Staff Perspective: An Unresolved Sleep Mystery - Segmented Sleep

Staff Perspective: An Unresolved Sleep Mystery - Segmented Sleep

Dr. Diana Dolan

You may have heard that sleep in the modern world is different. That due to our frenetic 24/7 society, we cram sleep into a shortened sleep window and then hurry on into our days. In contrast, or so I am told going around TikTok, historically people slept in “segmented sleep,” in which they would retire to bed early, awaken and stay awake for a period of an hour up to several hours, and then return to sleep until rise time. Where is the evidence for this concept? Is this some previously unknown historical discovery brought to light, or just a now-digitalized urban legend? I do not profess to be an expert in this area, so let’s explore together.

Years ago, when asked about the concept of segmented sleep, also known as polyphasic sleep, as historical ‘fact’ by a provider who attended a Cognitive Behavioral Therapy for Insomnia (CBT-I) training I was facilitating, I was flummoxed. In my years of training in the BSM field from graduate school on this was not addressed, nor mentioned at conferences, nor studied in any research projects of which I was aware. The only trail I could find for this was a historian, who since about twenty years ago, has promoted the concept as normative among medieval and pre-Industrial European societies, arguing that segmented sleep is the typical pattern of at least Western human sleep in the absence of artificial light (Ekirch, 2021a). In looking over respective publications on this topic, while I did not conduct an exhaustive deep dive, I saw repeated self-citations, brief, vague anecdotes from diaries and journals, or references to write-ups that included atypical societies that were neither located in Europe nor dated in the pre-Industrial time period.

One recent historical study evaluating documentation from justice records, specifically eyewitness reports, serves to provide a more wide-scale, quantified pool of data with over 2,000 observations (Verhoeven, 2021a). The study, based on case reports in Belgium in the 1700s, shows that instead of segmented sleep, these Europeans slept in one consolidated chunk of approximately seven hours starting at a median time of 10:30 p.m., with members of the upper class having longer sleep periods. This finding seems remarkably consistent with modern reports of sleep in the U.S. In fact, one study of southern U.S. adults found an average bedtime of 11 p.m. with approximately eight hours of reported average sleep (Thomas et al., 2014). Similarly, other historical studies referenced in Britain, Sweden, and Germany documented a similar sleep pattern. Additionally, recent anthropological data of hunter-gather societies referenced also show one consolidated chunk of sleep. The author concludes that “recent anthropological and historical research has cast some serious doubt on the hypothesis” that the “monophasic pattern may in fact be a (relatively) recent pathological development caused by chronic sleep deprivation” (Verhoeven, 2021a). In other words, there does not seem to be clear and convincing evidence that segmented sleep occurred on a large scale before the Industrial Revolution.

Interestingly, Ekirch (2021a) wrote a scathing editorial article in response, calling the examination of the historical Belgium sleep records “deeply flawed, substantively and methodologically.” He stated that “voluminous allusions in historical sources to segmented sleep” exist and referred readers to his website (Ekirch). He also cited a study from the early 1990s showing an eventual pattern of segmented sleep in a sample of present-day adults who had no access to artificial light (Ekirch). Perhaps even more interestingly, the original study author then wrote a response to the response noting “you conveniently brush aside the literature that is much more critical about the predominance of biphasic sleep” and reiterating that the evidence demonstrates monophasic sleep (Verhoeven, 2021b). It turns out that the 1990s study forced subjects - who were not allowed to exercise or engage in any activities - into pitch darkness for 14 hours daily with instructions to attempt to sleep (Verhoeven, 2021b), hardly a replication of typical behavior. The author then acknowledges that one cannot say that some people at some times in some places never slept in a segmented sleep pattern, as even sleep today is by no means consistent across the board. As such, I agree there may be no way to conclusively settle this debate. By the way, there was a response to the response to the response (Ekirch, 2021b). Yes, I am enjoying the academic repartee!

For my part, from the perspective of sleep physiology rather than historical texts, circadian rhythms reach the peak of sleepiness between the hours of 1:00 to 3:00 a.m. Assuming a bedtime of say 10:00 or 11:00 p.m. followed by several hours of ‘first sleep’, this would place the mid-night alert period described in hypotheses of segmented sleep during this nadir. It does not seem likely to me that human circadian rhythms, which are notoriously difficult to shift even in targeted behavioral treatment for phase disorders and shift work, would have differed so greatly in the past that they would allow for a period of such alertness. However, I have not seen this consideration addressed by proponents of the segmented sleep theory.

To be fair, some modern humans do indeed engage in polyphasic sleep. Perhaps the most ready example consists of swing and night shift workers, who are recommended to attempt to sleep after their shift but typically struggle to maintain sleep due to the rising circadian rhythm and presence of daylight such that they are also recommended to nap prior to the start of their next shift. For shift workers, this segmented sleep runs counter to their circadian rhythm and it is widely known that there are a range of negative health consequences to this desynchronization. Similarly, parents of newborns are often forced into broken sleep in what is generally described as a difficult if not non-functional time period. More generally, although not directly comparable to the theory of historically segmented sleep as it included a broad definition of polyphasic sleep (including up to four sleep episodes per 24-hour period), a National Sleep Foundation consensus panel concluded that polyphasic sleep is associated with a variety of negative health and functional outcomes (Weaver et al, 2021) compared to consolidated nighttime sleep patterns.

I hesitate to draw firm, sweeping conclusions which is fitting in the world of sleep where so often we say “it depends.” Perhaps if we go back further in time there might be different evidence. Perhaps the diaries and broad comments alluding to a period of wake in the middle of the night refer more to maintenance insomnia that was not seen as distressing to those individuals. I think what we can say, though, is that any hype on social media that segmented sleep is the default or normal pattern of human sleep regardless of the presence of artificial light is not supported, and that segmented sleep appears to have a negative impact – as does insomnia, I might add. I wonder if interest in this concept stems from a desire to validate the many that are struggling with insomnia and sleep problems. I only wish that social media would instead promote evidence-based effective treatments such as CBT-I to encourage people to seek help. Ultimately, the old adage to take what one hears with a grain of salt applies to sleep, too.

The opinions in CDP Staff Perspective blogs are solely those of the author and do not necessarily reflect the opinion of the Uniformed Services University of the Health Science or the Department of Defense.

Diana Dolan, Ph.D., CBSM, DBSM, is a clinical psychologist serving as an Assistant Director of Training & Education with the Center for Deployment Psychology at the Uniformed Services University of the Health Sciences in Bethesda, Maryland..

Ekirch, A.R. (2021a). Response to Dr Gerrit Verhoeven, ‘(Pre)modern sleep: new evidence from the Antwerp criminal court (1715–1795)’. Journal of Sleep Research 30:e13167. DOI: 10.1111/jsr.13167
Ekirch (2021b). Dear Dr Verhoeven. Journal of Sleep Research 30: e13185. DOI: 10.1111/jsr.13185
Thomas, S.J., Lichstein, K.L., Taylor, D.J., Riedel, B.W., & Bush, A.J. Epidemiology of bedtime, arising time, and time in bed: analysis of age, gender, and ethnicity. Behavioral Sleep Medicine 12: 169–182. DOI: 10.1080/15402002.2013.778202
Verhoeven, G. (2021a).(Pre) modern sleep. New evidence from the Antwerp criminal court (1715–1795). Journal of Sleep Research 30. DOI: 10.1111/jsr.13099
Verhoeven, G. (2021b). Groping in the dark: A response to Roger Ekirch. Journal of Sleep Research 30: e13177. DOI: 10.1111/jsr.13177
Weaver, M.D., Sletten, T.L., Foster, R.G., Gozal, D., Klerman, E.B., Rajaratnam, S.M.W., Roenneberg, T., Takahashi, J.S., Turek, F.W., Vitiello, M.V., Young, M.W., & Czeisler, C.A. (2021). Adverse impact of polyphasic sleep patterns in humans: report of the National Sleep Foundation sleep timing and variability consensus panel. Sleep Health 7: 293-302. DOI: 10.1016/j.sleh.2021.02.009