Do We Need Less Sleep than We Thought? What Hunter-Gatherers Tell Us (Podcast)
We’re getting much less sleep than we used to! Well, at least that’s been the story. So how did that story develop? Epidemiological research and national polls that ask people to report how much sleep they get have suggested that, indeed, Americans are getting less sleep than we were 50 to 100 years ago, and therefore, we are likely chronically sleep deprived (Bonnet and Arand, 1995; NSF, 2008; McCallister, 2010).
In the early 1900s, Americans reported to average almost 9 hours of sleep a night, but samples of self-reported sleep times over the later part of the 20th century indicated a steady decrease. In 2008, a poll by the NSF showed that on average, American’s now spend approximately 6 hours and 45 minutes per night in bed during weekdays (and about 10% more on weekends). Compared to the early 1900s, that’s a 20–30% decline! This wouldn’t be an issue if there weren’t negative consequences from insufficient sleep, but there most certainly are. The same NSF survey also noted that individuals who self-reported 6 hours a night or less on workdays were almost 50% more likely to be obese!
Reasons we’re getting less sleep include such things as stress from work, elongated work hours and commute times, and increased screen time at night. So the idea goes that we need a certain amount of sleep, but because of a variety factors of a part of modern living, we’re getting less than even our recent ancestors, and this has negative effects on our health and performance in life. While we know that there are real repercussions to inadequate sleep, do current estimates of average sleep time represent a widespread chronic, sleep deficiency?
One approach to understanding this subject better is to investigate what “natural” sleep looks like. Luckily, we now have more information on this due to a ground-breaking study recently published by Professor Jerry Siegel at UCLA, and colleagues in the journal Current Biology. In his study, he evaluated three societies living in natural conditions (e.g., modern day hunter-gatherers) to examine their sleep behaviors and physiology. He also analyzed external factors like natural light, ambient temperature and the season in which the data was collected. And by doing so, Dr. Siegel appeared to turn the wide-spread belief that we are chronically sleep deprived on its head. At least that seemed to be the emphasis of most news reports that came out describing the study. But to really understand what this data means, we need to discuss the study and interpret the findings with an emphasis to explain the difference between sleep period and sleep time. Here you can listen to my interview with Dr. Siegel where we discuss the study and his findings:
SoundCloud (or iTunes)
More on These Preindustrial Societies
The study participants were from three different, geographically isolated preindustrial societies, each located in a tropical latitude close to the equator, thought to be similar to conditions where early humans evolved. These groups included:
- The Hadza live in northern Tanzania, 2° south of the equator, in woodland- savannah habitats around Lake Eyasi. This group is wholly dependent on hunting and gathering each day for wild foods.
- The San people of the Kalahari live 20° south of the equator. This particular group of San, were hunter-gatherers.
- The Tsimane live close to the Maniqui River in Bolivia and 15° south of the equator. These people are hunter-horticulturalists. Horticulturalists grow plants for food and medicinal uses, and for personal and social needs.
These Groups Are Healthy and Live in Natural Settings
Health studies performed on the Hadza, San, and Tsimane people show that while child mortality is higher due to infectious diseases, adults can live into their 80s, and beyond with lower levels of heart problems and higher levels of physical fitness compared to industrialized populations. These societies do not use electricity and, therefore, are not exposed to artificial light sources. Furthermore, they do not use heating or cooling systems and are exposed to the seasonal and daily variation in temperatures and natural light. The range of daily temperature for these societies across the year is from a low of 50 oF (10 oC) to a high of 90 oF (32 oC) (sounds good to me!).
Methodology — How Did They Measure Sleep and Light Changes?
All participants wore Actiwatch-2 devices, which are validated to accurately report sleep times and changes in light. These devices, however, are limited when it comes to evaluating all aspects of sleep stages and quality, and therefore, given their lower resolution, important questions remain about the nature of sleep in these people.
To obtain better resolution of sleep physiology than what wrist actigraphy can yield, in-lab polysomnography is performed. In this case, however, there are major problems with attempting to use more robust sleep monitoring. While in-lab PSG is highly useful in detecting physiological irregularities necessary for determining sleep pathology — like narcolepsy or sleep apnea — it’s not ideal for observing the natural sleep state because it is so disruptive to it! Never is the observer effect more true: measurements of a system cannot be made without affecting the system. How can one sleep normally when 27 different wires and electrodes are attached to the head and body? Additionally, it’s impractical to bring a sleep lab into the environments where these communities live, thus we are left with a low friction, but lower resolution method to determine sleep in the setting of natural living communities. It’s important to stress the limitations of the monitoring system used because it narrows what we can glean from the data. This doesn’t mean the findings aren’t useful. It does mean that important questions remain unresolved. For example, aside from a better understanding of sleep time in these people, does the sleep itself look different than modern-living humans? Questions like this remain for future investigations.
Findings — Sleep Duration
How much sleep do the preindustrial groups get?
This study found that the sleep duration for all three of these societies was in the range of 5hrs, 42 mins to 7hrs, 6 mins. The sleep period, however, which measures the time between sleep onset and offset, was between 6hrs, 54mins and 8hrs, 30mins. It’s crucial to understand the distinction between sleep period and sleep duration to interpret these findings correctly. I will discuss this in detail later.
Did these groups sleep more in winter versus summer?
Yes! On average, both the Tsimane and the San slept for about an hour longer in the winter than they did in the summer (56 minutes and 53 minutes longer in winter, respectively). This is notable in light of the fact that we now live in what’s been called “perpetual summer” conditions due to our control of temperature and light (e.g., clothing technology, electricity, buildings, etc.). Are seasonable fluctuations in sleep time and light exposure important to our health? This question was not addressed in this study, but these findings of seasonal differences indicate that this may be something that is important to explore further.
Findings — Naps and Insomnia
Did the preindustrial societies take naps?
Since the watches worn by participants in the study are optimized to record nighttime sleep and not daytime naps, the researchers also looked at motor activity to calculate daytime sleep episodes. By identifying time periods during the day where the motor activity of participants dropped to that of night time for 15 minutes or longer (an imperfect estimate), Siegel calculated that naps might have occurred on 7% of afternoons in the winter and 22% of afternoons in the summer. Recognize that this method is just capturing periods of time where naps could be occurring, but actually, naps were likely occurring with less frequency than these figures suggest, possibly as low as 3% of days during the winter. The actual duration of the nap was estimated to average 32 minutes. All groups showed similar findings.
Do preindustrial groups suffer from insomnia?
The researchers interviewed the Tsimane and San people to assess the prevalence of insomnia in these groups. Interestingly, neither group has a word for it in their language — an observation indicating that the condition of sleeplessness might not occur with enough frequency to require its own unique term (or not be seen as a problem). The researchers then explained the concept of insomnia to the participants, and only 1.5–2.5% of them reported regularly having it. As a comparison, these numbers are far lower than the 10–30% chronic insomnia rate that is reported in industrial societies.
Findings — Sleep in Relation to Light and Temperature
When did they go to sleep and when did they wake up?
The study found that the members of these societies remained awake long after darkness had fallen with average sleep onset occurring between 2hrs, 30 min to 4hr, 24min after sunset. The groups had small fires in the evening, but the measured light levels were lower than the lower limit of the sensor throughout the night. The Tsimane and Hadza groups woke up on average 1 hour before sunrise. The San group also woke up before twilight in the winter, but woke up 1 hour after sunrise in the summer.
Since neither sleep nor wake onsets were related to solar light levels, the researchers then evaluated the relation to ambient temperature. Both the Hadza and Tsimane showed a trend in which sleep onset correlated with the environment getting colder. To study this relationship in detail, Siegel and company measured the ambient temperature at the sleeping sites along with the body temperature collected at the fingers and abdomen of the participants. As expected, the sleep onset in both the winter and summer occurred during falling ambient temperature.
Also of note, morning awakenings were accompanied by strong vasoconstriction, in the hands and in the feet. The authors suggest that the purpose of the observed vasoconstriction could be to increase the temperature of the brain and the core of the body to prepare for waking activity.
This is one of the most interesting findings of this study. In future posts, I will explore how changes in environmental temperature, along with core temperature levels, influence sleep onset, sleep depth, and arousal.
Putting It All Together
In the beginning of this article, I described the idea that sleep times have declined over the last 50 to 100 years. Before 1970, reports point to an average self-reported sleep time of 8.5 hours or more per night. Recent reports, however, indicate that these times have declined and that it is now common for people to report sleeping less than 7 hours during week days and about 10% longer on weekends.
Because this is a very serious issue, these findings promoted a more in-depth analysis of the subject by Arizona State University Professor, Shawn Youngstedt and colleagues. His group analyzed 168 studies and included 6,052 individuals ages 18–88 years to assess objective sleep duration in healthy normal-sleeping adults. Contrary to the earlier signal that sleep times have decreased drastically, they found that objectively recorded sleep durations in healthy sleepers have actually remained stable over the last half-century and that the data from self-reported sleep studies has not been consistent. Furthermore, Youngstedt also found that those who sleep seven hours a day are least likely to die early. On the other hand, those who sleep fewer hours, and those who sleep longer, have a greater risk for an early death.
In the Siegel study, these natural living communities have sleep times that are on the lower end of what’s considered normal for modern-living adults. This too suggests that modern sleep times are not unusually short, but this is where the conversation about sleep period and sleep time comes into play, and understanding the distinction is crucial to decipher what is going on here.
Sleep period vs. sleep time
When asked, ‘how much sleep did you get last night,’ people naturally report sleep period and not sleep time. This is part of the problem with the interpretation of Siegel’s findings by the media, and it is also part of the distinction between self-reported and objectively-monitored sleep times. When news articles on Siegel’s research state that these natural-living groups slept an average of 6hrs and 24mins, it is factual. It also suggests that, perhaps, we don’t need 8 hours of sleep after all. Importantly, Siegel’s group also found that the period of time in which these people got their sleep, known as ‘sleep period,’ averaged 7hrs and 42mins. Here is the best way to understand this distinction:
When you ask a person how much sleep they got last night, they will go through the exercise of saying something like this,
“I went to bed at 11pm, and I woke up at 7am, so I got about 8 hours of sleep.”
People do not report sleep like this:
“I went to bed at 11pm, and I woke up at 7am, and my sleep efficiency* was 85%, so I was in bed for 8 hours, but I slept 6 hrs and 48 mins.”
With a conventional sleep study, the monitoring equipment can detect sleep time within the sleep period, but that is a big reason for this ambiguity: term confusion. It’s often referred to broadly as “sleep time,” but we’re talking about two different things here: the amount of time in bed, and the amount of time in that period where a person actually slept. To a person reporting normal sleep, there won’t be a perceived significant difference between the two, so sleep time and sleep period will be the same. This is one reasons why people report getting more sleep than then actually get in the sleep lab.
The National Sleep Foundation recently published a guideline stating that most people require 7–9 hours of sleep per night. This guideline is communicated in a manner that is intuitive for a non-sleep expert. Siegel’s study found that the sleep period for people in these natural living communities was 7hrs to 8hrs, 30mins per night, which is right in the range of the National Sleep Foundation’s recommendations.
This is why I have always recommended focusing on the time in bed (i.e., sleep period) when you’re trying to get sleep, and not the sleep time itself. For one, the focus from a behavioral perspective is to plan for enough time in bed to get the sleep you need. Imperfect sleep efficiency is normal, meaning that we may feel that we were asleep for, let’s say, 8 hours, but it’s likely you were only asleep for 7 hours in that sleep period. But, that doesn’t mean you only need 7 hours in bed! When you get less time in bed, your sleep efficiency may increase somewhat, but you also may not be allowing enough time at night to get “complete sleep” (a term I use to describe adequate bedtime for maximal sleep satisfaction). If you rob yourself of even 30 minutes each night by not planning enough time in bed, it’s likely you’re going to perform sub-optimally in a host of mental functions that you rely upon to achieve your best at work and in life.
Siegel’s study is very interesting, and there are things he and I discuss in the podcast that might indicate other ways to optimize environmental conditions to enhance sleep at night. Two important conclusions are that these people don’t sleep more than modern humans, and they achieved their sleep in a period of 7–8.5 hrs per night, which is in line with the National Sleep Foundation’s recommendations for adequate sleep time (really sleep period) of 7 to 9 hours per night for most adults.
This is an incredibly interesting study, and some of the findings should stimulate new research into ways to optimize sleep in modern life, but overall, the findings regarding sleep time are less controversial than they appear.
Lastly, I also did an interview with Chris Kresser discussing the same subject. You can listen to that here.