Circadian science Sleeping troubles

Visualizing MESA, pt. 1

One of the things we’re interested in as scientists is what longitudinal, large-scale data collection can tell us about sleep. Along those lines, one of our research projects involves looking at how models of circadian rhythms, as well as different sleep regularity metrics, can help us understand different outcomes for different folks. And as part of all that, we wanted to make some pretty pictures.

We recently teamed up with Ryan Rezai, a data scientist and student at the University of Waterloo, to visualize some data from the Multi-Ethnic Study of Atherosclerosis (MESA). All the plots below were made by Ryan to showcase some of the high-level properties of the MESA dataset. We think that with beautiful, interactive datasets, the nuances of big data stories, like in MESA, can become a lot clearer. Let’s dive in!

Total sleep and outcomes of interest

How much does how much sleep you get correlate with your perceived sleepiness? Below shows average nightly sleep (from wrist actigraphy—a way of measuring sleep based on how much you move) and responses to the Epsworth Sleepiness Scale (ESS). Higher numbers mean more sleepiness. Right away there’s something interesting: this plot isn’t a line that starts high at low levels of sleep and goes straight down:

Instead, we see the famous U-shaped curves of sleep research. Short version of what we mean by that: lots of bad things (like sleepiness) are correlated with both short amounts of sleep and long amounts of sleep. One natural thing to think is that, for the extreme long sleepers, there’s something underlying both the bad thing and their propensity to sleep a lot (for instance, if you’re sick, you may sleep more and also generally feel sleepier all around). But it’s certainly the case that other things could be going on too, which is why it’s helpful to consider each case of a U-shape in isolation.

We’ll come back to that in a minute. In the meantime, here’s a case of a curve that looks… pretty flat:

This graph is showing total sleep from actigraphy, plus the fraction of respondents who said they had a diagnosis of a sleep disorder from a clinician. This is actually pretty weird: why is it highest for people who seem to have a lot of hours of sleep?

Our theory: This could be one of the pitfalls of wrist-acceleration sleep tracking. Any time you’re trying to measure someone’s sleepiness from their wrist, you run the risk of mistaking “them being very, very still but awake” for “them being asleep.” In this case, it may be that the people with the large amounts of sleep recorded are simply trying to fall asleep for longer (and staying mostly still for longer), but not actually managing to do it. That could mean that their recorded sleep is quite high, but their actual sleep is much lower. It could also be another explanation for what was going on with the U-shaped curve above: maybe some of the long sleepers in that plot weren’t true long sleepers.

Here’s one more interesting total sleep duration tidbit. The number of apnea events (from a night of polysomnography, or PSG) is much higher in people who were usually found to sleep a pretty long duration (around 10.5 hours):

You might expect this to be close to a straight line (with habitual short sleepers having shorter PSG nights and fewer hours in which to have an apnea event), but it doesn’t look particularly linear at all at the end there. Another suggestive hint that some people actigraphy is picking up as sleeping for a very long time probably aren’t sleeping very well after all. (And maybe also an indicator that there weren’t that many 10.5 hour sleepers in the dataset; see: widening standard error bounds around the line).

So let’s talk about actigraphic sleep

Disadvantages of using an acceleration-based device to track your sleep: It’ll probably mistake a lot of “awake but still” time for sleep. You’ll have questions, like the ones we had above, about how accurate it is for people who are immobile for a long time. Comes with the territory.

Advantages: it’s objective. It doesn’t care about being judged for saying you only slept two hours last night.

With these pros and cons in mind, a natural question to have is: how does an objective measure like actigraphy compare to subjective measures like “asking people how long they slept last night”? For starters, if they were perfectly correlated, you’d expect them to track along the line of slope one.

Spoiler alert: they don’t.

The orange curve shows self-reported versus actual sleep, while the gray line shows the line of slope one. Those lines are not the same! Instead, we see people who sleep a long time according to actigraphy saying they sleep less than actigraphy expects, and people who sleep only a short amount according to actigraphy saying they sleep more than actigraphy is saying. These differences are pretty wild.

And there’s another interesting difference to explore here. Among people who say they don’t get a lot of sleep (subjectively), ESS scores are really high, suggesting profound sleepiness. For people who don’t get a lot of sleep according to actigraphy, not so much.

Moral of the story

Whenever you want to understand data, make a picture out of it first. Are our long sleepers sleeping a long time because they’re otherwise not doing well, or are they not even sleeping a long time at all? What effect does the choice of sleep measure—subjective or objective—have on our results? Only way to know is to try it out and see. And that’s what we’ll be doing in the coming posts in this series. Stay tuned!

With thanks to these resources:

Zhang GQ, Cui L, Mueller R, Tao S, Kim M, Rueschman M, Mariani S, Mobley D, Redline S. The National Sleep Research Resource: towards a sleep data commons. J Am Med Inform Assoc. 2018 Oct 1;25(10):1351-1358. doi: 10.1093/jamia/ocy064. PMID: 29860441; PMCID: PMC6188513.

Chen X, Wang R, Zee P, Lutsey PL, Javaheri S, Alcántara C, Jackson CL, Williams MA, Redline S. Racial/Ethnic Differences in Sleep Disturbances: The Multi-Ethnic Study of Atherosclerosis (MESA). Sleep. 2015 Jun 1;38(6):877-88. doi: 10.5665/sleep.4732. PMID: 25409106; PMCID: PMC4434554.

The Multi-Ethnic Study of Atherosclerosis (MESA) Sleep Ancillary study was funded by NIH-NHLBI Association of Sleep Disorders with Cardiovascular Health Across Ethnic Groups (RO1 HL098433). MESA is supported by NHLBI funded contracts HHSN268201500003I, N01-HC-95159, N01-HC-95160, N01-HC-95161, N01-HC-95162, N01-HC-95163, N01-HC-95164, N01-HC-95165, N01-HC-95166, N01-HC-95167, N01-HC-95168 and N01-HC-95169 from the National Heart, Lung, and Blood Institute, and by cooperative agreements UL1-TR-000040, UL1-TR-001079, and UL1-TR-001420 funded by NCATS. The National Sleep Research Resource was supported by the National Heart, Lung, and Blood Institute (R24 HL114473, 75N92019R002).

 Interested in using acceleration to track sleep? We’ve got a software package for that.

Circadian science

Book Of The Month (March)

When by Daniel H. Pink

Another month, another Book of the Month! Starting with its title, When by Daniel Pink hones in on just how simple the idea of timing things better really is. The question of “when” is in the same fundamental category as where, why, how, and what.

Yet, as Pink notes:

“We simply don’t take issues of when as seriously as we take questions of what.”

That’s starting to change. One of our favorite article titles ever is “Medicine in the Fourth Dimension” by Cederroth et al. In it, the authors highlight the growing awareness of how circadian biology influences drug efficacy and tolerability. At Arcascope, we think that so many aspects of health and wellbeing could benefit from taking the dimension of time into account, and that dosing time could come to be seen to be as fundamental as dosing amount.

But we didn’t just enjoy this book because it recognizes that timing matters. We also liked it for highlighting something we experience every day as we work on product: Sometimes you’re alert, and sometimes you’re not, and you should take breaks when you’re not instead of powering through. As he writes in his conclusion:

I used to believe that lunch breaks, naps, and taking walks were niceties. Now I believe they’re necessities.

Of course, you can’t always take a break when you need one most. But there should be more recognition—from employers, schedulers, and from ourselves—that human beings aren’t constant in time over the course of a day. We change from dawn to dusk, and our needs change too. Maybe that means timing light or exercise to help yourself adjust to a shift faster, or maybe it means giving yourself a break so you’re not hammering away at a problem when your brain just isn’t having it.

Timing things right is what Arcascope is about. Want to see When we tell you to do stuff? Reach out about getting on our early app access list.

Get early App access!

Circadian science Technology

Scientific Tests

Algorithms are really easy to mess up. Take your pick for how: overfitting to training data, having bad training data, having too little training data, encoding human bias from your training data in the model and calling it “objective”. Feeding in new data that’s in the wrong format. Typos, subtle typos, nightmarishly subtle typos. Your cat stepping on the keyboard when you’re out of the room.

Having done this for a while, my first impulse any time I get amazing performance with an algorithm is to be deeply suspicious. This isn’t because algorithms can’t be incredibly powerful; you really can get amazing performance if you do them right. But you can also get seemingly amazing performance if you do them wrong, and there are a lot of ways to be wrong.

The core issue, I think, is that there are so many choices involved in the making and maintaining of an algorithm, and if the algorithm is trying to do something complicated, those choices can have complicated downstream effects. You can’t readily anticipate what these effects might be (this problem is hard enough that you’re building an algorithm for it, after all), but your brain tells you that there can’t be that much of a difference between a threshold being 0.3 and that same threshold being set to 0.4. So you blithely make the change to 0.4, expecting minimal effects, and then the whole thing just collapses underneath you.

I’m saying this because, while developers on the whole have gotten pretty on board with the concept of unit tests and test coverage for code, I’m not as sure about what currently exists around tests for data science algorithms in biology, medicine, and health. I’m not talking about tests that confirm a function gives the number we expect it to based on an old run of it (e.g. asserting that f(3.4) = 6.83 because we ran it once with 3.4 and got 6.83 as the answer). Those kinds of tests basically act like a flag that something’s changed, and if you change your function f, you can just paste in the new output from f to make the test pass.

I’m talking about using a lot of data — a representative sample of what you’ve collected— in your code’s tests, to assert that some macro-property of the algorithm’s output is preserved. If you make a change to the algorithm, and that macro-property changes, your code should let you know about it. If this sounds like one flavor of functional testing, that’s because it is. But the key thing I’m arguing for is that the performance of data science-components of a product— code two steps removed from sklearn or torch or R—be tested functionally as well.

Let’s talk through an example. Imagine you’ve got an algorithm for distinguishing sleep from wake over the course of the night. In my own experience, algorithms of this kind can have an unfortunate tendency to flip quickly, in response to a threshold being changed, from thinking a person slept a lot over the night, to thinking that person barely slept at all. There are ways to address this, and it’s not always an issue, but it’s still something to look out for.

So how could you look out for it? Take a representative sample of data from a group of sleeping people, and add a test that runs your sleep algorithm on all of them and asserts that everyone gets detected to have at least some baseline amount of sleep. This way you don’t have to worry that changes to the algorithm made some people better but other people unexpectedly, dramatically worse. You can rest easy because assurances of this kind are built into your development pipeline.

At Arcascope, a lot of our scientific tests center around the mean absolute error of our prediction of melatonin onset. We want to predict when somebody’s melatonin onset is happening so we can map it to other quantities of interest: minimum core body temperature, peak athletic performance, peak fatigue, you name it. But we also want to do continual development work on these algorithms, to keep bringing the mean absolute error down over time. How can we make sure that our changes actually make things better? How can we make sure that a change that improves performance for some people isn’t sabotaging others? Tests that confirm the properties we care about are preserved, no matter what we do in the backend.

I mentioned above that I don’t know what’s out there elsewhere in digital health, and I don’t. Maybe a lot of people are writing tests of this kind in their code! All I know is that when we realized we could do this— add fundamental performance checks for the algorithms that make up our backend, embedding deidentified human data from our studies directly into our testing suite— it was a very cool moment. It helps me sleep a lot better at night knowing those tests are there.

Which is good, because sleeping more means I’m introducing fewer nightmarishly subtle typos to our code. Wins all around.

Circadian science Lighting Sleeping troubles

Official Company Stance on Permanent DST



(To hear our actual stance on permanent DST, check out this blog post. Short version: we love getting rid of the seasonal time change, as long as we end up on permanent standard time, not permanent DST.)

Circadian science

Book Of The Month (February)

Sync by Steven Strogatz

For our second book of the month, we picked Sync: How Order Emerges from Chaos In the Universe, Nature, and Daily Life, by Steven Strogatz. This might seems like a bit of an oddball choice for a company that’s working on consumer apps in the health space. Why pick a math book if you’re a sleep, circadian rhythms, and well-being start-up?

We’ve got our reasons, but before we get into them, let’s back up a little. There are a lot of things in life we think of as incontrovertibly rhythmic. Walking, for instance. That’s rhythmic: there’s a beat to your steps. Swinging on a swing is another. Breathing, heartbeats, dancing to music—it’s weird to think of these without a rhythm. More bluntly, if these things don’t have a nice, clear rhythm, odds are pretty good that something’s pretty wrong.

For some reason, though, we don’t seem to care about the rhythms of our sleep. There’s this hyper-focus on eight hours of sleep a night, and nowhere near enough focus on when those eight hours are happening. Analogy time: imagine you’re listening to a weak, erratic heartbeat. You wouldn’t say that everything was fine, just so long as a certain number of beats happened each minute. You’d care that the rhythm was off.

Rhythm is a fundamental property of our bodies and our health. Literally fundamental: you can write down equations to describe how molecules at the smallest scale interact in the body and have rhythms arise spontaneously from the physics of how they bind and bounce off each other. And in the messy, chaotic conditions of the real world, rhythms often try to match up with other rhythms. There’s something very foundational about synchronizing.

Or, to quote Sync:

“For reasons we don’t yet understand, the tendency to synchronize is one of the most pervasive drivers in the universe, extending from atoms to animals, from people to planets.”

Your internal clock tries to sync up with the rhythms of the sun. The rhythms of the clock in your stomach try to sync up with the rhythms from your brain, as well as the rhythms of the food you eat. If the brain rhythms and the food rhythms are telling two different stories—think, two pieces of music with different tempos playing at the same time—your stomach clock can struggle to find the beat.

Modern life pretty much makes it impossible for us to keep all our circadian systems in sync: There are going to be times when you have to stay up late; when your work, or life, or just being really hungry one night make it so your brain falls out of sync with the sun and your stomach. The solution isn’t never losing synchrony: it’s recovering it quickly whenever you do.

So why did we pick Sync for our book of the month? Well, #1, we’re fans of Steven Strogatz and applied math in general. Reason #2, we love how it calls attention to the fact that rhythms aren’t some optional add-on to life; they’re at the very root of it.

Or, to quote the book:

“[T]he capacity for sync does not depend on intelligence, or life, or natural selection. It springs from the deepest source of all: the law of mathematics and physics.”

As for Reason #3? There are just some beautiful pieces of writing. Take this one:

“Synchronized chaos brings us face-to-face with a dazzling new kind of order in the universe, or at least one never recognized before: a form of temporal artistry that we once thought uniquely human. It exposes sync as even more pervasive, and even more subtle, than we ever suspected.”

When people stop thinking of sleep as something to count, and start thinking of it as one instrument in the complex orchestra of the body’s rhythms, we think they’ll feel benefits they weren’t expecting. Our bodies are hardwired for rhythms. Let’s bring them into sync.

Circadian science Technology

Biophysics for Better Living 2.0

In my last blog post, I talked about the power and potential for biophysics engines to contribute to clinical care and ragged on my own ability to play video games. The short version of it, if you don’t have time to circle back: If you’re going to make a digital twin of somebody, base it on a system of differential equations that captures how the physics of the human body work. Also, I’m terrible at video games.

I talked in that blog post about how you can use a biophysics engine—a.k.a. known properties of how the body works, codified into equations— to estimate things that are hard to measure because they’re hard to reach. Things like the firing rates in your ventral and dorsal suprachiasmatic nucleus (SCN), which are tucked pretty deeply inside your brain. You’re not going to be able to easily peek and see what the SCN is up to (at least not right now, in 2022), but you can figure out what your SCN is likely up to passing the same inputs your real SCN got into a model of the SCN.

Woke up really early and got a blast of light? That probably sped up your clock. Stayed up late and kept the lights on overhead? That probably slowed your clock down. Where did your clock wind up at the end of it all? That’s a question that hinges on the details of when, how long, and how bright your light exposures were. In other words, it’s a question for the mathematical model. 

But there are other questions you can ask a model, beyond “where am I now?” You can ask “what if?” What if I were to stay up the entire night in bright light? What if I dimmed the lights at 3:15 pm? How would these changes affect the state of my internal clock? And how would that changed state affect downstream outputs, like the timing of my peak fatigue or my peak performance?

When we talk about what we do at Arcascope, we say we do circadian tracking and circadian recommendations. The tracking is using a biophysics model to answer the question “where is your circadian clock right now?” The recommendations come from repeatedly asking the model “what would happen if you did (insert series of behaviors here)?” Asking a lot of questions gets you a lot of answers, which you can then pick from based on which ones best meet your goals, whether that’s sleeping more, adjusting faster, or being maximally alert at a specific time. 

It’s all still biophysics models, but in one use-case, we’re putting them to work to capture reality, while in the other, we’re trying to capture a whole swath of possible futures. A list of potential realities to choose from as you scope out the rest of your week. Better living by picking the best answers to “what if?”— I think that’s pretty neat. And it keeps me up at night thinking—what if we could do this for other systems, besides sleep and wake, too?

Interested in beta testing our app? Send us an email!

Circadian science Technology

Biophysics for Better Living

One of the highest stakes moments of my life was the time in graduate school when I was (badly) playing Donkey Kong at Pinball Pete’s Arcade, and a very, very good Donkey Kong player came over to watch me play in silence. I’d just spent six dollars worth in quarters dying repeatedly on the first floor of the first level, but something about his judging, appraising eyes summoned the fireball-hopper within. I made it past the first level, then the second, then the third—an absolute record for me—and I had almost cleared the fourth when the pressure finally got the best of me. Still, it felt like a tremendous achievement. I like to think the good player and I shared a nod of mutual respect as he elbowed me out of the way and proceeded to spend the next forty-five minutes on an uninterrupted victory run. 

Besides letting everyone know that I need to get out more, this story is an opportunity to talk about a kind of math that should be more integrated into clinical care, and isn’t— yet. There was a model of physics deciding what happened in that game, a system of differential equations for converting my button mashing into movement across a screen. And just like the inputs I gave the system could be transformed into outputs that captured the physics of motion, so too can the inputs we give our bodies be transformed by a realistic biophysics model to give us outputs that capture the ways our bodies work without the need for invasive tests. 

Let’s back up: At a very high level, what’s going on in a video game is that you’ve got prescribed rules for how to update all the positions of all the moving pieces from one frame to the next. In the next blink of an eye, where should Mario go? If I’m pushing the joystick to the left in that instant, nudge him a little to the left. If I’ve just hit the jump button, nudge him a little bit upwards. If I’m mid-jump, make the force of gravity and the speed he took off the ground fight it out to see if he should be nudged a little up or a little down. 

The whole set of rules for making these decisions on how to nudge can be called a physics engine. Passing the history of my joystick spins and button presses into this physics engine gets you an output that describes how Mario moves, how he jumps, and how far I make it in the game before an untimely demise. This type of math is all over the place in games, CGI, and user interaction elements (like views that scroll and bounce). But there’s no reason a physics engine has to be restricted to visuals, or video games, or the physics of movement. You can have a physics engine that captures ion channels open and close, for instance, or how neurons talk to each other in the brain. 

Which brings us to your body’s internal clock. Circadian rhythms are a massive pain to measure experimentally; it’s one of their defining properties. It’s hard to know what’s going on in the brain without looking at it, and it’s hard to get a glimpse of it without getting pretty invasive. But if you know the rules for how the clock part of your brain works— “the differential equations describing the suprachiasmatic nucleus”— you can pass through a history of inputs to the system, and get out an output that describes how the neurons are communicating with each other at a specific point in time. Pass in what my lighting history has looked like over the last four weeks, and I can tell you what time my brain would cue melatonin to start rising in my body tonight. 

So here’s why I think we should be using differential equations in medicine more: they can help you see things you otherwise wouldn’t be able to see. You can use a system of differential equations to track how the brain would react to the inputs you give it, even if you can’t readily experimentally measure brain state, just how a person with access to Donkey Kong’s code could track how far I got on my amazing run, even if someone blocked off the console screen with a tarp. Knowing the rules means you can measure things in the dark. 

And there’s another reason, too—personalization. By tweaking the rules of your video game physics engine, you can change the way the gameplay works. Cut gravity in half, and suddenly Mario’s soaring with every jump. Double it, and he’s a sitting duck for everything Donkey Kong throws at him. As the importance of individual differences in light sensitivity and fatigue management become better understood, biophysics engines are ready-made for customization. Want to better fit the data of someone who’s hyper-sensitive to light? Crank up the light sensitivity parameters. Got genetic data on them? Plug that into a molecular model of their rhythms

Biophysics engines can be the key to a true digital twin, helping us look inside a simulation of ourselves in ways we can’t look inside our real selves. There’s potential for these quantitative methods in medicine beyond sleep and circadian rhythms, too. Who knows? Maybe someday a model of this kind could be useful for understanding how the (fun) stress of my epic Donkey Kong run affected my mood and memory. Or at the very least, how my body processed the enormous bubble tea I drank in triumph right after. 

Want to learn why we don’t think the answer to circadian estimation in the real world is to just throw an artificial neural net at it? Check out this blog post by CTO Kevin Hannay. 

Circadian science Sleeping troubles

Naps, part one.

The first thing I want to say about naps is that I’m almost always for them. Naps can help you recover from sleep deprivation. Naps are good.

But let’s talk about that almost always. When might you want to avoid napping? 

Well, maybe you’re trying to shift your personal time zone and are at a point in your internal circadian day where getting light exposure will be very, very helpful to achieving that shift. Closing your eyes to take a nap will block photons from reaching your retina, which means your brain won’t have the photic momentum it needs to push through a shift in your rhythms. Probably not a big setback if the nap is short, but a multi-hour nap at the wrong time could end up slowing down how quickly you adjust.

Or maybe you really, really need to be alert right at the moment when you’d be waking up from a nap. In that case, you might worry about sleep inertia, the phenomenon of general grogginess and impaired performance that can persist for several hours after waking. This, too, might make you want to hold off on a nap. 

And then there is the classic “I napped, ergo, I cannot sleep now.” Your transition into sleep is driven both by your circadian clock as well as your accumulated “hunger for sleep” (or sleep drive). Feed that hunger for sleep right before bed, and you might not have enough sleep pressure built up to flip your switch from on to off. This is one of the reasons why avoiding naps in the evening is a common component of cognitive behavioral therapy for insomnia

So: You’re sleepy. You’ve got other stuff to do today. Do you take a nap? If so, how long? 

Answer: You probably want a ten minute nap. 

This, like all science boiled down to a single tidbit, is a big ole simplification. It matters what your internal time is (does your body think it’s day or night?) and what your recent sleep/wake history is. 

But multiple studies have found that a 10 minute nap during the day improves performance right off the bat, while longer naps mean that you have to sink time into recovering from your nap after you wake up. A 20 or 30 minute nap in these studies was still found to be better than staying awake, but participants could still be shrugging off the effects of sleep inertia more than two hours after waking, while a 5 minute nap was generally not enough for much of an effect.

Would you ever want to take a longer nap? You might, if your goal is not so much “perform better for the next two hours” as it is “don’t fall asleep in the next ten hours.” In a classic study from 1986, researchers kept subjects up all night, let them take a morning nap, and then measured how readily they fell asleep at different points over the rest of the day. Here, a 15 minute nap was barely better than no nap at slowing down how rapidly people fell back asleep, while a 60 minute nap had alerting effects that persisted 4 to 8 hours later. The benefits didn’t keep increasing past 60 minutes, though: a 120 minute nap didn’t get you anything more than a 60 minute one did. 

Another reason to consider a longer nap is memory. People who get a 60 minute nap do a better job at remembering words they’ve been exposed to than people who don’t get a nap. That said, a 6 minute nap is also enough to see a significant memory boost— for one tenth the time investment. 

In conclusion: You probably want a short nap. You might want a longer nap, though, if you don’t need to be super alert for the next few hours, but you do need to stay awake later in the day. 

Lastly, you might benefit from a nap, even if you don’t think of them as particularly helpful for you. The benefits of napping that show up in objective reaction time tests often aren’t reflected in how people subjectively rate their own sleepiness. You might get more of a boost from naps than you think, and you may also need less of a nap than you’d expect to see that boost.

Much of this blog post was helped along by this review. Thanks to the authors for the great resource!

Circadian science Sleeping troubles

Take a Break (From Social Jet Lag)

Why do you go to sleep when you do? 

Sure, there’s a big part of it that’s physical: You go to sleep because you’re sleepy. But you might also stay awake, even when you’re on the verge of collapsing from fatigue because you have work to get done. Or because your neighbor is practicing a percussion solo at 2:00 am. Or because there’s something mildly fun happening on the internet.

It’s like you’re caught in a tug of war, with your body on one side and eighteen different kinds of social pressure on the other. When your body finally triumphs and drags you into sleep, it’s winning out over incoming texts from friends, the snare drum next door, and that interesting passage in the book you’re reading. Team Body can get a boost from a stronger circadian signal for sleep, but it can also be helped along by that responsible part of our brains that tells Team Social Pressure to pack it up and go home, since “We have to get up early tomorrow, folks.”

The absence of this internal chiding is why people tend to stay up later on the weekend than on the weekdays, which shifts their circadian clock’s timekeeping, which makes it so they can’t fall asleep early enough on Sunday, which makes them end up feeling like a right proper Garfield on Monday. This is called social jet lag, and it’s been linked to lots of bad things

But what if we had a very, very long weekend? In other words: how do we sleep when we’re on break? 

I’m going to talk about this through the limited lens of a paper I was an author on in 2017, with collaborators at the University of Chicago. We looked at how people’s Twitter activity patterns changed as a function of where they were living and the time of the year. Tweeting isn’t the greatest proxy for sleep and wake, but we can at least conclude that if someone is posting a tweet, they’re probably—though not necessarily—awake.

You can make plots like the below, where the blue lines show the (normalized) amount of tweets at any point in time over the course of the day, on weekdays (dark blue) versus weekends (light blue). 

When the lines are low, less tweets are happening then. When the lines are high, more tweets. In the plot on the left that shows February tweeting, you can see a huge difference between weekday and weekend tweeting. In the plot on the right, for August, the minimum points for tweeting (the troughs) are nearly the same.

I remember seeing this and thinking, “Ah-ha! Seasonality effects! Human behavior is changing because the sun is changing over the year!” This, I thought, would be very interesting to report on. 

Then somebody on the team (one of my brilliant co-authors), wondered if it wasn’t just that people were more likely to be on vacation in August than in February. 

Reader, it was totally that.

Or, at the very least, we found some pretty compelling data to suggest that the reason why “Twitter social jet lag”— the difference between the trough in tweeting on weekends versus the weekdays— was higher in February than in August was not because of the sun being different in those two months, but people’s social responsibilities being different in those two months. 

Just look at the timing of the weekday tweeting trough versus the week of the year (red line), alongside the timing of big K-12 holidays (yellow) in Orange County, FL:

Sure looks to me like every time the kids go on break, people’s tweeting activity shifts later in the day. (You see this in other counties, too).

It’s true that tweeting in 2017 might have been disproportionately done by younger people. But this still means that, according to this proxy for sleep and wake, their social jet lag was decreased. They didn’t have to get up early for school or work, so they didn’t, and there wasn’t much of a difference between their weekdays and weekends. 

Which is a good thing! Don’t get me wrong: society is still set up in a way that punishes night owls more than early birds. But having your sleep be more consistent, and not jaggedly interrupted by the weekend, is healthy. 

This December, as we cruise towards that glorious week between Christmas and New Year’s, the sleep of those of us on vacation time will probably shift later, like we’re in one big, endless weekend. Wednesday will look like Saturday. Sunday night and Monday morning will be no big deal. 

The problem comes when we go back. Because if Twitter activity patterns are to be believed, the coming months are some of the worst in terms of “jet lagging” ourselves on the weekend.

So with the holidays coming up, take the time to rest and sleep in a nice, consistent way. Then keep it going into January. This will mean quashing down the parts of you that push for staying up late on Friday and Saturday as much as you can. But it will also mean that one of the best things about the way we seem to sleep on break will carry forward with you into the new year: better weekday-weekend sleep regularity. 

Oh, and by the way: this sleep regularity? In some ways, it might even be more important to your health than sleep duration. But more on that after the break. 

Circadian science Sleep Meme Review

We Rate Sleep Memes

Meme #1

Here we see Squidward staying up and reading instead of going to sleep. Squidward himself might smugly point out that he’s reading a book, not looking at a light-emitting screen, and use that as an excuse to feel superior. If so, he would be tragically mistaken. There are clearly lights on in the room that he’s using to read while staying up late, and that light will have an effect on his clock much in the same way light from a screen would. After all, most homes are bright enough in the evening to significantly mess up sleep-related processes, like melatonin production. Though Squidward would never accept it, his efforts to prevent circadian disruption pale in comparison to those of his neighbor Patrick, who blocks light by being a starfish who lives under a rock. 

Originality: 3 out of 5. Not being able to put your phone down is a classic meme topic. 

Quality: 4 out of 5. Slight cognitive dissonance caused by the “scrolling through social media” text coupled with the image of him reading a book, but it gave us a chance to talk about how room light exposure matters for circadian rhythms, which is what we’re all here for. 

Meme #2

Let us start by noting that Homer’s perception of his sleep here may be skewed: many people with insomnia overestimate how long it takes them to fall asleep, and underestimate how much sleep they actually get. It may be that a more accurate version of this meme would be “me all night vs me four hours before my alarm goes off”– which is still, to be perfectly clear, a miserable experience. It’s miserable even if you’re objectively getting more than 6.5 hours of sleep per night but perceiving that you’re not sleeping much at all (also known as “paradoxical insomnia”). We love targeting sleep improvements through light exposure over here, but if you’re relating hard to this meme, you’ll probably want to get yourself some cognitive behavioral therapy

Originality: 3/5. This, too, is a pretty typical sleep meme topic.

Quality: ⅘. He looks very cozy at the end there. 

Meme #3

Oh, Leo. Leo, no. This is a terrible idea.

For starters, we know what happens to people who get four hours of sleep a night. First, they have more and more “vigilance lapses” with every passing day (4 hours of sleep a night = circles in the below, black squares = no sleep, white squares = 6 hours, diamonds = 8 hours). 

A vigilance lapse means that something popped up on a screen in front of you for half a second and you didn’t even register it. This is bad if you are, for instance, driving a car. 

People on four hours of sleep a night also fail to get better at subtraction and addition tasks, despite days of practice (see: circles staying flat in the below):

And yes, caffeine can counteract “getting worse and worse at things” to an extent, but so can naps. As the authors of a recent review on fatigue and caffeine write, “It is important for caffeine consumers to understand that caffeine at any dose is not a chemical substitute for adequate healthy sleep.” 

Originality: 4.5/5. Nice shout out to shift workers.  

Quality: 2.5/5. Inscrutable indenting decisions. Objectively bad sleep practice.