REM sleep helps protect against predator attack


Rapid eye movement (REM) sleep brings about brief but periodic awakenings. In 1966, Dr. Frederick Snyder reported the “sentinel” function of REM could help animals prepare a fight or flight response against potential predator attacks.

However, to date there has been no experimental evidence for this hypothesis.

Now, a research team led by Dr. Wang Liping from the Shenzhen Institute of Advanced Technology (SIAT) of the Chinese Academy of Sciences has reported a common circuit regulating both innate fear and REM sleep, which has proved this hypothesis.

In the experiment, the animals slept in a sealed chamber and were exposed to trimethylthiazoline (TMT) odor, a stimulus indicating the presence of a predator.

“TMT triggered rapid arousal from REM sleep but not from non-rapid-eye-movement sleep. This suggests that REM sleep has specific properties that allow rapid arousal in response to predatory stimuli,” said Dr. Wang.

As REM sleep is generally characterized by higher arousal thresholds than non-rapid-eye-movement (NREM) sleep, the research team aimed to uncover the neural mechanisms underlying this REM-specific function. They examined the medial subthalamic nucleus (mSTN), a brain region that contains a high density of corticotropin-releasing hormone (CRH) neurons.

By combining in vivo neural activity recording and cell-type specific manipulations, they found that mSTN-CRH neurons produced a lowered arousal threshold during REM sleep for detecting predator threats, as well as increased defensive responses after awakening.

The results also showed that sustained predator exposure induced a significant increase in total REM sleep time but shorter durations of individual REM-sleep episodes and sleep architecture fragmentation. The mSTN-CRH neurons are required for this REM sleep adaptation to chronic threat exposure.

This is an example of how evolution leads to two distinct but related functions for the same set of neurons rather than two completely separate neuronal networks. “We may hypothesize that natural selection favors optimizing existing neural circuitry for efficiency in signal transduction and energy usage over metabolically more expensive solutions,” said Dr. Wang.

The co-occurrence of increased REM sleep and stress-related mood disorders has been observed in clinical studies. The new findings in this study offer a potential evolutionary explanation of this phenomenon and elucidate the underlying neurobiological mechanism.

“Our study raises the question of whether it is possible to treat mood disorders by targeting the common regulatory circuit of sleep and fear. We will continue working on this question,” said Dr. Yu-Ting Tseng, lead author of the study.

Figure thumbnail fx1

Human sleep

Human sleep has undergone additional changes from other great apes in several key features. An obvious feature is where we sleep, namely on the ground; among other apes, terrestrial sleep is rare, occurring only when predation risk is low, and typic- ally only by very large bodied males [48–51]. In con- trast, humans of both sexes habitually sleep on the ground, which could plausibly provide even more stable sleeping locations to achieve even deeper sleep. Predation represents a major tradeoff in this context, with risk of predator attack thought to in- crease for terrestrial primates [52, 53].

In relation to human ground sleep, Coolidge and Wynn [54] proposed the ‘tree-to-ground hypothesis’. They suggested that when hominins became fully terrestrial they gained the advantage of greater sta- bility than was possible in arboreal sleep. Freed from the disadvantages of arboreal sleep they could have achieved longer duration and higher quality sleep, which would have improved waking cognition.

Without terrestrial sleeping sites, they argue, fully human procedural memory consolidation for vis- ual-motor skills and visual-spatial locations could not have evolved. In addition, under the assumption that sleep plays a role in problem solving in social and other domains involving ‘threat simulation’ [55], they proposed that hominins would have been less primed for daily activity due to less sleep the previ- ous night [20, 54].

The controlled use of fire may have been an essential precursor to secure ground sleep [20]. Arboreal sleeping platforms reduce predation risk [56] and

minimize insect biting rates by masking host attract- ants or actually repelling insects [57, 58]. Sleeping platforms also provide some insulation for warmth [57], and give a stable and secure environment to enable higher quality sleep [39, 40]. A fire probably also reduces risk of predation and provides opportunities for thermoregulation, while smoke re- duces insect activity [59, 60]. Control of fire in early Homo erectus may therefore have enabled the night- time transition from trees to the ground [20, 61].
Quantitative characteristics of human sleep have also evolved along the human lineage. We consider here two major aspects: reduced total sleep and a higher percentage of rapid eye movement (REM) sleep [21]. Humans are empirically the shortest sleeping primates and have the highest percentage of REM (Fig. 1). New phylogenetic methods can rigorously investigate evolutionary change on a sin- gle branch, allowing a comparative biologist to in- vestigate whether an exceptional amount of evolutionary change has occurred [62, 63]. More spe- cifically, these methods compare actual sleep char- acteristics in humans to the predicted outcomes from a statistical model that includes both phyl- ogeny and a set of predictor variables that influence sleep characteristics. One can then test whether humans are a typical primate (our observed sleep duration falls within the predicted 95% credible interval) or a ‘phylogenetic outlier’ (our sleep dur- ation falls outside the predicted 95% credible interval).
Using this approach, Samson and Nunn [21] discovered that human sleep duration is extremely dif- ferent from phylogenetic predictions: our actual sleep duration falls outside the 95% credible inter- val, suggesting that we can be more than 95% cer- tain that human sleep differs from other primates. As we discuss below when considering the potential evolutionary drivers of shorter sleep along the human lineage, tradeoffs between sleep and other activities are likely to be important factors. When this same approach was applied to study the propor- tion of REM sleep in humans, the analyses revealed that humans pack a higher proportion of REM into their sleep than any other primate. It is worth noting, however, that some other primates have a longer absolute duration of REM sleep (see Fig. 1).

As a last point of comparison to other primates, humans may be more flexible in the timing of sleep than our closest living relatives. Evidence from small-scale societies and subtropical hunter-gath- erers [22], the historical record [64] and experiments in developed countries [65] suggest that humans show flexibility in their sleep. In a review of human sleep across cultures, Worthman [22] noted that, ‘Human nights are filled with activity and signifi- cance, and nowhere do people typically sleep from evening to dawn’ (p. 301).

Similarly, reflecting on his study of the Piraha˜ hunter-gatherers in South America, Everett (66) noted, ‘Piraha˜s take naps (fif- teen minutes to two hours at the extremes) during the day and night. There is loud talking in the village all night long’ (p. 79). Similar patterns appeared to occur in European and equatorial societies prior to the advent of cheap and effective lighting, with a historical analysis documenting extensive use of the concept of ‘first sleep’ and ‘second sleep’, con- sistent with a biphasic sleep pattern that differs rad- ically from what we consider ‘normal’ in Western societies today [64, 67].

Flexibility can also occur in the context of daytime sleep, i.e. the occurrence of napping or siestas. For example, Pennsylvanian Old Order Amish, a conservative Christian sect that avoids modern electrical conveniences, have been characterized as ‘common’ nap-takers, with 58% of the population recording a nap a least once per week [68].

Counter to these findings and suggestions, how

ever, a recent study of sleep in three hunter-gatherer populations [69] interpreted their actigraphy data as indicating consolidated sleep at night and with little napping during the day, and thus arguing against the flexibility of sleep. This presents a challenge, and calls for better methods of assessing sleep phasing using actigraphy, including through use of new al- gorithms, validation with reported episodes of sleep and wakefulness, and development of new methods to better assess sleep without reliance on actigraphy. It should be noted, however, that this study also re- vealed considerable heterogeneity in sleep onset time (but less in awakening), consistent with flexi- bility in the timing of sleep.
Given the global distribution of humans, adaptation
to local conditions may be expected for sleep, as seen for other human phenotypes. One obvious aspect of this involves latitude, and the effects of large changes in day-length throughout the year. Unfortunately, how- ever, sleep research in circumpolar environments has primarily focused on European populations [70, 71] and the effects of latitude on the physiology of military personnel [72]. Thus, little is known regarding the effects of seasonally variable day–night cycles on the sleep-wake patterns of nonindustrial indigenous populations [12]. Moreover, reports of sleep in

Figure 1. Duration of REM, NREM and total sleep in primates. Humans sleep the least compared to all other primates, yet have the greatest proportion of total sleep time dedicated to REM
Figure 2. Infant versus adult sleep. A sleep comparison between polyphasic human infant and consolidated sleep in an adult living in a post-industrial society (adapted from reference [75])

post-industrial societies have shown conflicting evidence and small effects with respect to sleep duration across seasons [73, 74]. Several factors may influence the outcome of such studies, including lack of direct exposure to changes in light and temperature among participants in la- boratory environments, or the environmental buffer provided by modern work and residential facilities.

In contrast, evidence supports the idea that sleep is modulated by season in traditional, equatorial societies; e.g. longer total sleep times (53–56 min increase) were associated with the ‘winter’ season in the San and Tsimane [69].

Sleep and human development
Ontogeny can also shed light on human sleep. As all parents know, babies sleep a lot, yet they are born without a regular sleeping rhythm (Fig. 2). The chaos of sleep phasing in the first days of life consolidates into a polyphasic sleep schedule consisting of at first two naps and one bout of night-time sleep, and even- tually one and then no naps (with longer consolidated sleep at night).

Furthermore, infant sleep is characterized by larger amounts of REM sleep, suggesting that REM sleep may have import- ant consequences for the developing brain [76]. Infant sleep is important to the evolutionary story of sleep in two other ways: one involves the role of infant-parent co-sleeping, and the other involves infant crying.

Infant-parent co-sleeping has attracted much attention in recent decades, with parents faced with the dilemma of sleeping with the baby versus putting the baby in a separate room. All discussions of co- sleeping should begin by appreciating how radically novel it is for dependent children to even have the option to sleep separately from their caregivers. Throughout evolutionary history, families slept to- gether, possibly with extended family members,

and the same is true in many traditional societies today [59, 77]. It is only in modern living conditions—with increased safety and availability of sep- arate bedrooms for parents and children—that the dilemma of infant–parent co-sleeping arises.
James McKenna was among the first anthropologists to investigate mother–infant night-time interactions empirically, often injecting an evolutionary perspective [78, 79].

In some of this research, the investigators found that bed-sharing resulted in less deep sleep for mothers and infants, but more sim- ultaneous awakenings by mothers and infants that were associated with more breastfeeding [24]. Thus, mothers would tend to awaken or transition between sleep states at times when babies were also likely to awaken, resulting in less disruption to the mothers’ sleep cycles and a higher feeding frequency for infants [80]. Overall, these studies demonstrate a mu- tually reinforcing relationship between mother– infant co-sleeping and feeding, probably reflecting correlated evolution among these behaviors.

This research has been used to inform the potential risks associated with solitary sleep practices; e.g. the lack of breastfeeding and solitary sleeping has been identified as a risk factor for sudden infant death syndrome (SIDS), suggesting that less deep sleep in infants who were co-sleeping and breast- feeding more regularly were at lower risk of SIDS [81, 82]. However, other studies have found that bed sharing also increases risk of SIDS, which may be amplified by factors such as infant age or use of alcohol or drugs [83].

The other insight to infant sleep comes in the con-text of infant crying, a feature not observed in chim- panzees [84]. Haig [23] revived and extended a hypothesis [85] that night-time arousal and crying by infants is an adaptive behavior to extend inter- birth intervals, benefiting the crying infant at the potential cost to parental reproductive success. Reviewing the literature, Haig [23] notes that shorter inter-birth intervals lead to greater offspring mortality, and that more night-time breastfeeding episodes results in longer postpartum amenorrhea.

Thus, ‘natural selection will have preserved suckling and sleeping behaviors of infants that suppress ovarian function in mothers because infants have benefited from delay of the next birth’ (p. 34). Additionally, Haig [23] incorporated modern perspectives of gen- omic conflict by considering how imprinted genes of maternal origin might favor more consolidated sleep, whereas genes of paternal origin promote greater wakefulness.

As noted by Haig [23], the explicit inter-gener- ational and intra-genomic conflicts in his proposal challenge the assumption of mother–infant cosleeping as a highly co-evolved and harmonious sys- tem that was suggested above in some of the re- search on co-sleeping. Instead, Haig’s [23] research suggests a need to appreciate that substantial parent-offspring conflict likely exists even in the context of sleep.

reference link :

More information: Liping Wang, The subthalamic corticotropin-releasing hormone neurons mediate adaptive REM sleep responses to threat, Neuron (2022). DOI: 10.1016/


Please enter your comment!
Please enter your name here

Questo sito usa Akismet per ridurre lo spam. Scopri come i tuoi dati vengono elaborati.