I wanted to capture what I thought was most exciting about our company.
Sure, you can sleep better by taking care of your circadian rhythms, but circadian rhythms are about a lot more than just sleep.
And yep, we’re a circadian rhythms company, but we’re one founded by people who are (primarily) mathematical biologists.
Rather than trying to figure out what time is in a person’s brain with a gold standard test (for instance, by taking repeated saliva samples for hours in the dark), we use math models to infer what’s most likely to be going on in their brain, given all the inputs that have gone into it.
Rather than trying to estimate their internal time from a single measurement or number, we use lots of data, collected on the timespan of weeks to months.
In a sense, we’re using our algorithms to see inside a black box: letting us predict what’s going on in your head, without the need for invasive testing. That’s what I think is most cool about our tech.
Sleep regularity is a gauge of how consistent a person’s sleep patterns are, based on the day-to-day variability in their sleep–wake times. In general, having poorer sleep regularity, or irregular sleep patterns, has been shown to lead to many adverse outcomes in metabolism, mental health, and cognitive performance. Low sleep regularity has even been linked to increased inflammation. In order to avoid these and other complications, you want to increase your sleep regularity by aiming to get into bed at the same time every night.
How can we score sleep regularity?
There are at least five different metrics that can be used to quantify sleep regularity, each capturing different aspects of it and useful in its own way. The five measures of sleep regularity that we’ll look at in this blog post are listed below:
Individual Standard Deviation (StDev)
Interdaily Stability (IS)
Social Jet Lag (SJL)
Composite Phase Deviation (CPD)
Sleep Regularity Index (SRI)
Traditionally, the most common overall metrics that have been used to assess sleep regularity are quantified by measuring sleep deviations in sleep patterns from an individual’s average. Examples of overall metrics are StDev and IS, both of which compare sleep from each day to an average sleep–wake pattern, and SJL, a metric that compares two average sleep patterns (workdays and free days).
StDev: lower is more regular / StDev⬇ = Sleep regularity⬆
This score is just the standard deviation of your sleep metric of choice, like sleep onset, sleep offset, or sleep midpoint. The standard deviation captures the variation of a quantity from its mean.
IS: higher is more regular / IS⬆ = Sleep regularity⬆
This metric uses sleep-wake data (can also use rest-activity data) over a period of days to measure the stability of a person’s sleep-wake rhythms. It does this by comparing the pattern of daily sleep activity to the average pattern across many days.
SJL: lower is more regular / SJL⬇ = Sleep regularity⬆
Social jet lag is a metric that quantifies the mismatch in the average mid-sleep timing between workdays and free days. Negative SJL values represent earlier mid-sleep timing on weekends than weekdays while positive values indicate the opposite.
Two newer measures of sleep regularity are CPD and SRI. These two fall under the category of consecutive metrics, which means they measure variability in sleep–wake patterns between consecutive days. The circadian system makes adjustments daily, and consecutive metrics were developed in order to utilize day-to-day information and more accurately predict circadian disruptions associated with poor sleep regularity.
CPD: Lower is more regular / CPD⬇ = Sleep regularity⬆
Composite phase deviation is a metric that was created with shift workers in mind. CPD quantifies circadian disruption where sleep is both irregular (rotating shifts) and mistimed (sleeping in daytime). This metric uses an individual’s chronotype to determine optimal timing of sleep. The chronotype then helps to quantify how “mistimed” they are. The regularity aspect is calculated using the difference between mid-sleep timing from one day to that of the prior day. In order for CPD to be derived it requires data that has one main sleep session per day or some other daily sleep variable, like sleep duration.
SRI: higher is more regular / SRI⬆ = Sleep regularity⬆
The sleep regularity index is a measure based on binary sleep-wake time series. It measures the similarity of a person’s sleep-wake pattern from one day to the next. The scale for this metric ranges from 0 (random) – 100 (perfectly regular) and it represents the percentage probability that an individual will be in the same sleep/wake state at any two time points. It’s important to note that this metric does not account for total sleep time so a person that (hypothetically) sleeps 0% of the time will still be able to get an SRI value of 100.
So I’m regular, that means I’m healthy right?
Well, not quite. Depending on the kind of variability you have in your sleep patterns and the method used to record your sleep, different metrics may tell you very different stories regarding your sleep regularity. Context is very important when making a decision about which sleep regularity metric to use.
Just think about what would happen if you increased the variability in your work week sleep timing, but maintained a consistent average. Your SJL score would stay the same, while your other metrics would likely shift to indicate greater variability. The ordering of days also matters. In Fischer et al. they shuffle days around to show how consecutive metrics can give you different stories on regularity than overall metrics do.
In order to properly assess sleep regularity for yourself or your patients, it is necessary to understand the little things that go into calculating each of these sleep metrics. A variety of unknowns, such as the type of data being gathered or the length of the data set, can cause these metrics to disagree with each other. The good news is that you’ve got lots of options to choose from.
I woke up feeling groggy and lazy almost every day in the past school semester leading up to this summer. I was already doing research on sleep at that point, and I had a general idea about what I was doing wrong, but when I tried to dive deeper into the hundreds of articles on sleep science, I just found myself getting tired and confused. The last few months working as an intern at Arcascope have taught me how to get in sync with my circadian rhythms and taught me of the many subtleties that are involved with supporting your circadian system.
When I started learning about sleep, I found myself asking friends about their habits and experiences with sleep out of curiosity. I quickly realized that almost everyone I talked to had experienced frustrating sleep problems at some point during their lives and had no idea how to deal with them because they were just never taught enough about sleep. I decided that I wanted to give them some of the knowledge that they will need to improve their sleep quality by reflecting on certain essential components of sleep that I have learned about throughout this past summer.
I want to begin this informational recap by stating what I think is the most important factor involved in getting good sleep: Get enough light and get it only when you are supposed to. Getting light, especially bright light, at the wrong times causes an advance or delay to your circadian clock which essentially means that you’re giving yourself jet lag without ever having to leave your room. Light is the strongest signal to the human circadian system and it can do some amazing things when used correctly. Bright light has the ability to affect the amplitude of your circadian clock, and if you time it right, can allow you to cross time zones faster than you otherwise would. However, this kind of entrainment schedule can be especially hard to follow given the pervasiveness of screen use at night. The presence of almost any light, particularly above 50 lux, has been shown to have a melatonin suppressing effect that can make the process of falling or staying asleep difficult.
Using light as effectively as possible is not just about timing and brightness but also about daily regularity. Previous light history has been shown to affect your circadian clocks current sensitivity to light, which makes paying attention to your light exposure important to maintain your healthy sleep habits. Poor sleep regularity has been linked to many health complications ranging from subdued cognitive performance all the way to inflammation. Recently, some new metrics have been created to quantify sleep regularity which may give clinicians the chance to make more well-informed decisions regarding recommendations for their patients’ sleep health.
The timing and brightness of light, as well as the regularity of your sleep schedule, are three components that have an immense influence on overall sleep health. My personal experience in school so far has left me lacking this key knowledge on how to promote my sleep efficiency. Learning about what to be cognizant of in regards to my sleep habits has greatly increased my alertness in the early morning, and I still have lots of room to improve upon my habits. The crux of my argument here is that just being aware of the things that can mess with your sleep will inherently allow you to avoid bad situations and improve the quality of your sleep.
This post was written by Arcascope’s intern, Ali Abdalla. Thanks Ali!
In this blog post, I want to point to some common symptoms of circadian disruption. The role of circadian rhythms in sleep is subtle, and many times issues that are attributed to other sources really have circadian factors as the root cause.
The two process model:
To understand the interaction between sleep and circadian rhythms we need to discuss the two primary drivers of sleep: the homeostatic sleep drive and the circadian drive to sleep.
The homeostatic sleep drive (called “Process H”) describes the build up of sleep pressure the longer you are awake. This drive builds up anytime we are awake, and the higher it gets, the harder it is to stay awake. In contrast, when we are asleep, the homeostatic sleep drive decreases.
Let’s imagine what would happen if this were the only component driving sleep. Then people would operate a bit like my iPhone: Active as long as your battery lasts, then asleep/recharging as long as it takes to get back to full battery… or at least until you get yanked off the charger.
If I wanted to adjust the “sleep schedule” of my iPhone I could just adjust the time that I pull it off the charger. I could also use it more (burn up the battery, turn the screen to full brightness) if I wanted to make it “go to sleep” sooner. This probably doesn’t match your experience with sleep. Being more active during the day doesn’t ensure that you will go to sleep earlier (although it can help). And staying up an hour later doesn’t necessarily mean that you’ll wake up an hour later the next morning. From experience, you’ve probably already learned that sleep duration isn’t just a function of how tired you were when you fell asleep.
This is because the homeostatic sleep drive is one of the two processes which control our natural sleep cycles. An iPhone has no issues with jet lag, shift work or sleeping on Sundays! That’s because it doesn’t have…
Circadian Rhythms: The Second Process
The more subtle process which controls our sleep cycle is the circadian clock, also called “Process C”. Circadian rhythms in humans act to help us sleep in a single block at night by modulating the sleep drive according to the body’s internal clock time. Much more about this process below.
So what are signs that your sleep issues are being driven by your circadian clock?
I wake up at 3am and can’t get back to sleep
One of the most common sleep disturbances is waking in the middle of the night and not being able to get back to sleep. Very often circadian rhythms play a role in this annoying occurrence.
We can think of the homeostatic sleep drive and circadian sleep drive as executing a delicate hand-off in the middle of the night. Let’s walk through what happens when everything is in sync. Since homeostatic sleep drive increases whenever you are awake, the hours near bedtime are when the homeostatic sleep drive is at its peak , while the circadian sleep drive is opposing sleep– or at least not promoting it. This push and pull helps keep you awake through the evening hours even if you have had an active day. This also keeps your bedtimes consistent (and in a natural environment aligned with sunset). However, once you fall asleep, the homeostatic sleep drive begins to decrease steadily, and soon it reaches levels similar to those you had during the daylight hours. So why do you stay asleep?
Well, as the night progresses, the circadian process begins to take over the job of promoting sleep. This maintains an overall drive for sleep throughout the night. Finally, around dawn, the circadian drive to sleep drops enough for you to wake up. The handoff in the middle of the night between the homeostatic sleep drive and the circadian sleep drive is what allows for one, contiguous block of sleep.
If your circadian rhythms are out of whack, this handoff can be fumbled, leading to the annoying episodes of waking in the middle of the night and not being able to get back to sleep.
Trouble getting to sleep on Sunday
Another very common sleep disturbance is “social jetlag”. This is caused by the likely familiar practice of staying up later on Friday and Saturday night and sleeping in the next mornings. This move to later light exposures tells our circadian clock to shift later, so it creates the same effect as jet lag without you ever leaving the couch.
If you stay up three hours later on Friday and Saturday and sleep in a commensurate amount, you have effectively traveled from New York to Los Angeles for the weekend– a three hour shift west. The pain comes when you need to perform the reverse trip to get back on your workweek schedule. When you try to go to sleep at 10pm on Sunday night then, as far as your body is concerned, it’s 7pm. Worse yet: this will often move the Sunday bedtime into the dreaded wake maintenance zone, discussed next.
I try to go to bed a few hours earlier and I just can’t fall asleep
This is caused by the so-called “wake-maintenance zone”: in the hours leading up to bedtime, there’s a period of time where it’s hard to fall asleep. From an evolutionary perspective, this wake maintenance zone, which would occur as the sun was setting, could have existed to ensure we’d be awake and active while we still had some light to make our way to a shelter (or into a tree) before nightfall. In modern life, bedtime is rarely sunset, and this wake maintenance zone can fall in the 9pm-11pm range.
In conclusion: Your inability to move your bedtime up by a few hours may not have anything to do with mindfulness and have everything to do with how much sunlight you got the previous morning.
Whenever I am on a break from (school/work/obligations), I end up going to bed at 3am
This typically happens when the societal constraints that are keeping the circadian clock tethered to the sun are removed. Often, someone on a spring break (and without any reason to set an alarm), will find that their schedule starts to drift later and later each night.
This phenomenon can originate from a feedback loop between behavior and the circadian rhythm. Staying up later one night will delay the circadian clock through light exposure, which will tend to move bedtime the next day later. This is compounded if you sleep in later, as you are missing the morning light which can counteract the extra evening light the night before.
This cycle keeps repeating, slowly driving the bedtime later (for me, this was something like 30 minutes each night). This progression can be curtailed by hitting the circadian wall where the circadian drive to sleep is maximal. This drive, combined with the homeostatic sleep drive which has been building up all day and night, can induce you to finally fall asleep. That doesn’t always happen, though: delay yourself enough, and you might find yourself cycling all the way back to a day schedule.
I sleep better when I go camping
Finally, one example where you may have experienced the benefits of having healthy circadian rhythms. Many people find that they sleep better when they are camping. This is especially surprising, since this typically means leaving comfy mattresses and other sleep aids behind. Personally, I can fall asleep easily much earlier in the night when I am camping, and– even if I wake several times during the night to roll over– I wake near sunrise feeling much more refreshed than normal.
A big part of this effect can be traced to the therapeutic effects that camping has on circadian rhythms.These results come from one of my favorite circadian rhythms papers which will be the subject of a future blog post: Stay tuned.
Lately, I’ve been watching clips from the Olympics, getting misty-eyed when the athletes hug at the end, and then stalking the winners on social media: a normal Saturday night in 2021. I have a shirt that looks like that, I think as I scroll through photos of them lifting approximately three times my body weight. We’re not so different, you and I, I think as they hit a five-inch target from a football field away.
Coming from the East Coast, this was a three-hour shift west. To adjust to California time, I needed to delay my circadian clock. One catch, though: my flight out was extremely early in the morning. That meant, like it or not, I was going to advance myself as I set out on the journey.
Let’s back up a little. We talk about directions your internal clock can shift as advances or delays. Think of advancing as hustling your clock along, making your circadian rhythms more like those of people in time zones east of you. Delaying, on the other hand, is like a temporary slowdown for your clock, making your rhythms more like people living to your west. Light at different times of the day advances or delays you, depending on your clock’s state when you’re exposed to it.
Shift work disorder, orSWD, is a type of circadian rhythm sleep disorder which is caused by working shifts that do not fall within the conventional working hours of around 9 am – 5 pm. These shifts overlap with periods of significant light sensitivity which can cause shift workers to be particularly vulnerable to having dysfunctional circadian rhythms.
One of the most common questions we get at Arcascope is…
“Can’t you just do circadian phase estimation using machine learning?”
Living in the data age, we have become used to thinking that big data and machine learning can do just about anything. In this post, I will break down some of the unique challenges for circadian phase estimation with an eye towards machine learning techniques. I’ll also do a brief review of the previous attempts to apply machine learning to this task.
The simple solution would seem to be to get rid of jobs that put people on work schedules that are brutal to their well-being. But society needs 24-hour emergency and healthcare workers to function, which means night shift work is here to stay.
It’d be great if there was a pill you could take to erase shift work’s negative effects. Unfortunately, there’s not much in the way of pharmacological solutions that work for shift workers at present. There is hope, however: non-pill solutions, like changing shift timings and light therapy, can offer relief to shift workers. In this blog post, we’ll cover what we know about what works for shift workers.
So you want to get somebody’s internal time from a wearable…
Let’s talk about wearable data. On the one hand, wearables are an incredible innovation, allowing self-quantification and anomaly detection with unprecedented ease, at unprecedented scales.
On the other hand, they’re a data science nightmare. Or three nightmares, really.
Nightmare #1: All the devices are different, and you have to use different ways to get raw data off them.
Sure, apps like Apple Health that act as clearinghouses make this easier for you. But you can’t use Apple Health for everything. Sometimes, wearables require permission to be granted for you to access their full data. Sometimes, wearable companies go out of business after you’ve built an infrastructure to work with them.
Can you process heart rate signals from two wearables using the same algorithm? What if they decide what counts as a “step” in different ways? What if the firmware changes? People have certainly thought about these questions, and that’s the whole point: you have to think about them. The effort of keeping track of everything adds up.
Today we had the pleasure of interviewing one of our own, Eric. He’s a skilled mathematician and AWS developer here at Arcascope. Everyone has their own unique experience when it comes to catching Zzz’s, or in Eric’s case, NOT doing that. We wanted to chat with him and learn more about these troubles and, more importantly, how he overcame him.
Let’s jump right in. How has your sleep been lately?
“Right now, my sleep schedule is pretty good, but this has changed over time and pretty dramatically over the last eight months or so. I feel like I have finally figured out some other things that my body needs for me to be able to fall asleep quickly. I have struggled to sleep for a lot of my life. I break it down to two components: One, once I was asleep, staying asleep, and two, also falling asleep to begin with– this has been the hardest part for me. I’m still not to the point where I can fall asleep straight away, but I have gotten it to under an hour.”
So, your sleep has improved recently?
“Yes, It’s a lot better. Much more regular in schedule, and it’s pretty easy to fall asleep and stay asleep throughout the night.”
You mentioned having trouble falling asleep your whole life. What has it been like, and what are the specific problems you’ve had?