Smart Sleep: Unpacking Consumer Sleep Technology

Joshua Lennon, MD

Thank you very much for having me. Appreciate the offer.

Audrey Wells, MD

It’s interesting because we both get asked a lot of questions about these wearables that are like in watch form, there’s ring form, and there’s even Mira balls that you can put next to your bed and a lot of sleep apps. We’re going to dive into this a little bit today with some questions that commonly come up from patients. I want to start with the idea that, okay, I’m in the market for purchasing a wearable sleep tracker. What kinds of things should I be looking for in terms of what I should buy?

Joshua Lennon, MD

Great question. The first thing is, what’s the motivation? Is the intent to get something just to track sleep, or is the intent to get something extra with it? For instance, a lot of the most commonly studied trackers are watches developed by Google, Garmin, Pixel, Samsung, and Apple. So if you’re in the market for a watch and you’re also looking for some added features, like some information about your sleep, tracking your sleep stages, and things like that, then those are decent options to do that. They also have all sorts of other non-sleep features to them as well. If you’re interested in something purely for sleep, this is where some of the Mira balls can be helpful, like the winding strip. These are under-the-mattress sensors that do nothing except track your sleep. There are no extra bells and whistles with them, just straight-to-sleep tracking if that’s just the intent.

Something like the offering is pretty similar in that regard as well, although it does offer a little bit more insight into things like daily rhythms based on your body temperature and circadian health and things like that. Then there is this whole other category of arable tools, which is this software. They’re not in any hardware. It’s just an app on your phone. It uses your microphone or,  ultrasonic sound coming from the speakers and listening to to response to make some determinations about sleep for that no added device that you have to get for something like that. But again, that’s just for your sleep alone. Are you willing to shell out a few hundred dollars just for your sleep? Do you want to spend a lot less money if you’re going for something like just a software solution? Or are you looking for something multipurpose, like an activity tracker that’s going to check your steps and your heart rate across the day? Your calories burned. It’s a watch. It has other apps. You can talk like a walkie-talkie with it. You also looking for a little insight into your sleep. All of these are the factors that play probably a more important role in the selection of one device over the other.

Audrey Wells, MD

It’s a very personal choice, isn’t it? There are a lot of things to consider. A question that I get asked a lot is: how accurate are these things, and how does that factor into making a purchase?

Joshua Lennon, MD

For me, the question is accurate compared to what? Our best understanding of sleep is this symphony of changes in the nervous system from neurons in your brainstem that turn on and off, and they’re communicating with the hypothalamus and the rest of the cortex, or there’s just an extremely complex symphony of changes going on. What we measure now as the gold standard sleep study in the clinical sleep lab is the equivalent of standing in the overflow parking lot and hearing some rumbling on the ground, which is a lot different than being close to the stage in human music itself. The way that we measure sleep clinically is with electrodes placed all over the scalp, a noninvasive but still bothersome way. We have the technological span role of using these 30-second segments to say this is one stage and this is a different stage. It’s more of a holdover of older technology that we never even use any longer. so it’s almost arbitrary. It’s a wonder why we even stick with this system when there are probably more interesting ways of saying what sleep is in the white stages, the medium stages and the deeper stages with things like the odds ratio product or a spectral analysis of these different frequencies picked up from the scalp.

All that being said, our gold standard is this: an arbitrary set of stages: three stages of non-REM and one stage of REM sleep. So a lot of the consumer technologies are being assessed against either subjective changes like a patient’s or a subjective sense of how well they slept compared to a sleep diary, sometimes compared to clinical grade act graphs or wrist-worn devices that have been around for several decades that measure movement alone to stage weight versus sleep. then most frequently against these clinical grades in lab polysomnograms or standard sleep studies. So all that being said, accuracy is all over the place. It depends a lot on how the devices where we have multi-generations, say the Fitbit devices, the Apple Watch, or the Samsung Galaxy watches, have changed over time, say from the early 20s, mid-2000s, and mid-teens over the last couple of years. How these things have improved and their accuracy depends on what you’re looking for. If you’re looking at, say, just how well does it track whether I’m sleeping or not? The accuracy is pretty good. It’s typically in the low nineties, which is not all that bad compared to, say, our Clinical Grade ActiGraph that we might use in a sleep clinic have an accuracy of in the mid-eighties, give or take. It seems to be more accurate at tracking wake versus sleep than something like that, which is impressive. However, if you dive a little bit deeper and you want to find out how accurate is it going to determine when I fell asleep versus when I was awake across the night versus when I woke up in the morning. It comes down to how are they assessing the accuracy here. 

If I have a stopwatch and I say go and you’re about to fall asleep now, and then that this is where I’m going to start making the assessments, the accuracy looks good. If I say, it’s getting close to your bedtime, I go ahead and you lay down and maybe it will just see what time you fall asleep maybe it’s five minutes; maybe it’s an hour. The accuracy of these devices falls off pretty significantly, so it depends on how you’re measuring it as well. If we look at segment by segment minute by minute, how accurate these devices are, it also tends to look a lot poorer compared to if you just sum up everything overall, especially as we started diving into sleep stages, light sleeping, deep sleep versus REM sleep, the accuracy is all over the place where we might say for the lighter stages of sleep, accuracy is probably mid-sixties, two to 70%, give or take. Being able to accurately say this person is in a lighter stage of sleep for deep sleep in REM sleep, it’s typically in the 50 to 60% range, which means if I’m in deep sleep, there’s a chance that my device will accurately say I am in deep sleep or it’ll say I’m awake and same in my sleep or something else instead.

As we get a little bit more finer green detail from these devices, that’s when they tend to be a little bit less hopeful. So that matters a great deal as well. If we look at some other features, like, say, sleep efficiency, how much time was spent sleeping in this window of time where recording was happening? It tends to be a little bit more helpful. They are pretty close, within a few percent, compared to a gold standard sleep setting. that’s helpful. The broader view you take of your sleep, the more helpful and accurate these devices can be, and the more detailed you try to get with them. That’s when they start to be a little bit less helpful.

Audrey Wells, MD

There is so much good information in your answer. I just want to attempt to summarize what I’m hearing in three tiers. If you’re looking at something broad like how much sleep a person gets, these wearable devices, which are on your wrist or your finger night after night, are 90% more accurate compared to an in-lab sleep study, which at this point is the best measure. But we’ll probably even improve on that in the future. The second tier is sleep-onset. When you fall asleep and wake up in the morning, they tend to be pretty accurate in determining that. But in between, when these trackers are making a stab at whether you’re in light sleep, deep sleep, or REM sleep, that’s when things start to get nebulous. I feel like there are a lot of marketing efforts that go into these wearable devices. But you’ve given a nice perspective on how to process the data that’s coming out of it. One thing that is a factor is that most of the time, and possibly all of the time, when these types of trackers are studied in populations of people to develop the algorithms that they use, these populations of people are not representative of the average person buying them. In other words, they tend to be young and healthy people.

Joshua Lennon, MD

As a researcher, this sounds like the ideal population. How am I going to make my product look the best? It also depends on who you think your market is. If I’m marketing to someone who’s going to shell out a few hundred dollars for a nice accessory, that might be a certain segment of the population. Whereas, my grandfather is probably not going to pay for a Garmin watch, an Aurora, or something like that. That’s, in their defense, one caveat. But, you know, most of the time, I should say we don’t know what population they’re basing their initial data on for these things to create their algorithms in the first place because this is all black box, proprietary technology.

We don’t have a peek behind the curve to know what the Oura ring is versus the Apple Watch versus the Samsung Galaxy or anything like that. When we look at the validation studies, most of them are done by third parties or by university labs. Most of them are using, again, healthy young subjects. You take a bunch of college kids who want some extra credit or a few bucks in their pocket. They’ll do these studies to try to figure out whether this is accurate or not. Some of the validation studies will also include a clinical population, which is, again, not necessarily representative of the general population or peers—people who are seeking out sleep care. here’s you. People are at the peak of their physical fitness, and then you mix those two and hope that somehow that averages out to a typical American adult, which is not necessarily the case. But we do see pretty significant discrepancies between the ideal user and people who tend to be a little bit less healthy.

Some of these studies will look at discriminating between normal weight and those that have a BMI that’s higher than 25 and are considered overweight or obese. There was a slight difference in accuracy between these two weight populations. You can also split based on those who either have normal or maybe mild disordered breathing during sleep versus those who have moderate or severe disordered breathing during sleep. Again, when we start to see some discrepancy between those two, and then most significantly, as we see the largest chasm between populations, are those that essentially have a normal sleep efficiency or those that don’t have insomnia, and those who are spending a lot more time awake when they are in bed, either on purpose or those who are struggling to fall asleep in the first place.

What we tend to see is that,  when it comes to how accurate these devices are, they’re pretty good at confirming when you are definitively asleep, but what they lack is confirming when you are awake because a lot of them rely on some combination of motion as well as the heart rate changes. When you are still, calm, and resting but wide awake, these devices have a hard time distinguishing that restful, weak state from sleep. The more time that somebody is awake, the less accurate the device has become. For individuals with insomnia, we see that the accuracy drops, say, from 90 to 93% down into the low seventies. For those who are typical normal sleepers and don’t have any trouble staying awake when they’re trying to be asleep, we tend to see an accuracy closer to the mid-nineties or higher for these devices.

There are some differences there between even just normal adults who might have different,  health conditions, whether it’s sleep-related or weight or anything else. We also don’t necessarily see these devices undergo any validation for older adults children or teenagers. So, if you’ve got,  a 14-year-old at home who was hoping for, his next birthday turn to get one of these fancy devices, you just keep in mind that the the numbers spat out from these apps about your sleep scores and this percentage of that have not been validated for that particular population doesn’t mean it’s not going to work. We just don’t have any data at least available to say whether that’s accurate in a teenage population or if a child is compared to a normal adult. Likewise, if you are,  aging gracefully and you want to give a bit more insight into your sleep quality, we don’t necessarily have any data to show that these devices are just as accurate in the 60-plus age group as they are in the 25-year-olds. You are currently attending college triennial, but extra credit for,  involving getting involved with these studies.

Audrey Wells, MD

It’s so interesting because, to your point, people who are aging gracefully are seeing their non-REM three or slow-wave deep sleep go down as a normal effect of aging. This is one of the components of sleep that wearable devices purport to measure, and I’ve had several episodes of insomnia. People with insomnia just come in and get focused on that data and wonder, How can they improve non-REM three sleep or slow wave deep sleep? How can they take action based on what this information is telling them? How do you advise people?

Joshua Lennon, MD

One, when I trained, what we were taught was that, hey, this is just a byproduct of getting old—that your sleep was terrible. so you just have to expect that. There’s been some information that’s come out since then that, essentially, those older studies that show non-REM three or slow wave sleep or deep sleep, whatever you want to call it, decline precipitously with age. It turns out that it is based on clinical data. It was patients, not normal subjects—this difference between someone seeking medical care versus a normal adult—that new patients were demonstrating a sharp decline. There was a paper that came out in the Lancet Respiratory back in 2019, a meta-analysis of all sorts of controls across clinical studies over several decades, ensuring that if you are not a patient and don’t have a sleep disorder, the change in deep sleep over time is minimal.

Something like 1.3% per decade. When we were trained, we were told the wrong thing. Yes, we do see a decrease in deep sleep with age, but if nothing else is going on, it should be quite insignificant, should be quite small. What we see day to day is that there are a lot of people with insufficient deep sleep. Either this, non-REM three or the third stage, non-REM sleep or deep sleep or somebody or whatever you want to call it, as well as decreases in REM sleep. These have implications, if that’s true, decreases in these particular stages of sleep are associated with all sorts of problems, from metabolic to mental health to cognitive and beyond. So if that is true, then yes,  individuals should be concerned. so, what can we do about this? Well, several things can be done. for the first, we need to find out what is possibly interfering with my normal development of these stages across the night. So this can be anything from,  a lot of medicines from some blood pressure pills to,  most medicines that people take, for mood or anxiety, the medicines that we might take for pain.

There are all sorts of pharmacological agents that can impact the depth of sleep and whether or not we can achieve enough or sufficient REM sleep as well. That doesn’t mean just because your REM is a little bit lower than you would like to see that somebody’s going to stop taking all your pills. It is just we understand that there can be a direct effect of these medications. So that needs to be taken into account before you stop any drugs you do talk to the prescribers and find out,  what are taking this for how long you have to take it, what timeline is there, make any adjustments if needed, and if you are concerned about the other effects of this agent on your sleep, try to find out, is there some substitution that we can make where I’m not going to have this effect? They seem to think looking for those interferences, is helpful. again, being very cautious about making any rash judgments about changing a medical regimen. Number two would be to look for things that you might be doing daily that could also interfere with the normal development of these deeper stages of sleep across the night. The two most common, I’m sure, and know people during the summer are going to hear this in almost every single session in this, notorious notion of sleep hygiene with the two most common substances that we talk about, alcohol and caffeine, because both of these substances can significantly interfere with the normal development of deep sleep and REM sleep at night.

One thing to keep in mind is that caffeine has a pretty long half-life. So just because you said, it’s been hours since my last espresso or since I’ve been chomping down on some chocolate, the effect of caffeine, whether from an energy drink or coffee or tea or chocolate, lasts for many, many hours. Again, this doesn’t mean you have to stop eating chocolate; it doesn’t mean you have to stop drinking coffee just to understand that these substances can have an impact on the depth and quality of your sleep. If you are concerned that your watch is telling you that you’ve not gotten enough deep sleep, that’ll be one thing to look at to find out how much caffeine I am getting throughout the day.

When am I getting the exposure to that caffeine? is there a signifier enough interval between my last caffeine consumption and when I’m trying to go to sleep? Because that can interfere with things that the other major thing that we see not as necessarily interfering with the development of deep sleep, but something that’s lacking potentially as we age, that’s not necessarily there in younger adult years or even children, which is how physically active we are. We know that one of the biggest drivers of deep sleep specifically is how active we are,  how much removed, how much energy they were spending, and how active our brain is. The more that our body is doing, the more and the stronger that drive to get sleep in general and deep sleep specifically becomes. 

So as we age and we’re retired and we’re not, commuting any longer, we’re not running around the office or chasing around young children any longer. We’re just less physically busy and even if we’re still engaged in a lot of things, but if we’re just less physically busy, we’re not creating the physiological drive in the brain that would push us into these deeper stages of sleep. It’s not just what is interfering potentially with our deep sleep, but what is it that we could be doing that should be pushing us into deeper stages of sleep that we might need to take a closer look at?

Audrey Wells, MD

Great explanation. You’ve just given people actionable ways to increase slow-wave deep sleep, which increases the amount of physical activity, even mental activity during the day, and also take a look at caffeine consumption, which is so pervasive in our society, and just recognize that because of the long half-life of caffeine, even if you are taking caffeine in the afternoon or evening, it could still be impacting your ability to get slow, wave deep sleep. One of the ways that I  look at this is with the pendulum swinging, in the daytime, you want to increase your activity, your social connection, your light exposure, and your mental effort, if you will. then as a result, when things swing the other way at nighttime, you’ll be in those deeper stages of sleep. You’ll have a better ability to stay asleep throughout the night and wake up feeling refreshed. The amplitude of that pendulum has a lot to do with your sleep quality.

Joshua Lennon, MD

Brings me back to my old physics days of, potential kinetic energy and translating between the two and activities are potential energy for sleep. your deep sleep is your actual kinetic energy for sleep. If this is a concern for somebody, if you say, well, whether I’ve heard from somebody else or I use one of these things once and you’re trying to figure out, what is a potentially good tracker for me, this should be considered because if you say, well, all I care about is just my sleep, then,  say one of these near blows, or when absent is only looking at sleep, could be a fine option. But if you’re trying to figure out, how there is a way for me to tie in my activity levels during the day, my heart rate changes during the day with the depth of my sleep or the quantity of deep sleep I’m getting at night.

Then a a wearable that you’re wearing all day long might provide a little bit more insight in that regard because they can get things like your step count and your burn calories and and things like that that you can use to try to compare, on this day when I was a little bit less active, it looks like my sleep was not quite as deep or as plentiful as this other day when I was running around like a mad person all day long. then, lo and behold, ended up,  sleeping like a baby that night. look how much better it was on this night comparatively. That information can be helpful. so if if that’s one of somebody’s concerns about their sleep, then,  a technology that you can involve activity tracking, not just sleep tracking, might be potentially more helpful.

Audrey Wells, MD

 I love that too, because you’ve just named the two reasons why these consumer consumable consumer wearables are useful. They elevate sleep into our daily discussion and give us something to get feedback from. They’re also tracking sleep night after night after night for the same person. When you do enact a change, you might see that play out in the data that your wearable is reporting.

Joshua Lennon, MD

They provide a little bit more context, say just an isolated sleep tracker alone. so that’s one of the benefits. 

Audrey Wells, MD

Are there differences between men and women when it comes to wearable devices?

Joshua Lennon, MD

That’s a great question. One would think probably with the data that we have so far that it is negligible and again,  these studies, range from,  a couple of people to, almost 100 people. then we have a meta-analysis. We combined all the data and looked at them. But as far as we can tell, there’s no significant difference in the accuracy for any one of the individual devices or software for men versus women. So I don’t think that’s necessarily a factor to consider, this one only gained four men in this mission, only is on the women or vice versa. A lot of the wearables specifically, most of them come in a variety of, colors, shapes, sizes, different bands, everything else. so even if the actual sensor itself,  can’t be swapped out, everything else about the look of it can be more customizable. It’s not like you have to get,  generic black or have to get this one that looks a bit too feminine or this one looks a bit too masculine, that there are quite a lot of options for most of these wearable devices.

Audrey Wells, MD

Good points. That was a factor when I was shopping for a wearable for myself. What is the aesthetic of the device? I want to pivot a little bit and talk about Mira balls, which encompasses a smart bed and other things to measure your sleep, not wearing something on your person. First question: are smart beds worth it?

Joshua Lennon, MD

It depends. The most unsatisfying answer. Smart beds come at a premium compared to your standard mattress box ring, your frame set.  if you look at your sleep as an investment, that might be worth it potentially. It depends on what you’re trying to get out of your bed. If the thing that you’re just trying to get is sleep tracking, it’s probably a waste of money to spend that extra premium on a smart bed compared to just any other bed. If you’re going for comfort. Some of the smart beds are designed not smart, as in how well they are at tracking your sleep or predicting anything else, but smart in the sense that they can respond to your sleeping to try to improve it in the moment itself. For instance, the common thing that they might do is they might change the firmness level. Rather than just having a little remote in your dial-in, from one to 10, this is stiff and this is soft. But ones that will adjust across the night—a certain degree of firmness when you’re lying on your back—can sense that your hip is positioned, so now you’re on your side, it might adjust the firmness level slightly because of that position. Changing that adaptability can be helpful to improve the comfort of the bed. Because we are not just some static creatures and we have a lot of dynamics with us, especially when we’re sleeping, the firmness level that we find comfortable is going to change depending on what position we’re lying in as well.

That can be quite helpful. They can also adjust the temperature across the night so that we might be a bit more sensitive to hot or cold at different times of the night and at different times of the morning. Some can even use temperature as an alarm. You want to wake up, you don’t want to wake up your bed, partner. You want something buzzing, vibrating, and making lots of cleaning noises so you can wake yourself up. You can have your bed, crank up the heat and it’s that extra heat in the morning on your side of the bed that could wake you up. those features of smart beds,  you’re interested. But again, that comes down to whether is it worth getting a smart bed for sleep tracking or it’s probably not. But if you want features of a smart bed that are designed and intended to improve the comfort of the bed itself, not just, does it feel good to lay down what you do that little you jump on the bed and when these mattresses store to figure out can I do this or not but is it going to adapt to me as well? Potentially some of these beds could be worth it.

Audrey Wells, MD

It’s interesting because when you’re talking about the temperature being here in Minnesota in the Wintertime, having the temperature crank up as a signal for getting out of bed and being awake sounds heavenly to me.

Joshua Lennon, MD

There’s a word of caution here because we have some mattresses that can be adapted to the temperature. Other devices are worn, like little slip covers, to cover any mattress, no matter who the maker is, that will circulate fluid to adjust the temperature. as a mattress cover. The issue there comes down to whether uniform heating or cooling is necessarily great for your sleep. You’re going back to this idea of being a dead horse with sleep hygiene. One of the things that we will tell people to help them fall asleep is that if you take a warm bath before you go to bed, that can help you.

When we say baths on purpose, we don’t mean like you go clean yourself because you’re dirty. Don’t take a shower, we lay in a tub of warm water because what that does is take all your peripheral blood vessels and dilate them because you’re in this hot environment. Your skin says we’re going to get rid of all this extra heat. So we dilate all the peripheral vessels as we’re trying to dump heat away from the body into this water. Then once we get up, because we’ve been preheated, essentially, it’s easy for all the heat from the body to escape to our course, skin, and everything else. So we can cool ourselves down by setting ourselves up first. In contrast to what we see, if you take somebody who’s at their normal temperature and normal blood dilation levels and put them on a cooling surface, what happens is that all the blood vessels on the skin start to constrict.

Something is called we don’t want to lose all that heat. So we try to retain heat. What you end up doing is preventing the core body temperature from dropping down, which is the main signal for falling asleep in the first place. To achieve a deeper state of sleep, cool down the core body temperature by a few degrees at peak across the night. If you’re just lying on it on a bed of ice, then what you’re doing is going to constrict all your blood vessels and retain more heat than you ought to. So this is where wearing heavy socks or layers of socks can be quite helpful. You wear mittens and socks in bed because you don’t want to lose heat through your hands, feet, or forehead.

You just want to lose heat from the core. That can be helpful for some of the smart beds, rather than just having the whole bed change temperature uniformly. It might change just in the center, and then down where your feet are lying, it might heat it. You might have a wider temperature differential between your extremities where you don’t want to be losing too much heat from in the core, where you want to cool things down, and you don’t necessarily get that with a lot of these mattress cover coolers. Even though the idea of cooling your body temperature down improves sleep, it helps you fall asleep faster, and you get better, deeper sleep. But the method matters a lot as well.

Audrey Wells, MD

It’s so true. One of my favorite tricks is taking baths at night and explaining to people how that works to pull down the core body temperature. Sometimes,  speaking personally, if it’s time to go to sleep and my feet are cold, cold feet equal no sleep. So sometimes a heating pad or one of those microwavable rice. Yes. Pillows can be helpful for that. We’ve just learned. Or a smart bed. I’ll have to start a little bit of a savings account for that. We covered a lot of ground here with wearables and Mira balls. There’s some great actionable advice for anybody who wants to better interpret the data that they’re getting. Sleep stages are not quite as accurate as we’d like at this point, but still paying attention to your sleep and watching that information daily can result in positive behavioral change. That’s why I would hope for people who are interested in this topic. Dr. Lennon, as we close down, I want you to tell people where they can learn more about you or even work with you if they’re interested.

Joshua Lennon, MD

You can find me at wellrestedmd.com, and I’ve got a podcast. I’ve got a course on cognitive-behavioral therapy for insomnia. If you want to work with me professionally and you’re in the Mid-South area, come and find this practice called the Neurology Clinic Sleep Center. We’re located in Memphis, Tennessee. We’ll be happy to see you in that way as well.

Audrey Wells, MD

Fantastic. Thank you so much for speaking with us today.

Joshua Lennon, MD

My pleasure. Thank you so much. Dr. Wells.