stay abreast of everything that’s happening in the field of longevity. So my message to all the listeners is get interested. Set tree debate.
Explore, find out what’s happening. And it’s not necessary that you need to follow any of these advice that’s floating around and you can take your time to find your own course, but at least stay informed, right? What I have started doing since I got associated with Deep Longevity is I’ve started intermittent fasting, right, I heard about this a lot, and never really did anything about it, but now I’m doing it quite frequently and I think that it’s doing working great for me. I don’t see that I am feeling tired – I can still do my workouts, I can still do my sports and I can feel that my health is getting better. So that’s one small improvement which I think I have done in my life and which I would attribute to the beginning of this relationship that I am forging with Deep Longevity, and that will be my message to all the listeners out there.