ArticlePDF Available

Can People Sleep Too Much? Effects of Extended Sleep Opportunity on Sleep Duration and Timing


Abstract and Figures

Many people are concerned about whether they are getting “enough” sleep, and if they can “sleep too much.” These concerns can be approached scientifically using experiments probing long-term (i.e., multi-night) sleep homeostatic processes, since homeostatic processes move the system toward its physiological setpoint (i.e., between “not enough” and “too much”). We analyzed sleep data from two human studies with sleep opportunities much longer than people usually stay in bed (i.e., conditions in which sleep homeostatic responses could be documented): sleep opportunities were 14–16 h per day for 3–28 days. Across the nights of the extended sleep opportunities, total sleep duration, Rapid Eye Movement (REM) sleep duration and non-REM sleep durations decreased and sleep latency increased. Multiple nights were required to reach approximately steady-state values. These results suggest a multi-day homeostatic sleep process responding to self-selected insufficient sleep duration prior to the study. Once steady state-values were reached, there were large night-to-night variations in total sleep time and other sleep metrics. Our results therefore answer these concerns about sleep amount and are important for understanding the basic physiology of sleep and for two sleep-related topics: (i) the inter-individual and intra-individual variability are relevant to understanding “normal” sleep patterns and for people with insomnia and (ii) the multiple nights of sleep required for recovery from insufficient sleep from self-selected sleep loss is important for public health and other efforts for reducing the adverse effects of sleep loss on multiple areas of physiology.
Content may be subject to copyright.
fphys-12-792942 December 17, 2021 Time: 12:6 # 1
published: 22 December 2021
doi: 10.3389/fphys.2021.792942
Edited by:
Sara Montagnese,
University of Padua, Italy
Reviewed by:
Vladyslav Vyazovskiy,
University of Oxford, United Kingdom
Malcolm Von Schantz,
University of Surrey, United Kingdom
Elizabeth B. Klerman
Present address:
Elizabeth B. Klerman,
Department of Neurology,
Massachusetts General Hospital and
Harvard Medical School, Boston, MA,
United States
Specialty section:
This article was submitted to
a section of the journal
Frontiers in Physiology
Received: 11 October 2021
Accepted: 15 November 2021
Published: 22 December 2021
Klerman EB, Barbato G,
Czeisler CA and Wehr TA (2021) Can
People Sleep Too Much? Effects
of Extended Sleep Opportunity on
Sleep Duration and Timing.
Front. Physiol. 12:792942.
doi: 10.3389/fphys.2021.792942
Can People Sleep Too Much? Effects
of Extended Sleep Opportunity on
Sleep Duration and Timing
Elizabeth B. Klerman1,2*, Giuseppe Barbato3, Charles A. Czeisler1and Thomas A. Wehr4
1Division of Sleep and Circadian Disorders, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA,
United States, 2Department of Neurology, Massachusetts General Hospital and Harvard Medical School, Boston, MA,
United States, 3Department of Psychology, University degli Studi della Campania Luigi Vanvitelli, Campania, Italy,
4Intramural Research Program, NIMH, Bethesda, MD, United States
Many people are concerned about whether they are getting “enough” sleep, and if
they can “sleep too much.” These concerns can be approached scientifically using
experiments probing long-term (i.e., multi-night) sleep homeostatic processes, since
homeostatic processes move the system toward its physiological setpoint (i.e., between
“not enough” and “too much”). We analyzed sleep data from two human studies with
sleep opportunities much longer than people usually stay in bed (i.e., conditions in
which sleep homeostatic responses could be documented): sleep opportunities were
14–16 h per day for 3–28 days. Across the nights of the extended sleep opportunities,
total sleep duration, Rapid Eye Movement (REM) sleep duration and non-REM sleep
durations decreased and sleep latency increased. Multiple nights were required to reach
approximately steady-state values. These results suggest a multi-day homeostatic sleep
process responding to self-selected insufficient sleep duration prior to the study. Once
steady state-values were reached, there were large night-to-night variations in total sleep
time and other sleep metrics. Our results therefore answer these concerns about sleep
amount and are important for understanding the basic physiology of sleep and for two
sleep-related topics: (i) the inter-individual and intra-individual variability are relevant to
understanding “normal” sleep patterns and for people with insomnia and (ii) the multiple
nights of sleep required for recovery from insufficient sleep from self-selected sleep loss
is important for public health and other efforts for reducing the adverse effects of sleep
loss on multiple areas of physiology.
Keywords: sleep, homeostasis, recovery sleep, sleep restriction, human, insomnia, sleep variability
“Can I sleep too much” is a common question. Epidemiologic data indicate that people who
habitually sleep more than 10 h at night are at greater risk of death and other effects [e.g., (Kripke,
2004)]. Because of these correlations between longer sleep duration and adverse effects, many
people are concerned that sleeping too much may be dangerous. In one study, once appropriate
control for medical conditions (e.g., sleep disorders or severe illness, both of which may increase
sleep duration) was conducted, however, the link between long sleep duration and adverse effects
was lost (Van Cauter et al., 2008). In addition, no causal links have been established and no
controlled laboratories of chronic extension of sleep duration have demonstrated adverse effects. In
contrast, controlled laboratory studies have demonstrated causal links between chronic insufficient
Frontiers in Physiology | 1December 2021 | Volume 12 | Article 792942
fphys-12-792942 December 17, 2021 Time: 12:6 # 2
Klerman et al. Sleep During Extended Sleep Opportunities
sleep and multiple adverse effects [e.g., (Scott et al., 2006;Van
Cauter et al., 2008;Mullington et al., 2009)]. Yet people still
ask sleep researchers and clinicians: can we sleep too much? Is
sleep like eating—such that you can eat too much and become
overweight? Can one over-sleep? Would over-sleeping have
negative health effects? To address these questions, we consider
“too much” to be a concept related to physiological homeostasis
(i.e., ability to maintain equilibrium).
Sleep timing and homeostasis are regulated within and across
multiple days. The physiology underlying this regulation can be
studied by shortening or extending sleep and wake opportunities
under controlled conditions and monitoring changes in the
amount and time course of changes in sleep metrics [e.g., total
sleep time (TST), Non-Rapid Eye Movement (NREM) sleep, Slow
Wave Sleep (SWS), or Rapid Eye Movement (REM) sleep]. To
date, most experiments have focused on shortening sleep and/or
extending wake; sleep extension protocols, however, can also
provide information about sleep homeostasis (e.g., sleep need,
sleep capacity, and sleep satiety).
In healthy individuals without a sleep disorder, the time
courses of sleep metrics during both chronic sleep restriction
(i.e., multiple days with insufficient sleep) and chronic sleep
extension (i.e., multiple days with more time available for sleep
than actual sleep duration) would be expected to depend on:
(i) prior sleep/wake history, including whether any preceding
sleep loss was from acute sleep deprivation (i.e., one extended
wake episode) or chronic sleep restriction, since the two have
different time-courses of recovery (Cohen et al., 2010); and (ii)
the timing, duration, and number of the bedrest episode(s).
The time-course and content of sleep during extended sleep
opportunities can be used to quantify sleep capacity, which we
define as whether someone can sleep when they are not “tired”
(e.g., when bored) or as recovery from prior insufficient sleep.
These questions are important for documenting basic physiology
and for public education about the consequences of sleep loss and
how to recover from sleep deficiency, which is formally defined as
“deficit in the quantity or quality of sleep obtained vs. the amount
needed for optimal health, performance and well-being” by the
2011 NIH Sleep Disorders Research Plan (National Center on
Sleep Disorders Research, 2011).
The design of previous protocols has not allowed full
documentation of recovery sleep after insufficient sleep for
multiple reasons. Firstly, most studies have only one or two
nights of recovery sleep, which may not be sufficient. Secondly,
many previous studies of extended sleep opportunities may not
provide sufficient time in bed (during each bedrest episode)
for some individuals to obtain complete recovery from their
prior sleep loss, such as scheduling only 10 h time in bed
(TIB) (Roehrs et al., 1989, 1996;Harrison and Horne, 1996;
Arnal et al., 2015;Skorucak et al., 2018) and/or allowing self-
selection of the extended sleep duration (Kamdar et al., 2004),
including one study (Webb, 1986) that allowed 12 h TIB for one
night, but only 4 of the 42 individuals stayed in bed that long.
Thirdly, nocturnal opportunities alone may not provide optimal
opportunity for obtaining recovery sleep; more sleep may be
obtained if both nighttime and daytime (e.g., nap) opportunities
are allowed. Finally, for the night or nights prior to the extended
sleep opportunity, either TIB (during which participants try to
sleep) or actual TST, have not been reported in most studies;
such information is necessary for quantifying the relationship
of recovery sleep metrics to prior sleep loss. When the amount
of sleep deprivation or restriction prior to the extended sleep
opportunity is not known, parameters of homeostatic drive
and recovery cannot be calculated. One of our studies, with
16 h of opportunity (12 h at night and 4 h during the day)
(Klerman et al., 2004;Klerman and Dijk, 2008) did document
each individual’s TIB history for approximately 3 weeks prior to
the intervention schedule; those results are examined in detail
below. In addition, alertness- or sleep-promoting substance use
have not been documented in many studies and so the sleep
obtained may differ depending on whether there is withdrawal
from those substances. Prior TIB and drug (e.g., caffeine, alcohol)
use may be highly variable within and across participants,
and may affect the amount of sleep they can obtain during
sleep opportunities. Given that individuals may choose to arise
from bed for physiological and psychological reasons before
their sleep homeostatic drive is fully dissipated, therefore, the
design of a study of the effects of extended sleep opportunity
should ideally include documentation of sleep history, use of
substances that may affect sleep or wake, and enforcement of
time in bed during the entire scheduled bedrest episode(s) for
each participant.
To address this question of whether it is possible to sleep
“too much,” therefore, we reanalyzed data from two relevant
studies that included most or all of the conditions cited above.
The first study was designed to assess photoperiod, and only
hence only scheduled people to sleep at night during an extended
episode of darkness each day (Wehr, 1991;Wehr et al., 1993); it
included 28 nights of inpatient 14-h scheduled TIB in darkness
from 6 pm–8 am. These studies documented increased total
sleep time including earlier sleep onset, and increased duration
of sleepiness and of melatonin secretion in individuals under
conditions of more hours of darkness and sleep opportunity
at night. The second study (Klerman et al., 2004;Klerman
and Dijk, 2008) protocol both extended the time in bed at
night and provided the opportunity for an afternoon nap: it
included 1 week at home of scheduled TIB based on the prior
1–2 weeks for each individual and 3–8 inpatient days of a
24-h schedule that included 12 h of scheduled TIB +4 h
wake +4 h scheduled TIB +4 h wake with midpoint of the
nighttime bedrest episodes the same as the midpoint of the first
inpatient bedrest episode that was based on each participant’s
at-home schedule. Caffeine, tobacco, and other alertness- or -
sleep-promoting substances were not allowed the 2 weeks before
or during the inpatient protocol. These studies also documented
increased total sleep time during both nocturnal and daytime
(nap) sleep episodes compared to an individual’s self-selected
sleep times at home; this sleep amount decreased over the days
of the protocol.
Study 1: 15 adults (20–36 years; 1F:14M) participated.
Healthy volunteers were screened with interviews, physical
Frontiers in Physiology | 2December 2021 | Volume 12 | Article 792942
fphys-12-792942 December 17, 2021 Time: 12:6 # 3
Klerman et al. Sleep During Extended Sleep Opportunities
examinations and routine laboratory tests and procedures.
None had medical or psychiatric illnesses, and none
had taken any medication for at least 3 weeks before the
Study 2: 35 adults (18–32 years; 18F:17M) participated. All
were healthy by history, physical exam, laboratory tests of
blood and urine, a clinical polysomnogram (PSG) for sleep
disorders, and a visit with a clinical psychologist. None were using
prescription or non-prescription medications for at least 3 weeks
before the start of the inpatient protocol and no caffeine, alcohol,
or tobacco were allowed 2 weeks before the start of the inpatient
protocol; urine toxicology screens were performed during those
2 weeks and at admission.
For Study 1, the protocol was approved by the NIH
Intramural Research Program IRB; written informed consent
was obtained. For Study 2, the protocol was approved by the
Partner’s Healthcare IRB; written informed consent was obtained.
Both protocols were conducted before the
Study 1: Participant sleep/wake schedule and caffeine or other
drug consumption before each inpatient segment of the protocol
was not regulated or recorded, except that daytime naps were
prohibited. The protocol included: (i) 7 nights of inpatient
8-h scheduled TIB in darkness from midnight—8 am, (ii)
a constant routine (i.e., enforced wakefulness with multiple
small meals under dim light conditions for 40 h), (iii)
2 weeks of unregulated and unreported sleep/wake schedule
at home; (iv) 28 nights of inpatient 14-h scheduled TIB in
darkness from 6 pm–8 am and unregulated activities outside
the laboratory from 8 am–6 pm; ending with (v) a constant
routine. For part (iv), the 6 additional hours of sleep opportunity
were added to the beginning of inpatient bedrest episode
compared to part (i) (Supplementary Figure 1). This analysis
only includes the data from the 28 nights of inpatient 14-h
TIB (section iv).
Study 2: Participants documented their self-selected
sleep/wake schedule for 2–3 weeks prior to the inpatient
stay using a paper-based log, a wrist-worn actigraph, and calls to
a time-stamped recording. The last week prior to the inpatient
stay and the first inpatient scheduled bedrest episode were at the
average onset and offset times of the prior 1–2 weeks for each
individual. The first inpatient day had a set of 5 multiple sleep
latency tests (MSLTs) (Carskadon, 1986) beginning 2 h after
awakening. The next 4 days had a 24-h schedule that included
12 h of scheduled TIB +4 h wake +4 h scheduled TIB +4 h
wake with midpoint of the nighttime bedrest episodes the same as
the midpoint of the first inpatient bedrest episode. Those 4 days
were followed by an intervention/control (no intervention) and
4 more days of 12-h TIB +4-h wake +4-h TIB +4-h wake
schedule. The inpatient protocol ended with a 12-h scheduled
TIB, a set of 5 MSLTs, and a final 8-h TIB (Supplementary
Figure 1). Only the inpatient data from days with 16-h of TIB
prior to the intervention or with no intervention (control)
are included in this report; 3–8 days of data are available for
each participant.
Quantification and Statistical Analysis
Polysomnography records were scored in 30 s epochs using
established criteria (Rechtschaffen and Kales, 1968) as non-
REM (NREM) sleep stages 1, 2, 3, or 4; Rapid Eye Movement
(REM) sleep, Wake, and Other (e.g., Movement Time or
Unscorable). Slow Wave Sleep (SWS) was defined as NREM
Sleep stages 3 and 4. Each site scored its own recordings.
The scored data from both studies were re-analyzed as
Bedrest episodes were defined as the time between lights out
and lights on; sleep episodes were defined as the time between
sleep onset and sleep offset (Czeisler et al., 1980a), which in this
study were calculated using Persistent Sleep (PersistSlp) onset
and Final Wake, respectively. PersistSlp onset was defined as
the first epoch of 10 consecutive minutes of any stage of sleep
after lights out time; Final Wake was defined as the number
of consecutive minutes of Wake before lights on; Sleep Episode
Duration was defined as the time between PersistSlp onset and
Final Wake onset; Wake within the Sleep Episode was the total
number of minutes of Wake within the Sleep Episode Duration;
Wake after persistent sleep onset (WAPSO) was calculated as
the number of minutes of Wake between PersistSlp onset and
lights on time; and total sleep time (TST) was defined as the
total number of minutes of NREM S1 +S2 +S3 +S4 and REM
sleep within the bedrest episode. Metrics were also computed as
the difference from the value for that individual on each day of
extended sleep opportunity and the first day of extended sleep
opportunity. Data from bedrest episodes with >5% missing
data or >5% stage Other were not included in analyses. For
Study 2, data from each 12-h bedrest episode and the following
4-h bedrest episode were combined to create a “daily” value;
only data from when both 12 and 4-h bedrest episodes were
usable (i.e., not excluded using above criteria) were used for
that metric.
For analysis of the time-course within a bedrest episode, data
were examined relative to lights out. Data were binned in 30 or
60 min bins and plotted relative to the middle of each bin.
Data were analyzed for all nights of extended sleep for
Experiments 1 and 2; for the first 7 and last 7 nights of Study 1;
and for the last 14 nights of Study 1.
For all analyses, data were combined within participants and
then across participants. 5th, 25th, 50th (median), 75th and 95th
percentiles were calculated due to the relatively small sample
size and non-Normal/Gaussian distributions of data. Formal
comparative statistics were not performed if data medians (50th
percentiles) of one protocol were between the 5th and 95th
percentile of data from the other protocol. All statistics were
performed using SAS for Windows.
Outpatient Sleep Durations
For Study 2, average-per-individual TIB durations prior to the
inpatient study were 6.1–10.3 h. Outpatient TIB or sleep duration
information was not available for Study 1.
Frontiers in Physiology | 3December 2021 | Volume 12 | Article 792942
fphys-12-792942 December 17, 2021 Time: 12:6 # 4
Klerman et al. Sleep During Extended Sleep Opportunities
Analysis by Entire Bedrest Episode as
Extended Sleep Opportunities Continue
Patterns of all sleep metrics were similar between the two
protocols. Daily (i.e., night episodes only for Study 1; night +day
episodes for Study 2) TST, NREM sleep, and REM sleep
but not SWS (Figure 1 and Supplementary Figure 2) were
markedly higher in the first days of extended sleep opportunity
(Table 1). The median values (across all participants) for
these three metrics decreased over the first 7 days and then
appeared to stabilize. Persistent Sleep (PersistSlp, 10 consecutive
minutes of any stage of sleep) latency and Sleep Episode
Duration (time between PersistSlp onset and Final Wake) both
increased over the first two nights (i.e., night bedrest episodes
only for both Study 1 and Study 2) and then appeared to
stabilize: after the first 7 days, latency to PersistSlp was
approximately stable at 90–120 min in Study 1 and 60 min
in Study 2 (Figure 1 and Supplementary Figure 2). The
time constant and steady-state values from an exponential
fit for TST, NREM sleep, and REM sleep were similar in
both studies (Table 2); of note, the fit steady-state value for
TST was more than 8.5 h (which would require a TIB of
more than 8.5 h). The time courses of other variables were
not appropriate for an exponential fit (see Figure 1). The
group median of TST, Latency to PersistSlp, NREM Sleep,
and REM sleep remained approximately constant after the
first 7 days through the end of the 28 days in Study
1 (Figure 1).
In Study 1, all (15/15) participants had higher total duration
of TST, NREM sleep, and Sleep Episode Duration and 12
of the 15 participants had higher total duration of REM
sleep in the first 7 nights than the last 7 nights of the
protocol (Figure 2). Total SWS duration was not higher
in the first 7 nights than the last 7 nights. The difference
between the first and last 7 nights in TST, Sleep Episode,
Duration and REM sleep duration was negatively correlated
TABLE 1 | Difference in minutes between values on night 8 vs. night 1 for Study 1.
Variable Mean (Std dev)
Pvalue Minimum,
Difference (min.)
TST 144.6 (91.4) <0.0001 353.5, 10.5
NREM sleep 113.1 (65.2) <0.0001 258.0, 15.0
REM sleep 31.5 (39.4) 0.008 114.0, 21.0
SWS 5.6 (33.3) 0.348 58.0, 38.5
WAPSO* 50.2 (106.6) 0.090 134.0, 172.5
LatPersistSlp* 89.2 (68.4) 0.0002 18.5, 192.0
Final wake** 39.2 (35.8) 0.0008 0.0, 104.5
SED* 128.4 (20.5) <0.0001 239.0, 12.5
*diff is bimodal, not normally distributed.
**diff is left skewed, not normally symmetrically distributed.
TABLE 2 | Approximate steady-state values from exponential fit.
Variable REM sleep NREM sleep
stages 2, 3, and 4
Study 1
Time constant (days) 0.8 3.7 3.1
Steady state value (hours) 2.0 5.9 8.6
Study 2
Time constant (days) (n.s.) 1.5 1.5
Steady state value (hours) (n.s.) 5.7 8.9
All parameters are at p <0.01 by fit.
Exponential fits were not appropriate for other sleep stages.
with the amount in the first 7 nights (i.e., greater amount in
first 7 nights was associated with less difference between first
and last 7 nights).
FIGURE 1 | (A) Patterns of TST, Latency to PersistSlp, NREM Sleep, and REM Sleep per 24-h across days of both Studies. Gray lines are data from Study 1; red
lines are data from Study 2. For each experiment, 5th, 50th (median), 7 and 95th percentiles are shown. Data for NREM Sleep Stage 1, NREM Sleep Stage 2, SWS,
Wake in PersistSlp. Final Wake Duration, and Sleep Episode Duration are in Supplementary Figure 2.(B) As in panel (A) except all values relative to the value for
each individual’s first night value.
Frontiers in Physiology | 4December 2021 | Volume 12 | Article 792942
fphys-12-792942 December 17, 2021 Time: 12:6 # 5
Klerman et al. Sleep During Extended Sleep Opportunities
Analysis by Changes Within Each
Bedrest Episode
Within each bedrest episode (Figure 3 and Supplementary
Figure 3), the highest group median amount of NREM sleep
and SWS occurred near the beginning of the bedrest episode
and shifted slightly later across the first 7 days, as the amount of
Wake at the onset of the bedrest episode increased over the first
3 nights and then appeared to stabilize in both protocols. REM
sleep remained highest near the end of the nocturnal bedrest
episode across all nights in Study 2. The group median did not
show a prominent Wake bout in the middle of the extended night
in Study 1, since extended mid-sleep-episode Wake bouts were
seen only in some individuals at variable times in some nights
(Supplementary Figure 4).
Analysis by Changes Across Bedrest
Episode for the Last 14 Nights of Study 1
For the last 14 nights of Study 1 (Figure 4 and Supplementary
Figure 5), there was large night-to-night variability in multiple
metrics: there was a range of 20–100% in single night values
relative to median values in TST (median range 185.0 min,
median of medians 518.3 min, median of ratio [=(max-
min)/median] across participants for 14 nights of 0.4), Latency
to PersistSlp (range 112.5 min, median value 112.5 min, average
ratio 1.0), NREM sleep (range 157.0, Median 392.5, Ratio
0.4) SWS (range 28.5 min, median 35.8 min, ratio 0.9), REM
sleep (range 71.0 min, median 119.5 min, ratio 0.6), WAPSO
(range 178.5 min, median 162.3 min, ratio 1.0) (Figure 4).
The variation between individuals seen in median TST appears
to be mostly due to due to changes in median NREM sleep
more than median REM sleep. There were relationships between
median value in the last 14 nights and variability, with negative
correlations between median and range (75–25%) for TST
(correlation = −0.54), NREM Sleep (correlation = −0.41)
and WAPSO (correlation = −0.41), a positive correlation
for REM sleep (correlation =0.29), but not LatPersistSlp
(correlation =0.01).
There was no obvious consistent 2-night or 3-night periodicity
in this variability, suggesting that short-term (e.g., one or two
night) sleep homeostasis was not a major factor (Figure 5).
However, analysis of the night-to-night differences demonstrates
homeostatic regulation is present: there is a negative correlation
for the night-to-night change in duration of TST, NREM sleep,
and REM sleep or of Latency to PersistSlp: a larger difference
between consecutive nights is followed by a smaller difference
and vice-versa (Supplementary Figure 5).
In response to the question: “Can I sleep too much?,” the
answer is “No,” since “too much” implies sleeping longer than is
biologically necessary. We demonstrate that the average healthy
young human can not chronically sleep over 10 h per day: the
average sleep duration was 8.6 h over the last 14 nights of the
Study 1 protocol. Individual night TST, however, ranged from
5.2–11.0 h. Two other studies support our observed homeostatic
set-point value of 8.6 h: in one study in which healthy individuals
had scheduled 8-h bedrest episodes (Van Dongen et al., 2003),
there was a progressive decrease in visual vigilant performance
(suggesting insufficient sleep), while participants with 9-h bedrest
episodes in another study (Belenky et al., 2003) did not
have this decrease.
In this report, we compared data from two conditions with
extended sleep opportunity: one with one 14-h nighttime bedrest
episode and one with 16 h of sleep opportunity split into 12 h
at night plus 4 h during the day. For both studies, the values of
sleep metrics in the first few days suggested rebound recovery
(i.e., a homeostatic process) from sleep loss: when scheduled to an
extended sleep opportunity, there were initially increased values
of TST, NREM sleep, and Sleep Episode Duration and decreased
sleep latency that lasted 1–3 days. This was followed by decreasing
values of TST, NREM sleep, and Sleep Episode Duration and
increasing sleep latency for the next few days. This pattern is
consistent with a homeostatic process, likely reflecting recovery
from prior behaviorally-induced insufficient sleep (or sleep debt).
If any extra sleep obtained were voluntary (e.g., due to extra time
in bed without any stimulation) and not because of a homeostatic
process, then the initial increased total sleep duration would be
expected to continue until the protocol ended, and not show the
time-course of decaying to a new set point. Indeed, given that
after the first few nights, participants were often awake in bed for
more than 1 h (total per night), such voluntary sleep might have
been preferred. The values and time-courses of recovery were
remarkably similar in both protocols. Consistent with our data,
another report documented increased TST and REM time during
10-h sleep extension when compared to the 8 h baseline sleep
opportunity (Skorucak et al., 2018). Previous reports of “banking”
sleep (Rupp et al., 2009), probably reflected recovery from prior
sleep restriction. A sleepy feeling after awakening, especially from
a nap, may occur in some people, but this does not mean that too
much sleep occurred; this feeling is associated with sleep inertia
and dissipates as wakening continues [e.g.,(Jewett et al., 1999)].
Total Sleep Time (i.e., Nighttime Sleep
There are multiple potential explanations as to why this
presumed sleep debt was not repaid in one bedrest episode.
(i) The duration of sleep depends on the circadian phase at
which the sleep episode is initiated (Czeisler et al., 1980b). The
circadian wakeup signal in the morning may therefore end the
sleep episode before sleep need is fully dissipated. Note that
individuals may be able to sleep past this circadian wakeup
signal if they still have a high sleep pressure at that time. (ii)
There is a limit to the amount of sleep a person can obtain
in a single sleep episode because other homeostatic processes
(e.g., hunger, bladder fullness) may interfere with continuing
to sleep. (iii) Allostasis (i.e., changing homeostatic level) is
present; one mechanism of this would be a change in receptor
sensitivity (Phillips et al., 2017). (iv) Previous reports suggested
the possibility of core vs. optional sleep (Horne, 1988). This
hypothesis is based on studies in which there was sleep restriction
Frontiers in Physiology | 5December 2021 | Volume 12 | Article 792942
fphys-12-792942 December 17, 2021 Time: 12:6 # 6
Klerman et al. Sleep During Extended Sleep Opportunities
FIGURE 2 | Comparison of total number of hours of TST, Latency to PersistSlp, NREM Sleep, and REM Sleep in first vs. last 7 nights of Study 1 with line of unity
(solid line) shown.
FIGURE 3 | Color heat maps of 30-min bins for the median (across all participants) for TST, NREM Sleep, REM sleep, and Wake for each day of the protocol. Top
row contains data from first 8 days of Study 1 and bottom row contains data from all 8 days of Study 2. For ease of viewing, data on each line are plotted as relative
to time of lights out for the main bedrest episode. Clock time for Study 1 time of lights out was 1800–0800. Study 2 had bedrest episodes scheduled relative to each
individual’s habitual bedtime. Color heat maps for the entire 28 days of Study 1 are in Supplementary Figure 3.
followed by ad lib sleep. However, in those studies the ad lib sleep
conditions included self-selected sleep durations in the real world
(i.e., not laboratory conditions) and therefore the amount of sleep
obtained may have been less than under the conditions of Studies
1 and 2. (v) If sleep restriction squeezed Evening (E) and Morning
(M) oscillators (Wehr, 2001) together, as if the individual were
living on an extreme photoperiod (as at home before inpatient
portion of Study 1), then several days may be required for the
two oscillators to expand to another phase relationship, with the
associated change in sleep duration. Further work is required to
test these explanations.
Sleep Latency
Study 1 participants had overall shorter Sleep Latency than
those in Study 2 for the nighttime bedrest episodes. This
may be due to the different prior wake durations in the
two experiments (10 vs. 4 h, respectively). Participants in
both experiments had increasing Sleep Latency over the first
few bedrest episodes. There are multiple possible reasons for
these results, including decreased sleep pressure as recovery
proceeds, and changes in timing of melatonin onset to a later
time, since both protocols include scheduled sleep in darkness
ending at a later time (i.e., later in the morning) than the
Frontiers in Physiology | 6December 2021 | Volume 12 | Article 792942
fphys-12-792942 December 17, 2021 Time: 12:6 # 7
Klerman et al. Sleep During Extended Sleep Opportunities
FIGURE 4 | Individual data points, 25th, 50th (median) and 75th values for hours of sleep TST, Latency to PersistSlp, NREM Sleep, REM Sleep in the last 14 nights
of Study 1. Data for Latency to PersistSlp, NREM Sleep, REM Sleep are organized by TST duration.
participants had before entering the inpatient portion of the
protocol. It is possible this delaying pattern also reveals the
inherent onset of the sleep propensity rhythm once sleep debt
is partially recovered. When individuals are at home in our
industrial world, such a circadian onset might not be seen
because of prior sleep debt and because indoor light from
TV/room/computers may suppress the circadian signal to go
to sleep and also phase delay the circadian system. Conversely,
symptoms associated with chronic sleep loss might be synergistic
with the circadian signal to sleep, thereby decreasing Sleep
Latency when individual are living at home with self-selected
sleep schedules.
Slow Wave Sleep
It is not known why there is almost no SWS in some of the
individuals in Study 1. All data in Study 1 were scored by
the same individual, though there were different scorers for
Study 1 and Study 2. One possibility is that, according to
Rechtschaffen and Kales (1968), SWS is defined by a conservative
amplitude criterion (0.5–2 Hz, peak to peak amplitude greater
than 75 µV), reduced “sleep pressure” during sleep extension
might have contributed to a decreased amplitude (below the
necessary threshold level) (Skorucak et al., 2018). SWS has
also a great inter-individual variability (Tan et al., 2001). Age,
gender, BMI and race have also been described as factors
affecting SWS (Tucker et al., 2007;Tomfohr et al., 2010); Study
1 did not have a large enough sample size to determine if
any of these variables were important. The very low amount
of SWS at levels unexpected in a healthy population in
many participants in Study 1 makes comparison between the
studies difficult.
In Study 2, SWS decreased over the first 7 days of study,
suggestive of recovery process, as seen in TST, above.
Rapid Eye Movement Sleep
There is a striking increase in REM sleep in Study 2 over
baseline that then decreases over the first few nights (similar
to TST and NREM sleep), suggesting a rebound phenomenon
from levels during sleep episodes prior to the study; a similar
decline is seen in data from Study 1 (for which baseline data
are not available). This raises the question of whether healthy
individuals in society are REM-sleep deprived in their habitual
sleep: within an 8-h (or shorter) sleep opportunity, there may
not be sufficient time for REM sleep or at least the amount
received during a longer sleep opportunity. Additionally, since
light at night delays circadian rhythms [e.g., (Chang et al., 2015)]
and there is strong circadian regulation of REM sleep such
that highest levels are expected in early morning (i.e., near the
end of a normally timed sleep episode) (Czeisler et al., 1980c),
there may be less time for REM sleep to be expressed before
the individual awakens after a self-selected 8 (or fewer) hours
by alarm and before sleep would otherwise end. A rebound
in REM Sleep was also observed after space shuttle missions,
during which sleep was 6.5 h in duration and had more
wakefulness and less SWS in the last third of the bedrest episodes
(Dijk et al., 2001).
It is interesting to contemplate the significance of individuals
living under conditions of never getting enough TST or REM
sleep, which may be a feature of modern life. Acute sleep
deprivation (Wehr, 1990;Leibenluft and Wehr, 1992;Giedke
and Schwärzler, 2002;Benedetti and Colombo, 2011) and REM
sleep deprivation (Vogel, 1980) are both known transient anti-
depressants, and most antidepressants suppress REM sleep
(Azumi and Shirakawa, 1982;Wilson and Argyropoulos, 2005).
On the other hand, in normal participants, chronic sleep
restriction (sometimes called partial sleep deprivation, such as
we hypothesize occurred prior to the inpatient studies) causes
mood deterioration (Motomura et al., 2017a). Participants in
Study 1 (Wehr et al., 1993) reported improved mood and energy,
and decreased fatigue during the 28 days of 14-h time in bed
compared with the 1 week of 8-h time in bed. Motomura et al.
(2017b), have recently shown that sleep extension (after self-
imposed chronic sleep restriction) improves mood regulation via
prefrontal suppression of amygdala activity.
Given the longer sleep durations in these studies, there was no
evidence of competition (Carskadon and Dement, 1980;Duncan
et al., 2009) between REM sleep and NREM sleep during the first
days of recovery sleep. In these two protocols, however, there are
longer bedrest episodes so both more NREM sleep and REM sleep
can occur after sleep loss.
Frontiers in Physiology | 7December 2021 | Volume 12 | Article 792942
fphys-12-792942 December 17, 2021 Time: 12:6 # 8
Klerman et al. Sleep During Extended Sleep Opportunities
FIGURE 5 | Individual daily variations in hours of TST, Latency to PersistSlp,
NREM Sleep, REM Sleep in the last 14 nights of Study 1. Consecutive current
day value vs. next day value (both in hours) are plotted for each individual.
Participant numbers are the same as in Figure 4.
Patterns Across the Bedrest Episodes
For Study 1, the schedule had extra sleep opportunity with
darkness beginning earlier than habitual times, which may have
advanced the circadian rhythms and/or onset of melatonin
secretion in those participants (Wehr, 1991). For Study 2, sleep
was scheduled such that the mid-sleep times were consistent
between outpatient and inpatient portions of the study. This
may have delayed circadian timing since this schedule delays the
time of lights on.
For Study 1, an evening wake maintenance zone (i.e., time
when bedrest episodes are not usually initiated) was observed in
some people [e.g., 2 people in the 1993 report (Wehr et al., 1993)]
and in the overall average. In some individuals on some
nights, there was prolonged Wake (plotted as 0 min of TST
in Supplementary Figure 4) in the middle of the night, after
Persistent Sleep had begun, but the timing of this waking
episode was not consistent within and across participants, and
therefore is not seen in overall averages. There are historical
references of such segmented sleep patterns (Ekirch, 2005). There
may have been an evolutionary advantage for a population in
which someone is awake at all times. This bimodal pattern
also may reflect different movements during the night of dual
evening and morning (i.e., E and M) circadian oscillators
(Wehr et al., 1995).
There were large inter-individual and intra-individual
differences in all sleep metrics (Figures 4,5and Supplementary
Figure 4). To what extent these differences in Study 1 are
due to different behaviors (e.g., caffeine use, light exposure)
during the 10 h each participant was not in the research
facility and/or differences in physiology is not known. Two
potential reason for this intra-individual variability are (i)
sleep intensity (e.g., as measured by delta power in the EEG)
and (ii) the combination of the variability of duration of
NREM-REM sleep cycles and the increased probably of
awakening from REM sleep (vs. NREM sleep) (Dijk et al.,
2001). Reasons for this intra-individual variability should be
investigated. Longer-term intra- and inter-individual variability
is affected by photoperiod in humans and other animals,
with longer sleep episodes across the month (with nighttime
light from the moon) (Casiraghi et al., 2021) and during
seasons with longer durations of darkness [e.g., (Wehr, 1991;
van Hasselt et al., 2020)].
Given that none of these participants had a history of
insomnia disorder, it is noteworthy that during this extended
sleep opportunity protocol, many of the participants exhibited
sleep patterns similar to those of insomnia patients (e.g.,
long sleep latency, substantial wake after sleep onset,
and early morning awakenings). Moreover, like many
insomnia patients, after night with those sleep patterns,
there were following nights with longer sleep. Therefore,
for some individuals, the night-to-night variability may
not in itself be pathologic. It is possible that individuals
with insomnia and an 8-h bedrest episode have similar
patterns (Bianchi, 2017) that may contribute to complaints of
short sleep duration.
Ecological Significance
The question arises as to whether any relatively stable on
average value in this sleep satiation protocol reflects sleep need
for an individual person. If so, do all physiological processes
Frontiers in Physiology | 8December 2021 | Volume 12 | Article 792942
fphys-12-792942 December 17, 2021 Time: 12:6 # 9
Klerman et al. Sleep During Extended Sleep Opportunities
(e.g., metabolism, immune function, mood regulation, subjective
alertness) require the same amount of TST or NREM sleep?
Or, does the TST or NREM sleep value reflect a ceiling of
what the physiology allows or a compromise among multiple
influences? Additional experiments are required to address these
questions, including the definition and physiology underlying
sleep need. One possibility is that sleep need is reflected by TST
and sleep pressure is reflected by SWS. However, the fact that
relative SWS is preserved in chronic sleep restriction studies
(Belenky et al., 2003;Van Dongen et al., 2003;McHill et al.,
2018) as neurobehavioral performance and glucose metabolism
deteriorate, suggests that TST may be a better marker of sleep
need for these physiological functions than SWS. Note, however,
both TST and SWS require the individual to fall asleep before they
can be expressed; if there are problems falling asleep [e.g., due
to age-related changes in neuron number (Lu et al., 2000;Lim
et al., 2014)], then these metrics may not accurately reflect sleep
need or pressure.
In summary, data from two studies with extended sleep
opportunities reveal consistent evidence of self-selected chronic
sleep restriction in individuals living at home, and a multi-
day time course of recovery from this chronic insufficient
sleep. These individuals may not have recognized that their
self-selected sleep patterns were not allowing for sufficient
sleep. Once steady state-values were reached, there were large
nightly variations in each metric. The steady-state value of
TST was more than 8–1/2 h; no individual consistently slept
more than 10 h. Therefore, there was no evidence that healthy
people can sustain over multiple days, sleeping throughout
long (e.g., >10 h) bedrest episodes. Consistent sleep durations
more than 10 h may have a pathophysiological basis (e.g.,
sleep disorder or other disease) that should be diagnosed and
treated. The long sleep episodes that some people have on
a free day and some people associate with over-sleeping are
probably a transient self-limiting phenomenon, suggesting that
people can not consistent over-sleep in the same way that they
can consistently over-eat. Additional research is required to
document inter-individual and intra-individual variability and
how many nights are required for recovery of other physiologic
and behavioral processes (e.g., mood, metabolism, learning) from
insufficient sleep.
The original contributions presented in the study are included
in the article/Supplementary Material, further inquiries can be
directed to the corresponding author/s.
The studies involving human participants were reviewed
and approved by NIH Intramural Research Program IRB
(Study 1) or Partner’s Healthcare IRB (Study 2). The
patients/participants provided their written informed consent
to participate in this study.
EK, GB, and TW conducted the experiments and were
responsible for the primary analyses. CC and TW obtained
the funding. EK conducted the secondary analyses. All authors
contributed to writing and editing of the manuscript.
Supporting grants or fellowships: NIH intramural funds. NIH
K24-HL105664 (EK), R01-GM-105018, P01-AG009975, R01-
HL128538, M01-RR02635, and M01-RR01066.
The authors would like to thank NIH, BWH, and Harvard
Catalyst research facilities, resources, and staff.
The Supplementary Material for this article can be found
online at:
Arnal, P. J., Sauvet, F., Leger, D., van Beers, P., Bayon, V., Bougard, C., et al.
(2015). Benefits of sleep extension on sustained attention and sleep pressure
before and during total sleep deprivation and recovery. Sleep 38, 1935–1943.
doi: 10.5665/sleep.5244
Azumi, K., and Shirakawa, S. (1982). Characteristics of spindle activity and their
use in evaluation of hypnotics. Sleep 5, 95–105. doi: 10.1093/sleep/5.1.95
Belenky, G., Wesensten, N. J., Thorne, D. R., Thomas, M. L., Sing, H. C., Redmond,
D. P., et al. (2003). Patterns of performance degradation and restoration during
sleep restriction and subsequent recovery: a sleep dose-response study. J. Sleep
Res. 12, 1–12. doi: 10.1046/j.1365-2869.2003.00337.x
Benedetti, F., and Colombo, C. (2011). Sleep deprivation in mood disorders.
Neuropsychobiology 64, 141–151. doi: 10.1159/000328947
Bianchi, M. (2017). Chronic insomnia. Semin. Neurol. 37, 433–438. doi: 10.1055/s-
Carskadon, M. A. (1986). Guidelines for the Multiple Sleep Latency Test (MSLT):
a standard measure of sleepiness. Sleep 9, 519–524. doi: 10.1093/sleep/9.4.519
Casiraghi, L., Spiousas, I., Dunster, G. P., McGlothlen, K., Fernández-Duque, E.,
Valeggia, C., et al. (2021). Moonstruck sleep: synchronization of human sleep
with the moon cycle under field conditions. Sci. Adv. 7:eabe0465. doi: 10.1126/
Carskadon, M. A., and Dement, W. C. (1980). Distribution of REM sleep on a 90
minute sleep-wake schedule. Sleep 2, 309–317.
Chang, A.-M., Aeschbach, D., Duffy, J. F., and Czeisler, C. A. (2015). Evening use
of light-emitting eReaders negatively affects sleep, circadian timing, and next-
morning alertness. Proc. Natl. Acad. Sci. U.S.A. 112, 1232–1237. doi: 10.1073/
Cohen, D. A., Wang, W., Wyatt, J. K., Kronauer, R. E., Dijk, D.-J., Czeisler,
C. A., et al. (2010). Uncovering residual effects of chronic sleep loss on
human performance. Sci. Transl. Med. 2:14ra3. doi: 10.1126/scitranslmed.300
Frontiers in Physiology | 9December 2021 | Volume 12 | Article 792942
fphys-12-792942 December 17, 2021 Time: 12:6 # 10
Klerman et al. Sleep During Extended Sleep Opportunities
Czeisler, C. A., Borbely, A. A., Hume, K. I., Kobayashi, T., Kronauer, R. E., Shulz, H.,
et al. (1980a). Glossary of standardized terminology for sleep-biological rhythm
research. Sleep 2, 287–288.
Czeisler, C. A., Weitzman, E. D., Moore-Ede, M. C., Zimmerman, J. C., and
Knauer, R. S. (1980b). Human sleep: its duration and organization depend
on its circadian phase. Science 210, 1264–1267. doi: 10.1126/science.74
Czeisler, C. A., Zimmerman, J. C., Ronda, J. M., Moore-Ede, M. C., and Weitzman,
E. D. (1980c). Timing of REM sleep is coupled to the circadian rhythm of body
temperature in man. Sleep 2, 329–346.
Dijk, D.-J., Neri, D. F., Wyatt, J. K., Ronda, J. M., Riel, E., Ritz-De Cecco, A., et al.
(2001). Sleep, performance, circadian rhythms, and light-dark cycles during
two space shuttle flights. Am. J. Physiol. Regul. Integr. Comp. Physiol. 281,
R1647–R1664. doi: 10.1152/ajpregu.2001.281.5.R1647
Duncan, W. C., Barbato, G., Fagioli, I., Garcia-Borreguero, D., and Wehr, T. A.
(2009). A biphasic daily pattern of slow wave activity during a two-day 90-
minute sleep wake schedule. Arch. Ital. Biol. 147, 117–130.
Ekirch, A. R. (2005). At Day’s Close: Night in Time’s Past. New York, NY: W.W.
Norton and Co.
Giedke, H., and Schwärzler, F. (2002). Therapeutic use of sleep deprivation in
depression. Sleep Med. Rev. 6, 361–377.
Harrison, Y., and Horne, J. A. (1996). Long-term extension to sleep -are we really
chronically sleep deprived? Psychophysiology 33, 22–30. doi: 10.1111/j.1469-
Horne, J. A. (1988). “Core and optional sleep,” in Why We Sleep: The Functions of
Sleep in Humans and Other Mammals, ed. J. A. Horne (New York, NY: Oxford
University Press), 180–217.
Jewett, M. E., Wyatt, J. K., Ritz-De Cecco, A., Khalsa, S. B., Dijk, D., and Czeisler,
C. A. (1999). Time course of sleep inertia dissipation in human performance
and alertness. J. Sleep Res. 8, 1–8. doi: 10.1111/j.1365-2869.1999.00128.x
Kamdar, B. B., Kaplan, K. A., Kezirian, E. J., and Dement, W. C. (2004). The impact
of extended sleep on daytime alertness, vigilance, and mood. Sleep Med. 5,
441–448. doi: 10.1016/j.sleep.2004.05.003
Klerman, E. B., Davis, J. B., Duffy, J. F., Dijk, D.-J., and Kronauer, R. E. (2004).
Older people awaken more frequently but fall back asleep at the same rate as
younger people. Sleep 27, 793–798. doi: 10.1093/sleep/27.4.793
Klerman, E. B., and Dijk, D.-J. (2008). Age-related reduction in the maximal
capacity for sleep–implications for insomnia. Curr. Biol. 18, 1118–1123. doi:
Kripke, D. F. (2004). Do we sleep too much? Sleep 27, 13–14.
Leibenluft, E., and Wehr, T. A. (1992). Is sleep deprivation useful in the treatment
of depression? Am. J. Psychiatry 149, 159–168. doi: 10.1176/ajp.149.2.159
Lim, A. S. P., Ellison, B. A., Wang, J. L., Yu, L., Schneider, J. A., Buchman,
A. S., et al. (2014). Sleep is related to neuron numbers in the ventrolateral
preoptic/intermediate nucleus in older adults with and without Alzheimer’s
disease. Brain 137, 2847–2861. doi: 10.1093/brain/awu222
Lu, J., Greco, M. A., Shiromani, P., and Saper, C. B. (2000). Effect of lesions of
the ventrolateral preoptic nucleus on NREM and REM sleep. J. Neurosci. 20,
3830–3842. doi: 10.1523/JNEUROSCI.20-10- 03830.2000
McHill, A. W., Hull, J. T., Wang, W., Czeisler, C. A., and Klerman,
E. B. (2018). Chronic sleep curtailment, even without extended
(>16-h) wakefulness, degrades human vigilance performance. Proc.
Natl. Acad. Sci. U.S.A. 115, 6070–6075. doi: 10.1073/pnas.170669
Motomura, Y., Katsunuma, R., Yoshimura, M., and Mishima, K. (2017a). Two
days’ sleep debt causes mood decline during resting state via diminished
amygdala-prefrontal connectivity. Sleep 40:zsx133. doi: 10.1093/sleep/zsx133
Motomura, Y., Kitamura, S., Nakazaki, K., Oba, K., Katsunuma, R., Terasawa, Y.,
et al. (2017b). Recovery from unrecognized sleep loss accumulated in daily
life improved mood regulation via prefrontal suppression of amygdala activity.
Front. Neurol. 8:306. doi: 10.3389/fneur.2017.00306
Mullington, J. M., Haack, M., Toth, M., Serrador, J. M., and Meier-Ewert,
H. K. (2009). Cardiovascular, inflammatory, and metabolic consequences of
sleep deprivation. Prog. Cardiovasc. Dis. 51, 294–302. doi: 10.1016/j.pcad.2008.
National Center on Sleep Disorders Research (2011). National Institutes of Health
Sleep Disorders Research Plan. Bethesda, MD.
Phillips, A. J. K., Klerman, E. B., and Butler, J. P. (2017). Modeling the adenosine
system as a modulator of cognitive performance and sleep patterns during sleep
restriction and recovery. PLoS Comput. Biol. 13:e1005759. doi: 10.1371/journal.
Rechtschaffen, A., and Kales, A. (1968). A Manual of Standardized Terminology,
Techniques and Scoring System for Sleep Stages of Human Subjects. Washington,
DC: United States Government Printing Office.
Roehrs, T., Shore, E., Papineau, K., Rosenthal, L., and Roth, T. (1996). A two-week
sleep extension in sleepy normals. Sleep 19, 576–582.
Roehrs, T., Timms, V., Zwyghuizen-Doorenbos, A., and Roth, T. (1989). Sleep
extension in sleepy and alert normals. Sleep 12, 449–457. doi: 10.1093/sleep/
Rupp, T. L., Wesensten, N. J., Bliese, P. D., and Balkin, T. J. (2009). Banking sleep:
realization of benefits during subsequent sleep restriction and recovery. Sleep
32, 311–321. doi: 10.1093/sleep/32.3.311
Scott, J. P. R., McNaughton, L. R., and Polman, R. C. J. (2006). Effects of sleep
deprivation and exercise on cognitive, motor performance and mood. Physiol.
Behav. 87, 396–408. doi: 10.1016/j.physbeh.2005.11.009
Skorucak, J., Arbon, E. L., Dijk, D.-J., and Achermann, P. (2018). Response to
chronic sleep restriction, extension, and subsequent total sleep deprivation
in humans: adaptation or preserved sleep homeostasis? Sleep 41, 1–17. doi:
Tan, X., Campbell, I. G., and Feinberg, I. (2001). Internight reliability and
benchmark values for computer analyses of non-rapid eye movement (NREM)
and REM EEG in normal young adult and elderly subjects. Clin. Neurophysiol.
112, 1540–1552. doi: 10.1016/S1388-2457(01)00570- 3
Tomfohr, L. M., Ancoli-Israel, S., and Dimsdale, J. E. (2010). Childhood
socioeconomic status and race are associated with adult sleep. Behav. Sleep Med.
8, 219–230. doi: 10.1080/15402002.2010.509236
Tucker, A. M., Dinges, D. F., and Van Dongen, H. P. A. (2007). Trait interindividual
differences in the sleep physiology of healthy young adults. J. Sleep Res. 16,
170–180. doi: 10.1111/j.1365-2869.2007.00594.x
Van Cauter, E., Spiegel, K., Tasali, E., and Leproult, R. (2008). Metabolic
consequences of sleep and sleep loss. Sleep Med. 9, S23–S28. doi: 10.1016/S1389-
Van Dongen, H. P. A., Maislin, G., Mullington, J. M., and Dinges, D. F.
(2003). The cumulative cost of additional wakefulness: dose-response effects on
neurobehavioral functions and sleep physiology from chronic sleep restriction
and total sleep deprivation. Sleep 26, 117–126. doi: 10.1093/sleep/26.2.117
van Hasselt, S. J., Rusche, M., Vyssotski, A. L., Verhulst, S., Rattenborg, N. C., and
Meerlo, P. (2020). Sleep time in the european starling is strongly affected by
night length and moon phase. Curr. Biol. 30, 1664–1671.e2. doi: 10.1016/j.cub.
Vogel, G. W. (1980). Improvement of depression by REM sleep deprivation: new
findings and a theory. Arch. Gen. Psychiatry 37:247. doi: 10.1001/archpsyc.1980.
Webb, W. B. (1986). Enhanced slow sleep in extended sleep. Electroencephalogr.
Clin. Neurophysiol. 64, 27–30. doi: 10.1016/0013-4694(86)90040- 4
Wehr, T. A. (1990). “Effects of wakefulness and sleep on depression and mania,
in Sleep and Biological Rhythms, eds J. Montplaisir and R. Godbout (London:
Oxford Press), 42–86.
Wehr, T. A. (1991). The durations of human melatonin secretion and sleep
respond to changes in daylength (photoperiod). J. Clin. Endocrinol. Metab. 73,
1276–1280. doi: 10.1210/jcem-73- 6-1276
Wehr, T. A. (2001). Photoperiodism in humans and other primates: evidence
and implications. J. Biol. Rhythms 16, 348–364. doi: 10.1177/0748730011290
Wehr, T. A., Moul, D. E., Barbato, G., Giesen, H. A., Seidel, J. A., Barker, C., et al.
(1993). Conservation of photoperiod-responsive mechanisms in humans. Am.
J. Physiol. 265, R846–R857. doi: 10.1152/ajpregu.1993.265.4.R846
Wehr, T. A., Schwartz, P. J., Turner, E. H., Feldman-Naim, S., Drake, C. L.,
and Rosenthal, N. E. (1995). Bimodal patterns of human melatonin secretion
consistent with a two-oscillator model of regulation. Neurosci. Lett. 194, 105–
108. doi: 10.1016/0304-3940(95)11740- N
Wilson, S., and Argyropoulos, S. (2005). Antidepressants and sleep: a qualitative
review of the literature. Drugs 65, 927–947. doi: 10.2165/00003495-200565070-
Frontiers in Physiology | 10 December 2021 | Volume 12 | Article 792942
fphys-12-792942 December 17, 2021 Time: 12:6 # 11
Klerman et al. Sleep During Extended Sleep Opportunities
Conflict of Interest: Financial Disclosure
EK (2019-present) has consulted for Pfizer Pharmaceuticals, the National Sleep
Foundation, Sanofi-Genzyme, and Circadian Therapeutics; received travel support
from Society for Reproductive Investigation, the Sleep Research Society, the
National Sleep Foundation; The World Conference of Chronobiology, and the
Gordon Research Conferences. She reviewed payment from the Puerto Rico Trust
for a grant review, and received a grant from the Australia Harvard Club. GB
(2019-present) has no outside funding to disclose. CC reports grants to BWH from
FAA, NHLBI, NIA, NIOSH, NASA, and DOD; is/was a paid consultant to AARP,
American Academy of Dental Sleep Medicine, Eisenhower Medical Center, Emory
University, Inselspital Bern, Institute of Digital Media and Child Development,
Klarman Family Foundation, M. Davis and Co., Physician’s Seal, Sleep Research
Society Foundation, State of Washington Board of Pilotage Commissioners,
Tencent Holdings Ltd., Teva Pharma Australia, UC San Diego, University of
Washington, and Vanda Pharmaceuticals Inc., in which Czeisler also holds
an equity interest; received travel support from Annenberg Center for Health
Sciences at Eisenhower, Aspen Brain Institute, Bloomage International Investment
Group, Inc., UK Biotechnology and Biological Sciences Research Council,
Bouley Botanical, Stanley Ho Medical Development Foundation, European
Biological Rhythms Society, German National Academy of Sciences (Leopoldina),
Illuminating Engineering Society, National Safety Council, National Sleep
Foundation, Society for Research on Biological Rhythms, Sleep Research Society
Foundation, Stanford Medical School Alumni Association, Tencent Holdings
Ltd., University of Zurich, and Vanda Pharmaceuticals Inc., Ludwig-Maximilians-
Universität München, National Highway Transportation Safety Administration,
Office of Naval Research, Salk Institute for Biological Studies/Fondation Ipsen;
receives research/education support through BWH from Cephalon, Mary Ann
& Stanley Snider via Combined Jewish Philanthropies, Harmony Biosciences
LLC, Jazz Pharmaceuticals PLC Inc., Johnson & Johnson, NeuroCare, Inc.,
Philips Respironics Inc.,/Philips Homecare Solutions, Regeneron Pharmaceuticals,
Regional Home Care, Teva Pharmaceuticals Industries Ltd., Sanofi SA, Optum,
ResMed, San Francisco Bar Pilots, Sanofi, Schneider, Simmons, Sysco, Philips,
Vanda Pharmaceuticals; is/was an expert witness in legal cases, including those
involving Advanced Power Technologies, Aegis Chemical Solutions LLC, Amtrak;
Casper Sleep Inc., C&J Energy Services, Catapult Energy Services Group, LLC,
Covenant Testing Technologies, LLC, Dallas Police Association, Enterprise
Rent-A-Car, Espinal Trucking/Eagle Transport Group LLC/Steel Warehouse
Inc., FedEx, Greyhound Lines Inc.,/Motor Coach Industries/FirstGroup America,
Pomerado Hospital/Palomar Health District, PAR Electrical Contractors Inc.,
Product & Logistics Services LLC/Schlumberger Technology Corp.,/Gelco
Fleet Trust, Puckett Emergency Medical Services LLC, South Carolina Central
Railroad Company LLC, Union Pacific Railroad, United Parcel Service/UPS
Ground Freight Inc., and Vanda Pharmaceuticals; serves as the incumbent of an
endowed professorship provided to Harvard University by Cephalon, Inc.; and
receives royalties from McGraw Hill, and Philips Respironics for the Actiwatch-2
and Actiwatch Spectrum devices. Czeisler’s interests were reviewed and are
managed by the Brigham and Women’s Hospital and Mass General Brigham
in accordance with their conflict of interest policies. TW (2019-present) has no
outside funding to disclose.
Non-Financial Disclosure
EK: partner owns Chronsulting. GB, CC, and TW: no non-financial disclosures.
Publisher’s Note: All claims expressed in this article are solely those of the authors
and do not necessarily represent those of their affiliated organizations, or those of
the publisher, the editors and the reviewers. Any product that may be evaluated in
this article, or claim that may be made by its manufacturer, is not guaranteed or
endorsed by the publisher.
Copyright © 2021 Klerman, Barbato, Czeisler and Wehr. This is an open-access
article distributed under the terms of the Creative Commons Attribution License
(CC BY). The use, distribution or reproduction in other forums is permitted, provided
the original author(s) and the copyright owner(s) are credited and that the original
publication in this journal is cited, in accordance with accepted academic practice.
No use, distribution or reproduction is permitted which does not comply with
these terms.
Frontiers in Physiology | 11 December 2021 | Volume 12 | Article 792942
... Regular sleep timing at the appropriate circadian time is hypothesized to be important for optimal sleep, functioning, and health, 6 yet variability in the timing of sleep and light exposure are common in modern society. 7 While we often consider sleep timing variability to be a consequence of artificial light availability, experimental studies simulating the short photoperiod that occurs during the winter season (ie, 14 hours of darkness/sleep opportunity per day) reveal that variability in sleep latency and duration may have even occurred under such conditions (ie, with more extended darkness and opportunities for sleep) prior to the widespread availability of artificial light, 8 consistent with historical records from that era. 9 In contrast to the stability of exposure to the solar light-dark cycle in that preindustrial era, in modern society variability in the timing of sleep is typically associated with variability in the timing of exposure to light, which can induce circadian misalignment. Growing evidence indicates that circadian misalignment is associated with adverse health outcomes, including metabolic disorders such as obesity and diabetes, cardiovascular disease, immune dysfunction, cancer, and impaired mental health. ...
Full-text available
Objective: To develop and present consensus findings of the National Sleep Foundation sleep timing and variability panel regarding the impact of sleep timing variability on health and performance. Methods: The National Sleep Foundation assembled a panel of sleep and circadian experts to evaluate the scientific evidence and conduct a formal consensus and voting procedure. A systematic literature review was conducted using the NIH National Library of Medicine PubMed database, and panelists voted on the appropriateness of 3 questions using a modified Delphi RAND/UCLA Appropriateness Method with 2 rounds of voting. Results: The literature search and panel review identified 63 full text publications to inform consensus voting. Panelists achieved consensus on each question: (1) is daily regularity in sleep timing important for (a) health or (b) performance? and (2) when sleep is of insufficient duration during the week (or work days), is catch-up sleep on weekends (or non-work days) important for health? Based on the evidence currently available, panelists agreed to an affirmative response to all 3 questions. Conclusions: Consistency of sleep onset and offset timing is important for health, safety, and performance. Nonetheless, when insufficient sleep is obtained during the week/work days, weekend/non-work day catch-up sleep may be beneficial.
... As stated by Scorucack et al., "in the search for the mechanisms underlying the negative consequences of insufficient sleep, the implication of a REM sleep deficit should be considered". Klerman et al. [84] have recently reported a striking increase in REM sleep during sleep extension over baseline, suggesting a rebound phenomenon from a possible REM deprivation that occurs in a habitual sleep opportunity. A rebound in REM sleep has been also observed after a space shuttle mission, during which sleep duration was approximately 6.5 h [85]. ...
Full-text available
Standard polysomnographic analysis of sleep has not provided evidence of an objective measure of sleep quality; however, factors such as sleep duration and sleep efficiency are those more consistently associated with the subjective perception of sleep quality. Sleep reduction as currently occurs in our 24/7 society has had a profound impact on sleep quality; the habitual sleep period should fit within what is a limited nighttime window and may not be sufficient to satisfy the whole sleep process; moreover, the use of artificial light during the evening and early night hours can delay and disturb the circadian rhythms, especially affecting REM sleep. The correct phase relationship of the sleep period with the circadian pacemaker is an important factor to guarantee adequate restorative sleep duration and sleep continuity, thus providing the necessary background for a good night’s sleep. Due to the fact that REM sleep is controlled by the circadian clock, it can provide a window-like mechanism that defines the termination of the sleep period when there is still the necessity to complete the sleep process (not only wake-related homeostasis) and to meet the circadian end of sleep timing. An adequate amount of REM sleep appears necessary to guarantee sleep continuity, while periodically activating the brain and preparing it for the return to consciousness.
Objective: Mental health and substance use disorders are commonly associated with disrupted sleep and circadian rest-activity rhythms. How these disorders in combination relate to sleep and circadian organization is not well studied. We provide here the first quantitative assessment of sleep and rest-activity rhythms in inpatients with complex concurrent disorders, taking into account categories of substance use (stimulant vs. stimulant and opioid use) and psychiatric diagnosis (psychotic disorder and mood disorder). We also explore how sleep and rest-activity rhythms relate to psychiatric functioning. Methods: A total of 44 participants (10 female) between the age of 20-60 years (median = 29 years) wore wrist accelerometers over 5-70 days and completed standardized questionnaires assessing chronotype and psychiatric functioning (fatigue, psychiatric symptom severity, and impulsiveness). To examine potential influences from treatment, we computed (1) length of stay; (2) days of abstinence from stimulants and opioids as a measure of withdrawal; and (3) a sedative load based on prescribed medications. Results: Participants exhibited a sustained excessive sleep duration, frequent nighttime awakenings, and advanced rest-activity phase related to sedative load. Sleep disruptions were elevated in participants with a history of opioid use. Patients with a psychotic disorder showed the longest sleep and most fragmented and irregular rest-activity patterns. Non-parametric circadian rhythm analysis revealed a high rhythm amplitude by comparison with population norms, and this was associated with greater psychiatric symptom severity. Psychiatric symptom severity was also associated with greater fatigue and later MCTQ chronotype. Conclusions: This pilot study provides initial information on the prevalence and severity of sleep and circadian rhythm disturbances in individuals with severe concurrent disorders are disturbed. The results underline the need for further studies to start to understand the role of sleep in the disease and recovery process in this understudied population.
This chapter reviews the physiology of sleep and outlines the neurocircuitry underlying sleep and wakefulness. We discuss sleep disorders such as REM-sleep behavior disorder, insomnia, hypersomnia, sleep-disordered breathing, and the role of circadian and sleep disturbances in the pathogenesis of neurodegenerative diseases.
Full-text available
Before the availability of artificial light, moonlight was the only source of light sufficient to stimulate nighttime activity; still, evidence for the modulation of sleep timing by lunar phases is controversial. Here, we use wrist actimetry to show a clear synchronization of nocturnal sleep timing with the lunar cycle in participants living in environments that range from a rural setting with and without access to electricity in indigenous Toba/Qom communities in Argentina to a highly urbanized postindustrial setting in the United States. Our results show that sleep starts later and is shorter on the nights before the full moon when moonlight is available during the hours following dusk. Our data suggest that moonlight likely stimulated nocturnal activity and inhibited sleep in preindustrial communities and that access to artificial light may emulate the ancestral effect of early-night moonlight.
Full-text available
Sleep loss causes profound cognitive impairments and increases the concentrations of adenosine and adenosine A1 receptors in specific regions of the brain. Time courses for performance impairment and recovery differ between acute and chronic sleep loss, but the physiological basis for these time courses is unknown. Adenosine has been implicated in pathways that generate sleepiness and cognitive impairments, but existing mathematical models of sleep and cognitive performance do not explicitly include adenosine. Here, we developed a novel receptor-ligand model of the adenosine system to test the hypothesis that changes in both adenosine and A1 receptor concentrations can capture changes in cognitive performance during acute sleep deprivation (one prolonged wake episode), chronic sleep restriction (multiple nights with insufficient sleep), and subsequent recovery. Parameter values were estimated using biochemical data and reaction time performance on the psychomotor vigilance test (PVT). The model closely fit group-average PVT data during acute sleep deprivation, chronic sleep restriction, and recovery. We tested the model’s ability to reproduce timing and duration of sleep in a separate experiment where individuals were permitted to sleep for up to 14 hours per day for 28 days. The model accurately reproduced these data, and also correctly predicted the possible emergence of a split sleep pattern (two distinct sleep episodes) under these experimental conditions. Our findings provide a physiologically plausible explanation for observed changes in cognitive performance and sleep during sleep loss and recovery, as well as a new approach for predicting sleep and cognitive performance under planned schedules.
Full-text available
Many modern people suffer from sleep debt that has accumulated in everyday life but is not subjectively noticed [potential sleep debt (PSD)]. Our hypothesis for this study was that resolution of PSD through sleep extension optimizes mood regulation by altering the functional connectivity between the amygdala and prefrontal cortex. Fifteen healthy male participants underwent an experiment consisting of a baseline (BL) evaluation followed by two successive interventions, namely, a 9-day sleep extension followed by one night of total sleep deprivation (TSD). Tests performed before and after the interventions included a questionnaire on negative mood and neuroimaging with arterial spin labeling MRI for evaluating regional cerebral blood flow (rCBF) and functional connectivity. Negative mood and amygdala rCBF were significantly reduced after sleep extension compared with BL. The amygdala had a significant negative functional connectivity with the medial prefrontal cortex (FCamg–MPFC), and this negative connectivity was greater after sleep extension than at BL. After TSD, these indices reverted to the same level as at BL. An additional path analysis with structural equation modeling showed that the FCamg–MPFC significantly explained the amygdala rCBF and that the amygdala rCBF significantly explained the negative mood. These findings suggest that the use of our sleep extension protocol normalized amygdala activity via negative amygdala–MPFC functional connectivity. The resolution of unnoticed PSD may improve mood by enhancing frontal suppression of hyperactivity in the amygdala caused by PSD accumulating in everyday life.
Sleep is considered to be of crucial importance for performance and health, yet much of what we know about sleep is based on studies in a few mammalian model species under strictly controlled laboratory conditions. Data on sleep in different species under more natural conditions may yield new insights in the regulation and functions of sleep. We therefore performed a study with miniature electroencephalogram (EEG) data loggers in starlings under semi-natural conditions, group housed in a large outdoor enclosure with natural temperature and light. The birds showed a striking 5-h difference in the daily amount of non-rapid-eye-movement (NREM) sleep between winter and summer. This variation in the amount of NREM sleep was best explained by night length. Most sleep occurred during the night, but when summer nights became short, the animals displayed mid-day naps. The decay of NREM sleep spectral power in the slow-wave range (1.1–4.3 Hz) was steeper in the short nights than in the longer nights, which suggests that birds in summer have higher sleep pressure. Additionally, sleep was affected by moon phase, with 2 h of NREM sleep less during full moon. The starlings displayed very little rapid-eye-movement (REM) sleep, adding up to 1.3% of total sleep time. In conclusion, this study demonstrates a pronounced phenotypical flexibility in sleep in starlings under semi-natural conditions and shows that environmental factors have a major impact on the organization of sleep and wakefulness.
Significance Millions of individuals obtain insufficient sleep on a daily basis, which leads to impaired performance. Whether these decrements are caused by short sleep duration or extended wakefulness is unknown. In this study, healthy volunteers were randomized into either a chronically sleep-restricted or control protocol while living on a 20-h “day,” thus enabling short sleep without extended wakefulness. We demonstrate that chronic insufficient sleep, even without extended wakefulness, leads to neurobehavioral performance decrements at all times of the day, even when the circadian system is promoting arousal. These findings have implications for the understanding of basic physiology, the substantial population who chronically obtains insufficient sleep, and all of us who depend on sleep-restricted individuals working in safety-sensitive occupations.
Sleep is regulated by a homeostatic process which in the two-process model of human sleep regulation is represented by EEG slow-wave activity (SWA). Many studies of acute manipulation of wake duration have confirmed the precise homeostatic regulation of SWA in rodents and humans. However, some chronic sleep restriction studies in rodents show that the sleep homeostatic response, as indexed by SWA, is absent or diminishes suggesting adaptation occurs. Here, we investigate the response to 7 days of sleep restriction (6 h time in bed) and extension (10 h time in bed) as well as the response to subsequent total sleep deprivation in 35 healthy participants in a cross-over design. The homeostatic response was quantified by analyzing sleep structure and SWA measures. Sleep restriction resulted primarily in a reduction of REM sleep. SWA and accumulated SWA (slow-wave energy) were not much affected by sleep extension/restriction. The SWA responses did not diminish significantly in the course of the intervention and did not deviate significantly from the predictions of the two-process model. The response to total sleep deprivation consisted of an increase in SWA, rise rate of SWA and SWE and did not differ between the two conditions. The data show that changes in sleep duration within an ecologically relevant range, have a marked effect on REM sleep and that SWA responds in accordance with predictions based on a saturating exponential increase during wake and an exponential decline in sleep of homeostatic sleep pressure during both chronic sleep restriction and extension.
Insomnia is characterized by difficulty falling asleep or staying asleep, with consequent daytime impairment of mental and/or physical function. A detailed clinical history reveals the relative impact of a variety of different contributing and perpetuating factors, which then informs prioritization among different treatment options. Nonpharmacological approaches, especially the validated approach of cognitive–behavioral therapy for insomnia, are preferred over hypnotic medications. If hypnotics are chosen, the goal is short-term interventions after a careful risk-benefit assessment and shared decision-making with the patient. Although objective testing via polysomnography is not routinely indicated, such investigations can be informative in those at risk for concurrent primary sleep disorders, and in those who are treatment refractory. Circadian rhythm disorders can present with insomnia complaints, but are managed with chronotherapy. Whatever management pathway is pursued, the response to therapy should be anchored in improvements in daytime function.
Study objectives Sleep debt has been suggested to evoke emotional instability by diminishing the suppression of the amygdala by the medial prefrontal cortex (MPFC). Here, we investigated how short-term sleep debt affects resting-state functional connectivity between the amygdala and MPFC, subjective mood, and sleep parameters. Methods Eighteen healthy adult men aged 29±8.24 years participated in a 2-day sleep control session (SC; time in bed, 9 h) and 2-day sleep debt session (SD; time in bed, 3 h). On day 2 of each session, resting-state functional magnetic resonance imaging (fMRI) was performed, followed immediately by measuring subjective mood on the State-Trait Anxiety Inventory-State subscale (STAI-S). Results STAI-S score was significantly increased and functional connectivity between the amygdala and MPFC was significantly decreased in SD compared with SC. Significant correlations were observed between reduced rapid eye movement (REM) sleep and reduced left amygdala-MPFC functional connectivity (FCL_amg-MPFC) and between reduced FCL_amg-MPFC and increased STAI-S score in SD compared with SC. Conclusions These findings suggest that reduced MPFC functional connectivity of amygdala activity is involved in mood deterioration under sleep debt, and that REM sleep reduction is involved in functional changes in the corresponding brain regions. Having adequate REM sleep may be important for mental health maintenance.
The multiple sleep latency test (MSLT) is used in the assessment and diagnosis of disorders of excessive somnolence and to evaluate daytime sleepiness in relation to various therapeutic or experimental manipulations, such as administering drugs and altering the length of timing of nocturnal sleep. The repeated measurement of sleep latency across a day provides direct access to the diurnal fraction of the sleep/wake interaction, which is of fundamental concern to the sleep specialist. Objective laboratory documentation of the clinical symptoms of slepiness well as abnormal sleep structure has greatly facilitated the diagnosis of narcolepsy, in particular, and has also been useful to determine the severity of somnolence and therapeutic response in other disorders. At the current level of clinical experience, a diagnosis of narcolepsy or other disorders of excessive somnolence usually has lifelong consequences for the patients, for example, chronic chemotherapy with psychoactive compounds, legal proscription from driving, or surgery. It therefore is incumbent upon the clinical sleep specialist to achieve as much diagnostic precision as possible. The MSLT greatly enhances the accurate diagnosis of disorders of excessive somnolence.