Content uploaded by Michelle Capozzoli
Author content
All content in this area was uploaded by Michelle Capozzoli on Jun 02, 2015
Content may be subject to copyright.
The effects of physical activity on sleep: a meta-analytic review
M. Alexandra Kredlow •Michelle C. Capozzoli •
Bridget A. Hearon •Amanda W. Calkins •
Michael W. Otto
Received: June 27, 2014 / Accepted: January 6, 2015
ÓSpringer Science+Business Media New York 2015
Abstract A significant body of research has investigated
the effects of physical activity on sleep, yet this research
has not been systematically aggregated in over a decade.
As a result, the magnitude and moderators of these effects
are unclear. This meta-analytical review examines the
effects of acute and regular exercise on sleep, incorpo-
rating a range of outcome and moderator variables. Pub-
Med and PsycINFO were used to identify 66 studies for
inclusion in the analysis that were published through May
2013. Analyses reveal that acute exercise has small bene-
ficial effects on total sleep time, sleep onset latency, sleep
efficiency, stage 1 sleep, and slow wave sleep, a moderate
beneficial effect on wake time after sleep onset, and a small
effect on rapid eye movement sleep. Regular exercise has
small beneficial effects on total sleep time and sleep effi-
ciency, small-to-medium beneficial effects on sleep onset
latency, and moderate beneficial effects on sleep quality.
Effects were moderated by sex, age, baseline physical
activity level of participants, as well as exercise type, time
of day, duration, and adherence. Significant moderation
was not found for exercise intensity, aerobic/anaerobic
classification, or publication date. Results were discussed
with regards to future avenues of research and clinical
application to the treatment of insomnia.
Keywords Sleep Physical activity Exercise
Insomnia Sleep quality
Introduction
Sleep problems, including chronic insomnia, represent a
significant public health concern. The prevalence of
insomnia symptoms ranges from 25 to 48 % (Mallon et al.,
2000; Quera-Salva et al., 1991) and insomnia diagnoses
from 4.4 to 9.5 % (Morin et al., 2006; Ohayon et al., 1997;
for review see Ohayon, 2002), indicating that sleep diffi-
culties are relatively common. Sleep problems are associ-
ated with poorer quality of life, as well as mental and
physical health issues (Kripke et al., 2002; Simon & Vo-
nKorff, 1997). Notably, multiple studies have demon-
strated an association between insomnia and heart disease,
the leading cause of death in the United States (Schwartz
et al., 1999). In addition to significant individual costs,
sleep difficulties have a negative impact on society; the
National Sleep Foundation (n.d.) estimates that insomnia
alone costs the United States $14 billion annually in direct
treatment-related costs and an additional $28 billion in
indirect costs, due to factors such as lost productivity and
absenteeism.
Despite the availability of efficacious pharmacologic
(Glass et al., 2005; Smith et al., 2002) and cognitive-
behavioral (Montgomery & Dennis, 2003; Morin et al.,
1999) sleep interventions, the majority of insomnia suf-
Electronic supplementary material The online version of this
article (doi:10.1007/s10865-015-9617-6) contains supplementary
material, which is available to authorized users.
M. A. Kredlow (&)B. A. Hearon M. W. Otto
Department of Psychological and Brain Sciences, Boston
University, 648 Beacon Street, 5th Floor, Boston, MA 02215,
USA
e-mail: kredlow@bu.edu
M. C. Capozzoli
Psychology Department, University of Nebraska-Lincoln, 238
Burnett Hall, Lincoln, NE 68588, USA
A. W. Calkins
The Center for Anxiety and Traumatic Stress Disorders,
Massachusetts General Hospital/Harvard Medical School, 1
Bowdoin Square, 6th Floor, Boston, MA 02114, USA
123
J Behav Med
DOI 10.1007/s10865-015-9617-6
ferers do not seek or have access to provider-directed care.
Studies suggest that only 30 % of patients with sleep dif-
ficulties have spoken to their primary care physician about
their difficulties (Shochat et al., 1999), and only 15 % of
individuals with chronic insomnia receive any type of
treatment (Mellinger et al., 1985). Researchers have pro-
posed strategies that have the potential to improve the
dissemination of evidence-based treatments for insomnia,
such as a stepped-care and technology assisted model of
cognitive-behavioral therapy delivery (Edinger, 2009; Es-
pie, 2009). An alternative and complementary strategy is to
promote self-help interventions, such as physical activity,
that have the potential to be easily and more widely dis-
seminated.
A significant body of research has investigated the
effects of physical activity on sleep. Results generally
suggest that physical activity is beneficial for sleep; how-
ever, it is unclear how large these benefits are, and to what
extent variables such as age and type and duration of
exercise moderate these benefits (Driver & Taylor, 2000).
Meta-analytic techniques allow researchers to combine and
synthesize data from multiple studies, accounting for
varying sample sizes and study designs, in order to produce
more reliable estimates of the effects of an intervention on
outcomes of interest. Two meta-analytical reviews exam-
ining this topic were published in the 1990s, one examining
32 studies of acute and 12 studies of regular exercise
(Kubitz et al., 1996) and the other examining 38 studies of
acute exercise (Youngstedt et al., 1997). Although the
magnitude of effects varied, these reviews provided con-
verging evidence that acute and regular exercise have a
positive impact on total sleep time, sleep onset latency, and
slow wave sleep and a negative impact on rapid eye
movement sleep. Results for other important measures of
sleep, such as wake time after sleep onset and sleep effi-
ciency, were either non-significant or inconclusive.
In the over 15 years since the publication of these meta-
analyses, more than 30 studies have been conducted
examining the effects of physical activity on sleep. In
addition, the methodological rigor of studies has increased
over time, with a large number of newer studies utilizing
randomized and controlled designs. Of these newer studies,
six were examined in a narrowly focused meta-analysis by
Yang et al. (2012), which investigated how regular exer-
cise impacts sleep quality in middle-aged and older adults
with sleep complaints. Results indicated moderate effects
in the favorable direction for regular exercise on sleep
quality. Yet, this meta-analysis focused on a narrow pop-
ulation and did not examine other important metrics of
sleep.
Previous meta-analyses (Kubitz et al., 1996; Youngstedt
et al., 1997) have also examined potential moderators of
the benefits of exercise for sleep. These studies have sug-
gested that sex, age, and fitness level of individuals as well
as exercise duration may moderate the effects of physical
activity on different sleep outcomes. At the same time,
conflicting results have been found with regard to other
moderators (e.g., fitness level, time of day; Buman & King,
2010) and additional moderators have not been thoroughly
explored (e.g., participant sleep complaints, exercise type).
In the current meta-analysis, we examine the effects of
acute and regular exercise on a range of sleep variables
and explore the impact of potential moderators of these
outcomes. We hypothesize that acute and regular exercise
will have beneficial effects on objective and subjective
sleep variables. We examine this question across a large
age range.
Method
Search strategy
The search engines PubMed and PsycINFO were used to
identify studies for inclusion in the analysis that were pub-
lished through May 2013. Combinations of the key terms
‘‘physical activity,’’ ‘‘exercise,’’ ‘‘sleep,’’ and ‘‘insomnia’’
were used. In addition, the reference sections of relevant
prior meta-analyses (Kubitz et al., 1996; Yang et al., 2012;
Youngstedt et al., 1997) were reviewed to identify additional
studies. The literature search was limited to studies con-
ducted with human participants that were published in the
English language. The first and second authors indepen-
dently conducted the literature search and screened titles,
abstracts, and manuscripts of potentially eligible studies.
Study selection
Studies identified through these methods were then selec-
ted based on the following inclusion criteria: (1) experi-
mental studies examining the effect of acute or regular
physical activity on a self-reported or biological measure of
total sleep time, sleep onset latency, wake time after sleep
onset, number of awakenings after sleep onset, sleep
maintenance, sleep efficiency, stage 1, 2, 3, and 4 sleep,
slow wave sleep, rapid eye movement sleep, rapid eye
movement sleep latency, insomnia severity, sleep quality;
(2) studies examining 4 or more adult participants (i.e.,
average age C18); and (3) the presence of a control con-
dition (i.e., control comparison day/s or control group). The
sample size threshold was chosen to be consistent with
previous meta-analyses (Youngstedt et al., 1997) and allow
for inclusion of studies of acute exercise conducted in sleep
laboratories with few participants. Acute exercise was
defined as less than one week of exercise, and regular
exercise was defined as equal to or greater than one week
J Behav Med
123
of exercise. For studies utilizing self-report to assess sleep
quality, we required the use of at least a full subscale of a
validated measure. Insomnia severity was assessed using
the Insomnia Severity Index (Bastien et al., 2001). For
studies that described delivery of multi-component inter-
ventions that included physical activity (e.g., Alessi et al.,
2005), inclusion in the meta-analysis was based on whether
physical activity was judged to be the dominant component
of the intervention (e.g., the majority of the duration of the
intervention).
Our exclusion criteria were chosen to ensure that studies
included were well-controlled and ecologically valid.
Studies meeting any of the following criteria were exclu-
ded: (1) studies examining atypical exercise regimens,
defined as greater than 3 h of moderate exercise in one day,
greater than 10 miles distance in one day, or exercise
taking place during typical sleeping hours; (2) studies
examining atypical sleep regimens, such as studies with
shift workers or utilizing experimentally-induced insomnia;
(3) studies of populations with medical and psychiatric
conditions other than sleep disorders; (4) studies utilizing
an active comparison group without a placebo or no-
intervention control.
Data abstraction
Articles meeting these criteria were collected and data were
abstracted for analysis by the first author and independently
checked for accuracy by the second author. The accuracy
rate for abstraction of effect size data was 98 %.
Effect size data
For the calculation of effect sizes, means and standard
deviations were favored over other statistics (pvalues,
tvalues, Fstatistics, odds ratios, and event and non-event
rates). In instances when insufficient data were presented to
calculate an effect size, attempts were made to contact the
authors of the study. For studies that did not report whether
ttests were one or two-sided, an assumption was made that
tests were two-sided. For studies that failed to report an exact
pvalue but reported a less than statement (e.g., p\.05), a
conservative assumption was made equating the pvalue to
the stated amount (e.g., p=.05). In order to calculate the
effect sizes for between-subject designs with pre and post
data, pre-post measure correlations were needed. Recom-
mendations for imputation of pre-post correlations vary
between r=0.5 (Follmann et al., 1992) and r=0.7 (Ro-
senthal, 1991). In the current analyses, an estimate between
these two recommendations (r=0.6) was utilized when
correlations were not provided in published reports. In
addition, sensitivity analyses were conducted per Cochrane
Collaboration (2011) recommendations, using values of 0.5
and 0.7, and this did not significantly change the main out-
comes. In some studies, participants were randomized to one
of multiple regular exercise programs (Elavsky & McAuley,
2007; Frye et al., 2007; Kline et al., 2012; Pinniger et al.,
2013) or completed multiple acute exercise conditions on
different experimental days (Baekeland & Lasky, 1966;
Desjardins et al., 1974; Flausino et al., 2012; Horne & Staff,
1983; Kupfer et al., 1985; Myllymaki et al., 2012; O’Connor
et al., 1998; Passos et al., 2010; Wong et al., 2013). In these
cases, effect sizes for all exercise conditions were calculated
and data for the control condition were used multiple times
in these calculations. Lastly, some studies presented sub-
jective and objective assessment outcomes for the same
sleep variable, in which case, data from objective assess-
ments were abstracted.
For ease of interpretation, the direction of effect sizes
were coded as positive if they represented a beneficial
change and were coded as negative if they represented a
deleterious change in sleep consequent to the exercise
condition. Increases in the following variables were con-
sidered to be beneficial: total sleep time, slow wave sleep,
sleep efficiency, rapid eye movement, stage 3 sleep, stage 4
sleep, and sleep quality. Decreases in the following vari-
ables were considered to be beneficial: sleep onset latency,
rapid eye movement sleep latency, stage 1 sleep, stage 2
sleep, wake time after sleep onset, number of awakenings
after sleep onset, and insomnia severity. These determi-
nations were made based on prior research regarding the
benefits of certain sleep stages and standard interpretation
of sleep variables (Carskadon & Dement, 2000). None-
theless, we acknowledge that there is debate as to the value
of changes in certain sleep stages or metrics.
Moderator variables
Data regarding participant, exercise, sleep, and study
characteristics were abstracted to be used in moderator
analyses.
Participant characteristics Average age, sex, baseline
physical activity level, and presence of sleep complaints
were abstracted. For studies that only provided an age
range rather than mean, the median of the range was used
as the average age. The percentage of female participants
was abstracted from each study. Baseline physical activity
level was coded as ‘‘low’’ if participants were identified as
‘‘sedentary’’ (referring to the American College of Sports
Medicine [ACSM], 2013 definition of low physical activ-
ity) or the article provided evidence that participants were
engaging in less than 30 min of moderate intensity exercise
3 days per week for at least 3 months at entry into the
J Behav Med
123
study (ACSM, 2013). Baseline physical activity level was
coded as ‘‘high’’ if participants were identified as ‘‘regular
exercisers’’ or ‘‘athletes.’’ Sleep complaints were evaluated
based on study inclusion criteria.
Exercise characteristics The intensity, aerobic/anaerobic
nature, and type of exercise were abstracted for each study.
Percent of maximum heart rate, percent of heart rate reserve,
percent of VO
2
max, ventilatory threshold, and METs were
used to classify exercise interventions as low, moderate, or
vigorous intensity per ACSM standards (Garber et al., 2011).
Exercise interventions were classified as aerobic or anaero-
bic based on whether the article explicitly stated such or there
was sufficient information to deduce this classification. Type
of exercise was noted as described by each study. In addition,
we classified exercise type as either mind/body or not mind/
body exercise based on ACSM definitions (Schroeder,
2012). Specifically, yoga, pilates, tai chi, and qigong were
classified as mind/body exercise.
We also examined variables related to exercise schedul-
ing. Time of day was coded as follows: greater than 8 h
before bedtime, 3–8 h before bedtime, and less than 3 h
before bedtime. These time categories were chosen in rela-
tion to bedtime in order to examine the validity of the com-
mon sleep hygiene recommendation (Stepanski & Wyatt,
2003) that exercising close to bedtime has a negative impact
on sleep. We examined exercise duration in the following
ways: duration of bouts in minutes, days per week, total
number of weeks (regular exercise only), and total duration
in minutes. If the duration of exercise bouts or days per week
varied across the exercise program, averages were calcu-
lated. For acute studies that had participants complete mul-
tiple exercise and control days (e.g. ABAB design) but
examined sleep outcomes on each day separately (not
cumulatively), outcomes were averaged across the exercise
days and control days. In addition, for studies of regular
exercise, we examined exercise adherence. Methods of
assessing adherence varied across studies; thus, we differed
to the authors of the studies for definitions of adherence that
were appropriate to their study designs and utilized the
adherence percentages that were reported. We did not
examine adherence as a factor in acute exercise studies, as
the majority of these studies were conducted in laboratories
where experimenters observed the assigned exercise.
Sleep characteristics Studies utilized various assessment
tools to collect sleep outcomes, including sleep diaries,
self-report questionnaires, electroencephalogram (EEG),
and polysomnography (PSG). We recorded whether the
assessment tool was subjective (e.g. sleep diary) or
objective (e.g. EEG). Although subjective reports of sleep
(e.g. sleep quality) are useful in their own right and mod-
erately correlated with objective measures of sleep (Lau-
derdale et al., 2008), the objectivity of the assessment tool
may impact the accuracy of the results of each study.
Study characteristics We extracted what type of study
design was utilized and the year of publication.
Details of analysis
Effect size analyses
Data analysis was conducted using the standard software
program Comprehensive Meta-Analysis (Version 2; Boren-
stein et al., 2005). The effects for acute and regular exercise
were analyzed separately as previous meta-analyses have
demonstrated distinctions between these classes of exercise
(Kubitz et al., 1996). All effects were calculated using ran-
dom-effects models. The random-effects method was chosen
as it is recommended over the fixed-effects method to assess
social science data, which are likely to have heterogeneous
population effect sizes (Field, 2003). In order to reduce the
likelihood of drawing conclusions from limited data, aggre-
gate effect sizes derived from less than five comparisons were
not reported. We used Cohen’s d(Cohen, 1977)andits95%
confidence interval as an indicator of effect size and inter-
preted effects as small (0.2), medium (0.5) or large (0.8).
Moderator analyses
We examined whether effect sizes varied as a function of
participant, exercise, sleep, or study characteristics outlined
above. Moderator analyses were conducted when a total of
eight or more comparisons and at least three comparisons
for any particular moderator grouping were available to
contribute to the analyses. For categorical moderators,
separate effect sizes for each group were computed utiliz-
ing mixed effects analysis and Cochran’s Qtest of heter-
ogeneity was used to evaluate significance between groups.
For continuous moderators, meta-regression analyses were
used to compute unstandardized regression coefficients and
z-test were applied to evaluate significance.
Publication bias
Although there is no exact method to assess and correct for
publication bias, in addition to our comprehensive search
strategy, we used three approaches to explore the potential
effects of publication bias on our results. Fail-safe Nvalues
(Rosenthal, 1991; Rosenthal & Rubin, 1988) were calcu-
lated to estimate the number of additional studies with null
results that would be needed to reduce the overall effect
J Behav Med
123
size to non-significance. Subsequently, Rosenthal’s (Ro-
senthal, 1991) criteria were utilized to examine the
robustness of the effects; for an effect to be considered
robust, the fail-safe Nmust be greater than 5 K+ 10,
where Kequals the number of comparisons included in the
analyses. In addition, we conducted a visual inspection of
funnel plots for each outcome to evaluate symmetry rela-
tive to the mean effect size. Lastly, we conducted Trim and
Fill (Duval & Tweedie, 2000) analyses, which are a con-
servative method to estimate effect sizes accounting for
publication bias. The Trim and Fill method assumes that
funnel plot asymmetry is due to publication bias, imputes
artificial studies to balance out asymmetric funnel plots,
and recalculates estimated effect sizes and confidence
intervals based on the newly symmetrical distribution.
Risk of bias
In addition to assessing the potential for publication bias,
two independent raters assessed the quality and risk of bias
of individual studies included in our analyses. To assess the
quality of non-randomized experimental studies of acute
exercise, we assessed whether studies met the quality cri-
teria outlined by Youngstedt et al. (1997). To assess the
quality of randomized controlled trials (RCTs), the Coch-
rane Collaboration’s tool for assessing risk of bias in RCTs
(Higgins et al., 2011) was utilized. This tool was developed
to replace quality scales and checklists, which have been
shown to be inconsistent and inadequate in appraising
clinical trials (Greenland & O’Rourke, 2001; Juni et al.,
1999). Four studies of acute exercise (Alessi et al., 2005;
Jennings, 1981; Passos et al., 2010; Viana et al., 2012)
were RCTs; these trials were evaluated using both
Youngstedt’s criteria and the Cochrane Collaboration tool.
Results
Trial flow
Using the search strategy described previously, we identified
3,291 unique articles. The vast majority of these articles
(n=3,007) were excluded by title or abstract alone. Full
text articles were obtained for the 284 remaining articles, of
which 221 were excluded for various reasons (see Fig. 1).
Ultimately, 63 articles containing 66 studies were identified
that met the inclusion criteria for this meta-analysis.
Study characteristics
Of the 66 studies included in the analysis, 41 were studies
of acute exercise and 25 were studies of regular exercise.
Of the acute studies, 37 utilized within-participant open
trial designs and compared sleep following days when
participants exercised to days when the same participants
did not exercise. The remaining four were RCTs of acute
exercise. The majority of the regular exercise studies were
RCTs (n=23); two studies of regular exercise identified
matched controls to compare to a group receiving an
exercise intervention (Naylor et al., 2000; Yeh & Chang,
2012). One study (Jennings, 1981) involved an acute and
regular exercise investigation and was thus utilized in both
totals. Some studies included multiple active exercise
interventions, which resulted in multiple comparisons as
described above; 60 comparisons were made across the 41
acute exercise studies and 30 comparisons across the 25
regular exercise studies. The sample sizes reported from
here after reflect number of comparisons unless stated
otherwise. In total, 457 effect sizes were calculated (333
for acute and 124 for regular exercise).
Across all studies, 2,863 participants were included,
with a mean age of 42.0 years (SD =20.4, range
18.3–88.5) and approximately equal representation of
women and men (45 % women; SD =38.4, range 0–100).
The majority of studies recruited individuals from a com-
munity sample with no specified sleep complaints; 11 % of
studies specifically recruited individuals with sleep com-
plaints. The studies examined individuals of varying
baseline physical activity levels: 24 % of studies examined
individuals of high baseline physical activity, 23 % of
studies examined individuals of low baseline physical
activity, and the remainder examined individuals of mixed
or unknown baseline physical activity. The duration of
exercise examined for the acute studies was predominantly
1 day (with one study examining cumulative effects after
5 days). Exercise duration for the regular studies ranged
from 2 to 52 weeks.
Predominantly objective measures of sleep were used to
assess sleep outcomes in acute studies. The modal assess-
ment method for the variables of total sleep time, sleep
onset latency, wake time after sleep onset, sleep efficiency,
stage 1 sleep, stage 2 sleep, slow wave sleep, rapid eye
movement sleep, and rapid eye movement sleep latency
was PSG, while the modal assessment method for all other
sleep variables was EEG. In studies of regular exercise,
predominantly objective measures of sleep were used to
assess the outcomes of total sleep time, sleep onset latency,
and sleep efficiency. Of the 10 total sleep time effect sizes,
two were derived from EEG, two from PSG, one from
accelerometery, one from actigraphy, and four from self-
report data. Of the nine sleep onset latency effect sizes, two
were derived from EEG, three from PSG, one from actig-
raphy, and two from self-report data. Of the six sleep
efficiency effect sizes, two were derived from PSG, one
from EEG, one from actigraphy, one from accelerometery,
and one from self-report data. Only subjective measures
J Behav Med
123
were used to assess the outcome of sleep quality. Complete
summaries of participant, exercise, sleep, and study char-
acteristics for acute and regular exercise studies are pro-
vided in Online Resource 1.
Quantitative data synthesis
Acute exercise
Individual study and aggregate effect sizes for acute
exercise studies are presented in Table 1. Overall, acute
exercise had nearly-equivalent, small, beneficial effects on
total sleep time [d=0.22, 95 % CI (0.10, 0.34), p\.001;
n=41], sleep onset latency [d=0.17, 95 % CI (-0.02,
0.32), p=.03; n=35), sleep efficiency [d=0.25, 95 %
CI (0.12, 0.39), p\.001, n=28], and slow wave sleep
[d=0.19, 95 % CI (0.02, 0.35), p=.03; n=41], indi-
cating that the total sleep time, sleep efficiency, and
duration of slow wave sleep were greater, and sleep onset
latency was lesser, on days after acute exercise compared
to control days of no exercise. Acute exercise also had
small-to-medium beneficial effects on wake time after
sleep onset [d=0.38, 95 % CI (0.21, 0.55), p\.001;
n=22] and stage 1 sleep [d=0.35, 95 % CI (0.18, 0.52),
p\.001; n=20] indicating that time awake after sleep
onset and duration of stage 1 sleep was better (i.e., shorter)
on days after acute exercise compared to control days.
Acute exercise also had a small effect on rapid eye
movement sleep [d=-0.27, 95 % CI (-0.45, -0.08),
p=.005; n=40] in the negative direction, indicating that
rapid eye movement sleep duration was shorter on days
after acute exercise compared to control days. No signifi-
cant effects were found for acute exercise on rapid eye
movement sleep latency, stage 2–4 sleep, or number of
awakenings after sleep onset (ps[.18) and there were
insufficient data (n\5) to analyze the effects of acute
exercise on sleep maintenance, sleep quality, and insomnia
severity.
Included
Eligibility
Identification
Screening
Records identified through
PubMed/PsychInfo searching
(n = 3,291)
Records excluded
by title/abstract
(n = 3,007)
Full-text articles assessed
for eligibility
(n = 284)
Full-text articles excluded (n = 221)
Reasons for exclusion:
Insufficient sleep outcome (n = 63)
Observational studies (n = 44)
No outcomes of interest (n = 29)
Relevant analyses not conducted (n = 24)
Non-natural exercise regimen (n = 10)
Uncontrolled trial (n = 10)
Non-experimental article (n = 9)
Contained active comparison group (n = 8)
Medical/psychiatric patients (n = 7)
Insufficient exercise intervention/
measurement (n = 6)
Combined treatment (n = 3)
Experimentally-induced insomnia (n = 3)
Duplicate report of study (n = 2)
Mean participant age < 18 (n = 2)
Less than 4 participants (n=1)
Studies included in
meta-analysis
(n = 66)a
Acute exercise studies
(n = 41)
Regular exercise studies
(n = 25)
Articles included in
meta-analysis
(n = 63)
Fig. 1 Consort diagram. This figure displays a consort diagram of the
article search and selection process.
a
One article (Montgomery et al.,
1988) contained three studies that were included in the meta-analysis.
One article (Jennings, 1981) contained one study that involved an
acute and regular intervention; as data from both were utilized this
study is counted in the acute exercise studies and the regular exercise
studies
J Behav Med
123
Table 1 Main outcomes—acute exercise studies
References n Study design Subgroup TST SOL WASO NAASO SE SM
Baekeland and Lasky (1966) 10 Open trial Afternoon exercise
Baekeland and Lasky (1966) (10) Open trial Evening exercise
Baekeland (1970) 14 Open trial 0.35
Zir et al. (1971) 5 Open trial Light exercise
Adamson et al. (1974) 12 Open trial 0.29 -0.29
Desjardins et al. (1974) 6 Open trial High intensity
Desjardins et al. (1974) (6) Open trial Low intensity
Griffin and Trinder (1978) 16 Open trial
Walker et al. (1978) 10 Open trial Nonrunners -0.10 0.08 0.14 0.32
Walker et al. (1978) 10 Open trial Runners 0.20 -0.50 0.32 -0.19
Browman (1980) 7 Open trial 0.14
Jennings (1981) 9 RCT 0.65 -0.32
Paxton et al. (1982) 24 Open trial -0.18 0.28 0.36 -0.41
Trinder et al. (1982) 12 Open trial
Bunnell et al. (1983a) 9 Open trial 0.49
Bunnell et al. (1983b) 4 Open trial Exercisers
Bunnell et al. (1983b) 4 Open trial Non-Exercisers
Horne and Staff (1983) 8 Open trial High intensity 0.44 0.42
Horne and Staff (1983) (8) Open trial Low intensity 0.96 0.89
Paxton et al. (1983) 8 Open trial Athletes
Matsumoto et al. (1984) 6 Open trial 1.47
Horne and Moore (1985) 6 Open trial 0.73 0.08
Kupfer et al. (1985) 10 Open trial Double exercise 0.29 0.45 0.57 0.39
Kupfer et al. (1985) (10) Open trial Regular exercise 0.12 0.20 0.26 0.22
Montgomery et al. (1985) 8 Open trial 0.37 -0.39
Shapiro et al. (1985) 4 Open trial Neutral environment 0.09
Driver et al. (1988) 9 Open trial 0.65 0.68
Montgomery et al. (1988) 7 Open trial Experiment 1
Montgomery et al. (1988) 22 Open trial Experiment 2
Montgomery et al. (1988) 10 Open trial Experiment 3 -0.22
Vein et al. (1991) 30 Open trial 0.69
Bevier et al. (1992) 14 Open trial 0.51
Driver et al. (1994) 8 Open trial 15 km run -0.44 -0.13
Montmayeur et al. (1994) 4 Open trial -0.22 -0.03 -0.60
O’Connor et al. (1998) 8 Open trial Low intensity -0.30 0.03 -0.11 -0.52
O’Connor et al. (1998) (8) Open trial Moderate intensity -0.14 0.12 0.35 0.00
Oda (2001) 8 Open trial 0.15 -0.01 0.28 0.20
Hague et al. (2003) 15 Open trial 2.27
Alessi et al. (2005) 118 RCT 0.28 0.54 -0.29 0.30
Esteves et al. (2009) 22 Open trial Experiment 1 0.36 -0.02 0.48 0.35
Passos et al. (2010) 24 RCT High aerobic -0.45 -0.12 -0.12 0.15 -0.14
Passos et al. (2010) 12(12) RCT Moderate aerobic 0.67 0.70 0.13 0.16 0.47
Passos et al. (2010) 12(12) RCT Moderate resistance -0.22 -0.11 -0.37 0.49 -0.27
Bulckaert et al. (2011) 9 Open trial 0.08
Morita et al. (2011) 42 Open trial 0.52 0.40 0.07’’ 0.18
Myllymaki et al. (2011) 11 Open trial 0.09 0.64 0.14
Flausino et al. (2012) 17 Open trial 30 min, VT1 0.52 -0.06 1.29 1.16
Flausino et al. (2012) (17) Open trial 60 min, VT1 0.24 0.11 0.81 0.67
J Behav Med
123
Table 1 continued
References n Study design Subgroup TST SOL WASO NAASO SE SM
Flausino et al. (2012) (17) Open trial 30 min, VT1.5 0.44 -0.40 1.11 0.88
Flausino et al. (2012) (17) Open trial 60 min, VT1.5 0.36 -0.19 0.86 0.68
Myllymaki et al. (2012) 14 Open trial 30 min, easy 0.02 0.06
Myllymaki et al. (2012) (14) Open trial 30 min, moderate -0.20 0.04
Myllymaki et al. (2012) (14) Open trial 30 min, vigorous 0.05 0.15
Myllymaki et al. (2012) (14) Open trial 60 min, moderate -0.05 0.31
Myllymaki et al. (2012) (14) Open trial 90 min, moderate 0.05 -0.07
Viana et al. (2012) 40 RCT -0.11 -0.31 0.18 -0.16
Wong et al. (2013) 12 Open trial 45 % VO2max 0.64 0.05 0.23 0.27
Wong et al. (2013) (12) Open trial 55 % VO2max 0.31 -0.20 0.27 0.13
Wong et al. (2013) (12) Open trial 65 % VO2max -0.13 0.01 -0.26 -0.22
Wong et al. (2013) (12) Open trial 75 % VO2max 0.37 0.55 0.29 0.42
Overall effect 670 0.22*** 0.17* 0.38*** -0.07 0.25***
References n Study design Subgroup S1 S2 S3 S4 SWS REM REM-L
Baekeland and Lasky (1966) 10 Open trial Afternoon exercise 0.51 0.90 -0.18
Baekeland and Lasky (1966) (10) Open trial Evening exercise 0.18 0.27 -0.04
Baekeland (1970) 14 Open trial -0.01 0.20 0.02 0.15 0.15 -0.08 -0.62
Zir et al. (1971) 5 Open trial Light exercise 0.27
Adamson et al. (1974) 12 Open trial -0.25 -1.02 -0.44
Desjardins et al. (1974) 6 Open trial High intensity -1.29
Desjardins et al. (1974) (6) Open trial Low intensity -1.29 -1.29
Griffin and Trinder (1978) 16 Open trial 0.13 0.05 0.16
Walker et al. (1978) 10 Open trial Nonrunners -0.14 -0.49 0.01 -0.12 -0.17 0.17
Walker et al. (1978) 10 Open trial Runners -0.21 0.01 -0.06 -0.05 -0.20 -0.51
Browman (1980) 7 Open trial 0.64 0.10 0.27 -0.05 0.27 0.39
Jennings (1981) 9 RCT 0.52 -0.80
Paxton et al. (1982) 24 Open trial -0.18 0.39 0.10 0.16 0.12 -0.20
Trinder et al. (1982) 12 Open trial -0.35
Bunnell et al. (1983a) 9 Open trial 0.52 0.12 0.26 0.54 0.68 -0.35
Bunnell et al. (1983b) 4 Open trial Exercisers 0.82 -1.52
Bunnell et al. (1983b) 4 Open trial Non-Exercisers -0.87 2.78
Horne and Staff (1983) 8 Open trial High intensity 0.11 0.77 0.26 0.67 -1.00 -0.43
Horne and Staff (1983) (8) Open trial Low intensity -0.63 -0.04 -0.02 -0.03 -0.65 -0.06
Paxton et al. (1983) 8 Open trial Athletes 1.08 0.33
Matsumoto et al. (1984) 6 Open trial 0.47 0.61 0.62 1.93 2.24 0.21 -0.63
Horne and Moore (1985) 6 Open trial -0.23 0.15 0.53 1.37 1.29 -1.09 -0.44
Kupfer et al. (1985) 10 Open trial Double exercise 0.07 0.18 -0.16 -0.01 0.53
Kupfer et al. (1985) (10) Open trial Regular exercise 0.24 -0.35 -0.03 0.54 1.13
Montgomery et al. (1985) 8 Open trial 0.25 -0.30 -0.60
Shapiro et al. (1985) 4 Open trial Neutral environment 0.07 0.05 0.00 0.98 0.67 -0.27
Driver et al. (1988) 9 Open trial -0.69 -0.46
Montgomery et al. (1988) 7 Open trial Experiment 1 -1.36
Montgomery et al. (1988) 22 Open trial Experiment 2 -0.61 -0.61
Montgomery et al. (1988) 10 Open trial Experiment 3 0.72
Vein et al. (1991) 30 Open trial 0.52 0.69 0.52
Bevier et al. (1992) 14 Open trial 0.39 -0.24 0.08 0.31
Driver et al. (1994) 8 Open trial 15 km run 0.11 0.28 -0.21 0.16 0.45
Montmayeur et al. (1994) 4 Open trial
J Behav Med
123
Regular exercise
Individual study and aggregate effect sizes for regular
exercise are presented in Table 2. Overall, regular exercise
had small beneficial effects on total sleep time [d=0.25,
95 % CI (0.07, 0.43), p=.005; n=10] and sleep effi-
ciency [d=0.30, 95 % CI (0.06, 0.55), p=.02; n=6]
indicating that individuals who participated in regular
exercise training on average had significantly greater total
sleep time and sleep efficiency than individuals in control
conditions. Regular exercise also had a small-to-medium
beneficial effect on sleep onset latency [d=0.35, 95 % CI
(0.00, 0.70), p\.05; n=9] indicating that individuals
who participated in regular exercise training had signifi-
cantly better sleep onset latency (i.e., shorter) than indi-
viduals in control conditions. Lastly, regular exercise
approached a large beneficial effect on overall sleep quality
[d=0.74, 95 % CI (0.48, 1.00), p\.001; n=19] indi-
cating that individuals who participated in regular exercise
training had significantly better sleep quality than indi-
viduals in control conditions.
Sleep quality was most commonly assessed using the
Pittsburgh Sleep Quality Index (PSQI) which provides a
total score, as well as seven subscale scores (daytime
sleepiness, sleep disturbance, sleep duration, sleep effi-
ciency, sleep latency, sleep medication use, and subjective
Table 1 continued
References n Study design Subgroup S1 S2 S3 S4 SWS REM REM-L
O’Connor et al. (1998) 8 Open trial Low intensity
O’Connor et al. (1998) (8) Open trial Moderate intensity
Oda (2001) 8 Open trial 0.30 -0.13 -0.25 0.32
Hague et al. (2003) 15 Open trial 2.17 -2.92 -3.23
Alessi et al. (2005) 118 RCT
Esteves et al. (2009) 22 Open trial Experiment 1 0.36 0.22 0.03 0.13 0.44 0.13
Passos et al. (2010) 24 RCT High aerobic -0.10 -0.25 0.09 -0.54 0.28
Passos et al. (2010) 12(12) RCT Moderate aerobic 0.08 -0.08 0.29 -0.17 0.32
Passos et al. (2010) 12(12) RCT Moderate resistance -0.11 0.02 0.51 -0.54 0.26
Bulckaert et al. (2011) 9 Open trial -0.61 0.47 0.16
Morita et al. (2011) 42 Open trial
Myllymaki et al. (2011) 11 Open trial 0.23 -0.13 0.22 -0.42
Flausino et al. (2012) 17 Open trial 30 min, VT1 0.91 0.06 -0.11 0.45 0.60
Flausino et al. (2012) (17) Open trial 60 min, VT1 0.53 -0.36 -0.22 0.00 0.04
Flausino et al. (2012) (17) Open trial 30 min, VT1.5 0.75 0.06 -0.05 0.43 0.22
Flausino et al. (2012) (17) Open trial 60 min, VT1.5 0.79 -0.35 -0.09 -0.14 -0.01
Myllymaki et al. (2012) 14 Open trial 30 min, easy
Myllymaki et al. (2012) (14) Open trial 30 min, moderate
Myllymaki et al. (2012) (14) Open trial 30 min, vigorous
Myllymaki et al. (2012) (14) Open trial 60 min, moderate
Myllymaki et al. (2012) (14) Open trial 90 min, moderate
Viana et al. (2012) 40 RCT 0.52 0.05 0.14 0.23 0.71
Wong et al. (2013) 12 Open trial 45 % VO2max -0.76 -0.25 -0.29
Wong et al. (2013) (12) Open trial 55 % VO2max -0.08 -0.21 -0.36
Wong et al. (2013) (12) Open trial 65 % VO2max -0.58 -0.75 -0.61
Wong et al. (2013) (12) Open trial 75 % VO2max -0.60 -0.76 -0.27
Overall effect 670 0.35*** -0.01 0.14 0.14 0.19* -0.27** -0.11
Studies conducted with participants with sleep disturbance in bold
Participants utilized multiple times are noted in parentheses
Overall effect only presented for outcomes with n[5 comparisons
Sleep Outcomes: TST total sleep time, SOL sleep onset latency, WASO wake time after sleep onset, NAASO number of awakenings after sleep
onset, SM sleep maintenance, SE sleep efficiency, S1 stage 1, S2 stage 2, S3 stage 3, S4 stage 4, SWS slow wave sleep, REM rapid eye movement
sleep, REM-LREM sleep latency
*p\.05; ** p\.01; *** p\.001
J Behav Med
123
Table 2 Main outcomes—regular exercise studies
References n Study design Subgroup TST SOL WASO NAASO SE SM S1 S2
Jennings (1981) 9 RCT 1.27
Guilleminault et al. (1995) 20 RCT 0.50 0.56 -0.60
King et al. (1997) 43 RCT 0.68 0.71 0.65
Alessi et al. (1999) 29 RCT 0.72
Naylor et al. (2000) 23 Matched CT 0.32 0.50 1.26 1.15 0.01 -0.12
King et al. (2002) 85 RCT
Tworoger et al. (2003) 173 RCT 0.12
de Jong et al. (2006) 181 RCT
Elavsky and McAuley (2007) 102 RCT Walking
Elavsky and McAuley (2007) 61(39) RCT Yoga
Frye et al. (2007) 49 RCT Low impact exerise
Frye et al. (2007) 23(21) RCT Tai Chi
Littman et al. (2007) 167 RCT 0.20
King et al. (2008) 66 RCT -0.23 0.12 0.44 0.03 0.59 -0.40
Chen et al. (2009) 128 RCT
Manzaneque et al. (2009) 39 RCT
Chen et al. (2010) 55 RCT
Reid et al. (2010) 17 RCT 1.09 0.86
Hosseini et al. (2011) 56 RCT
Richards et al. (2011) 102 RCT 0.32 -0.38 0.45
Chen et al. (2012) 55 RCT
Kalak et al. (2012) 51 RCT 0.42 0.63 -0.48 0.45 -0.01 0.36 0.06
Kline et al. (2012) 243 RCT 4 KKW
Kline et al. (2012) 99(92) RCT 8 KKW
Kline et al. (2012) 95(92) RCT 12 KKW
Nguyen and Kruse (2012) 73 RCT
Yeh and Chang (2012) 70 Matched CT
Oudegeest-Sander et al. (2013) 21 RCT 0.07 0.00 -0.25 0.00
Pinniger et al. (2013) 35 RCT Exercise
Pinniger et al. (2013) 23(18) RCT Tango
Overall effect 2,193 0.25** 0.35* 0.30*
References n Study design Subgroup S3 S4 SWS REM REM-L IS SQ
Jennings (1981) 9 RCT 0.81 -1.57
Guilleminault et al. (1995) 20 RCT
King et al. (1997) 43 RCT 1.10
Alessi et al. (1999) 29 RCT
Naylor et al. (2000) 23 Matched CT 0.76 0.02
King et al. (2002) 85 RCT
Tworoger et al. (2003) 173 RCT -0.05
de Jong et al. (2006) 181 RCT 0.26
Elavsky and McAuley (2007) 102 RCT Walking 0.37
Elavsky and McAuley (2007) 61(39) RCT Yoga 0.12
Frye et al. (2007) 49 RCT Low impact exerise 0.39
Frye et al. (2007) 23(21) RCT Tai Chi 0.60
Littman et al. (2007) 167 RCT
King et al. (2008) 66 RCT -0.15 0.39
Chen et al. (2009) 128 RCT 0.72
J Behav Med
123
sleep quality). Effect sizes for the PSQI subscales were also
calculated. Regular exercise had a significant beneficial
effect on all subscales (see Online Resource 2), ranging
from moderate-to-large in magnitude (0.56 \d\1.02, all
ps\.01; 7 BnB9), except sleep medication use
[d=0.19, 95 % CI (-0.02, 0.39), p=.07; n=6]. There
were insufficient data (n\5) to examine the average
effects of regular exercise on any other sleep outcomes.
Moderator analyses
Participant characteristics Moderator analyses revealed
more beneficial effects of acute exercise for men than
women on certain sleep outcomes. Specifically, participant
sex significantly moderated the effect of acute exercise on
stage 1 sleep (B=-0.69, SE =0.27, p=.01). For every
10 % increase in the percentage of the sample that was
female, the effect size for the benefit of acute exercise on
decreasing stage 1 sleep decreased by .07 standard devia-
tion units. Sex also significantly moderated the effect of
acute exercise on wake time after sleep onset (B=-0.46,
SE =0.22, p=.04). For every 10 % increase in the per-
centage of the sample that was female, the effect size for
the benefit of acute exercise on wake time after sleep onset
was reduced by .05 standard deviation units. Sex did not,
however, significantly moderate the relationship between
acute exercise and total sleep time, slow wave sleep, sleep
onset latency, sleep efficiency, stage 2–4 sleep, number of
awakenings after sleep onset, rapid eye movement sleep,
and rapid eye movement sleep latency, nor did it moderate
the relationship between regular exercise and any sleep
outcomes (ps[.10).
Examining age as a moderator revealed significantly less
beneficial effects of regular exercise on the sleep onset
latency of older individuals (B=-0.02, SE =0.01,
p=.02). Specifically, a 10-year increase in mean age of
the sample was associated with a 0.15 standard deviation
unit decrease in the beneficial effect of regular exercise for
reducing sleep onset latency. Age did not, however, sig-
nificantly moderate the relationship between regular
exercise and total sleep time, sleep efficiency, or sleep
quality, nor did it moderate the relationship between acute
exercise and any sleep outcomes (ps[.05).
Examining baseline physical activity level revealed
more beneficial effects of acute exercise on slow wave
sleep for individuals with high rather than low baseline
activity (Q=9.73, df =1, p=.002). Specifically, mod-
erate and significant effects were observed for samples of
individuals with high baseline physical activity (d=0.51,
p=.008) and non-significant effects were observed for
Table 2 continued
References n Study design Subgroup S3 S4 SWS REM REM-L IS SQ
Manzaneque et al. (2009) 39 RCT 0.21
Chen et al. (2010) 55 RCT 1.15
Reid et al. (2010) 17 RCT 2.23
Hosseini et al. (2011) 56 RCT 0.47
Richards et al. (2011) 102 RCT 0.67
Chen et al. (2012) 55 RCT 1.73
Kalak et al. (2012) 51 RCT -0.29 0.65 -0.62 -0.67 1.10
Kline et al. (2012) 243 RCT 4 KKW 0.55
Kline et al. (2012) 99(92) RCT 8 KKW 0.56
Kline et al. (2012) 95(92) RCT 12 KKW 0.59
Nguyen and Kruse (2012) 73 RCT 1.45
Yeh and Chang (2012) 70 Matched CT 2.54
Oudegeest-Sander et al. (2013) 21 RCT
Pinniger et al. (2013) 35 RCT Exercise 0.00
Pinniger et al. (2013) 23(18) RCT Tango 0.41
Overall effect 2,193 0.74***
Studies conducted with participants with sleep disturbance in bold
Participants utilized multiple times are noted in parentheses
Overall effect only presented for outcomes with n[5 comparisons
Sleep Outcomes: TST total sleep time, SOL sleep onset latency, WASO wake time after sleep onset, NAASO number of awakenings after sleep
onset, SM sleep maintenance, SE sleep efficiency, S1 stage 1, S2 stage 2, S3 stage 3, S4 stage 4, SWS slow wave sleep, REM rapid eye movement
sleep, REM-LREM sleep latency
*p\.05; ** p\.01; *** p\.001
J Behav Med
123
samples of individuals with low baseline physical activity
(d=-0.31, p=.09). We did not observe this pattern,
however, for the beneficial effect of acute exercise on other
sleep outcomes such as sleep onset latency, sleep effi-
ciency, and total sleep time (ps[.26). Nor did level of
baseline physical activity moderate the effect of acute
exercise on rapid eye movement sleep or rapid eye
movement sleep latency (ps[.23). No studies of regular
exercise examined participants of high baseline physical
activity; therefore, we were unable to examine baseline
physical activity level as a moderator of the effects of
regular exercise on sleep.
As only 11 % of studies examined individuals who were
identified as having sleep complaints, this was not formally
examined as a moderator.
Exercise characteristics Many publications did not pro-
vide specific details about the exercise in which partici-
pants engaged; hence, moderator analyses are limited by
these omissions. We examined exercise intensity as a
moderator of the effect of acute exercise on all outcomes
and found no significant differences in effect sizes across
light, moderate, and vigorous intensity (ps[.08). As only
four studies examined light exercise, we also compared
acute studies of moderate and vigorous intensity exercise
and found no significant differences in effect sizes for all
outcomes (ps[.10). Due to power constraints, we were
not able to examine intensity as a moderator of the effects
of regular exercise on sleep.
Although few studies (n=4acute studies) strictly
examined anaerobic exercise and many utilized mixed
aerobic/anaerobic interventions, we were able to compare
the effects of aerobic and anaerobic exercise for some
outcome variables. No significant differences in effect sizes
were found between acute aerobic and anaerobic exercise
for the outcome variables of total sleep time, slow wave
sleep, stage 2 sleep, stage 1 sleep, and rapid eye movement
sleep (ps[.08). Due to power constraints, we were not
able to examine aerobic/anaerobic classification as a
moderator of the effects of regular exercise on sleep.
There was a large variety of types of exercise that were
examined across studies of acute and regular exercise,
however, there were few types of exercise for which there
were a critical mass of studies (n[3) examining one type.
There were a significant number of acute studies that
examined running (n=19) and cycling (n=6), thus we
were able to compare mean effect sizes for these two types
of exercise. We found that exercise type significantly
moderated the beneficial effect of acute exercise on slow
wave sleep (Q=4.19, df =1, p=.04), in that cycling
resulted in a significant moderate beneficial effect
(d=0.49, p=.04), while running resulted in a non-sig-
nificant effect (d=-0.04, p=.71). We did not, however,
find any significant differences in effect sizes for cycling
versus running for other sleep outcomes (ps[.15). For
studies of regular exercise, we were able to compare
studies that examined mind/body exercise (n=9) to
studies examining more traditional exercise (n=10).
Although the effect size for regular mind/body exercise on
sleep quality was large and significant, it did not signifi-
cantly differ from that of non-mind/body exercise on sleep
quality (d=0.98 vs. 0.48, ps\.001; Q=3.37, df =1,
p=.07).
In addition to examining aspects of the type of exercise,
we examined characteristics of the exercise schedule. Time
of day of exercise in relation to bed time significantly
moderated the beneficial effects of acute exercise on wake
time after sleep onset (Q=8.47, df =1, p=.004) and
stage 1 sleep (Q=9.31, df =2, p=.01). Exercising less
than 3 h before bedtime was significantly associated with
less disturbed sleep (lower wake time after sleep onset),
whereas exercising 3–8 h before bedtime was not. Exer-
cising less than 3 h before bedtime and greater than 8 h
before bedtime were significantly associated with less time
spent in light sleep (stage 1 sleep), however, exercising
3–8 h before bedtime was not. Time of day also signifi-
cantly moderated the effect of acute exercise on rapid eye
movement sleep (Q=6.72, df =2, p=.03). Specifically,
exercising 3–8 before bedtime was significantly associated
with a decrease in rapid eye movement sleep; however, less
than 3 h before bedtime and greater than 8 h before bed-
time were not. Time of day did not moderate the effects of
acute exercise on total sleep time, sleep efficiency, slow
wave sleep, sleep onset latency, stage 2 sleep, or rapid eye
movement sleep latency (ps[.07). There were insufficient
studies of regular exercise that noted or controlled for time
of day, precluding our ability to examine exercise time as a
moderator of regular exercise on sleep.
For acute exercise, duration of individual exercise bouts
(mins) significantly moderated the beneficial effect of
acute exercise on total sleep time (B=.005, SE =0.00,
p=.006), slow wave sleep (B=.011, SE =0.00,
p\.001), sleep onset latency (B=.009, SE =0.00,
p=.002), stage 4 sleep (B=.007, SE =0.00, p=.04),
and the effects of acute exercise on rapid eye movement
sleep latency (B=-.016, SE =0.00, p\.001), and rapid
eye movement sleep (B=-.008, SE =0.00, p=.009).
Increased duration was associated with an increase in
magnitude of beneficial effect sizes for total sleep time,
slow wave sleep, sleep onset latency, and stage 4 sleep.
Consistent with our main effect finding regarding rapid eye
movement sleep and previous findings regarding rapid eye
movement sleep latency (Youngstedt et al., 1997),
increased duration of acute exercise was also associated
with effect sizes representing larger decreases in rapid eye
J Behav Med
123
movement sleep and larger increases in rapid eye move-
ment sleep latency. Duration of individual exercise bouts
did not moderate the effect of acute exercise on wake time
after sleep onset, sleep efficiency, stage 3 sleep, stage 2
sleep, stage 1 sleep, or number of awakenings after sleep
onset (ps[.15). For regular exercise, duration of indi-
vidual exercise bouts (mins) significantly moderated the
beneficial effect on sleep onset latency (B=.037,
SE =0.02, p=.02), in that beneficial effects were larger
for longer durations. We did not find that the duration of
individual exercise bouts significantly moderated the effect
of regular exercise on any other sleep outcomes (ps[.16).
Number of days per week of exercise did not significantly
moderate the effect of regular exercise on any sleep out-
comes (ps[.07). We did find that the duration of the
exercise program in total number of weeks (range
2–52 weeks) and total duration significantly moderated the
beneficial effect of regular exercise on total sleep time
(B=-0.01, SE =0.00, p=.04 and B=-0.00,
SE =0.00, p\.05, respectively), in that effect sizes were
slightly smaller for longer programs of exercise. Total
number of weeks and total duration did not significantly
moderate the beneficial effects of regular exercise on sleep
quality, sleep onset latency, or sleep efficiency (ps[.15).
Adherence metrics were obtained for over half of the
regular exercise study comparisons (63 %). The average
adherence to exercise programs was 84 % (range
63–107 %). For the studies that reported adherence,
adherence significantly moderated the effects of regular
exercise on sleep quality (B=0.02, SE =0.01, p=0.02).
As would be expected, more beneficial effects were seen in
studies with higher adherence rates. Adherence was not
confounded by (i.e. not correlated with) participant age or
exercise program duration (ps\.59).
Sleep characteristics The majority of studies of acute
exercise were conducted in sleep labs and thus utilized
objective methods of assessing sleep. Only two studies of
acute exercise were known to have used subjective methods,
thus, we did not compare objective and subjective assess-
ment type for the effect of acute exercise on sleep. We were
able to examine whether type of objective assessment (EEG,
PSG, or actigraphy) moderated the effects of acute exercise
on sleep. The only outcome for which there was a significant
moderating effect of objective assessment type was stage 4
sleep; studies utilizing PSG showed a moderate positive
effect of acute exercise on stage 4 sleep (d=0.58) whereas
studies utilizing EEG did not show a significant effect
(Q=3.96, df =1, p\.05). Results did not vary across
assessment type for all other outcomes (ps[.05).
More variation existed in assessment type in studies of
regular exercise; six studies utilized objective assessments
and four utilized subjective assessments of total sleep time.
Assessment type did not, however, significantly moderate
the beneficial effect of regular exercise on total sleep time
(p=.49). As we noted apparent differences in the effect
sizes for the duration and efficiency subscales of the PSQI
(d=0.63, 0.85 respectively) which is subjective data, and
our main effect sizes for total sleep time and sleep effi-
ciency (d=0.25, 0.30 respectively) which were predom-
inantly based off objective data, we used ttests to
statistically compare these effects. No significant differ-
ences, however, were not found between the PSQI subscale
effect sizes and our main total sleep time and sleep effi-
ciency outcomes (p=.07, p=.11, respectively).
Study characteristics The average effect sizes for the four
RCTs of acute exercise were compared to those of the open
trials of acute exercise. Study design did not moderate the
beneficial effects of acute exercise on wake time after sleep
onset, total sleep time, slow wave sleep, sleep onset
latency, sleep efficiency, and stage 1–2 sleep (ps[.11),
nor did it moderate the effects of acute exercise on number
of awakenings after sleep onset or rapid eye movement
sleep (p=.69). Study design did appear to moderate the
effect of acute exercise on rapid eye movement sleep
latency (Q=5.13, df =1, p=.02), with more beneficial
effects found in acute RCTs compared to open trials. As all
but two studies of regular exercise were RCTs, study
design was not examined as a moderator of regular exer-
cise.
Year of publication did not moderate the effects of acute
exercise or regular exercise on any sleep outcomes
(ps[.06).
Publication bias
Evaluation of publication bias suggested that the results
varied in robustness. The beneficial effects of regular
exercise on sleep quality (fail-safe N=710) and the PSQI
subscales (64 \fail-safe Ns\196) are robust according to
Rosenthal’s (1991) standards. Fail-safe Nvalues for the
effects of acute exercise on wake time after sleep onset
(fail-safe N=90), total sleep time (fail-safe N=75), slow
wave sleep (fail-safe N=56), sleep onset latency (fail-safe
N=24), sleep efficiency (fail-safe N=52), S1 (fail-safe
N=56), rapid eye movement sleep (fail-safe N=124)
and regular exercise on total sleep time (fail-safe N=17),
sleep onset latency (fail-safe N=11), and sleep efficiency
(fail-safe N=5) do not meet Rosenthal’s criteria for
robustness. Individual effect sizes for acute exercise on
wake time after sleep onset and regular exercise on sleep
quality are presented in forest plots in Figs. 2and 3,
respectively. The majority of the funnel plots for the sig-
J Behav Med
123
nificant outcomes presented above appeared to be roughly
symmetrical or favoring negative small-study effects by
visual inspection (see Online Resource 3). The funnel plots
for the effects of acute exercise on slow wave sleep and
total sleep time appeared to be asymmetrical, displaying
more positive small-study effects than negative small-study
effects. Conservative Trim and Fill analyses (Duval &
Tweedie, 2000) resulted in an adjusted effect size for acute
exercise on slow wave sleep of d=0.06 and no impact on
the effect size of acute exercise on total sleep time
Study name Subgroup within study Outcome Statistics for each study Std diff in means and 95% CI
Std diff Lower Upper
in means limit limit p-Value
Flausino et al 2012 30 mins, VT1 WASO 1.29 0.55 2.03 0.00
Flausino et al 2012 30 mins, VT1.5 WASO 1.11 0.39 1.84 0.00
Flausino et al 2012 60 mins, VT1 WASO 0.81 0.11 1.51 0.02
Flausino et al 2012 60 mins, VT1.5 WASO 0.86 0.15 1.56 0.02
Wong et al 2013 45% VO2max WASO 0.23 -0.58 1.03 0.58
Wong et al 2013 55% VO2max WASO 0.27 -0.54 1.07 0.52
Wong et al 2013 65% VO2max WASO -0.26 -1.07 0.54 0.52
Wong et al 2013 75% VO2max WASO 0.29 -0.51 1.10 0.48
A
lessi et al 2005 None WASO 0.44 0.08 0.81 0.02
Bevier et al 1992 None WASO 0.51 -0.24 1.26 0.18
Jennings et al 1981 None WASO -0.43 -1.83 0.97 0.55
Oda et al 2001 None WASO 0.28 -0.70 1.26 0.58
Paxton et al 1982 None WASO 0.36 -0.21 0.93 0.22
Viana et al 2012 None WASO 0.18 -0.44 0.81 0.56
Esteves et al 2009 Experiment 1 WASO 0.48 -0.12 1.08 0.11
Passos et al 2010 High aerobic WASO -0.12 -0.93 0.68 0.76
Passos et al 2010 Moderate aerobic WASO 0.13 -0.67 0.93 0.75
Passos et al 2010 Moderate resistance WASO -0.31 -1.11 0.50 0.46
O'Connor et al 1998 Low intensity WASO -0.11 -1.09 0.87 0.82
O'Connor et al 1998 Moderate intensity WASO 0.35 -0.64 1.34 0.49
Walker et al 1978 Nonrunners WASO 0.14 -0.74 1.02 0.76
Walker et al 1978 Runners WASO 0.32 -0.57 1.20 0.48
0.38 0.21 0.54 0.00
-2.00 -1.00 0.00 1.00 2.00
Fig. 2 Acute exercise and wake time after sleep onset (WASO) forest plot. This figure presents a forest plot of individual study effects of acute
exercise on wake time after sleep onset. Std diff in means =Cohen’s d
Study name Subgroup within study Outcome Statistics for each study Std diff in means and 95% CI
Std diff Lower Upper
in means limit limit p-Value
Chen et al 2009 None SQ 0.72 0.36 1.08 0.00
Chen et al 2010 None SQ 1.15 0.57 1.72 0.00
Chen et al 2012 None SQ 1.73 1.11 2.35 0.00
de Jong et al 2006 None SQ 0.26 -0.03 0.56 0.08
Elavsky et al 2007 Walking SQ 0.37 -0.04 0.77 0.07
Elavsky et al 2007 Yoga SQ 0.12 -0.29 0.52 0.57
Frye et al 2007 Low Impact Exercise SQ 0.39 -0.18 0.96 0.18
Frye et al 2007 Tai Chi SQ 0.60 -0.01 1.20 0.05
Hosseini et al 2011 None SQ 0.47 -0.06 1.00 0.08
King et al 1997 None SQ 1.10 0.45 1.74 0.00
King et al 2008 None SQ 0.39 -0.10 0.88 0.12
Kline et al 2012 12 KKW SQ 0.59 0.18 1.01 0.00
Kline et al 2012 4 KKW SQ 0.55 0.18 0.92 0.00
Kline et al 2012 8 KKW SQ 0.56 0.15 0.98 0.01
Manzaneque et al 2009 None SQ 0.21 -0.43 0.84 0.52
Nguyen et al 2012 None SQ 1.45 0.93 1.97 0.00
Reid et al 2010 None SQ 2.23 1.01 3.45 0.00
Tworoger et al 2003 None SQ -0.05 -0.30 0.19 0.67
Yeh et al 2012 None SQ 2.54 1.91 3.17 0.00
0.74 0.48 1.00 0.00
-2.00 -1.00 0.00 1.00 2.00
Fig. 3 Regular exercise and sleep quality (SQ) forest plot. This figure presents a forest plot of individual study effects of regular exercise on
sleep quality. Std diff in means =Cohen’s d
J Behav Med
123
(d=0.22). The funnel plots for the effects of regular
exercise on total sleep time, sleep onset latency, sleep
efficiency, and sleep quality also appeared to be asym-
metrical, displaying more positive small-study effects.
Conservative Trim and Fill analyses resulted in an adjusted
effect size for regular exercise on total sleep time of
d=0.17, regular exercise on sleep onset latency of
d=0.17, and regular exercise on sleep efficiency of
d=0.26. The effect size for regular exercise on sleep
quality remained the same even after Trim and Fill anal-
yses (d=0.74). In considering these analyses, it is
important to note that funnel plot asymmetry could be due
to factors other than publication bias and because of this
the Trim and Fill method may lead to overcorrection
(Vevea & Woods, 2005).
Risk of bias
A summary of the quality assessment of the experimental
studies of acute exercise using Youngstedt and colleagues’
(Youngstedt et al., 1997) criteria is outlined in the Online
Resource 4. The inter-coder agreement was 91 % at the
first round of assessments, after which 100 % consensus
was reached through discussion between the assessors and
the authors. On average, studies met 6.5 out of the 10
criteria. The majority of studies provided a description of
subject fitness levels and/or activity history (95 %), mea-
sured sleep on the night before experimental conditions to
minimize first night effects (81 %), provided a detailed
description of the exercise stimulus (88 %), adhered to
standard procedures for recording and scoring sleep data
(98 %), and set bedtimes according to customary bedtimes
(93 %). About half of the studies provided a description of
participants’ sleep history (44 %), counterbalanced or
randomized experimental days/conditions (44 %), con-
trolled for napping, caffeine, and outside experiment
exercise (38 %), and utilized blind scoring of the data
(51 %). A small percentage utilized two or more experi-
mental nights in each condition (27 %). A significant
percentage of criteria could not be assessed and were
marked as unknown due to a lack of reporting (11 %).
A summary of the risk of bias assessment for RCTs
utilizing the Cochrane Collaboration criteria (Higgins
et al., 2011) is presented in Online Resource 4. Inter-coder
agreement was 83 % at the first round of assessments, after
which 100 % consensus was reached through discussion
between the assessors and the authors. The overall risk of
bias for many studies was unclear mainly owing to the fact
that the majority of studies failed to report how the ran-
domization sequence was generated (72 %) and allocation
concealed (79 %). Given that the nature of exercise studies
precluded the blinding of participants and most key study
personnel to the intervention allocation, this criterion was
not assessed. Furthermore, for studies utilizing self-report
measures, participants were not blind to what the outcome
assessment was measuring. Because we only included
studies utilizing validated self-report measures and most
participants were non-treatment seeking, we judged that the
outcome measurement was not likely to have been influ-
enced by the lack of blinding and rated the risk of bias from
this source as low. All studies appeared to be free of
selective outcome reporting, and the majority of trials
properly addressed incomplete outcome data (97 %) and
appeared to be free of other biases (83 %) as judged by the
coders using the Cochrane Collaboration guidelines.
Discussion
To our knowledge, this meta-analysis is the first in over a
decade to provide a comprehensive review of the effects of
acute and regular exercise on sleep. Our findings are
generally consistent with past meta-analyses in demon-
strating that exercise has a beneficial effect on sleep. Our
analyses, however, provide a more nuanced perspective on
this matter and allow us to consider moderating variables
that may impact the sleep benefits derived from exercise.
Results suggest that acute exercise has a beneficial
effect on many objective indices of sleep; however, these
effects are small and vary in robustness. Consistent with
past research (Kubitz et al., 1996; Youngstedt et al., 1997),
acute exercise had beneficial but small effects on total
sleep time, slow wave sleep, and sleep onset latency. We
also found that acute exercise resulted in a decrease in
rapid eye movement sleep, although the effect was small
rather than moderate as reported by a past meta-analysis
(Youngstedt et al., 1997), and was influenced by time of
day, with larger reductions for exercise completed 3–8 h
before bedtime than other time periods. In addition, we
found small beneficial effects for acute exercise on sleep
efficiency, wake time after sleep onset, and stage 1 sleep,
which have not been previously reported. Our results for
wake time after sleep onset provide strong evidence that
acute exercise reduces sleep disturbance. In contrast to past
findings, we did not find that acute exercise had specific
significant effects on rapid eye movement sleep latency,
stage 2, stage 3, or stage 4 sleep (Kubitz et al., 1996;
Youngstedt et al., 1997).
Our results extend Yang et al. (2012) findings regarding
the effect of regular exercise on sleep quality in older
individuals with sleep complaints to a wider age range of
individuals with and without sleep complaints. Specifi-
cally, we found moderate and robust positive effects of
regular exercise on sleep quality overall. We furthermore
found moderate-to-large robust positive effects for regular
J Behav Med
123
exercise on all subscales of the PSQI, except sleep medi-
cation use. In addition, our results suggest that regular
exercise has small beneficial effects on total sleep time and
sleep efficiency and a small-to-moderate beneficial effect
on sleep onset latency. However, these results are lacking
in robustness and additional studies are needed to confirm
these findings.
Our findings regarding sleep onset latency, wake time
after sleep onset, and sleep efficiency are hopeful as these
are metrics of sleep disturbance that have been associated
with increased mortality (Dew et al., 2003) and are often
targeted in the treatment of insomnia (Morin et al., 1999).
In addition, robust positive findings regarding the effects of
regular exercise on sleep quality are noteworthy consid-
ering that perceived sleep quality has been associated with
daytime functioning, life satisfaction, and mental health
symptoms in healthy individuals (Pilcher et al., 1997;
Zeitlhofer et al., 2000). Perceived sleep quality has also
been shown to have an impact on daytime functioning in
insomnia patients, regardless of objective quality of sleep
(Semler & Harvey, 2005). In addition, negative beliefs
about sleep quality may contribute to the maintenance of
insomnia and therefore are important treatment targets
(Harvey, 2002).
As the functions of different sleep stages are still
unclear, we are cautious in drawing conclusions from our
results regarding specific stages of sleep. Nonetheless,
stage 1 sleep, the first cycle entered after falling asleep, is
often referred to as ‘‘light sleep’’ and is associated with a
low arousal threshold. It is generally accepted that an
increase in the amount of stage 1 sleep is a sign of sleep
disruption, and thus a decrease in stage 1 sleep, as dis-
played in our results, is beneficial (Carskadon & Dement,
2000). Research also suggests that slow wave sleep and
rapid eye movement sleep may be especially important for
memory consolidation (Diekelmann & Born, 2010; Peig-
neux et al., 2004) and rapid eye movement sleep addi-
tionally important for emotional memory processing
(Walker & van der Helm, 2009). Thus, the decrease in
stage 1 sleep and increase in slow wave sleep as a result of
acute exercise, suggest a shift of the sleep cycle away from
light sleep towards more restorative sleep. These results
should be interpreted with caution, however, as our finding
for the effect of acute exercise on slow wave sleep and
stage 1 sleep were not robust and could possibly be influ-
enced by publication bias according to our conservative
analysis. In addition, whereas our findings regarding slow
wave sleep are promising, our findings regarding the
reduction in rapid eye movement sleep as a result of acute
exercise may suggest that the sleep benefits of exercise
come with a cost. The decrease in rapid eye movement
sleep, however, is not necessarily detrimental as it could
also reflect consolidation of rapid eye movement sleep or
be a consequence of slow wave sleep augmentation (Lucidi
et al., 1996). Nonetheless, this highlights the need for
research on the impact of regular exercise on slow wave
sleep and rapid eye movement sleep to determine if these
effects are long lasting or subside with continued engage-
ment in physical activity. At the same time, our findings
indicate that the impact of acute exercise on rapid eye
movement sleep may not be as large as previously reported
(Youngstedt et al., 1997) and similar reductions in rapid
eye movement sleep are seen with some common phar-
macological treatments for insomnia (Borbe
´ly et al., 1984).
Exercise has been proposed as a potential alternative
intervention for insomnia (Sarris & Byrne, 2011). Although
samples in our meta-analysis were predominantly selected
without attention to sleep complaints or insomnia, it is
nonetheless useful to compare the effect sizes we found to
those of treatments for insomnia to characterize the
potential benefit of exercise interventions. Specifically, the
effects of regular exercise on the PSQI subscales for
latency (d=0.75), duration (d=0.63), and quality
(d=1.02) compare well to results from a meta-analysis of
pharmacologic and behavioral therapy interventions for
insomnia (Smith et al., 2002): mean effect sizes derived
from subjective sleep diary data ranged from moderate to
very large for sleep onset latency (pharmacotherapy
d=0.45, behavior therapy d=1.05), total sleep time
(pharmacotherapy d=0.84, behavior therapy d=0.46),
and sleep quality (pharmacotherapy d=1.20, behavior
therapy d=1.44). Due to the small number of studies of
participants with sleep complaints in our meta-analysis, we
refrained from statistically examining the presence of sleep
complaints as a moderator of the effects of exercise on
sleep. We did observe, however, that the effect size of
regular exercise on sleep quality in samples with sleep
complaints was large and significant (d=1.09, n=3,
p=.02). Because exercise interventions may differentially
affect individuals with and without clinical insomnia, fur-
ther research with clinical populations is greatly needed.
Likewise, research in this area on unselected samples
should make a greater effort to assess and describe the
baseline sleep characteristics of participants to better
understand the degree and influence of sleep disturbances
in those samples.
With regards to moderators, our results indicate that sex,
age, baseline physical activity level as well as exercise
type, time of day, adherence, and duration moderate the
relationship between exercise and sleep. For the majority of
outcomes, there is no difference for the benefits of exercise
on sleep depending on age or sex. We did, however, find
significant differences for certain sleep variables. Specifi-
cally, the benefits of acute exercise did not vary by age and
for some variables (stage 1 sleep and wake time after sleep
onset), appeared to be stronger for men than women. In
J Behav Med
123
contrast, the benefits of regular exercise did not vary by
sex and for one variable (sleep onset latency), appeared to
be stronger for younger than older individuals. Interest-
ingly, this finding is in line with and could possibly be
reflective of general age-related changes in sleep; studies
suggest that in general, sleep onset latency slightly
increases with age (Ohayon et al., 2004). Lastly, contrary
to past research (Kubitz et al., 1996), we did not find that
exercise has a larger negative impact on the rapid eye
movement sleep of older individuals.
Our results indicate that individuals of high baseline
physical activity reap more slow wave sleep benefits from
acute exercise. No significant differences, however, were
found for the effect of acute exercise on other sleep vari-
ables when comparing individuals with high versus low
baseline physical activity level. This somewhat contradicts
the hypothesis of a ceiling effect for the benefits of exercise
for fit individuals (Kubitz et al., 1996). Discrepancies with
past results may be due to differences between engaging in
high baseline physical activity and being ‘‘fit.’’ The
advantage for individuals who are already exercising,
however, may be related to the stress involved in starting
an exercise when one is not a regular exerciser and a
potential ramp up time before benefits from exercise are
realized. This could also indicate that fitness, in addition to
exercise itself, may be especially beneficial for slow wave
sleep (Uchida et al., 2012).
We also found some preliminary evidence that exercise
duration, adherence, time of day, and type moderate sleep
outcomes. Our results regarding duration support that
longer exercise bout duration is associated with better
outcomes and are consistent with past meta-analyses (Ku-
bitz et al., 1996; Youngstedt et al., 1997). In addition, for
regular exercise, we noted a relationship between adhering
to the prescribed exercise program and sleep outcomes,
suggesting that increased engagement in exercise leads to
greater benefit. Time of day results, however, are incon-
sistent with past research and sleep-hygiene recommenda-
tions indicating that exercising within a few hours of
bedtime may be detrimental to sleep (Stepanski & Wyatt,
2003). In fact, our findings suggest that acute exercise
within a few hours of bedtime may actually be beneficial
and exercising 3–8 h before bedtime detrimental for some
sleep outcomes. We were only able to examine time of day
for acute studies, however, thus it is unclear whether
repeated exercise close to bedtime would negatively impact
sleep.
Of note, we did not find any significant differences in
the sleep benefits provided by different intensities of acute
or regular exercise. This is encouraging and in line with
research suggesting that vigorous intensity is not necessary
to obtain the mental health benefits of physical activity
(Ekkekakis et al., 2008). We also did not find any signifi-
cant differences in the sleep benefits provided by aerobic
versus anaerobic exercise. We were, however, only able to
examine this moderator for acute exercise studies; this
distinction could prove to be important with regards to the
sleep benefits derived from regular exercise. When com-
paring different types of acute exercise, cycling appeared
to be more beneficial than running. This may be due to the
fact that cycling is low impact and tends to result in fewer
injuries than running (Hootman et al., 2001; Requa et al.,
1993). Lastly, the large significant effect we observed for
mind/body exercise on sleep quality is promising given the
recent increase in focus on mind–body treatments for
psychological and physical health problems (Astin et al.,
2003; Ives & Sosnoff, 2000). Additional research is needed
to explore this further and examine the extent to which
benefits derive from the physical or mental aspects of
mind–body exercise.
In almost all cases, assessment type, study design, and
year of publication did not significantly moderate our
results. Our results were generally consistent across
objective and subjective assessments of sleep although
there were some inconsistencies within objective assess-
ment type for acute exercise. In general, effect sizes were
consistent across open trials and RCTs of acute exercise,
with the exception of rapid eye movement sleep latency.
Acute exercise appeared to have beneficial effects on rapid
eye movement sleep latency in RCTs, but negative effects
in open trials. This may explain why our results for rapid
eye movement sleep latency vary from previous meta-
analyses (Youngstedt et al., 1997), which did not include
more recently published RCTs. Although the literature
represented in this meta-analysis spans almost 50 years,
year of publication did not significantly moderate any of
our results, which indicates that findings are consistent
across time.
Mechanisms
Various pathways have been proposed to explain the
relationship between exercise and sleep, yet the underlying
mechanisms are still uncertain. Proposed mechanisms (for
review see (Buman & King, 2010; Uchida et al., 2012)
include body temperature changes (McGinty & Szymusiak,
1990), cytokine concentration changes (Santos et al.,
2007), increased energy consumption/metabolic rate
(Morselli et al., 2012), CNS fatigue (Uchida et al., 2012),
changes in mood/anxiety symptoms (Buman & King, 2010;
Uchida et al., 2012), changes in heart rate and heart rate
variability (Sandercock et al., 2005), growth hormone
secretion (Kanaley et al., 1997), brain derived neurotropic
factor (BDNF) secretion (Zagaar et al., 2013), improved
fitness level (Shapiro et al., 1984), and body composition
change (Uchida et al., 2012). Adding further complexity to
J Behav Med
123
this issue, research suggests that mechanistic pathways
may differ for acute versus regular exercise (Uchida et al.,
2012).
Our findings regarding moderators of the effects of
exercise on sleep provide insight into some of these
potential mechanisms. The hypothesis that improved fit-
ness level and body composition change may mediate the
relationship between exercise and sleep is somewhat sup-
ported by our finding that individuals with high baseline
physical activity reap more sleep benefits from acute
exercise. As the impact of regular exercise on sleep was
not examined in individuals with high baseline levels of
activity, it is unclear how fitness level may mediate the
relationship between regular exercise and sleep. Given that
this meta-analysis utilized non-psychiatric samples, we
excluded large-scale changes in mood or anxiety as an
explanation from benefits, although changes in the normal
range may still have an effect. The body temperature
hypothesis is called into question by the fact that we found
beneficial effects of exercise across multiples times of day
(greater than 8 h before and less than 3 h before bedtime).
This suggests that body temperature changes cannot be the
sole mechanisms of action. There is some evidence that it
may play a role, however, given that exercising less than
3 h before bedtime was particularly beneficial for wake
time after sleep onset. Interestingly, the theoretical basis
for the body temperature hypothesis came from research
indicating that slow wave sleep leads to a reduction in body
temperature and that in turn, increases body temperature
may promote slow wave sleep (McGinty & Szymusiak,
1990). We did not, however, find that exercising less than
3 h before bedtime was particularly beneficial for slow
wave sleep. Further research on whether exercising close to
bedtime results in body temperature elevation and how this
impacts sleep is needed as the current literature is incon-
sistent (Flausino et al., 2012; Horne & Staff, 1983). With
regards to changes in heart rate and heart rate variability, it
is unclear whether this mechanism can explain the bene-
ficial effects of acute exercise on sleep. Research has
shown that chronic exercise affects HR and HRV (Dixon
et al., 1992; Furlan et al., 1993; Routledge et al., 2010),
however, studies of acute exercise have found HR eleva-
tions but not significant changes in HRV and have only
examined late-night exercise (Myllymaki et al., 2011,
2012; Uchida et al., 2012). This mechanism would not
explain the beneficial effects we found for acute exercise
conducted earlier in the day (greater than 8 h before bed-
time).
Limitations
This meta-analysis is not without limitations. First, we
examined baseline levels of physical activity, rather than
utilizing objective metrics of fitness. We decided not to
examine objective metrics of fitness because few studies
provided this data and there is much debate over what is an
appropriate measure of fitness (Vanhees et al., 2005).
Although many studies provided body mass index (BMI),
BMI is not a reliable indicator of fitness (Prentice & Jebb,
2001). Peak oxygen uptake in response to maximal exer-
cise testing is considered to be the gold standard assess-
ment for exercise tolerance (Vanhees et al., 2005);
however, few studies utilized exercise testing or provided
this metric. Second, we were unable to examine baseline
levels of sedentary behavior (e.g. sitting, TV watching;
Rhodes et al., 2012) given that few articles characterized
participants based on this newer definition of sedentary. In
addition to studying baseline level of physical activity,
future studies should examine the impact of exercise on
sleep in groups of individuals who are engaging in greater
than average amounts of seated behaviors. Third, this meta-
analysis utilized slightly more restrictive exclusion criteria
than past meta-analyses. Although these exclusion criteria
resulted in some studies that were included in previous
meta-analyses being excluded from our analyses, we
believe that this served to increase the validity of our
results. For example, we excluded studies conducted in
individuals with medical or psychological conditions (other
than sleep disorders). As psychological disorders and
medical conditions can have a large impact on sleep
(Ohayon & Roth, 2003; Roth, 2007), we wanted to deter-
mine whether physical activity had a positive impact on
sleep without the confounding influence of these condi-
tions. Given our findings, further research should explore
whether exercise can help address sleep issues comorbid to
medical conditions and psychological disorders. Fourth, the
risk of bias of many studies included in the analyses was
unclear given that many studies did not disclose aspects of
their procedures. Lastly, as with all meta-analysis, we are
limited to data that have been published. In order to
account for this, we were very inclusive in our search
strategy and utilized multiple methods to evaluate the
likelihood of publication bias.
Conclusions
In summary, our meta-analysis presents compelling evi-
dence supporting exercise as an evidence-based interven-
tion to improve perceived and objective metrics of sleep in
healthy individuals. Our results indicate that the benefits of
exercise for sleep are realized immediately, with exercise
having an acute positive impact on many important
objective metrics of sleep. These immediate benefits pro-
vide the right contingency to motivate continued exercising
to improve sleep (Lattal, 2010), and should not suffer from
J Behav Med
123
potential demotivating delay discounting effects (Madden
& Bickel, 2010) as in the case of exercising for weight loss
or cardiovascular fitness. Furthermore, our results suggest
that regular exercise leads to even greater subjective and
objective sleep benefits over time, with subjective benefits
being comparable to those produced by behavior therapy or
pharmacotherapy for insomnia. In light of this evidence
and the high prevalence of sleep disturbance in the general
population, there is support for the use of exercise as a
prescriptive to improve sleep quality, with expectations for
immediate benefits that have the potential to grow over
time.
Conflict of interest M. Alexandra Kredlow, Michelle C. Capozzoli,
Bridget A. Hearon and Amanda W. Calkins declare that they have no
conflicts of interest. In the past 2 years, Michael W. Otto has served
as a paid consultant for MicroTransponder Inc., Concert Pharma-
ceuticals, and ProPhase; provided expert consensus opinion for Ot-
suka Pharmaceuticals, received royalty support for use of the SIGH-A
from ProPhase, and received book royalties from Oxford University
Press, Routledge, and Springer.
Human and Animal Rights and Informed Consent The authors
of this manuscript conform to the Helsinki Declaration concerning
informed consent and human rights and follow correct procedures
concerning experimental studies involving humans.
References
References marked with an asterisk indicate studies
included in the meta-analysis
*Adamson, L., Hunter, W. M., Ogunremi, O. O., Oswald, I., & Percy-
Robb, I. W. (1974). Growth hormone increase during sleep after
daytime exercise. Journal of Endocrinology, 62, 473–478.
*Alessi, C. A., Martin, J. L., Webber, A. P., Cynthia Kim, E., Harker,
J. O., & Josephson, K. R. (2005). Randomized, controlled trial of
a nonpharmacological intervention to improve abnormal sleep/
wake patterns in nursing home residents. Journal of the
American Geriatriatric Society, 53, 803–810. doi:10.1111/j.
1532-5415.2005.53251.x
*Alessi, C. A., Yoon, E. J., Schnelle, J. F., Al-Samarrai, N. R., &
Cruise, P. A. (1999). A randomized trial of a combined physical
activity and environmental intervention in nursing home resi-
dents: do sleep and agitation improve? Journal of the American
Geriatriatric Society, 47, 784–791.
American College of Sports Medicine (2013). ACSM’s guidelines for
exercise testing and prescription. USA: Lippincott Williams &
Wilkins.
Astin, J. A., Shapiro, S. L., Eisenberg, D. M., & Forys, K. L. (2003).
Mind-body medicine: State of the science, implications for
practice. Journal of the American Board of Family Practice, 16,
131–147.
*Baekeland, F. (1970). Exercise deprivation. Sleep and psychological
reactions. Archives of General Psychiatry, 22, 365–369.
*Baekeland, F., & Lasky, R. (1966). Exercise and sleep patterns in
college athletes. Perceptual and Motor Skills, 23, 1203–1207.
Bastien, C. H., Vallieres, A., & Morin, C. M. (2001). Validation of the
Insomnia Severity Index as an outcome measure for insomnia
research. Sleep Medicine, 2, 297–307.
*Bevier, W. C., Bliwise, D. L., Bliwise, N. G., Bunnell, D. E., &
Horvath, S. M. (1992). Sleep patterns of older adults and the
effects of exercise. Journal of Clinical and Experimental
Gerontology, 14, 1–15.
Borbe
´ly, A. A., Mattmann, P., Loepfe, M., Strauch, I., & Lehmann, D.
(1984). Effect of benzodiazepine hypnotics on all-night sleep
EEG spectra. Human Neurobiology, 4, 189–194.
Borenstein, Michael, Hedges, Larry, Higgins, Julian, & Rothstein,
Hannah. (2005). Comprehensive meta-analysis version 2 (p.
104). Englewood, NJ: Biostat.
*Browman, C. P. (1980). Sleep following sustained exercise.
Psychophysiology, 17, 577–580.
*Bulckaert, A., Exadaktylos, V., Haex, B., De Valck, E., Verbraec-
ken, J., & Berckmans, D. (2011). Elevated variance in heart rate
during slow-wave sleep after late-night physical activity. Chro-
nobiology International, 28, 282–284. doi:10.3109/07420528.
2011.552820
Buman, Matthew P., & King, Abby C. (2010). Exercise as a treatment
to enhance sleep. American Journal of Lifestyle Medicine, 4,
500–514.
*Bunnell, D. E., Bevier, W. C., & Horvath, S. M. (1983a). Nocturnal
sleep, cardiovascular function, and adrenal activity following
maximum-capacity exercise. Electroencephalography and Clin-
ical Neurophysiology, 56, 186–189.
*Bunnell, D. E., Bevier, W., & Horvath, S. M. (1983b). Effects of
exhaustive exercise on the sleep of men and women. Psycho-
physiology, 20, 50–58.
Carskadon, Mary A., & Dement, William C. (2000). Normal human
sleep: An overview. Principles and Practice of Sleep Medicine,
4, 13–23.
*Chen, K. M., Chen, M. H., Chao, H. C., Hung, H. M., Lin, H. S., &
Li, C. H. (2009). Sleep quality, depression state, and health
status of older adults after silver yoga exercises: Cluster
randomized trial. International Journal of Nursing Studies, 46,
154–163. doi:10.1016/j.ijnurstu.2008.09.005
*Chen, K. M., Chen, M. H., Lin, M. H., Fan, J. T., Lin, H. S., & Li, C.
H. (2010). Effects of yoga on sleep quality and depression in
elders in assisted living facilities. Journal of Nursing Research,
18, 53–61. doi:10.1097/JNR.0b013e3181ce5189
*Chen, M. C., Liu, H. E., Huang, H. Y., & Chiou, A. F. (2012). The
effect of a simple traditional exercise programme (Baduanjin
exercise) on sleep quality of older adults: A randomized
controlled trial. International Journal of Nursing Studies, 49,
265–273. doi:10.1016/j.ijnurstu.2011.09.009
Cochrane Collaboration (2011). Cochrane handbook for systematic
reviews of interventions version 5.1.0. Retrieved from http://
handbook.cochrane.org/
Cohen, J. (1977). Statistical power analysis for the behavioral
sciences (rev). UK: Lawrence Erlbaum Associates, Inc.
*de Jong, J., Lemmink, K. A., Stevens, M., de Greef, M. H., Rispens,
P., King, A. C., & Mulder, T. (2006). Six-month effects of the
Groningen active living model (GALM) on physical activity,
health and fitness outcomes in sedentary and underactive older
adults aged 55–65. Patient Education and Counseling, 62,
132–141. doi:10.1016/j.pec.2005.06.017
*Desjardins, J., Healey, T., & Broughton, R. (1974). Early evening
exercise and sleep. Sleep Research, 3, 447–450.
Dew, M. A., Hoch, C. C., Buysse, D. J., Monk, T. H., Begley, A. E.,
Houck, P. R., et al. (2003). Healthy older adults’ sleep predicts
all-cause mortality at 4 to 19 years of follow-up. Psychosomatic
Medicine, 65, 63–73.
Diekelmann, S., & Born, J. (2010). The memory function of sleep.
Nature Reviews Neuroscience, 11, 114–126. doi:10.1038/nrn2762
Dixon, E. M., Kamath, M. V., McCartney, N., & Fallen, E. L. (1992).
Neural regulation of heart rate variability in endurance athletes
and sedentary controls. Cardiovascular Research, 26, 713–719.
J Behav Med
123
*Driver, H. S., Meintjes, A. F., Rogers, G. G., & Shapiro, C. M.
(1988). Submaximal exercise effects on sleep patterns in young
women before and after an aerobic training programme. Acta
Physiologica Scandinavica. Supplementum, 574, 8–13.
*Driver, H. S., Rogers, G. G., Mitchell, D., Borrow, S. J., Allen, M.,
Luus, H. G., & Shapiro, C. M. (1994). Prolonged endurance
exercise and sleep disruption. Medicine and Science in Sports
and Exercise, 26, 903–907.
Driver, H. S., & Taylor, S. R. (2000). Exercise and sleep. Sleep
Medicine Reviews, 4, 387–402. doi:10.1053/smrv.2000.0110
Duval, S., & Tweedie, R. (2000). Trim and fill: A simple funnel-plot-
based method of testing and adjusting for publication bias in
meta-analysis. Biometrics, 56, 455–463.
Edinger, J. D. (2009). Is it time to step up to stepped care with our
cognitive-behavioral insomnia therapies? Sleep, 32, 1539–1541.
Ekkekakis, P., Hall, E. E., & Petruzzello, S. J. (2008). The
relationship between exercise intensity and affective responses
demystified: To crack the 40-year-old nut, replace the 40-year-
old nutcracker! Annals of Behavioral Medicine, 35, 136–149.
doi:10.1007/s12160-008-9025-z
*Elavsky, S., & McAuley, E. (2007). Lack of perceived sleep
improvement after 4-month structured exercise programs. Men-
opause, 14, 535–540. doi:10.1097/01.gme.0000243568.70946.
d4
Espie, C. A. (2009). ‘‘Stepped care’’: A health technology solution for
delivering cognitive behavioral therapy as a first line insomnia
treatment. Sleep, 32, 1549–1558.
*Esteves, A. M., de Mello, M. T., Pradella-Hallinan, M., & Tufik, S.
(2009). Effect of acute and chronic physical exercise on patients
with periodic leg movements. Medicine and Science in Sports and
Exercise, 41, 237–242. doi:10.1249/MSS.0b013e318183bb22
Field, Andy P. (2003). The problems in using fixed-effects models of
meta-analysis on real-world data. Understanding Statistics:
Statistical Issues in Psychology, Education, and the Social
Sciences, 2, 105–124.
*Flausino, N. H., Da Silva Prado, J. M., de Queiroz, S. S., Tufik, S., &
de Mello, M. T. (2012). Physical exercise performed before
bedtime improves the sleep pattern of healthy young good
sleepers. Psychophysiology, 49, 186–192. doi:10.1111/j.1469-
8986.2011.01300.x
Follmann, D., Elliott, P., Suh, I. L., & Cutler, J. (1992). Variance
imputation for overviews of clinical trials with continuous
response. Journal of Clinical Epidemiology, 45, 769–773.
doi:10.1016/0895-4356(92)90054-Q
*Frye, B., Scheinthal, S., Kemarskaya, T., & Pruchno, R. (2007). Tai
chi and low impact exercise: Effects on the physical functioning
and psychological well-being of older people. Journal of Applied
Gerontology, 26, 433–453. doi:10.1177/0733464807306915
Furlan, R., Piazza, S., Dell’Orto, S., Gentile, E., Cerutti, S., Pagani,
M., & Malliani, A. (1993). Early and late effects of exercise and
athletic training on neural mechanisms controlling heart rate.
Cardiovascular Research, 27, 482–488.
Garber, C. E., Blissmer, B., Deschenes, M. R., Franklin, B. A.,
Lamonte, M. J., Lee, I. M. et al. (2011). American College of
Sports Medicine position stand. Quantity and quality of exercise
for developing and maintaining cardiorespiratory, musculoskele-
tal, and neuromotor fitness in apparently healthy adults: Guidance
for prescribing exercise. Medicine Science in Sports and Exercise,
43, 1334–1359. doi:10.1249/MSS.0b013e318213fefb
Glass, J., Lanctot, K. L., Herrmann, N., Sproule, B. A., & Busto, U. E.
(2005). Sedative hypnotics in older people with insomnia: Meta-
analysis of risks and benefits. BMJ, 331, 1169. doi:10.1136/bmj.
38623.768588.47
Greenland, S., & O’Rourke, K. (2001). On the bias produced by
quality scores in meta-analysis, and a hierarchical view of
proposed solutions. Biostatistics, 2, 463–471. doi:10.1093/
biostatistics/2.4.463
*Griffin, S. J., & Trinder, J. (1978). Physical fitness, exercise, and
human sleep. Psychophysiology, 15, 447–450.
*Guilleminault, C., Clerk, A., Black, J., Labanowski, M., Pelayo, R.,
& Claman, D. (1995). Nondrug treatment trials in psychophys-
iologic insomnia. Archives of Internal Medicine, 155, 838–844.
*Hague, J. F., Gilbert, S. S., Burgess, H. J., Ferguson, S. A., &
Dawson, D. (2003). A sedentary day: Effects on subsequent
sleep and body temperatures in trained athletes. Physiology &
Behavior, 78, 261–267.
Harvey, A. G. (2002). A cognitive model of insomnia. Behaviour
Research and Therapy, 40, 869–893.
Higgins, J. P., Altman, D. G., Gotzsche, P. C., Juni, P., Moher, D.,
Oxman, A. D., et al. (2011). The Cochrane Collaboration’s tool
for assessing risk of bias in randomised trials. BMJ, 343, d5928.
doi:10.1136/bmj.d5928
Hootman, J. M., Macera, C. A., Ainsworth, B. E., Martin, M., Addy,
C. L., & Blair, S. N. (2001). Association among physical activity
level, cardiorespiratory fitness, and risk of musculoskeletal
injury. American Journal of Epidemiology, 154, 251–258.
*Horne, J. A., & Moore, V. J. (1985). Sleep EEG effects of exercise
with and without additional body cooling. Electroencephalog-
raphy and Clinical Neurophysiology, 60, 33–38.
*Horne, J. A., & Staff, L. H. (1983). Exercise and sleep: Body-heating
effects. Sleep, 6, 36–46.
*Hosseini, H., Esfirizi, M. F., Marandi, S. M., & Rezaei, A. (2011).
The effect of Ti Chi exercise on the sleep quality of the elderly
residents in Isfahan, Sadeghieh elderly home. Iranian Journal of
Nursing and Midwifery Research, 16, 55–60.
Ives, J. C., & Sosnoff, J. (2000). Beyond the mind-body exercise
hype. The Physician and Sportsmedicine, 28, 67–81. doi:10.
3810/psm.2000.03.776
*Jennings, S. E. (1981). The Effects of a 9-week conditioning program
on slow wave sleep and mood states of physically inactive women.
(Doctor of Philosophy), Pennsylvania State University. Retrieved
from http://search.proquest.com.ezp-prod1.hul.harvard.edu/
docview/303122521?accountid=11311
Juni, P., Witschi, A., Bloch, R., & Egger, M. (1999). The hazards of
scoring the quality of clinical trials for meta-analysis. JAMA,
282, 1054–1060.
*Kalak, N., Gerber, M., Kirov, R., Mikoteit, T., Yordanova, J., Puhse,
U., et al. (2012). Daily morning running for 3 weeks improved
sleep and psychological functioning in healthy adolescents
compared with controls. Journal of Adolescent Health, 51,
615–622. doi:10.1016/j.jadohealth.2012.02.020
Kanaley, J. A., Weltman, J. Y., Veldhuis, J. D., Rogol, A. D.,
Hartman, M. L., & Weltman, A. (1997). Human growth hormone
response to repeated bouts of aerobic exercise. Journal of
Applied Physiology (1985), 83, 1756–1761.
*King, A. C., Baumann, K., O’Sullivan, P., Wilcox, S., & Castro, C.
(2002). Effects of moderate-intensity exercise on physiological,
behavioral, and emotional responses to family caregiving: A
randomized controlled trial. Journals of Gerontology. Series A,
Biological Sciences and Medical Sciences, 57, M26–M36.
*King, A. C., Oman, R. F., Brassington, G. S., Bliwise, D. L., &
Haskell, W. L. (1997). Moderate-intensity exercise and self-rated
quality of sleep in older adults. A randomized controlled trial.
JAMA, 277, 32–37.
*King, A. C., Pruitt, L. A., Woo, S., Castro, C. M., Ahn, D. K.,
Vitiello, M. V., et al. (2008). Effects of moderate-intensity
exercise on polysomnographic and subjective sleep quality in
older adults with mild to moderate sleep complaints. Journals of
Gerontology. Series A, Biological Sciences and Medical Sci-
ences, 63, 997–1004.
J Behav Med
123
*Kline, C. E., Sui, X., Hall, M. H., Youngstedt, S. D., Blair, S. N.,
Earnest, C. P., & Church, T. S. (2012). Dose-response effects of
exercise training on the subjective sleep quality of postmeno-
pausal women: Exploratory analyses of a randomised controlled
trial. BMJ Open, 2. doi: 10.1136/bmjopen-2012-001044
Kripke, D. F., Garfinkel, L., Wingard, D. L., Klauber, M. R., &
Marler, M. R. (2002). Mortality associated with sleep duration
and insomnia. Archives of General Psychiatry, 59, 131–136.
Kubitz, K. A., Landers, D. M., Petruzzello, S. J., & Han, M. (1996).
The effects of acute and chronic exercise on sleep. A meta-
analytic review. Sports Medicine, 21, 277–291.
*Kupfer, D. J., Sewitch, D. E., Epstein, L. H., Bulik, C., McGowen,
C. R., & Robertson, R. J. (1985). Exercise and subsequent sleep
in male runners: Failure to support the slow wave sleep-mood-
exercise hypothesis. Neuropsychobiology, 14, 5–12.
Lattal, K. A. (2010). Delayed reinforcement of operant behavior.
Journal of the Experimental Analysis of Behavior, 93, 129–139.
doi:10.1901/jeab.2010.93-129
Lauderdale, D. S., Knutson, K. L., Yan, L. L., Liu, K., & Rathouz, P.
J. (2008). Self-reported and measured sleep duration: How
similar are they? Epidemiology, 19, 838–845. doi:10.1097/EDE.
0b013e318187a7b0
*Littman, A. J., Vitiello, M. V., Foster-Schubert, K., Ulrich, C. M.,
Tworoger, S. S., Potter, J. D., et al. (2007). Sleep, ghrelin, leptin
and changes in body weight during a 1-year moderate-intensity
physical activity intervention. Int J Obes (Lond), 31, 466–475.
doi:10.1038/sj.ijo.0803438
Lucidi, F., Devoto, A., Violani, C., De Gennaro, L., Mastracci, P., &
Bertini, M. (1996). Rapid eye movements density as a measure of
sleep need: REM density decreases linearly with the reduction of
prior sleep duration. Electroencephalography and Clinical Neu-
rophysiology, 99, 556–561. doi:10.1016/S0013-4694(96)95671-0
Madden, G. J., & Bickel, W. K. (2010). Impulsivity: The behavioral
and neurological science of discounting: Washington, DC:
American Psychological Association.
Mallon, L., Broman, J. E., & Hetta, J. (2000). Relationship between
insomnia, depression, and mortality: A 12-year follow-up of
older adults in the community. International Psychogeriatrics,
12, 295–306.
*Manzaneque, J. M., Vera, F. M., Rodriguez, F. M., Garcia, G. J.,
Leyva, L., & Blanca, M. J. (2009). Serum cytokines, mood and
sleep after a qigong program: is qigong an effective psychobi-
ological tool? Journal of Health Psychology, 14, 60–67. doi:10.
1177/1359105308097946
*Matsumoto, K., Saito, Y., Abe, M., & Furumi, K. (1984). The effects
of daytime exercise on night sleep. Journal of Human Ergology
(Tokyo), 13, 31–36.
McGinty, D., & Szymusiak, R. (1990). Keeping cool: A hypothesis
about the mechanisms and functions of slow-wave sleep. Trends
in Neurosciences, 13, 480–487.
Mellinger, G. D., Balter, M. B., & Uhlenhuth, E. H. (1985). Insomnia
and its treatment. Prevalence and correlates. Archives of General
Psychiatry, 42, 225–232.
*Montgomery, I., Trinder, J., Paxton, S., Fraser, G., Meaney, M., &
Koerbin, G. L. (1985). Sleep disruption following a marathon.
Journal of Sports Medicine and Physical Fitness, 25, 69–74.
*Montgomery, I., Trinder, J., Paxton, S., Harris, D., Fraser, G., &
Colrain, I. (1988). Physical exercise and sleep: the effect of the
age and sex of the subjects and type of exercise. Acta
Physiologica Scandinavica. Supplementum, 574, 36–40.
Montgomery, P., & Dennis, J. (2003). Cognitive behavioural inter-
ventions for sleep problems in adults aged 60+. Cochrane
Database of Systematic Reviews, 1, CD003161. doi:10.1002/
14651858.cd003161
*Montmayeur, A., Buguet, A., Sollin, H., & Lacour, J. R. (1994).
Exercise and sleep in four African sportsmen living in the Sahel.
A pilot study. International Journal of Sports Medicine, 15,
42–45. doi:10.1055/s-2007-1021018
Morin, C. M., Hauri, P. J., Espie, C. A., Spielman, A. J., Buysse, D. J.,
& Bootzin, R. R. (1999). Nonpharmacologic treatment of
chronic insomnia. An American Academy of Sleep Medicine
review. Sleep, 22, 1134–1156.
Morin, C. M., LeBlanc, M., Daley, M., Gregoire, J. P., & Merette, C.
(2006). Epidemiology of insomnia: Prevalence, self-help treat-
ments, consultations, and determinants of help-seeking behav-
iors. Sleep Medicine, 7, 123–130. doi:10.1016/j.sleep.2005.08.
008
*Morita, E., Imai, M., Okawa, M., Miyaura, T., & Miyazaki, S.
(2011). A before and after comparison of the effects of forest
walking on the sleep of a community-based sample of people
with sleep complaints. Biopsychosoc Med, 5, 13. doi:10.1186/
1751-0759-5-13
Morselli, L. L., Guyon, A., & Spiegel, K. (2012). Sleep and metabolic
function. Pflugers Archiv, 463, 139–160. doi:10.1007/s00424-
011-1053-z
*Myllymaki, T., Kyrolainen, H., Savolainen, K., Hokka, L., Jakonen,
R., Juuti, T., et al. (2011). Effects of vigorous late-night exercise
on sleep quality and cardiac autonomic activity. Journal of Sleep
Research, 20, 146–153. doi:10.1111/j.1365-2869.2010.00874.x
*Myllymaki, T., Rusko, H., Syvaoja, H., Juuti, T., Kinnunen, M. L.,
& Kyrolainen, H. (2012). Effects of exercise intensity and
duration on nocturnal heart rate variability and sleep quality.
European Journal of Applied Physiology, 112, 801–809. doi:10.
1007/s00421-011-2034-9
National Sleep Foundation (n.d.). Sleep aids and insomnia. Retrieved
2014, from http://sleepfoundation.org/sleep-disorders-problems/
insomnia/sleep-aids-and-insomnia/page/0,1/.
*Naylor, E., Penev, P. D., Orbeta, L., Janssen, I., Ortiz, R., Colecchia,
E. F., et al. (2000). Daily social and physical activity increases
slow-wave sleep and daytime neuropsychological performance
in the elderly. Sleep, 23, 87–95.
*Nguyen, M. H., & Kruse, A. (2012). A randomized controlled trial
of Tai chi for balance, sleep quality and cognitive performance
in elderly Vietnamese. Clinical Interventions in Aging, 7,
185–190. doi:10.2147/cia.s32600
*O’Connor, P. J., Breus, M. J., & Youngstedt, S. D. (1998). Exercise-
induced increase in core temperature does not disrupt a
behavioral measure of sleep. Physiology & Behavior, 64,
213–217.
*Oda, S. (2001). The effects of recreational underwater exercise in
early evening on sleep for physically untrained male subjects.
Psychiatry and Clinical Neurosciences, 55, 179–181. doi:10.
1046/j.1440-1819.2001.00816.x
Ohayon, M. M. (2002). Epidemiology of insomnia: What we know
and what we still need to learn. Sleep Medicine Reviews, 6,
97–111.
Ohayon, M. M., Caulet, M., & Guilleminault, C. (1997). How a
general population perceives its sleep and how this relates to the
complaint of insomnia. Sleep, 20, 715–723.
Ohayon, M. M., Carskadon, M. A., Guilleminault, C., & Vitiello, M.
V. (2004). Meta-analysis of quantitative sleep parameters from
childhood to old age in healthy individuals: Developing norma-
tive sleep values across the human lifespan. Sleep, 27,
1255–1274.
Ohayon, M. M., & Roth, T. (2003). Place of chronic insomnia in the
course of depressive and anxiety disorders. Journal of Psychi-
atric Research, 37, 9–15.
*Oudegeest-Sander, M. H., Eijsvogels, T. H., Verheggen, R. J.,
Poelkens, F., Hopman, M. T., Jones, H., & Thijssen, D. H.
(2013). Impact of physical fitness and daily energy expenditure
on sleep efficiency in young and older humans. Gerontology, 59,
8–16. doi:10.1159/000342213
J Behav Med
123
*Passos, G. S., Poyares, D., Santana, M. G., Garbuio, S. A., Tufik, S.,
& Mello, M. T. (2010). Effect of acute physical exercise on
patients with chronic primary insomnia. Journal of Clinical
Sleep Medicine, 6, 270–275.
*Paxton, S. J., Montgomery, I., Trinder, J., Newman, J., & Bowling,
A. (1982). Sleep after exercise of variable intensity in fit and
unfit subjects. Australian Journal of Psychology, 34, 289–296.
doi:10.1080/00049538208254724
*Paxton, S. J., Trinder, J., & Montgomery, I. (1983). Does aerobic
fitness affect sleep? Psychophysiology, 20, 320–324.
Peigneux, P., Laureys, S., Fuchs, S., Collette, F., Perrin, F., Reggers,
J., et al. (2004). Are spatial memories strengthened in the human
hippocampus during slow wave sleep? Neuron, 44, 535–545.
doi:10.1016/j.neuron.2004.10.007
Pilcher, J. J., Ginter, D. R., & Sadowsky, B. (1997). Sleep quality
versus sleep quantity: Relationships between sleep and measures
of health, well-being and sleepiness in college students. Journal
of Psychosomatic Research, 42, 583–596.
*Pinniger, R., Thorsteinsson, E. B., Brown, R. F., & McKinley, P.
(2013). Tango dance can reduce distress and insomnia in people
with self-referred affective symptoms. American Journal of
Dance Therapy, 35, 60–77. doi:10.1007/s10465-012-9141-y
Prentice, A. M., & Jebb, S. A. (2001). Beyond body mass index.
Obesity Reviews, 2, 141–147.
Quera-Salva, M. A., Orluc, A., Goldenberg, F., & Guilleminault, C.
(1991). Insomnia and use of hypnotics: Study of a French
population. Sleep, 14, 386–391.
*Reid, K. J., Baron, K. G., Lu, B., Naylor, E., Wolfe, L., & Zee, P. C.
(2010). Aerobic exercise improves self-reported sleep and
quality of life in older adults with insomnia. Sleep Medicine,
11, 934–940. doi:10.1016/j.sleep.2010.04.014
Requa, R. K., DeAvilla, L. N., & Garrick, J. G. (1993). Injuries in
recreational adult fitness activities. American Journal of Sports
Medicine, 21, 461–467.
Rhodes, R. E., Mark, R. S., & Temmel, C. P. (2012). Adult sedentary
behavior: A systematic review. American Journal of Preventive
Medicine, 42, 3–28. doi:10.1016/j.amepre.2011.10.020
*Richards, K. C., Lambert, C., Beck, C. K., Bliwise, D. L., Evans, W.
J., Kalra, G. K., et al. (2011). Strength training, walking, and
social activity improve sleep in nursing home and assisted living
residents: Randomized controlled trial. Journal of the American
Geriatrics Society, 59, 214–223. doi:10.1111/j.1532-5415.2010.
03246.x
Rosenthal, R. (1991). Meta-analytic procedures for social research
(Vol. 6). Beverly Hills, CA: Sage.
Rosenthal, R., & Rubin, D. B. (1988). [Selection models and the file
drawer problem]: Comment: Assumptions and procedures in the
file drawer problem. Statistical Science, 3, 120–125.
Roth, T. (2007). Insomnia: Definition, prevalence, etiology, and
consequences. Journal of Clinical Sleep Medicine, 3, S7–S10.
Routledge, F. S., Campbell, T. S., McFetridge-Durdle, J. A., &
Bacon, S. L. (2010). Improvements in heart rate variability with
exercise therapy. Canadian Journal of Cardiology, 26, 303–312.
Sandercock, G. R., Bromley, P. D., & Brodie, D. A. (2005). Effects of
exercise on heart rate variability: Inferences from meta-analysis.
Medicine and Science in Sports and Exercise, 37, 433–439.
Santos, R. V., Tufik, S., & De Mello, M. T. (2007). Exercise, sleep
and cytokines: Is there a relation? Sleep Medicine Reviews, 11,
231–239. doi:10.1016/j.smrv.2007.03.003
Sarris, J., & Byrne, G. J. (2011). A systematic review of insomnia and
complementary medicine. Sleep Medicine Reviews, 15, 99–106.
doi:10.1016/j.smrv.2010.04.001
Schroeder, J. (2012). Mind/body exercise: What is it? American
College of Sports Medicine. Retrieved from http://www.acsm.
org/access-public-information/articles/2012/01/19/mind-body-
exercise-what-is-it-
Schwartz, S., McDowell Anderson, W., Cole, S. R., Cornoni-Huntley,
J., Hays, J. C., & Blazer, D. (1999). Insomnia and heart disease:
A review of epidemiologic studies. Journal of Psychosomatic
Research, 47, 313–333.
Semler, C. N., & Harvey, A. G. (2005). Misperception of sleep can
adversely affect daytime functioning in insomnia. Behaviour
Research and Therapy, 43, 843–856. doi:10.1016/j.brat.2004.06.016
*Shapiro, C. M., Bortz, R., Mitchell, D., Clarke, R. A. I., & Jooste, P.
L. (1985). Effect of exercise in a hot environment on human
sleep patterns. South African Journal of Science, 81, 624–627.
Shapiro, C. M., Warren, P. M., Trinder, J., Paxton, S. J., Oswald, I.,
Flenley, D. C., & Catterall, J. R. (1984). Fitness facilitates sleep.
European Journal of Applied Physiology and Occupational
Physiology, 53, 1–4.
Shochat, T., Umphress, J., Israel, A. G., & Ancoli-Israel, S. (1999).
Insomnia in primary care patients. Sleep, 22, S359–S365.
Simon, G. E., & VonKorff, M. (1997). Prevalence, burden, and
treatment of insomnia in primary care. American Journal of
Psychiatry, 154, 1417–1423.
Smith, M. T., Perlis, M. L., Park, A., Smith, M. S., Pennington, J.,
Giles, D. E., & Buysse, D. J. (2002). Comparative meta-analysis
of pharmacotherapy and behavior therapy for persistent insom-
nia. American Journal of Psychiatry, 159, 5–11.
Stepanski, E. J., & Wyatt, J. K. (2003). Use of sleep hygiene in the
treatment of insomnia. Sleep Medicine Reviews, 7, 215–225.
*Trinder, J., Bruck, D., Paxton, S. J., Montgomery, I., & Bowling, A.
(1982). Physical-fitness, exercise, age and human sleep. Austra-
lian Journal of Psychology, 34, 131–138. doi:10.1080/
00049538208257806
*Tworoger, S. S., Yasui, Y., Vitiello, M. V., Schwartz, R. S., Ulrich,
C. M., Aiello, E. J., et al. (2003). Effects of a yearlong moderate-
intensity exercise and a stretching intervention on sleep quality
in postmenopausal women. Sleep, 26, 830–836.
Uchida, S., Shioda, K., Morita, Y., Kubota, C., Ganeko, M., &
Takeda, N. (2012). Exercise effects on sleep physiology.
Frontiers in Neurology, 3, 48. doi:10.3389/fneur.2012.00048
Vanhees, L., Lefevre, J., Philippaerts, R., Martens, M., Huygens, W.,
Troosters, T., & Beunen, G. (2005). How to assess physical
activity? How to assess physical fitness? European Journal of
Cardiovascular Prevention & Rehabilitation, 12, 102–114.
*Vein, A. M., Sidorov, A. A., Murtazaev, M. S., Karlov, A. V., et al.
(1991). Physical exercise and nocturnal sleep in healthy humans.
Human Physiology, 17, 391–397.
Vevea, Jack L., & Woods, Carol M. (2005). Publication bias in
research synthesis: Sensitivity analysis using a priori weight
functions. Psychological Methods, 10, 428.
*Viana, V. A., Esteves, A. M., Boscolo, R. A., Grassmann, V.,
Santana, M. G., Tufik, S., & de Mello, M. T. (2012). The effects
of a session of resistance training on sleep patterns in the elderly.
European Journal of Applied Physiology, 112, 2403–2408.
doi:10.1007/s00421-011-2219-2
*Walker, J. M., Floyd, T. C., Fein, G., Cavness, C., Lualhati, R., &
Feinburg, I. (1978). Effects of exercise on sleep. Journal of
Applied Physiology: Respiratory, Environmental and Exercise
Physiology, 44, 945–951.
Walker, M. P., & van der Helm, E. (2009). Overnight therapy? The
role of sleep in emotional brain processing. Psychological
Bulletin, 135, 731–748. doi:10.1037/a0016570
*Wong, S. N., Halaki, M., & Chow, C. M. (2013). The effects of
moderate to vigorous aerobic exercise on the sleep need of
sedentary young adults. Journal of Sports Sciences, 31, 381–386.
doi:10.1080/02640414.2012.733823
Yang, P. Y., Ho, K. H., Chen, H. C., & Chien, M. Y. (2012). Exercise
training improves sleep quality in middle-aged and older adults
with sleep problems: A systematic review. Journal of Physio-
therapy, 58, 157–163. doi:10.1016/s1836-9553(12)70106-6
J Behav Med
123
*Yeh, S. C., & Chang, M. Y. (2012). The effect of Qigong on
menopausal symptoms and quality of sleep for perimenopausal
women: A preliminary observational study. Journal of Alterna-
tive and Complementary Medicine, 18, 567–575. doi:10.1089/
acm.2011.0133
Youngstedt, S. D., O’Connor, P. J., & Dishman, R. K. (1997). The
effects of acute exercise on sleep: A quantitative synthesis.
Sleep, 20, 203–214.
Zagaar, M., Dao, A., Alhaider, I., & Alkadhi, K. (2013). Regular
treadmill exercise prevents sleep deprivation-induced disruption
of synaptic plasticity and associated signaling cascade in the
dentate gyrus. Molecular and Cellular Neuroscience, 56,
375–383. doi:10.1016/j.mcn.2013.07.011
Zeitlhofer, J., Schmeiser-Rieder, A., Tribl, G., Rosenberger, A.,
Bolitschek, J., & Kapfhammer, G. (2000). Sleep and quality of
life in the Austrian population. Acta Neurologica Scandinavica,
102, 249–257.
*Zir, L. M., Smith, R. A., & Parker, D. C. (1971). Human growth
hormone release in sleep: Effect of daytime exercise. Journal of
Clinical Endocrinology and Metabolism, 32, 662–665. doi:10.
1210/jcem-32-5-6
J Behav Med
123
