Sonic Sleep in the News
In our own podcast, the two Dans discuss how sleep is studied in a clinical study, how sleep monitoring devices have advanced over the past ten years, potential pitfalls in how these devices are used, ways to augment deep sleep and REM sleep, and much more. To learn about the future of consumer technology and sleep enhancement, check out the interview below!
Before tuning in, we do need to add a small caveat: Please note that the manipulation of sleep architecture, while new and exciting, may have unforeseen and undesirable consequences that we are simply unaware of yet.
Furthermore, this form of enhancement may not yield positive results for all populations equally. For instance, in a forthcoming show, Dan talks with Dr. Jennifer Goldschmied about how slow-wave sleep suppression (as opposed to enhancement) might be a strategy for ameliorating symptoms of negative affect in individuals with major depressive disorder. Like with all new types of novel technological interventions, caution is warranted.
As per our consistent message here at humanOS: first exhaust natural techniques (e.g, daily light optimization) to pursue the health benefits you seek.
Daniel Gartenberg: 00:00 We’ve shown that we could improve your sleep architecture by playing these sounds basically throughout the night.
Dan Pardi: Sleep is universal and essential. All of us needs sleep in order to live and good sleep is crucial for us to recover from stresses of the day and perform at our very best on a regular basis. But sleep is in some ways a mysterious process which is part of what makes it so fascinating, in particular aspects of sleep are somewhat difficult to assess outside of a lab. In recent years an array of trackers and devices have emerged which report to be able to gauge the duration and architecture of your sleep. But the information that we receive from these trackers is limited by their precision. How accurate are these devices? Can we rely on them for useful feedback to inform our future behavior? This is why I am pleased to have Dr. Daniel Gartenberg on the show. Daniel has dedicated his life to helping people sleep better. He has studied sleep and cognition and has investigated how to accurately track sleep quality through wearable technology. He’s also developed several apps including the Sonic Sleep Coach. Daniel welcome to HumanOS Radio.
Daniel Gartenberg: 00:08 Dan thanks for having me, really appreciate it.
Dan Pardi: Tell us about your background. How did you get interested in sleep in the first place?
Daniel Gartenberg: 01:15 Sleep was always something I struggled with in high school when they would make you wake up at an ungodly early hour that is the exact wrong time for your circadian rhythm at that age. I remember passing out in first and second period quite often. I think I had to wake up at like 5:30 to catch my bus.
Dan Pardi: Wow.
Daniel Gartenberg: 01:18 So that always stuck with me and then in college I was actually part of a brainstorming company and one of the ideas was an alarm clock that measured your sleep and it struck me that sleep is something we do so much of and if you can improve that process just a little bit it would have a massive global health impact. So that general idea has been driving me for 10 years basically. So I ended up getting a PhD in this trying to understand can you not only measure sleep, but actually improve people’s sleep quality? I was one of the first apps on the store back in the day. Did some validation work for some major wearable companies and saw how inaccurate some of these devices were first hand and then I was big in quantified self back in 2010. We thought these devices would change healthcare back then and sadly they didn’t.
Daniel Gartenberg: But what’s happening now is they’re finally getting the sensor accuracy where they’re almost as good as clinical grade devices and it’s just a matter of the clinical and other communities to embrace some of these technologies and hardware companies are just starting to open up their raw data which is enabling scientists and researchers and companies to actually use the raw data from these devices, more accurately predict the stage and actually manipulate you into deeper states of sleep or enhance your sleep in various ways with sounds, lights and temperature. Those are the tools that we think of when we think of how to actually augment someone’s sleep such as eight hours feels more like eight and a half and that would be the golden goose if we could actually scientifically demonstrate that we can do that.
Dan Pardi: 01:22 Let’s start by talking about how sleep is measured in the clinic. If somebody has the suspicion or their doctor has the suspicion that there’s some issue going on with the person’s sleep, what do they do and how has this been historically done to assess if a person’s sleep is normal or abnormal?
Daniel Gartenberg: It’s a complicated question and this is part of the problem at least if someone has insomnia we can start with that is typically what will happen is you go to your general practitioner. The GP doesn’t have the necessary tools or training to give you the recommended treatment which is cognitive behavioral therapy for insomnia. That’s actually what various clinicians and academics have designated as the first line treatment for insomnia. But the GP doesn’t have the time to do that or the training. There’s basically like 1,000 people trained in this in the United States so there’s a big issue with just finding people trained in this behavior therapy. So typically what’ll happen is the doctor will prescribe you a drug. Actually anti-anxieties are the most commonly used drug to help with insomnia. That’s currently how medical practice operates for dealing with that. They’re not even actually necessarily measuring your sleep that much before intervening and the crazy part is the current gold standard for measuring insomnia is a paper pencil sleep diary.
Daniel Gartenberg: 01:42 Right now we just submitted a proposal to do a clinical trial and we actually have to use the paper pencil sleep diary as our primary outcome measure because that’s the gold standard for how insomnia is measured and what we’re trying to do is demonstrate that subjectively how you feel you slept last night, there is some value in that information actually and clinicians know how to deal with and understand that. But we’re wanting to augment that information with objective truth from wearable devices regarding how well you slept. That’s insomnia and then we can get into the whole chain of apnea if you’d like.
Dan Pardi: Is there a clear example, but generally speaking there’s a variety of different clinical sleep issues insomnia and sleep apnea are most prevalent amongst those issues and yet if you do finally get to a sleep clinician and now there are as opposed to 30 years ago, 20 years ago far more sleep clinics around the world and in the United States. So it’s easier to find one than ever before, but it’s still because of the lack of knowledge of sleep at the primary care doctor there might be a delay in getting the treatment you need because they’re not as well trained on how to identify and treat the problem. So much of that model is recommending medications first.
Daniel Gartenberg: 01:43 There’s a small distinction that I want to draw there and the clinics often times will treat insomnia, but they’re largely focused on apnea because that’s really how they make money on the billing codes generally. So a lot of times the people trained in behavioral sleep health, you actually get a certificate in this. It’s not necessarily even doctors like social workers, psychologists, mental health professionals basically have the certificate in behavioral sleep medicine and they can help with insomnia.
Dan Pardi: Well I don’t think we need to go into all the different ways that different sleep disorders are actually characterized. Tell us what a night of getting sleep monitored in a sleep clinic, what does that entail?
Daniel Gartenberg: 02:42 So this is the problem that a lot of people have keyed into because if you say have apnea, or think you might have apnea, what will typically happen is first they’ll give you something called a pulse oximeter and it’ll screen you as something you wear on your finger. It’s pretty noninvasive and a lot of times I’ll recommend this to people just because it’s a first easy noninvasive indicator of whether or not you have apnea.
Dan Pardi: What is the pulse ox measuring?
Daniel Gartenberg: 03:39 The pulse ox is measuring the oxygenation of your blood basically and you put it over your finger typically. Garmin has some kind of loose measure of it. It would be hard to get over your finger and basically when you have apnea literally you’re not breathing. Your oxygenation of your blood goes down. So the pulse ox can detect that.
Dan Pardi: So if it is detected that oxygenation of the blood varies from what would be considered normal. What would the follow up be in that case?
Daniel Gartenberg: 03:59 The follow up would be you got to go to your sleep lab now. The pulse ox has said you’re not getting oxygen throughout the night and we’re going to hook you up to 16 channel electrodes which is called polysomnography and that’s electrodes on your head, under your eyes. There’s what’s called an EOG to get the motion of your ocular. When you’re in REM your eyes are shooting around and that’s what they use to detect that. An EKG as well is used to detect sleep wake often times. So you’re hooked up to this whole barrage and it’s a little bit of a stretch, but I like to think about the psychology version of the Heisenberg Uncertainty Principle when it comes to this which is called the Hawthorne Effect which is basically this idea that the act of measuring something can impact what you’re measuring. That’s the problem with this because you go to this sleep lab, it’s a weird environment, you’re primed not to sleep in that environment and then they hook you up to all these uncomfortable equipments and that’s how they measure your sleep but it’s actually impacting the quality of your sleep just being in that environment in it of itself.
Dan Pardi: I have to chuckle because I have spoken about that specifically in talks and used the image of Heisenberg from breaking bad.
Daniel Gartenberg: 05:10 Oh, really?
Dan Pardi: Yeah so that’s right. In order to get increasingly accurate measurement you’re also disturbing sleep and so you have this paradox of getting closer to identifying some sort of pathognomic characteristic of sleep that you could then say, ah, this is REM sleep behavior disorder or restless leg or whatever. But then in doing so you’re definitely disrupting normal sleep by wearing 27 different leads coming out of your body and on their waves. I’ve actually had a sleep study done. I don’t know how anybody sleeps under the sleep study. Of course I’m not saying it’s not important, I’m just saying that it’s not ideal.
Daniel Gartenberg: 05:48 Yeah that’s part of the problem and that’s what a lot of people in field are trying to solve for. It’s such a clear problem. One thing I will say in defense of the sleep labs, when you are asleep you can detect these apneas they pop out and even if you’re in a nasty environment, you will be able to see at least for sleep apnea there are really clear indicators when you’re asleep if you have it or not and you don’t necessarily even need to sleep well the whole night in order to evaluate that. What they’re also looking for is there’s two main types of apnea actually. There’s central apnea. There’s actually more than that. But largely speaking there’s central apnea and there’s obstructive. Almost everyone has obstructive sleep apnea if they have apnea at all. That’s actually a physical thing and that’s why if you have obstructive, weight loss is so important in all this stuff. But what they’re looking at in central apnea is actually your brain producing some of these apneas. A lot of times it goes hand in hand with obstructive and that’s called complex.
Daniel Gartenberg: There’s a whole rabbit hole and this is why as a cognitive psychologist I find this stuff so interesting. Sleep is the intersection of so many things.
Dan Pardi: 06:30 Not only are there different categories of sleep disorders, but within a category, let’s say insomnia, there are many different types. The ability to detect a signal accurately could lead to not only the broader classification of your have insomnia or sleep apnea, but possibly what type you have to see if you get a better intervention to get you to a better place.
Daniel Gartenberg: I love the train of thinking that you’re on right now. I riff on this with my professor often times. I work with a professor at Penn State Orfeu Buxton and I think you might know him actually from some of your past work.
Dan Pardi: 06:58 I do yeah.
Daniel Gartenberg: We’ll often chat about maybe there’s even more types of insomnia than are currently clinically known. Right now there’s morning insomnia and evening insomnia. I think what we’re going to find with a lot of chronic disease in the next 10 years whether it’s Alzheimer’s, insomnia, apnea is as humans we like to have these broad categories for things because it gives us a sense of control. But I think there might be 10 types of Alzheimer’s disease. They’re finding stuff like this. I think the same thing can be said for insomnia. Maybe there’s actually 15 underlying pathologies and we come to call them all insomnia but it’s a combination of actually your sleep environment, first night effects sleeping in a new environment messes with you. Temperature, sounds disrupting you. Circadian rhythm issues. These are all underlying pathologies and we call it insomnia, but maybe it’s much more complicated than that. What the wearables are going to enable us to do is for the first time in healthcare it’s not classifying you based on some population thing where we run a clinical trial and we show between group differences that this drug works.
Daniel Gartenberg: 07:11 I want to give you the intervention that is best for your specific brand of insomnia. By the way, I’m not doing this yet. This is a future goal, but that’s a real inspiration for what we’re doing.
Dan Pardi: That is an exciting future that’s possible. That’s a great primer on how sleep is generally measured. The different conditions that it detects, so even though you are disadvantaged in terms of the discomfort of the sleep measurement, it still is good at detecting characteristics that then let you know you have this condition. But because of the limitations and the disruption of the sleep monitoring and the limitations of only getting it one night versus 30 nights in a row we are undoubtedly not assessing the variety of different types of sleep issues as accurately as we could. For that you need a certain level of accuracy for some diagnostic ability. So since 2010 would you say that the accuracy of sleep monitoring devices across the big players in the wearable space, has that improved in the last 10 years?
Daniel Gartenberg: 07:38 That’s what’s exciting is dramatically improves and the main reason for that is frankly motion based sensors to measure sleep have been around since the 70s. What’s really interesting in 2014 when I reinvigorated my effort to improving people’s sleep is the heart rate monitors and the PPG. That if you have an apple watch you sometimes see that little green light coming out. I’m not an engineer necessarily but basically they shine the light and they can see the blood pulsing in and out and get your heart rate from that. Having that heart rate sensor changes the game in terms of sleep detection accuracy, especially for the sleep cycles actually. The signal for sleep wake is largely motion based. So the heart rate sensor also helps with that. But at the same time it seems like a small thing, but technology-wise it’s a big deal.
Daniel Gartenberg: In the past were these devices, a scientist or a third party app developer like myself couldn’t get the data in real time. You would only be able to see all the sleep data after the person wakes up. But now with apple watch or with the Oura ring or with the Fitbit, all of these devices there’s some differences but just from what I’ve seen comparing it to truth data which is polysomnography the electrodes, the sensors are pretty similar between all of these devices. I’m generalizing here, but most of the sensors, most of them they’re made in China it’s a very similar technology. The sensors are all almost on the same playing field. The form factor differ. A ring makes a difference compared to a wrist and also the battery life changes. But the biggest thing is the algorithms that sit on top of these sensor devices and what we’re trying to do is refine and improve those algorithms such that they’ll work with all of the best devices and be validated scientifically for actually measuring your sleep.
Dan Pardi: 07:40 Originally we were talking about clinical sleep monitoring in a lab. They’re looking at eye muscle movement to detect rapid eye movement sleep, REM sleep. They’re looking at how much blood oxygenated, EEG or electroencephalography which is monitoring brain activity. You have to measure multiple different parameters to then triangulate that signal from that signal assess what you’re looking at. That only occurs through people getting together, saying these are our diagnostic criteria. This rules you in, this rules you out of this condition if you have these features on your overnight sleep study. In the consumable wearable space it started with movement detection. They’ve been adding additional parameters, heart rate and different signals that you can also triangulate or better signal detection. But what’s been agreed upon as the best way to approach this we’re still in that time period where that’s been assessed out. Is that right?
Daniel Gartenberg: That’s a very fair assessment and just one thing I would also add that makes this all possible is the granular data. So to do the accuracy for the machine learning stuff you want as granular data as possible. This gets a little nerdy. There’s a lot of bang for the buck when you start getting multiple data streams. Like for example the Oura has some measures of internal body temperature too which is something that we’re interested in. When you have that big data approach and you start getting multiple data streams, the accuracy can dramatically improve. It’s just going to get better and better in the next 10 years. Who knows when Apple finally puts an EEG sensor in their earbuds or whatever. There’s a lot of exciting things pending in the future.
Dan Pardi: 08:01 So every time you add some other relevant signal you can leverage that signal to increase the reliability and the validity and accuracy of what you’re measuring. So we’re at a better place than we’ve ever been, but we still have to figure out the right algorithms that interpret all of this data and then say this is what’s closest to what our current or previous gold standard is and had been.
Daniel Gartenberg: That’s exactly right. So our approach is I like to not have a horse in the race. I just want to use the best wearables and just build the algorithm on whatever wearable you might have at the time. So one algorithm that’ll work on every single wearable out there that’s the scientific community can get behind that’s what I’m working on right now.
Dan Pardi: 08:09 It’s not that the signals that we’ve been detecting before from the clinic have been perfect, it’s just that they’ve been worked with and people know how to then work with them and know how reliable those signals are.
Daniel Gartenberg: That’s the sad thing in science sometimes as you’ve seen you have some archaic device and it’s clinically validated because it has all this research behind it, the scientific and clinically community can’t get beyond that sometimes.
Dan Pardi: 09:13 Yeah, yeah slow moving ship at times.
Daniel Gartenberg: Yep.
Dan Pardi: 09:19 Then you have consumers who have less pressure on somebody who is needing a certain degree of accuracy and acceptedness of a measure to then say okay well I trust the diagnosis that I’m getting from a device which is future thinking here. Then that is true and real. But people who are buying these devices for the last 10 years and continue to do so now, they’re not all necessarily trying to identify if they have a sleep issue.
Daniel Gartenberg: Yeah so healthy sleep is a way that we like to think about this problem even more so than unhealthy sleep and how do we actually make it healthier? I don’t have any underlying pathology for example. But I want to optimize myself as much as I can. I run a company, it’s pretty stressful. I know that I’m kind of a dick when I’m sleep deprived and I make poor decisions, so how do I just as a healthy person evaluate the things that are negatively impacting me and seeing that change in your sleep data after drinking alcohol is something that a lot of clients that I’ve spoken to have actually stopped drinking because they’ve seen that. So one of the questions is how do healthy people use these devices to improve themselves and that’s been a hot topic for debate in clinical environments as well.
Daniel Gartenberg: 09:20 I just got back from a conference at the society of behavioral sleep medicine where there was a lot of concern amongst the clinicians that people are using this data and don’t necessarily know how to interpret it and it can actually have negative unintended consequences.
Dan Pardi: Yes absolutely.
Daniel Gartenberg: 09:51 This is something as technologists that I’m really thinking about and trying to be careful around the New York Times also published an article a little while ago about this new concept orthosomnia which is basically people who are really focusing on the metrics of some of these devices to an unhealthy degree and it’s producing anxiety and actually making their sleep worse. So that’s something to think about, but I think there is a way to design a system that gives people this data in a way that is fair and accurate and doesn’t produce anxiety for people and a lot of this comes down to the interpretation. At the end of the day I’m an humanist. I was an English major and a Psych major and undergrad. I don’t think that some AI is just going to replace the human. I see these devices as facilitating humans helping other humans with informed metrics of what’s actually going on with that individual and I think when the technology is used in that way and more of a coaching, more of a facilitator model that’s when we can actually start seeing some positive health outcomes.
Dan Pardi: Some of the sensationalism of the idea that sleep trackers ruin your sleep. It is possible for some. I don’t think it’s representative for all. Anybody can take any piece of health equipment and misuse it in a way where it doesn’t become healthy and that might sometimes be from an underlying psychiatric issue of anxiety and so they develop anxiety around whatever they’re interacting with. The more interesting question for me is can this take somebody who doesn’t have existing anxiety or a history of that and create that issue commonly? Generally the mission of getting objective feedback on yourself understanding factors that might influence sleep satisfaction, helping you wake up and feel better and then serving as an engagement tool to stay true to your ideals about what you think that you could sleep and learning over time. We know that drinking caffeine and alcohol before bed is going to disrupt your sleep. We can tell you that. But if you have the personalized experience where you see the data it might resonate in a totally different way.
Daniel Gartenberg: 10:51 Exactly.
Dan Pardi: What do you think is the future of the next stage? You mentioned some things earlier. We start with tracking. But we might be intervening during sleep to enhance things. What does that look like in today’s world and what will it look like five or ten years from now?
Daniel Gartenberg: 10:59 I’ve made some weird patents and inventions around a speaker mounted in your bed and we’ve really focused on sound in our research. So the professor I work with who’s one of the first people to observe how sounds in a hospital environment negatively impact your sleep quality. So it’s kind of ridiculous that you go to a hospital, you’re expected to recover in this noisy environment and sleep is one of those things that helps your immune system and it helps your body recover yet we present people with this disruptive sleep environment when we’re trying to help them get better. In our lab we have systematically delivered sounds to people and we’ve done this for a few different reasons. One is to understand how sounds disrupt sleep and another is to understand how to mask those sounds with something we call pink noise just as snoring bedtime partner will negatively impact your sleep because the abruptness of the sound is really what’s disruptive.
Dan Pardi: Mm-hmm (affirmative).
Daniel Gartenberg: 11:20 Then the cool part of that is when you start getting into how can we deliver sounds such that the brain can process it, but it doesn’t wake you up. That’s the core technology that we have gotten some expertise in. So we’ll deliver hundreds of sounds to people throughout the night in a sleep lab using EEG while they’re connected to all the best wearables and from this we’re able to understand how to play a sound so that your brain processes it, but it doesn’t wake you up.
Dan Pardi: Mm-hmm (affirmative), yeah.
Daniel Gartenberg: 11:34 Sort of like a tight rope walk. If you push too hard you’ll awaken the person, that’s not good. We’ll deliver this hundreds throughout the night they’ll have no conscious awareness that we’re playing these sounds to them and what we’re trying to do, there’s a couple of things. First in 2013 I submitted a patent using sounds to query your sleep stage and it actually finally almost seems like it’s going to get accepted. It takes a long time. So basically the idea is you play a sound and since people have different heart rate responses to the sound based on their sleep stage you actually use it to figure out what stage sleep they’re in. Once we know which stage sleep they’re in with a higher degree of certainty then we try to prime their brain into different stages of sleep and basically what we’re doing here is we’re playing a delta wave which is the same frequency as your brain waves when they’re in the deep state of sleep.
Daniel Gartenberg: We’ve shown that we can improve your sleep architecture by playing these sounds basically throughout the night. Other things we’re exploring is manipulating dreams.
Dan Pardi: 11:34 Let me ask something about that. There’s been research playing pink noise while somebody sleeps for the whole night and in both young and olds it showed to augment memory pretty cool. But sleep one of the challenging parts about it and manipulating it is that it’s defined by such distinct neurophysiological signatures across the night and even behavioral patterns. The body is clearly in different states as it’s in different sleep stages across the night. So for drugs if you’re optimizing for one sleep stage you’re automatically probably sacrificing another if that can’t be targeted to a very specific moment in time. So it sounds like what you’re doing is you’re looking to evaluate what sounds work best at different times?
Daniel Gartenberg: We’re trying to figure out how to play a sound at the right volume to get your brain to produce more delta waves.
Dan Pardi: 12:45 And testing different sounds in that moment?
Daniel Gartenberg: We’ve systematically varied the quality of the sounds and we have a good understanding of which sound is good in that moment.
Dan Pardi: 12:59 Okay.
Daniel Gartenberg: That sound is universal across people.
Dan Pardi: 13:48 If I happen to be in deep sleep and my brain is manifesting these very slow high amplitude rhythmic waves, your system will start to play those sounds in that stage. Will the brain entrain itself so that it gets synced up with the sounds that are being played even if they occur slightly askew of do they have to be perfectly in sync in order for it to have an augmenting benefit to the sleeper?
Daniel Gartenberg: So this gets really sciency and I love going down this pathway with you. I hope your viewers appreciate it. But this is a big discussion in the literature right now. What invigorated me to return to this subject was some findings related to this delta frequency where some researchers out of Germany showed the authors NGO is the name, but what they did is they phase locked the sound, meaning they had an EEG, they could measure the up and down oscillations when you’re in deep sleep and they played the sounds at the up oscillations in order to prime the brain state. What if we can do this without an EEG? So that’s what the basis or our grants were and we got some grants from the national institute of aging actually because as you get older your deep sleep decreases and is thought to be associated with cognitive decline. We’re looking at inversion of mild cognitive repairment right now as an area. It’s also related to human growth hormones, cell recovery all these good things.
Daniel Gartenberg: 14:42 You want more deep sleep. You want more REM basically. Those are the two things you want at the expense of light sleep is how I usually think about it. What we tried to show is that we could produce this delta response with the sounds without phase locking, just by playing it at the right volume level at the right timing of your sleep and we’ll actually play it in both light sleep and deep sleep and play different volumes during those and we were able to show that we could induce these delta waves and we’re publishing a paper on this right now in a nonphase locked manner.
Dan Pardi: I don’t doubt that that’s possible and I like that approach. It’s perhaps a little bit less sophisticated but might be infinitely more implementable.
Daniel Gartenberg: 15:47 Exactly.
Dan Pardi: What is the potential capacity to augment slow wave sleep? It sounds like you’re looking to augment REM sleep too?
Daniel Gartenberg: 16:42 Yes. That’s a nerdy finding that we focused on but just broadly speaking and you’ve talked about this in your humanOS content, there’s a lot of ways to improve people’s deep sleep and if I’m talking to someone I’m trying to help them, I’m not just thinking about this one aspect of technology for doing it. Whether it’s in training your circadian rhythm, making sure the person is getting more sunlight during the day. You’ve posted and Matthew Walker has posted some stuff around raising your body temperature throughout the day can result in more deep sleep. I’m actually curious your insight in some of these other things.
Dan Pardi: Rocking, physical activity, light, possibly activation of ground fat from cold. There’s potentially a lot of things that are secondary measures feeding into the homeostat or the collector of signal that builds up pressure over the day that then would deepen your sleep at night if you look at hunter gathers they don’t necessarily get more sleep than modern sleeping humans, but they have very robust circadian amplitude. Probably largely driven by just getting natural light exposure across the day. I’m always one for optimizing lifestyle parameters before looking to intervene in a more technological way. But can you do everything right and still augment it with technology. Obviously it’s great to see that there is an impact that’s statistically significant in terms of deepening the way of sleep here and there, but does that also lead to outcomes that are desirable? Improvement of memory, improvement of vigilance without any untoward side effects that are [inaudible 00:30:10]. That is the potential and it’s exciting and I think we need it because we’re not living outside all day long.
Dan Pardi: 17:27 Can you prevent mild cognitive impairment and delay it for 15 years or longer or prevent it altogether by having some really sophisticated technologies that’s easy to implement for many people. I love that pursuit.
Daniel Gartenberg: Exactly. So when we think about this also we’re thinking about light and you tapped on that too. With these new wearable devices you hook up your Phillips Hews, the sleep detection system or your LifeX bulbs and then all of a sudden you’re entrenching your circadian rhythm so while even though we’re inside all day, with the technology you can noninvasively optimize people’s sleep and in various ways and the behavior change stuff can be basically augmented with some of these queues from your environment. That’s what we’re trying to do with this next phase of funding. There’s a lot of money in the federal government for solving Alzheimer’s right now because we’ve tried a lot of drugs. It’s hard to get past the brain, blood barrier. A lot of them haven’t worked. Recent evidence has shown that deep sleep in particular cleans out your beta amyloid plaques which is thought to be associate with developing Alzheimer’s disease. That’s one of our primary focuses right now.
Dan Pardi: 17:51 Sounds great. Yeah I interviewed Bryce Mander previously on the podcast show works in old Matt Walker lab. At Berkley we focused in on the aging brain and his recent publications in that field specifically. We used to think that sleep was just a very common symptom to all people that are experiencing mild cognitive impairment, particularly all forms of dementia like Alzheimer’s disease. We now think it’s actually causative. So it’s not just the symptom that tracks with the development of dementia, but it’s actually something that is probably driving the process. Makes sense that these certain pathologies might build up, interfere with the generation of slow aid sleep and then the lack of slow aid sleep accelerates the pathogenesis.
Daniel Gartenberg: Exactly. Yeah they’ve done some really captivating scientific work with FMRI and doing sleep deprivation for people and actually showing some of the bio markers for Alzheimer’s increasing. They’ve also started out with the mouse and done some really interesting neuroscience studies showing the role of slow aid sleep and cleaning out these maladaptive plaques that form.
Dan Pardi: 18:12 The technology doesn’t need to do all of the work. It can do some of the work. But those who are wearing it, it can serve again as the behavioral lever to go then do other things. It can serve as a catalyst to then go do other things that you can do simply from a behavioral perspective. That potential of some additional augmentation and the augmentation of the right behavioral set that puts you in a far better place than you would have been before and without it.
Dan Pardi: I’m all about how do we maximally leverage the existing technology and then build on that as new signals come on board, more accuracy comes on board and we can have more intervention in the middle of the night that potentially could make the whole process even better than it was previously. Now where are you currently? Do you have any apps in the store that people can go try and use? What sort of tech have you built so far that is available to consumers?
Daniel Gartenberg: 18:23 We’ve made the sonic sleep coach, that’s an app on the app store that also connects with the apple watch. So we have this algorithm that we think is one of the most accurate for measuring sleep and if you have an apple watch it’ll work. Also if you don’t have an apple watch we measure the sound in the room to figure out things like snoring and we have sleep diary and the smart alarm clock that wakes you up gradually which I firmly believe is the right way to wake up. If you have the apple watch it will deliver the deep sleep simulation sounds and also I’ve been doing this thing recently that I’m kind of doing a mad scientist thing on myself and this is very quantified self-esque fro the bio hackers listening you might be interested.
Daniel Gartenberg: I’ve been meditating on my ten year vision for myself while I’m playing a specific audio track, in my case it’s 528 hertz which is in our software as well and we have a few wind downs. The clinical community loves progressive muscle relaxation so we have that in there as a way to relax your mind as well. So I’ll meditate on my 10 year vision while listening to the sound and then our system since we know how to play sounds so that it doesn’t wake you up, will actually replay the sound when you’re in a REM state and what I’m trying to see is if I can entrain this way of thinking about my 10 year vision into my subconscious mind almost as a way of inceptioning myself. It’s very Joe Dispenza-y. You create your reality riffing off the placebo effect. So that’s another train of thinking that I’m on right now.
Dan Pardi: 18:39 There is a lot of potential to engrain some sort of behavioral practice around the last hour of time awake before you go to bed thinking perhaps on some challenges that you’re trying to solve, putting that into your mind right before bed. So it’s not necessarily watching some entertainment that allows you to relax which could be good too and it’s what most people do. You actually think really clearly on some challenge and then that is populating your mind as you go to sleep which might actually lead to developing creative ideas faster on that subject.
Daniel Gartenberg: Totally. There’s a lot of really captivating research on just the act of imagining playing basketball helps you perform better at playing basketball. One of the other things I’ve been doing is just in the same vein as what you were just mentioning is doing a little gratitude meditation before bed. I find that I’m kind of nicer to people the next day. So that’s another one.
Dan Pardi: 18:41 I love it. So if you have an apple watch you can download the app and is this on iOS and android?
Daniel Gartenberg: It is iOS and android yep.
Dan Pardi: 18:42 Okay so you download it. It’s both on your phone and it then goes onto the watch. Right now apple watch has got 18 hour battery life, so it’s not a full day. What do you recommend people do if they’ve warn it from the morning when they wake up, let’s do our best practice process for charging it before you go to bed.
Daniel Gartenberg: This is a little bit of a struggle and I can get into the nuances of this, but apple watch four we can get it up to 24 hours with some tweaking. So usually what I’ll do is well honestly for me I don’t really use the watch that much in the day and I just use it at night because I’m very interested in my sleep. But what you can also do is just charge it in the morning when you wake up or throughout the day.
Dan Pardi: 19:10 Do you recommend setting an alarm to say charge your watch let’s say at eight, or between let’s say that last hour when you’re in bed meditating? Then you’ll have a full battery before you go to bed?
Daniel Gartenberg: That’s a good idea Dan I should do that actually. Thanks for bringing that up. I should suggest that while they’re doing the wind down is a perfect time to charge. It only takes a half an hour to charge the new devices. So it’s not that bad.
Dan Pardi: 20:00 It’s frustrating. It’s another barrier, but those who are engaged could probably do it. It’s just a small curve to step over.
Daniel Gartenberg: But that’s why I love Oura and so we’re trying to collaborate with them and work with their device soon too.
Dan Pardi: 20:18 I’m very impressed with the frictionless ness of the form factor of Oura. I didn’t know that I was going to enjoy wearing it because I’ve never worn rings. But I’ve replaced my wedding ring with wear Oura and so it’s my wedding ring now.
Daniel Gartenberg: It actually is a nice … it can function as a wedding ring. I’m wearing one right now and it kind of looks like a wedding ring. I aesthetically appreciate how it looks.
Dan Pardi: 20:19 Oh, for sure. I appreciate that about my Fitbit too. So I’m always testing various devices because I’d like to have personal experience with them. My Fitbit lasts about seven days as well. I was doing something similar to what you were doing, I’d wear my apple watch during the day because I was doing some testing around that and then when I put it on the charger I put my Fitbit on at night and then in the morning I would just switch them. I would take maybe a grand total of 20 seconds over the 24 hour period to do that and it’s quite easy. I know other people that do that too. So I could measure my sleep and then I’d have my apple watch during the day. But what I found was that it was easier to then keep sticking with my Fitbit because it’s seven day battery life. So it’s similar to the Oura ring. I think when you have at least multiple days, hopefully a week it’s going to be more convenient for you to then do the charging. That’s an advantage that Fitbit or a Garmin have over apple watch at the moment and I have less experience with Samsung. But I think they have longer battery life as well.
Dan Pardi: The Garmins are really impressive. Some of them have three week battery lives which is crazy.
Daniel Gartenberg: 21:29 I mean you get a lot of bang for the buck if you don’t have a fancy interface on top of the app. I think apple watch will just keep getting better and that’s the hope in terms of the battery and in the meantime another solution is just to be a nerd like you and have eight devices that you just swap in and out. I’m actually wearing two apple watches right now. But I’m not a normal person.
Dan Pardi: There are people out there that are going to on their own wear a lot of different devices and are looking to get a lot of different information and work with it and I totally respect that. I’m mostly testing these devices for what I would consider our average user. So not necessarily a wonk or somebody that has extra special interest in what we do. But does this provide a nice experience for those who are trying to do their job really well and care about self care and their own health and are looking to leverage some of these technologies to do a little bit better. Each one of these devices has their strengths. Each one of them has their weaknesses. Each form factor has its strengths and its weaknesses and what I typically tell people is that unless there is some clear advantage for something specific that you would like to do and sometimes yes one of the different technologies will do something better, then find the one you’re most motivated to use.
Dan Pardi: 22:29 You are the willful participant in making it actually beneficial to you. Does it change your behavior versus it acting upon you and I’ve seen plenty of that. When I wore it, it didn’t do anything for me. What did you do with it?
Daniel Gartenberg: I like that perspective, it makes a lot of sense and compliance when it comes to some of these things is one of the biggest predictors if it’s actually going to help you or not.
Dan Pardi: 22:29 Yeah absolutely. Okay so if you have an android wearable or you have an apple watch, you can go to the store right now, download the app, try it for a night. See what kind of feedback you get. What do you think is a reasonable trial period to get a sense of how the app works and how you can work with the app?
Daniel Gartenberg: I always like to not make money unless it’s actually helping people. So we actually just offer it for free for a seven day period.
Dan Pardi: 22:45 That’s great.
Daniel Gartenberg: To see if it’s working for you. After that if it’s working then it’s 9.99 a month or 39.99 a year.
Dan Pardi: 23:43 With that you get a lot of the stuff that we mentioned, but it also looks like you’ve got email coaching and you can add in addition to that baseline price, augmented services.
Daniel Gartenberg: Exactly. So one of the things that I really believe in and this is also relevant for any bio hackers listening is you own your own data. So we’re collecting very granular heart rate motion and sound data from the room and you’re able to actually export all that information.
Dan Pardi: 23:44 Oh, that’s cool.
Daniel Gartenberg: From our web portal. Actually researchers could be interested in this as well. Since we have that on our web portal I’ve been training pre-med students on how to actually give people relevant feedback on the data that is being collected. Some of the phone calls I’ll be on the call and walking you through the data and trying to give you actionable solutions to improve your sleep health. You can basically email exchange with us after you collect seven days of data, that’s enough to really get a sense of what’s going on with you. We try to deliver this personalized feedback whether it’s get sunlight exposure at this time. Shift your bedtime back two hours. A lot of these solutions are very individual. So we try to make sure that you have the opportunity to say what’s going on in an email or over the phone and then look at the data and try to give customized feedback that’s most relevant to the person.
Dan Pardi: 24:18 I love it right? Science could say go to bed two hours earlier and the mom could say well my son gets home late every night from work I can’t sleep until he’s home.
Daniel Gartenberg: Exactly.
Dan Pardi: 24:20 So you have to work with that and say okay, then how can we do better then given the constraints of your real life.
Daniel Gartenberg: Exactly and something that I really appreciate and just from talking to so many people about their sleep problems you pick up on these little things on what the person is willing to do or not willing to do that are really important for actually giving someone a recommendation.
Dan Pardi: 25:15 Yeah.
Daniel Gartenberg: Just the subtleties of human communication.
Dan Pardi: 25:26 It would be very hard to have an algorithm that picked up on all of those potential nuances of a person’s real life and work that into the system. A coach could figure that out within a week.
Daniel Gartenberg: And at the end of the day just to add Dan, part of the effect is me having empathy for your situation and selling a solution in a very human way where you have trust for what I’m saying. Having that trust with a human and I studied this in grad school a little bit in human computer interaction. Trying to build that trust with an AI is not easy. No one’s done it really yet.
Dan Pardi: 26:08 The uncanny valley as they say which is that distance between some digital representation of humanness and real humanness. It might be a gap that seems across the river but is exceedingly difficult to cross.
Daniel Gartenberg: Yes, I would say right now it seems insurmountable but we’ll see what happens in the future.
Dan Pardi: 26:15 Well Daniel thank you for your time. One thing I haven’t mentioned so far is that you have a tremendous Ted Talk, what is it five minutes?
Daniel Gartenberg: Yeah.
Dan Pardi: 26:18 It’s short, it does a really nice explanation of a lot of what we’ve talked about. But I’m going to put it into our blog that accompanies this podcast so people can go and watch it you did a great job for sure and how many views has it gotten now?
Daniel Gartenberg: Three and a half million.
Dan Pardi: 26:24 That’s great. That’s fantastic. That says a lot too. It’s been liked and shared by many people and it clearly articulates this aspiration of sleep’s a problem and how do we do better? Thank you for all your work in this area. It’s always a pleasure to connect with you and hear what you’re up to and see how you are advancing things and it’s going to be fun to stay in touch and monitor and benefit from all of the contributions you make over your career.
Daniel Gartenberg: Awesome Dan, thanks so much.