Sleep Deprivation and Memory Meta-Analytic Reviews of Studies On Sleep Deprivation Before and After Learning

Psychological Bulletin
© 2021 The Author(s) 2021, Vol. 147, No. 11, 1215–1240

ISSN: 0033-2909 https://doi.org/10.1037/bul0000348
Sleep Deprivation and Memory: Meta-Analytic Reviews of Studies

on Sleep Deprivation Before and After Learning
Chloe R. Newbury, Rebecca Crowley, Kathleen Rastle, and Jakke Tamminen
Department of Psychology, Royal Holloway, University of London
Research suggests that sleep deprivation both before and after encoding has a detrimental effect on memory
for newly learned material. However, there is as yet no quantitative analyses of the size of these effects. We
conducted two meta-analyses of studies published between 1970 and 2020 that investigated effects of total,
acute sleep deprivation on memory (i.e., at least one full night of sleep deprivation): one for deprivation
occurring before learning and one for deprivation occurring after learning. The impact of sleep deprivation
after learning on memory was associated with Hedges’ g = 0.277, 95% CI [0.177, 0.377]. Whether testing
took place immediately after deprivation or after recovery sleep moderated the effect, with significantly
larger effects observed in immediate tests. Procedural memory tasks also showed significantly larger effects
than declarative memory tasks. The impact of sleep deprivation before learning was associated with Hedges’
g = 0.621, 95% CI [0.473, 0.769]. Egger’s tests for funnel plot asymmetry suggested significant publication
bias in both meta-analyses. Statistical power was very low in most of the analyzed studies. Highly powered,
preregistered replications are needed to estimate the underlying effect sizes more precisely.
Public Significance Statement

The health risks associated with lack of sleep are well known, but the consequences of missing one or
more nights of sleep for learning and memory are less well appreciated. In two meta-analyses pooling
studies across 5 decades of research, we found that total sleep deprivation before learning as well as after
learning had a detrimental impact on memory for the newly learned materials. These data suggest sleep
supports learning and memory in two ways: It prepares the brain for learning over the next day, and it
helps strengthen new memories learned during the previous day.
Keywords: sleep, sleep deprivation, learning, memory, meta-analysis
Supplemental materials: https://doi.org/10.1037/bul0000348.supp
There is a growing body of evidence suggesting a critical role of wake (Klinzing et al., 2019). On the other hand, memory encoding
sleep in learning and memory (Diekelmann & Born, 2010). On the capacity has been argued to saturate gradually during wake, with
one hand, offline memory consolidation during sleep benefits both sleep restoring this capacity (Cirelli & Tononi, in press; Tononi &
declarative and procedural memories acquired during preceding Cirelli, 2012). These theoretical advances have been accompanied
by practical societal concerns regarding the prevalence of poor
sleep, especially for students (Twenge et al., 2017) and shift workers
(Vidya et al., 2019). The proposed importance of sleep for memory
Chloe R. Newbury https://orcid.org/0000-0003-3515-5566 processes has been supported by many studies showing detrimental
Rebecca Crowley https://orcid.org/0000-0002-8629-8690 effects of total sleep deprivation on the learning and retrieval of new
Kathleen Rastle https://orcid.org/0000-0002-3070-7555 information. Yet, the effect sizes associated with the total sleep
Jakke Tamminen https://orcid.org/0000-0003-1929-3598
deprivation impairment are variable, and some studies have failed to
This work was funded by Economic and Social Research Council grant
find significant effects altogether (e.g., Diekelmann et al., 2008).
ES/P001874/1 awarded to Kathleen Rastle and Jakke Tamminen.
Chloe R. Newbury and Rebecca Crowley contributed equally to this work. Therefore, we used a meta-analytic approach to estimate the effect
The data and analysis scripts are available at osf.io/5gjvs/. size associated with the impact of sleep deprivation, both when the
This article has been published under the terms of the Creative Commons deprivation occurs before learning and when it occurs after learning.
Attribution License (http://creativecommons.org/licenses/by/3.0/), which
permits unrestricted use, distribution, and reproduction in any medium,
provided the original author and source are credited. Copyright for this article Impact of Total Sleep Deprivation After Learning
is retained by the author(s). Author(s) grant(s) the American Psychological and Potential Moderators of the Effect
Association the exclusive right to publish the article and identify itself as the
original publisher.
The active systems consolidation theory suggests that sleep after
Correspondence concerning this article should be addressed to Chloe R. learning strengthens new memories (e.g., Klinzing et al., 2019;
Newbury or Jakke Tamminen, Department of Psychology, Royal Hollo- Kumaran et al., 2016; McClelland et al., 1995). Information learned
way, University of London, Egham TW20 0EX, United Kingdom. Email: during wakefulness is initially encoded rapidly in the hippocampus,
chloe.newbury@rhul.ac.uk or jakke.tamminen@rhul.ac.uk where memories are stored separately from existing memory stores.
1215
1216 NEWBURY, CROWLEY, RASTLE, AND TAMMINEN
Repeated reactivation of these new memories, primarily during studies, participants are allowed to sleep for a shorter duration than
slow-wave sleep (SWS), supports the strengthening of memory they would do otherwise, and the manipulation typically continues for
representations and leads to the integration of these memories in the multiple nights. This chronic sleep restriction can have a detrimental
neocortex. Such neocortical representations are less liable to dis- impact on learning and memory (e.g., Cousins et al., 2018) although
ruption and form interrelated semantic networks with existing not always (e.g., Voderholzer et al., 2011). In selective sleep restric-
memories, yielding memory representations that allow abstraction, tion studies, participants are only deprived of the first half of the night,
generalization, and discovery of statistical patterns across discrete rich in SWS, or the second half of the night, rich in rapid eye
memories (Lewis & Durrant, 2011; Lewis et al., 2018; Stickgold & movement (REM) sleep (e.g., Plihal & Born, 1997). These studies can
Walker, 2013). Notably, since the mechanisms outlined in this reveal important information about the precise brain mechanisms
theory relate only to hippocampal-dependent memory consolidation underlying benefits of sleep on memory. There are two theoretically
(Inostroza & Born, 2013), these mechanisms may primarily support motivated reasons for excluding both types of sleep restriction from
the consolidation of declarative (or explicit) memory. our meta-analyses. Standard sleep restriction studies still allow
Effects of sleep on nondeclarative memory have also been observed; participants to sleep for several hours each night, and it may be
for example, sleep enhances motor skills such as finger-tapping sufficient for sleep-associated memory consolidation to occur. Thus,
sequence learning (Korman et al., 2007; Walker et al., 2002; see these manipulations do not provide a strong test of the relevant
King et al., 2017 for a review). However, this beneficial effect of theories, and their inclusion might lead us to underestimate the effect
sleep on procedural memories may be evident only when learning is size associated with sleep deprivation. Selective sleep restriction
intentional (explicit memory), rather than unintentional (implicit mem- studies on the other hand are designed to address the more fine-
ory). Robertson et al. (2004) found that awareness of learning a finger- grained question of which specific sleep stages are the most beneficial
tapping task led to a sleep benefit, whereas improvements in implicit for memory. Different theories make different predictions in this regard:
learning performance, when participants had little awareness of the For example, the active systems consolidation theory emphasizes the
task, were similar regardless of whether the retention interval contained role of SWS (Klinzing et al., 2019); the theory of Lewis et al. (2018)
sleep or wakefulness. Thus, there are suggestions that the mechanisms emphasizes REM (at least for learning involving creative thought); and
involved in the consolidation of hippocampal-dependent declarative the sequential theory of Giuditta (2014) proposes that an interaction
memories may also facilitate the consolidation of some procedural between SWS and REM is key. However, all current theories make the
tasks that rely on explicit learning and thus show some hippocampal- common prediction that sleep should benefit memory more than wake.
dependency (Schönauer et al., 2014; Walker et al., 2005). We do not attempt to adjudicate between the competing theories and
However, this latter theory does not take into account observa- therefore restrict our analysis to studies using total sleep deprivation
tions that some procedural tasks that do not rely on explicit learning where a clear prediction is made by all theories.
or an intact hippocampus still show superior performance after Where possible we use moderators to establish whether the effect
sleep. Stickgold et al. (2000) found that a period of sleep after size is modulated by variables that have been hypothesized to be
learning a visual discrimination task benefited later performance, important. For example, it is not clear whether sleep deprivation
and Schönauer et al. (2015) found that improvements in perfor- after learning impacts declarative and procedural memories simi-
mance on a mirror-tracing task were only observed after a period of larly. If the neural and cognitive mechanisms associated with the
offline consolidation. Recent studies in both animals (Sawangjit consolidation of declarative and procedural memories differ, one
et al., 2018) and humans (Schapiro et al., 2019) have suggested that may observe different effect sizes in studies targeting the two
the hippocampus may be involved in the sleep-dependent consoli- memory types. To test this hypothesis, we entered declarative versus
dation of memories that do not rely on the hippocampus during procedural memory type as a moderator in our meta-analysis.
encoding. For example, Schapiro et al. (2019) trained amnesic Some studies have tested the effects of sleep deprivation after one
patients with hippocampal damage on the motor sequence task, a or more nights of recovery sleep while others have tested immedi-
classic procedural memory task typically considered to be non- ately after a night of sleep deprivation with no intervening recovery
hippocampus-dependent. The patients were able to learn the task sleep. The primary reason for allowing recovery sleep before testing
equally well compared to matched controls, suggesting that the is that sleep deprivation has well-established impacts on attention
hippocampus is not required for learning of the task. However, while (e.g., Lim & Dinges, 2008; Vargas et al., 2017) that may compro-
the controls showed the expected overnight consolidation benefit, mise performance in an immediate memory test and make it difficult
the patients did not, leading the authors to conclude that the to distinguish between effects of sleep deprivation due to fatigue and
hippocampus may be involved in the consolidation of procedural effects due to disruption to memory consolidation processes. Using
tasks that do not require it for initial learning. recovery sleep to rule out potential effects of fatigue at test assumes
As reviewed above, theories of memory consolidation predict that that the first night of sleep following learning is of critical impor-
depriving participants of sleep after learning should impair memory tance in memory consolidation and that consolidation may not occur
for the information encoded before sleep deprivation, relative to in subsequent sleep periods or has a weaker impact after the first
control conditions where participants are allowed to sleep normally missed sleep opportunity. Although we are not aware of any studies
after learning. In our first meta-analysis, we analyze the current that have explicitly tested this assumption, for example, by system-
research into both declarative and procedural memories to estimate atically manipulating the number of nights of sleep following
the size of this sleep deprivation after learning effect. We focus on the learning, some support has been derived from studies such as
literature using manipulations of total sleep deprivation, as this is the that of Gais et al. (2007) who observed effects of one night of
strongest and most direct manipulation to test theories of sleep- sleep deprivation after two recovery nights and even 6 months later.
associated memory consolidation. In doing so, we exclude from Recent studies have cast doubt on the privileged status of the first
our analyses studies of sleep restriction. In standard sleep restriction night of sleep, however. Schönauer et al. (2015) found only a short-
META-ANALYSES OF SLEEP DEPRIVATION AND MEMORY 1217
term cost of sleep deprivation after learning for hippocampal- for a review) suggesting that sleep’s impact on recognition memory
dependent memories, with sleep deprivation after learning impairing may be modulated by emotionality. Given the inconsistency in the
retrieval of word pairs after one night of deprivation, but no existing literature on the extent to which sleep after learning benefits
difference in retrieval between the sleep and sleep deprivation recognition memory tasks, it is important to quantify and compare the
conditions after two nights of recovery sleep. They suggested size of the sleep benefit across recognition and recall tasks. Thus, we
that for such hippocampal-dependent memories, the hippocampus entered memory type as a moderator in our meta-analysis.
may act as a temporary buffer, storing memories until the first sleep Sleep has previously been found to improve emotional episodic
opportunity, even if that opportunity is delayed. Thus, sleep depri- memories more than neutral memories, with some studies suggest-
vation after learning would only have a detrimental effect on ing a specific role of REM sleep in the consolidation of emotional
memory performance if there is no sleep opportunity before testing. memories (e.g., Groch et al., 2015; Payne & Kensinger, 2010;
In contrast, they found that procedural memories that do not rely on Wagner et al., 2001; see Kim & Payne, 2020 for a review). Two
the hippocampus suffered a more long-term effect of sleep depriva- recent meta-analyses investigated the preferential role of sleep in
tion, supporting previous evidence that the first night of sleep is emotional memory consolidation. Schäfer et al. (2020) found that
crucial for improving performance on procedural tasks (Stickgold sleep improved both emotional and neutral memory equally, with no
et al., 2000). Therefore, for procedural memories, the loss of the first evidence for preferential impact on emotional memory; in fact, the
night of sleep may be critical. Given the mixed evidence in the difference between emotional and neutral memory was larger after
literature on the impact of recovery sleep, we entered the presence or wake than sleep. However, Schäfer et al. (2020) did find that when
absence of recovery sleep as a moderator in the meta-analysis. If the their analysis was restricted to experiments that contrasted SWS and
first night of sleep after learning is of critical importance, on the one REM sleep, sleep that consisted primarily of REM did show a
hand, we should see a sleep deprivation impairment both in studies preferential consolidation effect for emotional memory, although
that test immediately after deprivation and in studies that test after the number of studies included in this analysis was small. Lipinska
recovery sleep. If, on the other hand, consolidation processes can be et al. (2019) also found no meta-analytic evidence in favor of sleep’s
delayed, we should see a sleep deprivation impairment only in preferential impact on emotional memory. However, they did find
studies that test immediately (i.e., they provide no opportunity for that the preferential effect was larger in recall tasks compared to
delayed consolidation). The latter scenario is also consistent with the recognition tasks and in studies that controlled for initial learning.
possibility that the sleep deprivation impairment is due entirely to In the current meta-analysis, we shed new light on these issues by
fatigue at test (although see Schönauer et al., 2015, for data focussing on sleep deprivation manipulations. To investigate
suggesting fatigue does not impair memory recall). whether the effect of sleep deprivation on memory performance
Some studies of recognition memory have failed to find a is modulated by emotionality of the to-be-remembered stimuli, we
beneficial effect of sleep on memory performance leading to a entered emotionality as a moderator in our meta-analysis.
debate about whether sleep has no or only limited impact on
recognition memory (e.g., Drosopoulos et al., 2005; Hu et al., Impact of Total Sleep Deprivation Before Learning
2006; Morgan et al., 2019; Rauchs et al., 2004; Stepan et al.,
and Potential Moderators of the Effect
2017). The difference between recall and recognition tasks is
particularly stark in the literature using the Deese–Roediger– Recent research has also proposed a role for sleep before learning.
McDermott (DRM) paradigm of false memory formation. Here, According to the synaptic homeostasis hypothesis (Cirelli & Tononi, in
studies using recognition tasks have reported that sleep may reduce press; Tononi & Cirelli, 2003, 2012), learning occurs during wake
false memories (Fenn et al., 2009; Lo et al., 2014), whereas studies when a neuron detects a statistical regularity in its input and begins to
using recall tasks have reported that sleep may increase false fire in response to this regular input. In other words, successful learning
memories (Diekelmann et al., 2010; McKeon et al., 2012; requires neurons to be able to fire selectively in response to statistically
Pardilla-Delgado & Payne, 2017; Payne et al., 2009). Consequently, regular patterns observed in the environment. To do so, strength of the
a recent meta-analysis of sleep studies using the DRM paradigm synapses carrying these inputs must be increased. However, the neuron
concluded that the impact of sleep on false memory is restricted to now faces the plasticity-selectivity dilemma. As an increasing number
recall tasks (Newbury & Monaghan, 2019). This meta-analysis also of input lines become strengthened, a larger range of input patterns can
found that the sleep benefit for studied words (i.e., veridical make the neuron fire, reducing the neuron’s ability to fire selectively.
memory) was dramatically larger in recall tasks than in recognition This loss of selectivity corresponds to reduced ability to encode new
tasks (g = 0.407 vs. g = 0.005). The discrepancy between recall and information. During sleep, the brain spontaneously activates both new
recognition tasks might be explained by dual-process accounts, information encoded during previous wake and information encoded in
which suggest that recognition memory has both an explicit recol- the past. Over the course of this activation, those synapses that are
lection element as well as an implicit familiarity element (e.g., activated most strongly and consistently during wake survive, while at
Jacoby, 1991; Yonelinas, 2002), and these two elements may rely on the same time, those synapses that were less activated are weakened.
different neural structures, with only recollection depending on the This weakening occurs primarily during the transitions between intra-
hippocampus. Support for this account has come from findings that cellular up and down states experienced during SWS. This competitive
sleep only facilitates recognition memory based on recollection, not down-selection of weaker synapses restores memory encoding ability.
familiarity (e.g., Drosopoulos et al., 2005). On the other hand, studies The restorative function of sleep is supported by evidence
that have directly compared memory for emotionally negative and showing decreased episodic learning ability across a 6-hr retention
neutral stimuli appear to suggest that sleep benefits recognition memory interval in which participants remained awake, whereas encoding
for emotional stimuli but not for neutral stimuli (Alger et al., 2018; capacity was restored after a daytime nap (Mander et al., 2011).
Hu et al., 2006; Payne et al., 2008, 2015; see Kim & Payne, 2020 Further, neuroimaging evidence suggests sleep deprivation prior to
learning is associated with disrupted encoding-related functional between-subjects design, and random allocation to conditions),
activity in the bilateral posterior hippocampus (Yoo et al., 2007; for and questions on data analysis (e.g., preregistration and a priori
similar findings, see Alberca-Reina et al., 2015; Drummond & power analyses). We used similar questions to other meta-analyses
Brown, 2001; Van Der Werf et al., 2009). Thus, sleep deprivation in the sleep literature (Lim & Dinges, 2010; Lo, Groeger, et al.,
before learning may be detrimental specifically for the encoding of 2016; Schäfer et al., 2020) and examined methodological quality as
hippocampal-dependent declarative memories. Our second meta- a continuous moderator in the analysis. The full methodological
analysis seeks to estimate the effect size associated with this quality checklist is provided in Supplemental Appendix A.
impairment. As only two studies have looked at the impact of sleep It has been suggested that psychological science more broadly is
deprivation on procedural learning when it occurs after deprivation, currently suffering from a replication crisis due to low power,
we were not able to assess the potential moderating effect of publication bias, selection biases, and analysis errors (Nosek et al.,
declarative versus procedural memory. We were also not able to 2015). Low power limits potential to detect genuine effects but also
use emotionality as a moderator here due to the low number of results in Type I errors and exaggerated effect sizes (Ioannidis,
relevant studies. Yet, there are several studies that have used recall 2005; Pollard & Richardson, 1987). Szucs and Ioannidis (2017)
tasks and recognition memory tasks, so we were able to evaluate the conducted an analysis of almost 4,000 cognitive neuroscience and
moderating effects of recall versus recognition, as in the first meta- psychology papers and found that the overall mean power to detect
analysis. small, medium, and large effects was 17%, 49%, and 71%, respec-
tively, with even lower power in the subfield of cognitive neurosci-
ence. Given the convention that power to detect an effect size should
The Present Meta-Analyses
be at least 80% (Di Stefano, 2003), it is clear that a large number of
Despite the breadth of evidence for an effect of total sleep studies within psychology are underpowered (see Button et al.,
deprivation both before and after learning on memory performance, 2013; for further evidence of low statistical power within neurosci-
there is no comprehensive review and analysis of the strength of the ence). In the sleep literature, sample sizes tend to be low, potentially
effect of sleep deprivation on long-term memory. Previous reviews due to the resource intensity of conducting these experiments. Thus,
and meta-analyses investigating a role of sleep deprivation have we investigated whether the low power seen more broadly in
focused on tasks that are likely more susceptible to fatigue. Pilcher psychological science and neuroscience is also evident in the sleep
and Huffcutt (1996) conducted a meta-analytic review of the effect deprivation literature. For each individual effect size entered into the
of sleep deprivation on cognitive and motor task performance in 19 meta-analysis, we calculated the study’s power (defined as power to
primary studies and found that sleep deprivation had a significant detect our meta-analytic effect size) and investigated whether there
impact on performance. Still, this meta-analysis does not address is an association between a study’s power and the effect size
long-term memory performance. Similarly, Lim and Dinges (2010) observed in the study.
found an effect of sleep deprivation on a range of cognitive tasks
including attention, working memory, and short-term memory,
though the size of the effect varied depending on the task (e.g., a Method
nonsignificant effect on reasoning accuracy, but a large effect on
Search Strategy
attention). Finally, Harrison and Horne (2000b) in a review found
that sleep deprivation impacted decision-making ability. The tasks For study selection, we generated the Boolean search term “Sleep
studied in these reviews are often repetitive and monotonous (e.g., AND (deprivation OR restriction OR loss) AND (learning OR
the Psychomotor Vigilance Task; Dinges & Powell, 1985, the go/no memory OR conditioning)” and conducted a search in the electronic
go paradigm, and tests of serial addition), and they tend to probe databases EBSCOhost (included PsycARTICLES, PsycEXTRA,
lower-level functions that are particularly susceptible to fatigue, PsycINFO, and PsycTESTS) and PubMED on July 29th, 2020.
such as reaction times and processing speed. Therefore, the condi- This search yielded 2,213 empirical articles published between
tions of these studies are arguably better suited to finding adverse January 01, 1970 and July 29, 2020 in peer-reviewed journals in
effects of sleep deprivation on performance than studies looking at English using human participants.
higher-level learning and long-term memory. Thus, a review of the In line with best practice guidelines (Rothstein et al., 2005;
effects of sleep deprivation on such high-level, long-term memory is Siddaway et al., 2019), we ran several searches on July 13th,
required. 2020, using the same search terms as above, to identify gray literature
Taking a meta-analytic approach will permit not only a quantita- in an attempt to mitigate against publication bias. These searches
tive assessment of the size of the main effect of sleep deprivation and yielded a total of 553 items. Specifically, we widened our search
its moderators but also an investigation of methodological quality criteria in EBSCOhost and PubMED to include unpublished disserta-
within this literature including the statistical power of studies tions and theses, conference materials, and preprints; we searched the
proposing to find a sleep deprivation effect. Variety in sample bioRxiv and PsyArXiv repositories for preprints; and we searched the
selection and methodological designs used within this literature ProQuest and OpenGrey (a European database in which national
raises the possibility of variations in methodological quality. Such libraries submit unpublished studies) databases for unpublished dis-
variations could lead to biases in the meta-analysis by overestimat- sertations and theses, conference materials, and for research grants and
ing or underestimating the effect size (Higgins et al., 2011). Thus, fellowship awards. Additionally, we contacted all authors who had
we developed a checklist to assess multiple aspects of methodolog- published data included in our initial screening results to ask for
ical quality, including questions specifically relevant to the assess- unpublished data that fit our inclusion criteria, and this yielded one
ment of sleep effects (e.g., excluding participants with sleep preprint article. We also had one in-press article during the search
disorders), questions on study design (e.g., within-subject vs. period that fits our inclusion criteria (Tamminen et al., 2020) and was
therefore included in our search results. In sum, we identified both assessing the same manipulation, we chose only one outcome
published and unpublished data with search strategies spanning (a) measure for calculating an effect size in these instances, according
peer-reviewed published articles, (b) in-press articles, (c) preprints to the following hierarchy from most to least preferred outcome
uploaded to repositories, (d) unpublished dissertations and theses, (e) measure: accuracy as measured by retention performance (i.e.,
conference materials, and (f) research grants and fellowship awards. performance change from training to test), accuracy at test only,
We scanned the abstracts and full texts of all articles according to reaction time measured by retention performance (i.e., perfor-
our inclusion and exclusion criteria, and separated them into articles mance change from training to test), and reaction time at test
that investigated the effect of sleep deprivation after learning, and only. Further, in recognition tasks, if both signal detection analyses
those that investigated sleep deprivation before learning. Figure 1 and analyses based on proportion correct were reported, we chose
displays a screening process flowchart showing that after exclusions to include the signal detection measure only. For example, if a
were removed, 130 effect sizes (extracted from 45 reports) were study reported both d-prime and reaction times in a recognition
included in the sleep deprivation after learning meta-analysis and 55 memory task (e.g., Tamminen et al., 2020), we only used the
effect sizes (extracted from 31 reports) were included in the sleep d-prime data.
deprivation before learning meta-analysis. The number of effect A list of studies included in the two meta-analyses and their key
sizes included in each meta-analysis is greater than the number of properties can be found in Supplemental Appendix B (studies
full-text articles that fit our inclusion criteria. The reason for this is investigating sleep deprivation after learning) and Supplemental
that several studies measured performance differences between a Appendix C (studies investigating sleep deprivation before learn-
sleep deprivation and sleep control group using multiple tasks, ing), as well as at osf.io/5gjvs/.
multiple conditions (e.g., stimulus valence or procedural instruc-
tions), and across multiple time points. Effect sizes were calculated
for each of these data sets within an article because each variation Inclusion/Exclusion Criteria
represents a different, yet correlated, measurement of the impact of To select relevant studies, we applied the following inclusion/
sleep deprivation on memory. However, there were some studies exclusion criteria.
that used multiple outcome measures to assess performance in a
single task (e.g., accuracy and reaction time). Given that multiple a. Participants had to be healthy adults aged 18 years
outcome measures within the same task are different ways of and older.
Figure 1
Flowchart Displaying Literature Search Process for the Deprivation After Learning and Deprivation Before Learning Meta-Analyses
b. Studies must have included, as a primary independent Fischer et al., 2002 vs. complete control; Chatburn et al., 2017). For
variable, a manipulation of sleep deprivation that was a this reason, we assessed the methodological quality of each study
minimum of one night of total sleep deprivation with an entered into our meta-analyses and included this in our moderator
appropriate sleep control condition consisting of one analyses.
normal night of sleep. Residency studies (studies con- To assess methodological quality, we developed a 22-item check-
ducted in a medical setting) were excluded due to the lack list based on criteria for standard sleep deprivation experiments (e.g.,
of control over whether total sleep deprivation occurred preexperimental sleep monitored using actigraphy and exclusion of
(sleep deprivation is often reported despite naps having sleep disorders) and more general experimental psychology experi-
occurred on shift; e.g., Bartel et al., 2004; Guyette et al., ments (e.g., a priori power analysis and study design). For each item
2013). Additionally, studies using sleep restriction proto- on the checklist, studies were scored with either a zero or a one
cols, which involve multiple nights of limited sleep dura- according to whether they satisfied the criterion. To transform the
tion rather than one or more nights of no sleep, were total methodological quality score for each study into a risk of bias
excluded because the neural and cognitive effects of sleep that reflects a rank of all the studies on a common scale, we
restriction may differ from those caused by total sleep normalized the total scores by dividing each study’s total methodo-
deprivation (Banks & Dinges, 2007; Lowe et al., 2017). logical quality score by the maximum total methodological quality
c. Studies must have included, as a primary dependent score that was achieved among all studies (Stone et al., 2020). Lower
variable, at least one measure of learning or long-term values imply lower ranked studies (minimum score of 0) and higher
memory where the task was described in sufficient detail to values imply higher ranked studies (maximum score of 1) relative to
ascertain which cognitive function it assessed. the best study. The full methodological quality checklist can be found
in Supplemental Appendix A.
d. For the meta-analysis investigating sleep deprivation after Given that the checklist items form a multidimensional scale, the
learning, the cognitive task must have had a single en- items were clustered according to the Downs and Black’s (1998)
coding phase and a retrieval phase(s) that were temporally instrument for assessing methodological quality, which assesses
separated by either a period of sleep deprivation or an five types of bias: “reporting,” “internal validity—bias,” “internal
equivalent period of sleep. For the meta-analysis investi- validity—confounding,” “power,” and “external validity.” The
gating sleep deprivation before learning, the single encod- “reporting” cluster determines whether sufficient information is
ing phase and the retrieval phase(s) must have been provided to make an unbiased assessment of study findings. In
temporally separated by a retention interval that had a our methodological quality checklist, the items in this cluster
minimum duration of at least 1 min, rather than being part referred to the reporting of exclusion criteria for participant char-
of the same task session. The reason for this criterion is that acteristics (e.g., “Did the study exclude participants with a history of
our meta-analyses aimed to investigate effects of sleep sleep disorders?”). The “internal validity—bias” cluster assesses
deprivation on learning and long-term memory. The inclu- whether biases were present in the intervention or outcome measure
sion of studies with temporally indistinct encoding and that would favor one experimental group [e.g., “Was interference for
retrieval phases would have included short-term and the sleep deprivation group low (nondemanding activities given and
working memory tasks that form a separate body of monitored in the lab)?”]. The “internal validity—confounding”
literature (Lim & Dinges, 2010), the analysis of which cluster assesses whether biases were present in the selection and
was beyond the scope of these meta-analyses. allocation of participants (e.g., “For within-group studies, was the
e. In cases in which studies assessed the effects of other order of deprivation and control conditions counterbalanced?”). The
interventions (e.g., caffeine; Kilpeläinen et al., 2010) in “power” cluster assesses whether a study used a priori power
ameliorating sleep deprivation effects, studies were analyses to avoid Type II errors (e.g., “Did the study report an a
included only if data could be obtained from the control priori power analysis with power set at 80% or higher and an α at
sleep deprivation and control sleep groups. This criterion .05 or lower?”). The Downs and Black’s (1998) “external validity”
was included because the goal of the current meta-analyses cluster determines the extent to which findings can be generalized
was to assess effects of sleep deprivation in the absence of to the population from which a sample was taken (e.g., “Were the
alertness-promoting strategies. staff, places, and facilities where the patients were treated, repre-
sentative of the treatment the majority of patients receive?”;
f. Studies must have reported sufficient statistical detail to Downs & Black, 1998, p. 383). Since the items in this cluster
calculate effect sizes (means, SD, F, and t). When statisti- were designed for clinical intervention studies with nontypical
cal details were not reported in the text, we either contacted populations, we dropped the external validity cluster from our
corresponding authors to request relevant data or we checklist. In line with Cochrane Collaboration recommendations
extracted the data needed from published figures in the (Higgins et al., 2011), the four clusters in our methodological
article using WebPlotDigitizer (Rohatgi, 2019). quality checklist (hereon referred to as reporting, bias, confound-
ing, and power) were then included in moderator analyses. The
Methodological Quality
percentage of studies that passed on each item of the quality
Through our survey of the literature, it became clear that sleep checklist for both Meta-Analysis 1 and Meta-Analysis 2 can be
deprivation studies differ considerably in various aspects of meth- found in Supplemental Appendix D. Total methodological quality
odological rigor (e.g., lack of control over adherence to sleep scores for each study, as well as the item-level ratings, can be found
manipulations in the sleep deprivation and sleep control groups; at osf.io/5gjvs/.
Effect Size Calculation between-study variance. This allowed us to determine whether it

was necessary to account for both within- and between-study
Information on study means, standard deviation, and effect sizes
variances within our model.
for each item, as well as formulas used to calculate effect sizes, can
Assink and Wibbelink (2016) suggest that such log-likelihood
be found at osf.io/5gjvs/. We report the standardized mean differ-
ratio tests may be subject to the issues of statistical power when the
ence in task performance between a sleep deprivation and sleep
data set comprises a small number of effect sizes. Low statistical
control group, with positive values indicating that sleep deprivation
power may lead to nonsignificant effects of heterogeneity when in
influenced learning and memory such that performance was signifi-
fact there is variance within or between studies. To account for this,
cantly worsened compared to a sleep control group. For studies with
it is recommended to also calculate the I2 statistic, which indicates
independent samples (between-subjects designs), we computed
the percentage of variation across studies that is due to heterogeneity
Cohen’s ds based on the means and variance reported in each study
and that which is due to random sampling error (Higgins &
for the sleep and sleep deprivation group. For within-subject de-
Thompson, 2002). Hunter and Schmidt (2004) suggest the 75%
signs, in which participants took part in both the sleep deprivation
rule, such that if less than 75% of overall variance is attributed to
and sleep control conditions, we calculated Cohen’s dav, as recom-
sampling error, then moderating variables on the overall effect size
mended by Lakens (2013).
should still be examined. Using the formula of Cheung (2014), we
calculated the percentage of variance that can be attributed to each
Data Analysis level of our model.
However, although I2 reports the proportion of variation in
Overall Meta-Analytic Effect Size
observed effect sizes, it does not provide us with absolute values
All analysis code can be found at osf.io/5gjvs/. To analyze that tell us the variance in true effects (Borenstein et al., 2017). Thus,
whether there was an overall meta-analytic effect of sleep depriva- as recommended by Borenstein et al. (2011), we report the τ2, which
tion versus overnight sleep on memory performance, we fitted a provide an estimate of the true effect size, and we report prediction
multilevel random-effects model using the R package metafor intervals, which indicate that 95% of the time, effect sizes will fall
(Viechtbauer, 2010). A random-effects model allows for inconsis- within the range of those prediction intervals.
tencies between effect sizes from varying study designs, assuming
systematic variability between effect sizes in addition to random Publication Bias
sampling error. A random-effects model therefore provides more
conservative effect size estimates than a fixed-effect model To assess publication bias, we first examined a contour enhanced
(Borenstein et al., 2010). A multilevel model allows for the inclusion funnel plot. Funnel plots show each effect size plotted against its
of both within-study effect sizes and between-study effect sizes standard error, with contour lines corresponding to different levels
(Assink & Wibbelink, 2016). Many experiments included in the of statistical significance. If studies are missing almost exclusively
meta-analysis report multiple dependent effect sizes, such as results from the white area of nonsignificance, then there may be publica-
from multiple test sessions, multiple within-group experimental tion bias. If studies are missing from areas of statistical significance,
conditions (e.g., performance on emotional vs. neutral stimuli), the bias is likely due to other causes such as poor methodological
or multiple outcomes (e.g., a procedural and declarative memory quality, true heterogeneity, chance, or the bias may be artifactual
task). Including multiple dependent effect sizes from the same (Johnson, 2021; Sterne et al., 2011). We also conducted a variation
experiment violates the assumption of data independence assumed of Egger’s regression test for funnel plot asymmetry (Egger et al.,
in a typical random-effects model. A multilevel meta-analysis 1997) that can be conducted with multilevel meta-analyses.
accounts for such dependencies by modeling both within-study
and between-study effects. Thus, we were able to model variance Results
accounted for by (a) random error, (b) within-study differences Meta-Analysis 1: Sleep Deprivation After Learning
among effect sizes within the same experiment, and (c) between-
study differences across different experiments. This meta-analysis summarizes research from 45 reports investi-
gating effects of sleep deprivation after learning (130 effect sizes)
Heterogeneity published in English between 1994 and 2020 across a total of 1,616
participants. All reports used healthy adult populations and deprived
To investigate whether moderating variables may influence the participants of one night of sleep postlearning. Notably, two reports
size of the effect of sleep deprivation, we examined heterogeneity from this meta-analysis also report data that are relevant to the sleep
within the data set using the Q test (Cochran, 1954). The Q test deprivation before learning meta-analysis (Fischer et al., 2002;
indicates whether there is heterogeneity within the data set and is Tamminen et al., 2020). See Table 1 for central tendencies and
calculated by the weighted sum of the squared deviations of frequency data for moderator and descriptive variables of studies
individual study effect estimates and the overall effect across included. The table shows that this literature is heavily biased
studies. Significant heterogeneity suggests that some of the variance toward young adults, severely limiting conclusions that can be
within the data set may not be due to random sampling error, and drawn about older age groups. The literature predominantly uses
thus moderating variables may influence the effect. Since we were between-groups designs rather than the statistically more powerful
interested in both the within-study and between-study variance, we within-group designs, partly explaining and exacerbating the low
ran two separate one-sided log-likelihood-ratio tests. As such, the power highlighted later in our analysis. Recognition memory and
fit of the overall multilevel model was compared to the fit of a recall memory tasks are the most often employed measures of
model with only within-study variance and to a model with only memory, and most of the literature probes declarative memory
Table 1
Features of Included Sleep Deprivation After Learning Interventions (k = 130)
Study feature k M ± SD Mdn Mode Range
Publication date 130 NA 2012 2015 1994–2020

Sample size 130 31.84 ± 13.15 28 28 6–78
Age 54 22.15 ± 1.41 22.30 23.30 18.10–24.30
% Females 95 56.37 ± 19.04 56.25 50.00 0.00–85.71
Study design 130
Between-groups 104 NA NA NA NA
Within-group 26 NA NA NA NA
Paradigm 130
Motor skill 15 NA NA NA NA
Recognition 60 NA NA NA NA
Recall 42 NA NA NA NA
Route learning 4 NA NA NA NA
Temporal order 2 NA NA NA NA
Mere exposure effect 1 NA NA NA NA
Categorization 6 NA NA NA NA
Stimuli 130
Words 36 NA NA NA NA
Scenes 30 NA NA NA NA
Images 47 NA NA NA NA
Sequence 15 NA NA NA NA
Instruction 1 NA NA NA NA
Trajectory 1 NA NA NA NA
Emotionality of stimuli 130
Neutral 20 NA NA NA NA
Emotional 31 NA NA NA NA
Not reported 79 NA NA NA NA
Memory type 130
Procedural 21 NA NA NA NA
Declarative 109 NA NA NA NA
Sleep deprivation compliance check 130
Human observation of night 107 NA NA NA NA
Human observation of day and night 12 NA NA NA NA
No human observation 11 NA NA NA NA
Recovery sleep 130
Yes 85 NA NA NA NA
No 45 NA NA NA NA
Nights recovery sleep 74a 3.18 ± 3.02 2.00 2.00 1–13
Statistical power to detect meta-analytic effect sizeb 130 13.95% ± 5.00 13% 11% 7%–30%
Quality—reporting 130 0.56 ± 0.21 0.50 0.50 0.00–1.00
Quality—bias 130 0.74 ± 0.19 0.71 0.71 0.29–1.00
Quality—confounding 130 0.77 ± 0.15 0.83 0.83 0.33–1.00
a
Ten effect sizes had 6 months of recovery sleep and were excluded from the table as outliers. One study reported one to six nights of recovery sleep and was
excluded from the table due to lack of precision. b Statistical power to detect the meta-analytic effect size of Hedges’ g = 0.277, with α at .05.
rather than procedural memory. We return to these memory type See Supplemental Appendix B for a summary of all studies included
distinctions in our moderator analyses. Most studies used human in the meta-analysis.
observation to ensure participants in the sleep deprivation condition Some of the variance within the data set could not be explained by
did not sleep during the night, but few ensured that they did not sleep random error, highlighted by overall significant heterogeneity,
during the day to the same standard. Low compliance during the day Q(129) = 244.891, p < .001. An analysis of heterogeneity of
could possibly dilute the effect size. Finally, most but not all studies between-study variance (Level 2) revealed a significant difference
allowed recovery sleep after deprivation. We return to this in the between a full and a reduced model ( p < .001), suggesting signifi-
moderator analyses. cant variability between studies. An analysis of heterogeneity of
within-study variance (Level 3) also revealed a significant difference
between a full model and a reduced model ( p < .001), suggesting
Overall Meta-Analytic Effect Size
significant variability between within-study effect sizes. We further
The overall effect size for the mean difference in memory calculated the I2 statistic, which indicates the percentage of variance
performance between the sleep deprivation and sleep control that could be attributed to each level of the model. Using the formula
group, measured by Hedges’ g, was 0.277 (SE = 0.050), indicating from Cheung (2014), approximately 52% of variance can be
a small-to-medium effect according to Cohen’s categorization, attributed to sampling error, 14% to within-study variance, and
and a significant difference from zero, 95% CI [0.177, 0.377], 34% to between-study variance. Next, we calculated τ2, which
p < .001. Figure 2 displays a forest plot of the effect sizes. provides a measure of the variance of the true effects; τ2 = .026
Figure 2
Forest Plot Containing Effect Sizes and 95% Confidence Intervals for the Difference
in Performance Between a Sleep Deprivation and Sleep Control Group on Memory
Note. Effect sizes to the right indicate an effect of sleep deprivation after learning such that
memory was significantly worse than in a sleep control group.
for within-study variance, and τ2 = .061 for between-study variance. uses a multilevel approach, with dependencies between some effect
Prediction intervals indicated that 95% of effect sizes would fall in sizes. The adjusted effect size should therefore be considered a
the range of −0.316 and 0.870. preliminary estimate. The trim-and-fill method estimated 12 missing
Based on the significant heterogeneity between studies, the large studies from the left-hand side of the funnel plot. With these effect
prediction intervals, as well as the 75% rule, such that moderators sizes included, the adjusted overall meta-analytic effect size was
should be explored if less than 75% of the variance can be attributed smaller than the original effect size of g = 0.277, although it was still
to random sampling error (Hunter & Schmidt, 2004), we therefore significantly greater than zero, Hedges’ g = 0.166, SE = 0.041, 95%
explored the effect of potential moderating variables on the direction CI [0.086, 0.247], p < .001.
of the effect.
Outlier Analysis
Publication Bias We explored whether outliers and influential cases may have
significantly influenced the meta-analytic effect size. To identify the
Figure 3 shows a funnel plot of the effect sizes. Visual inspection
presence of outliers, we identified any effect sizes with studentized
of the funnel plot indicates that effect sizes are not evenly distributed
residuals greater than or smaller than three, which identified one
across the funnel plot, raising the potential for publication bias in
effect size as an outlier (Albouy, Sterpenich, et al., 2013). However,
which studies reporting a positive effect are more likely to be
an outlier may not necessarily influence the size of the overall effect
published. Egger’s regression test reveals significant funnel plot
(Viechtbauer & Cheung, 2010). Therefore, based on suggestions by
asymmetry (z = 2.297, p = .022), supporting this assessment of the
Viechtbauer and Cheung, we conducted influential case analyses, to
funnel plot. Further visual inspection of the funnel plot reveals two
identify any effect sizes that exerted a significant influence on the
potential outlier effect sizes in the area of high statistical signifi-
size of the overall meta-analytic effect. We measured Cook’s
cance; these potential outliers are characterized by large effect sizes
distance to examine the influence of deleting each study on the
and large standard error (therefore smaller sample sizes). These large
overall size of the effect, and DFBETAs to examine the effect of
effect sizes on the right-hand side of the plot suggest that there may
deleting each study on individual parameter estimates. Cook’s
be a bias in this literature toward publishing significant effects,
analysis identified a further one effect size that was found to be
regardless of the precision with which the study effect size can be
an influential case (Darsaud et al., 2011, Know judgements).
estimated. However, the presence of multiple studies in the area of
Removal of the outlier and influential case reduced the overall
nonsignificance suggests other biases may also contribute to the
meta-analytic effect size to 0.271 (from 0.277). Since moderator
asymmetry.
analyses examine smaller subsets of effect sizes, we removed these
Because Egger’s test indicates the presence of publication bias,
two specific effect sizes from all moderator analyses conducted.
we sought to quantify the impact of this bias on the estimated effect
size. We conducted a trim-and-fill analysis (Duval & Tweedie,
2000), which calculates potential missing effect sizes to create a Moderator Analysis
symmetric funnel plot and then provides an adjusted overall meta-
We introduced four categorical moderating variables and analyzed
analytic effect size based on this funnel plot symmetry. Although
the effect of each moderator separately on the size of the effect of
this is a well-used method to assess publication bias, it assumes that
sleep deprivation on learning and memory: (a) whether it was a
effect sizes are independent of each other. The current meta-analysis
declarative (k = 108) or procedural (k = 20) memory task, (b) for
declarative tasks, whether task type was recall (k = 42) or recognition
Figure 3 (k = 59), (c) whether participants received one or more recovery
Contour Enhanced Funnel Plot Showing the Hedges’ g Effect Size nights of sleep (k = 83) or no recovery sleep (k = 45), and (d) for those
on the x-Axis, and the Standard Error of Hedges’ g Effect Size on the studies that investigated emotionality, whether the stimuli were
y-Axis emotional (k = 31) or neutral (k = 20). Supplemental Appendix B
shows the classification of each study on these dimensions.
Whether participants received a recovery night of sleep before the
test session as a moderator had a significant effect, Q(1) = 10.496,
p < .001. Thus, we ran separate effect size analyses for those studies
where participants had a night of recovery sleep and those that did
not. For those studies that had one or more nights of recovery sleep,
there was a small effect of sleep deprivation on learning and
memory, Hedges’ g = 0.176 (SE = 0.058), which is significantly
different from zero, 95% CI [0.060, 0.292], p = .003; Q(82) =
133.766, p < .001. For those studies that did not have a night of
recovery sleep, the effect size was larger, Hedges’ g = 0.410, SE =
0.044, 95% CI [0.320, 0.499], p < .001; Q(44) = 50.042, p = .246.
Whether the task probed declarative or procedural memory also
had a significant moderating effect, Q(1) = 5.301, p = .021. We
therefore ran separate effect size analyses for those studies that
implemented a declarative memory task and those that implemented
a procedural memory task. For those studies with a declarative
memory task, there was a small effect of sleep deprivation on p = .005; Q(19) = 25.789, p = .136, for studies using neutral stimuli,
learning and memory, Hedges’ g = 0.218 (SE = 0.055), which is and an overall effect size of Hedges’ g = 0.207, SE = 0.085, 95% CI
significantly different from zero, 95% CI [0.109, 0.327], p < .001; [0.033, 0.380], p = .021; Q(30) = 41.888, p = .073, for emotional
Q(107) = 174.802, p < .001. For those studies with a procedural stimuli.
memory task, the effect size was larger, Hedges’ g = 0.449, SE = We then introduced four continuous moderating variables and
0.083, 95% CI [0.276, 0.623], p < .001; Q(19) = 24.794, p = .167. analyzed the effect of each moderator separately on the size of the
Since we found a significant moderating effect of both recovery effect of sleep deprivation on learning and memory: (a) methodo-
sleep and task type (declarative or procedural), we ran a further logical quality reporting cluster, (b) methodological quality bias
analysis to investigate whether there was an interaction between the cluster, (c) methodological quality confounding cluster, and (d)
two significant moderators. The analysis revealed no significant statistical power to find the mean effect size established in the meta-
interaction between recovery sleep and memory type, Q(1) = 0.804, analysis (g = 0.277). Since methodological quality was divided into
p = .370, suggesting that whether participants received recovery clusters in our methodological quality checklist, we introduced these
sleep or not affected declarative and procedural memory task clusters (reporting, bias, and confounding) as moderators. We did
performance in a similar way. Whether studies used a recall or not include the power cluster as a moderator, since the majority of
recognition task did not have a significant effect on the size of the studies did not calculate power and thus scored zero on this cluster,
effect of sleep deprivation on learning and memory. Studies using a with only one study (contributing eight effect sizes) providing a
recall task had a mean effect size of Hedges’ g = 0.209, SE = 0.082, power analysis. For each of the methodological quality clusters, we
95% CI [0.045, 0.374], p = .014; Q(41) = 76.317, p < .001, and created a meta-analytic scatter plot using the metafor package
studies with a recognition task had an overall effect size of Hedges’ (Viechtbauer, 2010), showing the Hedges’ g of each individual
g = 0.175, SE = 0.077, 95% CI [0.021, 0.330], p = .027; Q(58) = study plotted against each moderator (see Figure 4). The figure
91.950, p = .003. Effect sizes associated with emotional and neutral shows that an effect size of zero falls within the 95% confidence
stimuli were also not significantly different, with an overall effect interval in the reporting cluster for studies scoring 0.8 or higher,
size of Hedges’ g = 0.251, SE = 0.080, 95% CI [0.084, 0.417], suggesting that these studies show no effect of sleep deprivation
Figure 4
Meta-Analytic Scatter Plot With Methodological Quality of the Reporting, Bias, and Confounding
Clusters Plotted Against Individual Study Effect Size for Meta-Analysis 1
Note. The size of each point is proportional to the weight the study received in the analysis, with larger size
indicating larger weight. The solid regression lines represent the effect size predicted by the meta-regression
model as a function of each cluster score, with corresponding 95% confidence intervals.
while the lower scoring studies do. Consistent with this observation, Figure 5
the reporting cluster was a significant moderator, Q(1) = 9.214, p = Distribution of Power to Find the Mean Meta-Analytic Effect Size
.002. Visual inspection of the plot for the confounding cluster (Green), the Lower Bound of the 95% Confidence Interval Around
suggests that studies with scores of 0.3 or lower on this cluster the Mean (Pink), and the Upper Bound of the 95% Confidence
may not show an effect of sleep deprivation. However, this cluster Interval Around the Mean (Blue)
did not show a statistically significant moderating effect, perhaps
because there were no studies that scored below 0.3 on this cluster.
Bias did not show a significant moderating effect either.
To assess the achieved statistical power of each individual
experiment to detect the mean meta-analytic effect size, we con-
ducted a post hoc power analysis in G*Power (Faul et al., 2007). For
each study, we calculated the power to detect the mean meta-
analytic effect size (g = 0.277), as well as the power to detect
the 95% upper and lower confidence intervals of the effect size. The
distribution of the mean power and lower and upper confidence
interval bounds of the power estimate are plotted in Figure 5.
We found the mean power to find the average meta-analytic effect
to be 13.98% (SD = 4.60%, range = 7.2%–30.2%). The moderator
analysis revealed no significant effect of power on the size of the
meta-analytic effect. All moderator analyses are reported in Table 2.
Meta-Analysis 2: Sleep Deprivation Before Learning

This meta-analysis summarizes research from 31 reports investi-
gating effects of sleep deprivation before learning (55 effect sizes)
published in English between 1989 and 2020 across a total of 927
Note. See the online article for the color version of this figure.
participants. All reports used healthy adult populations and deprived
participants of one night of sleep prior to learning. Notably, two
Overall Sleep Deprivation Effect Size
reports from this meta-analysis also include data that are relevant to
the sleep deprivation after learning meta-analysis (Fischer et al., The overall effect size for the mean difference in memory perfor-
2002; Tamminen et al., 2020). See Table 3 for central tendencies and mance between the sleep deprivation and sleep control group, mea-
frequency data for moderator and descriptive variables of reports sured by Hedges’ g, was 0.621 (SE = 0.074), indicating medium to
included. Again, the literature mostly involves young adults, leaving large effect according to Cohen’s categorization, and a significant
a gap in our understanding of how the effects of interest change with difference from zero, 95% CI [0.473, 0.769], p < .001. Figure 6
age. The discrepancy between use of between- and within-group provides a forest plot of the effect sizes. See Supplemental Appendix C
designs is lower here than in the first meta-analysis, and recognition for a summary of all studies included in the meta-analysis.
memory and recall memory tasks are roughly equally represented. Some of the variance within the data set could not be explained by
However, nearly all studies look at declarative memory, suggesting random error, highlighted by overall significant heterogeneity,
that more work on procedural memory is needed. While most Q(54) = 118.166, p < .001. An analysis of heterogeneity of
studies ensured compliance with the sleep deprivation manipulation between-study variance (Level 2) also revealed a significant differ-
with direct observation at night, few did so during the preceding day, ence between a full and a reduced model ( p < .001), suggesting
potentially diluting the impact of sleep deprivation. significant variability between studies. An analysis of heterogeneity
Table 2
Effect of Each Moderator on the Overall Meta-Analytic Effect of Sleep Deprivation After Learning on Memory Performance
Moderator Variable type df Heterogeneity (Q) p
Recovery sleep (yes vs. no) Categorical 1 10.496 <.001*

Task type (declarative vs. procedural) Categorical 1 5.301 .021*
Recall versus recognition Categorical 1 0.115 .734
Emotionality (emotional vs. neutral) Categorical 1 0.169 .681
Quality—reporting cluster Continuous 1 9.337 .002*
Quality—bias cluster Continuous 1 0.989 .320
Quality—confounding cluster Continuous 1 0.049 .825
Power Continuous 1 0.611 .434
Recovery sleep × Task type Categorical 1 0.804 .370
* p < .05.
Table 3
Features of Included Sleep Deprivation Before Learning Interventions (k = 55)
Study feature k M Mdn Mode Range
Publication date 55 NA 2010 2000/2020 1989–2020

Sample size 55 30.75 ± 13.82 26 26 12–58
Age 27 24.55 ± 5.89 22.19 20.70 19.50–47.83
% Female 42 47.78 ± 22.16 50.00 50.00 00.00–74.36
Study design 55
Between-groups 34 NA NA NA NA
Within-group 21 NA NA NA NA
Paradigm 55
Recognition 23 NA NA NA NA
Cued recall 14 NA NA NA NA
Free recall 12 NA NA NA NA
Texture discrimination 1 NA NA NA NA
Recency discrimination 2 NA NA NA NA
Finger tapping 1 NA NA NA NA
Categorization 2 NA NA NA NA
Stimuli 55
Words 24 NA NA NA NA
Images 23 NA NA NA NA
Prose 3 NA NA NA NA
Numbers 1 NA NA NA NA
Instruction 4 NA NA NA NA
Memory type 55
Procedural 2 NA NA NA NA
Declarative 53 NA NA NA NA
Sleep deprivation compliance 55
Human observation of night 42 NA NA NA NA
Human observation of day and night 10 NA NA NA NA
No human observation 3 NA NA NA NA
Recovery sleep 55
Yes 13 NA NA NA NA
No 42 NA NA NA NA
Nights recovery sleep 12a 6.04 2.00 2.00 2–13
Statistical power to detect meta-analytic effect sizeb 55 54.77 ± 20.79 54 86 21–98
Quality—reporting 55 0.61 ± 0.28 0.50 0.50 0.00–1.00
Quality—bias 55 0.71 ± 0.23 0.71 0.86 0.14–1.00
Quality—confounding 55 0.81 ± 0.16 0.83 0.83 0.50–1.00
a b
One study using recovery sleep only had a 90-min nap opportunity as recovery sleep and was excluded from the table as an outlier. Statistical power to
detect the meta-analytic effect size of Hedges’ g = 0.621, with α at .05.
of within-study variance (Level 3) revealed a significant difference the right side of the funnel plot. While many of these studies fall in
between a full model and a reduced model ( p < .001), suggesting the area of nonsignificance, there appear to be studies missing from
significant variability between within-study effect sizes. The I2 the left-hand side of the plot. It is possible these missing studies are
statistic indicates that approximately 41% of variance can be due to researchers being unable to publish findings that contradict
attributed to sampling error, 9% to within-study variance, and their hypotheses, and that the bias indicated by Egger’s test may
50% to between-study variance. τ2 = .096 for between-study therefore be due to publication bias rather than other types of bias.
variance, and τ2 = .017 for within-study variance, and prediction Because Egger’s test indicates the presence of publication bias, we
intervals indicated that 95% of effect sizes would fall in the range of conducted a trim-and-fill analysis (Duval & Tweedie, 2000), in the
−0.069 and 1.312. Based on this evidence for heterogeneity, we same way as in the first meta-analysis. The trim-and-fill method
explored the effect of potential moderating variables on the direction estimated seven missing studies from the left-hand side of the funnel
of the effect. plot. With these effect sizes included, the adjusted overall meta-
analytic effect size was smaller than the original effect size of g =
0.621, although it was still significantly greater than zero, Hedges’ g =
Publication Bias 0.463, SE = 0.070, 95% CI [0.326, 0.601], p < .001.
Figure 7 shows a contour enhanced funnel plot of effect sizes.
Similar to Meta-Analysis 1, visual inspection of the funnel plot Outlier Analysis
indicates that effect sizes are not evenly distributed across the funnel
plot, raising the potential for publication bias. Egger’s regression In the same way as the first meta-analysis, we explored whether any
test supports this conclusion, indicating significant funnel plot outliers and influential cases significantly influenced the size of the
asymmetry (z = 3.363, p < .001). Further inspection of the funnel effect. We identified any effect sizes with studentized residuals greater
plot indicates that the majority of effect sizes are clustered toward than or smaller than three (k = 1; Tempesta et al., 2016, Recognition
Figure 6
Forest Plot Containing Effect Sizes and 95% Confidence Intervals for the Difference in Performance Between a Sleep
Deprivation and Sleep Control Group on Memory
Note. Effect sizes to the right indicate an effect of sleep deprivation before learning such that memory was significantly worsened compared to a
sleep control group.
Task). Following recommendations by Viechtbauer and Cheung Removal of the one outlier and two influential cases reduced the
(2010), influential case analysis (Cook’s distance and DFBETAs) overall meta-analytic effect size to 0.525 (from 0.621). We removed
identified two further effect sizes that may have had a significant these three specific outliers and influential cases from all moderator
influence on the results (Poh & Chee, 2017; Yoo et al., 2007). analyses.
Figure 7 meta-analysis (M = 54.77%, SD = 20.81%, range = 21.22%–

Contour Enhanced Funnel Plot Showing the Hedges’ g Effect Size 98.06%). The moderator analysis revealed that power to find the
on x-Axis, and Standard Error of Hedges’ g Effect Size on the y-Axis mean meta-analytic effect size did not moderate the effect of sleep
deprivation on memory, Q(1) = 3.179, p = .075. All moderator
analyses are reported in Table 4.
Discussion
The two meta-analyses presented here aimed to quantify the size
of the effect of sleep deprivation after learning and before learning
on memory performance. Based on previous evidence for an effect
of sleep on both declarative and procedural memories (Klinzing
et al., 2019), we predicted that sleep deprivation would have a
detrimental effect on learning and memory. We found that sleep
deprivation after learning was associated with a mean effect size of
g = 0.277. The effect size is positive, indicating that sleep depriva-
tion has a detrimental rather than facilitatory impact on memory, as
predicted by theory. Furthermore, the 95% confidence intervals
around the mean do not cross zero, indicating that the effect size is
statistically significantly higher than zero. For sleep deprivation
before learning, we found an average effect size of g = 0.621. The
effect size is positive, indicating that sleep deprivation before
Moderator Analysis learning impairs rather than facilitates memory, as predicted by
theory (e.g., Cirelli & Tononi, in press; Tononi & Cirelli, 2012). The
We introduced the categorical moderating variable task type, 95% confidence intervals around the mean do not cross zero, again
recall (k = 26) versus recognition (k = 20). We excluded studies that indicating that the effect size is statistically significantly higher than
used a different task type, including a recency discrimination task zero. Following Cohen’s guidelines for categorizing effect sizes a
(k = 3), a prototype learning task (k = 2), and a finger-tapping task small (0.20), medium (0.50), and large (0.8), the effect sizes above
(k = 1). Analysis of the moderator recall versus recognition revealed would correspond to small-to-medium and medium. However,
that the type of task used did not moderate the size of the effect, given that these cutoff points are wholly arbitrary and were only
Q(1) = 0.028, p = .868. There were only two studies of procedural ever intended to be used as a last resort, many now argue that effect
memory (Fischer et al., 2002; McWhirter et al., 2015) and only five sizes should be interpreted in the context of typical effect sizes
entries where the participants were given a night of recovery sleep observed in the relevant literature (Correll et al., 2020; Funder &
(two further entries did not report whether recovery sleep was Ozer, 2019). According to Brysbaert (2019), an effect size of d =
given). Likewise, only two studies investigated the effects of sleep 0.40 represents an average effect size in experimental psychology
deprivation on emotional memory (Kaida et al., 2015; Tempesta and has practical and theoretical relevance. Putting our meta-
et al., 2016). Thus, there was insufficient variability within the data analytic effect sizes into this context, it appears that sleep depriva-
set to assess whether one or more nights of recovery sleep, the type tion before learning has an effect size somewhat larger than the
of memory (declarative, procedural), and emotionality moderated average effect size in experimental psychology, while sleep depri-
the size of the sleep deprivation effect. vation after learning has a somewhat smaller than average effect
We tested the influence of the three continuous moderating size, although the latter varies as a function of both recovery sleep
variables assessing methodological quality (reporting, bias, and and memory type (declarative vs. procedural), as discussed in
confounding), as well as statistical power to detect the meta-analytic detail below.
effect size, on the size of the sleep deprivation effect. No methodo-
logical quality cluster had a significant moderating effect. For each
Theory-Based Mediators
of the methodological quality clusters, we created a meta-analytic
scatter plot (metafor package; Viechtbauer, 2010; see Figure 8). We Despite the wide range of literature examining an effect of sleep
did not include the power cluster as a moderator, since the majority deprivation on memory performance, this is, to our knowledge, the
of studies did not calculate power and thus scored zero on this first time that the size of this effect has been formally quantified. A
cluster, with only two studies (contributing a total of 10 effect sizes) benefit of meta-analyses is that they allow for the investigation of
providing a power analysis. potential moderating factors that may differentially influence the
We then investigated whether statistical power to find the meta- size of the meta-analytic effect. For deprivation after learning, we
analytic effect moderated the size of the effect of sleep deprivation were able to investigate whether the first night of sleep is essential
on memory. In the same way as in the first meta-analysis, for each for the consolidation of newly acquired memories or whether a later
study, we calculated the power to find the mean meta-analytic effect sleep opportunity can compensate for the first night of sleep
size, as well as the power to detect the upper and lower confidence deprivation. We found that studies where memory was tested
interval bounds around the mean. The distribution of power to immediately after one night of sleep deprivation and before recovery
detect the three estimates is plotted in Figure 9. We found the mean sleep showed a significant sleep deprivation associated memory
power to find the meta-analytic effect to be larger than in the first deficit (g = 0.410). Critically, those studies that had one or more
Figure 8
Meta-Analytic Scatter Plot With Methodological Quality Clusters of Reporting, Bias, and
Confounding, Plotted Against Individual Study Effect Size for Meta-Analysis 2
Note. The size of each point is proportional to the weight the study received in the analysis, with larger size
indicating larger weight. The solid regression lines represent the effect size predicted by the meta-regression
model as a function of each cluster score, with corresponding 95% confidence intervals.
nights of recovery sleep prior to retrieval also showed a small but previously encoded and consolidated word pairs. Furthermore,
statistically significant memory deficit (g = 0.176). Thus, memory despite a difference in the size of the effect of sleep deprivation,
impairments caused by sleep deprivation during the first postencod- it is important to note that we still see a significant detrimental effect
ing night were still present but less severe when recovery sleep of sleep deprivation after learning even when a later sleep opportu-
occurred before testing. On the one hand, this finding suggests that nity is permitted, albeit a smaller effect. An account based on fatigue
the first night of sleep after learning is important as its disruption is alone is insufficient to explain this finding. Another potential
still felt even after recovery sleep. On the other hand, it also suggests alternative explanation for the decrease in effect size after recovery
that recovery sleep can to some extent mitigate the disruption of sleep could be based on interference. When tested after one night of
the first night of sleep by reducing the effect size by about 50%. sleep or sleep deprivation, participants in the sleep group will have
While these data are consistent with theories arguing that, for experienced little interference from subsequent cognitive activity
hippocampal-dependent memories at least, the hippocampus may after learning. A large difference between the groups at this point
act as a buffer, retaining newly learned information until an offline could be due to sleep protecting new memories from interference
consolidation opportunity is available (Schönauer et al., 2015), the rather than due to active consolidation processes. After one or more
idea that consolidation processes can be spread over multiple nights nights of recovery sleep, both groups will have experienced some
of sleep is yet to be explicitly tested. degree of interference, and this could explain the reduction in the
It is also difficult to establish the extent to which the smaller effect effect size. Further research is needed to adjudicate between these
of sleep deprivation after recovery sleep on memory is due to the different accounts that could both contribute to the effect sizes we
occurrence of a delayed consolidation opportunity or due to effects have observed.
of fatigue being diminished. Recent work has suggested that fatigue For deprivation after learning, we also found that whether the task
at time of test might have little or no detrimental impact on tasks type was declarative or procedural had an effect on the size of the
assessing long-term memory. Schönauer et al. (2015) found that deprivation effect. Although both declarative and procedural tasks
sleep deprivation before a recall task did not impair memory for elicited a significant effect, a moderator analysis indicated that those
Figure 9 consolidation theory, hippocampal-dependent declarative memories

Distribution of Power to Find the Mean Meta-Analytic Effect Size benefit from repeated reactivation of newly learned memories
(Green), the Lower Bound of the 95% Confidence Interval Around during sleep, supporting the strengthening of memory representa-
the Mean (Pink), and the Upper Bound of the 95% Confidence tions in the neocortex (Born & Wilhelm, 2012; Walker & Stickgold,
Interval Around the Mean (Blue) 2006). However, procedural memories that rely on implicit learning
are unlikely to be dependent on such hippocampal–neocortical
representations. It has been theorized that such implicit memories
require more immediate offline consolidation to see a beneficial
effect of sleep (Schönauer et al., 2015; Stickgold et al., 2000).
Thus, it may be that without an immediate sleep opportunity, the
detrimental effects of sleep deprivation have a larger impact on
procedural memory consolidation, whereas declarative memory
consolidation is less impacted by the lack of an immediate sleep
opportunity. However, we found no significant interaction between
recovery sleep and task type, suggesting that for both declarative and
procedural tasks, lack of an immediate sleep opportunity increased
the size of the effect of sleep deprivation. Thus, recovery sleep had a
similar impact on both procedural and declarative task performance,
and procedural tasks elicited larger effect sizes than declarative
tasks, regardless of whether recovery sleep occurred.
For deprivation after learning, we found no effect of emotional
versus neutral memory on the size of the meta-analytic effect.
Although some studies do suggest a preferential effect of sleep for
emotional memories (e.g., Payne & Kensinger, 2010; Wagner et al.,
2001), our findings join two recent meta-analyses that report no
overall preferential effect of sleep on emotional memory consolida-
tion (Lipinska et al., 2019; Schäfer et al., 2020). These existing meta-
analyses focussed on emotional memory and were able to uncover
Note. See the online article for the color version of this figure.
mediators that may reveal boundary conditions for the preferential
effect; however, the number of studies in this domain is still low and
more research is needed to establish the reliability of the effect.
studies implementing procedural memory tasks had significantly For both deprivation after learning and deprivation before learn-
larger effect sizes on average (g = .449) than declarative tasks (g = ing, we found no effect of the recall versus recognition moderator on
.218). That both declarative and procedural memory tasks showed the size of the meta-analytic effect. This is in contrast to some
detrimental effects of sleep deprivation was unsurprising, given that previous studies investigating the beneficial role of sleep on memory
a benefit of sleep has been observed for both declarative memories that have found a differential effect of recall versus recognition
(e.g., Gais & Born, 2004; Talamini et al., 2008; Wagner et al., 2006) testing, and in contrast to the meta-analysis of Newbury and
and procedural memories (Korman et al., 2007; Schönauer et al., Monaghan (2019), which looked at sleep studies using the DRM
2014; Walker et al., 2005). Our findings are also consistent with paradigm. Although performance on recall tasks repeatedly benefits
recent studies showing that sleep is beneficial even in tasks that do from sleep, performance on recognition tasks has sometimes been
not require the hippocampus at learning (e.g., Schapiro et al., 2019). found to show little or no offline consolidation benefit (Ashton et al.,
Although the current meta-analysis indicates a detrimental effect of 2018; Diekelmann et al., 2009; Drosopoulos et al., 2005; Gais et al.,
sleep deprivation after learning on both declarative and procedural 2006; Hu et al., 2006). It is posited that, although recall tasks rely on
memories, the exact mechanisms that drive these effects are still explicit, hippocampal-dependent memory, recognition tasks could
debated, and thus it is unclear why procedural memories may show include both an explicit recollection and implicit familiarity element
larger sleep deprivation effects. According to active systems (Jacoby, 1991), only the former of which benefits from sleep-
associated consolidation. Thus, the mechanisms by which these
Table 4 two types of memories are consolidated may be different. Despite
Effect of Each Moderator on the Overall Meta-Analytic Effect of this, the present meta-analyses provide no evidence to suggest that
Sleep Deprivation Before Learning on Memory Performance performance on recall and recognition tasks are differentially
affected by sleep deprivation either before or after sleep. Whether
Variable Heterogeneity this finding extends to sleep paradigms other than total sleep
Moderator type df (Q) p deprivation remains to be established.
Recall versus recognition Categorical 1 0.261 .610 The null effects from our moderator analyses should be treated
Quality—reporting cluster Continuous 1 0.868 .351 with caution, however, as we may not have adequate statistical
Quality—bias cluster Continuous 1 0.064 .800 power to detect smaller moderator effect sizes. Hempel et al. (2013)
Quality—confounding Continuous 1 0.867 .352 suggest that power to detect moderator effects is dependent on a
cluster
3.110 .078
combination of the amount of residual heterogeneity within the data
Power Continuous 1
set, the number of studies in the data set, the number of participants
in the included studies, and the ratio of studies in the conditions We must be cautious in the way that we interpret the effects of
compared against each other. Based on power simulations, Hempel methodological quality on the size of the effect of sleep deprivation.
et al. (2013) provide estimations of the approximate number of Valentine (2009) argues that methodological scales of this nature
studies and participants required to detect categorical moderator frequently lack operational specificity (e.g., that each item deserves
effects of different effect sizes. We used data from these simulations equal weight) and include questions that are unclear. In an attempt to
to retrospectively assess the power of our moderator analyses to increase the validity of our methodological quality checklist, we
detect an effect. Since Hempel et al.’s simulations are not based on designed our checklist based on the Downs and Black checklist
multilevel meta-analyses, the below estimates should be treated with (1998), with modified questions relevant to sleep studies. For the
caution when applied to our analyses and only considered as first meta-analysis, we found a significant mediating effect of the
indicative. reporting cluster of our methodological quality checklist on the size
For our deprivation after learning analysis, residual heterogeneity of the sleep deprivation effect, Figure 4 shows that studies scoring
was τ2 = .026 for within-study variance, and τ2 = .061 for between- highest on this cluster show no effect of sleep deprivation while the
study variance. Therefore, based on a τ2 of between 0 and 0.1, the lower scoring studies do. The items in the reporting cluster are
simulations suggest that the moderator recall versus recognition was predominantly concerned with the number and nature of exclusion
powered to detect an effect of around 0.2–0.3 (based on 100 trials, and inclusion criteria used in the study. It therefore appears that the
20 participants per study, at 80% power). The emotionality moder- studies showing higher effect sizes may have employed fewer such
ator was powered to detect only large moderator effects of 0.3–0.4 criteria. However, some of the scores on this cluster may be
(based on 50 trials, 20 participants per study, 80% power). For the underestimated due to incomplete reporting. For example, studies
deprivation before learning analysis, residual heterogeneity was stating that they only recruited healthy participants may have used
τ2 = .096 for between-study variance, and τ2 = .017 for within- other sleep-related inclusion and exclusion criteria, such as exclud-
study variance. The simulations suggest that the moderator analysis ing participants who were taking medication that affects sleep or
of recall versus recognition was powered to detect only a large effect those who had recently traveled between time zones, without
size of between 0.3 and 0.4 (based on 50 trials, 20 participants per reporting these and may have scored higher on this cluster had
study, at 80% power). Therefore, it appears that our moderator these criteria been reported. We found no mediating effect of any
analyses were not sufficiently powered to detect small moderator cluster of methodological quality on the size of the effect of sleep
effects and the null findings in these analyses should be considered deprivation for the second meta-analysis. Given that only one cluster
preliminary. These analyses need to be repeated as more evidence of the quality score influenced the size of the effect in the first meta-
accumulates over time. analysis, and no clusters had a significant effect in the second meta-
There are other moderators that would be valuable to account for analysis, our effect size estimates are unlikely to be substantially
to increase the precision of our meta-analytic effect size, but that we biased by variation in methodological quality.
could not include in our analyses due to the small number of studies Taking a broader qualitative view of our quality checklist, we note
available. For example, some studies included in the meta-analysis that only one of the analyzed studies was preregistered, and only
involved manipulations that the authors expected to reverse or three justified their sample size with an a priori power analysis.
eradicate the detrimental effect of sleep deprivation. Kolibius et al. Given that preregistration has become a mainstream practice only in
(2021) predicted that large amounts of encoded information (640 the past few years (Nosek & Lindsay, 2018), and that an a priori
word pairs) would increase forgetting in the sleep group compared power analysis is part of the preregistration process, the low
to a sleep-deprived group. Similarly, Feld et al. (2016) hypothesized numbers here are unsurprising and are likely in line with the current
broader field of experimental psychology. The key quality measures
that those in a high memory load condition (360 word pairs) should
on study design were met by the clear majority of studies (e.g., equal
no longer show a sleep benefit compared to a sleep-deprived
group sizes, random allocation to groups or counterbalancing of
condition. Vargas et al. (2019) examined memory for emotionally
conditions).
negative and neutral objects and backgrounds, but only predicted an
impact of sleep deprivation on neutral objects. It is possible that the
inclusion of studies such as these (or conditions within those studies Power-Based Moderators
where no sleep deprivation effect is predicted) may have artificially
In the current meta-analyses, we calculated statistical power to
reduced our meta-analytic effect size. A mediator analysis would be
find the meta-analytic effect size for each experiment and assessed
the appropriate solution to establish whether this was the case, but
whether statistical power significantly influenced the size of the
the small number of relevant studies prevents this for now.
effect of sleep deprivation. For sleep deprivation after learning,
mean statistical power to find the meta-analytic effect size was just
Quality-Based Moderators 14%; for sleep deprivation before learning, it was higher though still
far less than optimal at 55%. Given that power is a function of the
The meta-analyses in this article suggest that there is a detrimental effect size, sample size, and the statistical test being employed, the
effect of sleep deprivation on learning and memory, and it is difference in obtained power across the two research questions is
observed across a range of methodologies. However, our meta- understandable: As the effect size decreases, power to detect it
analyses identified a number of potential limitations of the available decreases if sample size is held constant. Overall, these figures are
data sets in this domain. Methodological quality scores ranged from closely in line with the broader field: For example, Szucs and
4 to 19 out of 22 in the studies investigating deprivation after Ioannidis (2017) found that within psychology and cognitive neu-
learning; and they ranged from 7 to 19 in the studies investigating roscience, mean power to detect small, medium, and large effects
deprivation before learning. (in Cohen’s terms) was 17%, 49%, and 71%, respectively.
Given the current convention that statistical power to find an are not always completely uninformative; we return to this debate in
effect is at 80% or higher (Di Stefano, 2003), it is evident that the the Conclusions section.
majority of the studies in these meta-analyses are underpowered.
This is problematic as it increases the uncertainty around our meta-
analytical effect sizes. To better understand the consequences of the Publication Bias
uncertainty introduced by low power in the studies included in our Conducting a meta-analysis allows for an estimation of publica-
meta-analyses, we investigated whether statistical power to find the tion bias within the literature. Publication bias is evident when there
mean meta-analytic effect size influenced the size of the sleep are a large number of published studies in the direction of the
deprivation effect by entering obtained power as a moderator. hypothesis, with few nonsignificant published studies (Rosenthal,
For example, it might be the case that it is only low-powered studies 1979). This can lead to overestimation of the size of the effect. We
that show an impact of sleep deprivation, while high-powered found statistically significant evidence of publication bias in both
studies might show no impact. Such a pattern would suggest that meta-analyses. Adjusting the deprivation after learning effect size
our meta-analytic effect size might be overestimated as a conse- for publication bias using the trim-and-fill method changed the
quence of low power. The opposite pattern would suggest that our effect size from 0.277 to 0.166, and changed the deprivation before
effect size has been underestimated due to low power. For depriva- learning effect size from 0.621 to 0.463, although these adjusted
tion after learning, we found no moderating impact of statistical effect size should be treated with caution given that the trim-and-fill
power on the size of the effect. In other words, both low- and high- method was not designed for a multilevel approach. Nonetheless,
powered studies yielded similar effect sizes. However, the validity both estimates remained significantly different from zero after the
of this analysis is reduced by the fact that there were no studies in adjustment. To allow for more accurate effect size estimates in
this meta-analysis where power exceeded 33%, and therefore we future meta-analyses, we suggest researchers in this field should
have no way of knowing what effect sizes could be expected when adopt registered reports as an effective way of ensuring all results
power is higher. For deprivation before learning, we found a broader find their way into published literature.
range of power extending from about 20% to over 90%. However,
once again we found no statistically significant moderating impact
of power on the size of the effect. Limitations
To gain a more precise estimate of the true effect size, future
We focused specifically on the effects of total (overnight) sleep
studies should use the current meta-analytic effect size as a guide in
deprivation, and thus future meta-analyses are needed to establish
determining sample sizes that will yield high power. Studies plan- whether the effect size is similar in studies using sleep restriction.
ning to look at sleep deprivation after learning and running a two- We chose to concentrate on studies using total sleep deprivation
tailed t-test for a between-subjects design with a sleep versus sleep because depriving a participant of all sleep is a stronger test of the
deprivation manipulation would require a sample size of approxi- hypothesis that sleep benefits memory than depriving them of a
mately 410 to have 80% power to detect the meta-analytic effect single stage of sleep or restricting their sleep for some hours over a
size. For a within-subjects design, the sample size required would be period of time, as discussed in the Introduction section. An alterna-
105. For studies examining deprivation before learning, a two-tailed tive approach could have been to include restriction studies and to
t-test with a between-subjects design would require a sample size of conduct a moderator analysis to establish whether they lead to
82, whereas a within-subjects design would require a much smaller similar effect sizes as total deprivation. However, many sleep
sample size of 23, to have 80% power to detect the meta-analytic restriction studies in the literature are field studies that lack the
effect size. The above numbers are rough indications only, and rigorous controls we include in our inclusion criteria (e.g., lack of
lower or higher sample sizes may be appropriate depending on the control over hours slept, Deary & Tait, 1987; inappropriate sleep
specific design of the experiment and the statistical analysis control condition, Piérard et al., 2004), and therefore the number of
approach (Brysbaert, 2019; Lakens & Caldwell, 2021). eligible restriction studies would have been smaller than the number
It is clear that there is a significant discrepancy between the of total deprivation studies. As discussed earlier, such imbalance in
high-power sample sizes we have estimated above and the sample number of studies can make moderator analyses insensitive (Hempel
sizes found in the majority of the studies included in the current et al., 2013).
meta-analyses. This discrepancy is important as there are severe Our search focused solely on English language reports, thus
limitations to the strength of conclusions that can be drawn from risking a mono-language bias (Johnson, 2021). This restricts our
underpowered individual studies (see, e.g., Brysbaert, 2019, for a ability to generalize the results of our meta-analyses to non-English
detailed discussion). Fraley and Vazire (2014) described three language literature. In particular, by using English language sources
limitations: (a) underpowered studies are less likely than properly only, there is the possibility that our search missed much of the gray
powered studies to detect a true effect; (b) underpowered studies are literature such as PhD theses and conference abstracts written in
more likely to yield false-positive findings than properly powered other languages. The use of solely English language sources limits
studies; and (c) underpowered studies are less likely than properly our understanding of any possible cross-cultural differences in
powered studies to produce replicable findings. Wilson et al. (2020) effects of sleep on memory. Indeed, there are many cross-cultural
further demonstrate that underpowered studies are likely to yield differences in sleep habits (e.g., Cheung et al., 2021), and although
inflated effect sizes. Therefore, the results of any single underpow- we are not aware of any studies that have systematically compared
ered study should be treated with caution, and a meta-analytic sleep-associated memory consolidation effects across cultures, our
approach such as ours may be the more useful approach for reliance on English language literature means that we would not
extracting information from these studies. Yet, small-scale studies have captured such differences if they do exist.
We acknowledge that our inclusion criteria restrict our ability to needed studies, while also allowing better informed sample size
draw conclusions beyond healthy, typical populations. We excluded choice for continuing original science efforts.
studies that included participants under the age of 18 and studies that
involved participants suffering from sleep disorders or psychiatric
disorders. There is growing interest in understanding how sleep- References
associated memory consolidation in these groups might differ from References marked with *1 indicate studies included in the first meta-
healthy adults, however (e.g., Hoedlmoser, 2020; Manoach & analysis (sleep deprivation after learning) and references marked with *2
Stickgold, 2019), and future meta-analyses addressing these ques- indicate studies included in the second meta-analysis (sleep deprivation
tions will be valuable both for theoretical development and practical before learning).
reasons. Finally, we note that there were 17 studies that we were
unable to include in the analyses as the required statistical informa- Alberca-Reina, E., Cantero, J. L., & Atienza, M. (2015). Impact of sleep loss
tion was not reported (see Figure 1); unfortunately, the authors of before learning on cortical dynamics during memory retrieval. Neuro-
Image, 123, 51–62. https://doi.org/10.1016/j.neuroimage.2015.08.033
these papers were unable to provide with the necessary data when *1Albouy, G., Sterpenich, V., Vandewalle, G., Darsaud, A., Gais, S.,
contacted. Nonetheless, these studies made up a small proportion of Rauchs, G., Desseilles, M., Boly, M., Dang-Vu, T., Balteau, E.,
the studies we identified as eligible and would be unlikely to change Degueldre, C., Phillips, C., & Degueldre, C. (2013). Interaction between
the conclusions we have drawn. hippocampal and striatal systems predicts subsequent consolidation of
motor sequence memory. PLOS ONE, 8(3). Article e59490. https://
doi.org/10.1371/journal.pone.0059490
*1Albouy, G., Vandewalle, G., Sterpenich, V., Rauchs, G., Desseilles, M.,
Conclusions
Balteau, E., Degueldre, C., Phillips, C., Luxen, A., & Maquet, P. (2013).
To conclude, the two meta-analyses presented here provide a Sleep stabilizes visuomotor adaptation memory: A functional magnetic
comprehensive analysis of the impact of sleep deprivation after resonance imaging study. Journal of Sleep Research, 22(2), 144–154.
learning and before learning. The results of the first meta-analysis https://doi.org/10.1111/j.1365-2869.2012.01059.x
suggest that depriving participants of the first night of sleep after Alger, S. E., Kensinger, E. A., & Payne, J. D. (2018). Preferential consoli-
dation of emotionally salient information during a nap is preserved in
encoding new information results in lower performance at test,
middle age. Neurobiology of Aging, 68, 34–47. https://doi.org/10.1016/j
supporting the theories of sleep-associated memory consolidation .neurobiolaging.2018.03.030
(e.g., Diekelmann & Born, 2010; McClelland et al., 1995). This *1Ashton, J., Harrington, M. O., Langthorne, D., Ngo, H. V., & Cairney,
effect was larger before than after recovery sleep and larger in S. A. (2020). Sleep deprivation induces fragmented memory loss. Learn-
procedural memory tasks compared to declarative memory tasks. ing and Memory. 27(4), 130–135. https://doi.org/10.1101/lm.050757.119
The results of the second meta-analysis suggest that sleep-deprived Ashton, J. E., Cairney, S. A., & Gaskell, M. G. (2018). No effect of targeted
participants are able to encode less information than rested controls, memory reactivation during slow-wave sleep on emotional recognition
supporting the theories that propose that sleep restores memory memory. Journal of Sleep Research, 27(1), 129–137. https://doi.org/10
encoding capacity (e.g., Saletin & Walker, 2012; Tononi & .1111/jsr.12542
Cirelli, 2012). Assink, M., & Wibbelink, C. J. (2016). Fitting three-level meta-analytic
models in R: A step-by-step tutorial. The Quantitative Methods for
We found that levels of statistical power tended to be low,
Psychology, 12(3), 154–174. https://doi.org/10.20982/tqmp.12.3.p154
particularly in those studies looking at sleep deprivation after *1Atienza, M., & Cantero, J. L. (2008). Modulatory effects of emotion and
learning in which there was a small estimated effect size. Given sleep on recollection and familiarity. Journal of Sleep Research, 17(3),
that underpowered studies are ubiquitous across disciplines that use 285–294. https://doi.org/10.1111/j.1365-2869.2008.00661.x
human participants (Dumas-Mallet et al., 2017), new ways of Banks, S., & Dinges, D. F. (2007). Behavioral and physiological conse-
interpreting low-powered studies are emerging. One particularly quences of sleep restriction. Journal of Clinical Sleep Medicine, 3(5),
insightful interpretation has recently been offered by Wilson et al. 519–528. https://doi.org/10.5664/jcsm.26918
(2020). In short, Wilson and colleagues draw a distinction between Bartel, P., Offermeier, W., Smith, F., & Becker, P. (2004). Attention and
“original science” and “replication science.” Original science is working memory in resident anaesthetists after night duty: Group and
individual effects. Occupational and Environmental Medicine, 61(2),
roughly science as practiced today, in that it uses Null Hypothesis
167–170. https://doi.org/10.1136/oem.2002.006197
Significance Testing combined with study designs whose power *2Basner, M., Savitt, A., Moore, T. M., Port, A. M., McGuire, S., Ecker, A. J.,
falls far short of the gold standard of high-N studies. Original Nasrini, J., Mollicone, D. J., Mott, C. M., McCann, T., Dinges, D. F., & Gur,
science in this formulation serves an important and inexpensive R. V. (2015). Development and validation of the cognition test battery for
screening function to identify effects that may be true and would spaceflight. Aerospace Medicine and Human Performance, 86(11), 942–
therefore benefit from further, more costly examination of replica- 952. https://doi.org/10.3357/AMHP.4343.2015
tion science. Replication science consists of high-N, highly pow- *2Bensimon, G., Benoit, D., Lacomblez, L., Weiller, E., Warot, D., Weil,
ered, and costly direct replications of the key studies from original J. S., & Puech, A. J. (1991). Antagonism by modafinil of the psychomotor
science, vital for verifying the preliminary results of original and cognitive impairment induced by sleep-deprivation in 12 healthy
science. Applying this framework to the literature targeted in our volunteers. European Psychiatry, 6(2), 93–97. https://doi.org/10.1017/
S0924933800000201
meta-analyses, we propose that there is now sufficient evidence *2Blagrove, M., Cole-Morgan, D., & Lambe, H. (1994). Interrogative
from original science to warrant a move to replication science in this suggestibility: The effects of sleep deprivation and relationship with field
field. No highly powered, preregistered direct replications looking at dependence. Applied Cognitive Psychology, 8(2), 169–179. https://
the role of sleep deprivation in learning and memory have been doi.org/10.1002/acp.2350080207
conducted thus far. The meta-analytic estimates of the relevant Borenstein, M., Hedges, L. V., Higgins, J. P., & Rothstein, H. R. (2010).
effect sizes provided here will facilitate the design of such urgently A basic introduction to fixed-effect and random-effects models for
meta-analysis. Research Synthesis Methods, 1(2), 97–111. https://doi.org/ Deary, I. J., & Tait, R. (1987). Effects of sleep disruption on cognitive
10.1002/jrsm.12 performance and mood in medical house officers. British Medical Journal
Borenstein, M., Hedges, L. V., Higgins, J. P., & Rothstein, H. R. (2011). (Clinical Research Ed.), 295(6612), 1513–1516. https://doi.org/10.1136/
Introduction to meta-analysis. Wiley. bmj.295.6612.1513
Borenstein, M., Higgins, J. P., Hedges, L. V., & Rothstein, H. R. (2017). *1Deliens, G., Gilson, M., Schmitz, R., & Peigneux, P. (2013). Sleep unbinds
Basics of meta-analysis: I2 is not an absolute measure of heterogeneity. memories from their emotional context. Cortex, 49(8), 2221–2228. https://
Research Synthesis Methods, 8(1), 5–18. https://doi.org/10.1002/ doi.org/10.1016/j.cortex.2012.11.014
jrsm.1230 *1Deliens, G., Schmitz, R., Caudron, I., Mary, A., Leproult, R., & Peigneux,
Born, J., & Wilhelm, I. (2012). System consolidation of memory during P. (2013b). Does recall after sleep-dependent memory consolidation
sleep. Psychological Research, 76(2), 192–203. https://doi.org/10.1007/ reinstate sensitivity to retroactive interference? PLOS ONE, 8(7), Article
s00426-011-0335-6 e68727. https://doi.org/10.1371/journal.pone.0068727
*1Borragán, G., Urbain, C., Schmitz, R., Mary, A., & Peigneux, P. (2015). Di Stefano, J. (2003). How much power is enough? Against the development of
Sleep and memory consolidation: Motor performance and proactive an arbitrary convention for statistical power calculations. Functional Ecol-
interference effects in sequence learning. Brain and Cognition, 95, ogy, 17(5), 707–709. https://doi.org/10.1046/j.1365-2435.2003.00782.x
54–61. https://doi.org/10.1016/j.bandc.2015.01.011 Diekelmann, S., & Born, J. (2010). The memory function of sleep. Nature
*2Boyle, J., Stanley, N., James, L. M., Wright, N., Johnsen, S., Arbon, Reviews Neuroscience, 11(2), 114–126. https://doi.org/10.1038/nrn2762
E. L., & Dijk, D. J. (2012). Acute sleep deprivation: The effects of the *1Diekelmann, S., Born, J., & Wagner, U. (2010). Sleep enhances false
AMPAKINE compound CX717 on human cognitive performance, alert- memories depending on general memory performance. Behavioural Brain
ness and recovery sleep. Journal of Psychopharmacology, 26(8), 1047– Research, 208(2), 425–429. https://doi.org/10.1016/j.bbr.2009.12.021
1057. https://doi.org/10.1177/0269881111405353 *1Diekelmann, S., Landolt, H. P., Lahl, O., Born, J., & Wagner, U. (2008).
Brysbaert, M. (2019). How many participants do we have to include in Sleep loss produces false memories. PLOS ONE, 3(10). Article e0250671.
properly powered experiments? A tutorial of power analysis with refer- https://doi.org/10.1371/journal.pone.0003512
ence tables. Journal of Cognition, 2(1), Article 187. https://doi.org/10 Diekelmann, S., Wilhelm, I., & Born, J. (2009). The whats and whens of
.5334/joc.72 sleep-dependent memory consolidation. Sleep Medicine Reviews, 13(5),
Button, K. S., Ioannidis, J. P., Mokrysz, C., Nosek, B. A., Flint, J., Robinson, 309–321. https://doi.org/10.1016/j.smrv.2008.08.002
E. S., & Munafò, M. R. (2013). Power failure: Why small sample size *1Diekelmann, S., Wilhelm, I., Wagner, U., & Born, J. (2013). Sleep to
undermines the reliability of neuroscience. Nature Reviews Neuroscience, implement an intention. Sleep, 36(1), 149–153. https://doi.org/10.5665/
14(5), 365–376. https://doi.org/10.1038/nrn3475 sleep.2322
*2Chatburn, A., Kohler, M. J., Payne, J. D., & Drummond, S. P. (2017). The Dinges, D. F., & Powell, J. W. (1985). Microcomputer analyses of perfor-
effects of sleep restriction and sleep deprivation in producing false mance on a portable, simple visual RT task during sustained operations.
memories. Neurobiology of Learning and Memory, 137, 107–113. Behavior Research Methods, Instruments, and Computers, 17(6), 652–
https://doi.org/10.1016/j.nlm.2016.11.017 655. https://doi.org/10.3758/BF03200977
*2Chee, M. W., Tan, J. C., Parimal, S., & Zagorodnov, V. (2010). Sleep Downs, S. H., & Black, N. (1998). The feasibility of creating a checklist
deprivation and its effects on object-selective attention. Neuroimage, for the assessment of the methodological quality both of randomised and
49(2), 1903–1910. https://doi.org/10.1016/j.neuroimage.2009.08.067 non-randomised studies of health care interventions. Journal of Epidemi-
Cheung, B. Y., Takemura, K., Ou, C., Gale, A., & Heine, S. J. (2021). ology & Community Health, 52(6), 377–384. https://doi.org/10.1136/jech
Considering cross-cultural differences in sleep duration between Japanese .52.6.377
and Canadian university students. PLOS ONE, 16(4), Article e0250671. *2Drake, C. L., Roehrs, T. A., Burduvali, E., Bonahoom, A., Rosekind, M.,
https://doi.org/10.1371/journal.pone.0250671 & Roth, T. (2001). Effects of rapid versus slow accumulation of eight
Cheung, M. W. L. (2014). Modeling dependent effect sizes with three-level hours of sleep loss. Psychophysiology, 38(6), 979–987. https://doi.org/10
meta-analyses: A structural equation modeling approach. Psychological .1111/1469-8986.3860979
Methods, 19(2), 211–229. https://doi.org/10.1037/a0032968 Drosopoulos, S., Wagner, U., & Born, J. (2005). Sleep enhances explicit
*2Chuah, L. Y., Chong, D. L., Chen, A. K., Rekshan, W. R., III, Tan, J. C., recollection in recognition memory. Learning and Memory, 12(1), 44–51.
Zheng, H., & Chee, M. W. (2009). Donepezil improves episodic memory https://doi.org/10.1101/lm.83805
in young individuals vulnerable to the effects of sleep deprivation. Sleep, Drummond, S. P., & Brown, G. G. (2001). The effects of total sleep
32(8), 999–1010. https://doi.org/10.1093/sleep/32.8.999 deprivation on cerebral responses to cognitive performance. Neuropsy-
Cirelli, C., & Tononi, G. (in press). The why and how of sleep-dependent chopharmacology, 25(5), S68–S73. https://doi.org/10.1016/S0893-
synaptic down-selection. Seminars in Cell and Developmental Biology. 133X(01)00325-6
https://doi.org/10.1016/j.semcdb.2021.02.007 *2Drummond, S. P., Brown, G. G., Gillin, J. C., Stricker, J. L., Wong, E. C.,
Cochran, W. G. (1954). The combination of estimates from different & Buxton, R. B. (2000). Altered brain response to verbal learning
experiments. Biometrics, 10(1), 101–129. https://doi.org/10.2307/ following sleep deprivation. Nature, 403(6770), 655–657. https://
3001666 doi.org/10.1038/35001068
Correll, J., Mellinger, C., McClelland, G. H., & Judd, C. M. (2020). Avoid *2Drummond, S. P., Meloy, M. J., Yanagi, M. A., Orff, H. J., & Brown, G. G.
Cohen’s “small”,“medium”, and “large” for power analysis. Trends in (2005). Compensatory recruitment after sleep deprivation and the rela-
Cognitive Sciences, 24(3), 200–207. https://doi.org/10.1016/j.tics.2019 tionship with performance. Psychiatry Research: Neuroimaging, 140(3),
.12.009 211–223. https://doi.org/10.1016/j.pscychresns.2005.06.007
Cousins, J. N., Sasmita, K., & Chee, M. W. (2018). Memory encoding is Dumas-Mallet, E., Button, K. S., Boraud, T., Gonon, F., & Munafò, M. R.
impaired after multiple nights of partial sleep restriction. Journal of Sleep (2017). Low statistical power in biomedical science: A review of three
Research, 27(1), 138–145. https://doi.org/10.1111/jsr.12578 human research domains. Royal Society Open Science, 4(2), Article
*1Darsaud, A., Dehon, H., Lahl, O., Sterpenich, V., Boly, M., Dang-Vu, T., 160254. https://doi.org/10.1098/rsos.160254
Desseilles, M.,Gais, S., Matarazzo, L., Peters, F., Schabus, M., Schmidt, Duval, S., & Tweedie, R. (2000). Trim and fill: A simple funnel-plot–based
C., Tinguely, G., Vandewalle, G., Luxen, A., Maquet, P., & Collette, F. method of testing and adjusting for publication bias in meta-analysis.
(2011). Does sleep promote false memories?. Journal of Cognitive Biometrics, 56(2), 455–463. https://doi.org/10.1111/j.0006-341X.2000
Neuroscience, 23(1), 26–40. https://doi.org/10.1162/jocn.2010.21448 .00455.x
Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in meta- *2Giacobbo, B. L., Corrêa, M. S., Vedovelli, K., de Souza, C. E. B., Spitza,
analysis detected by a simple, graphical test. British Medical Journal, L. M., Gonçalves, L., Gonçalves, L., Paludo N., Molina, R. D., da Rosa,
315(7109), 629–634. https://doi.org/10.1136/bmj.315.7109.629 E. D., Argimon, I. L., & Bromberg, E. (2016). Could BDNF be involved in
*1Ertelt, D., Witt, K., Reetz, K., Frank, W., Junghanns, K., Backhaus, J., compensatory mechanisms to maintain cognitive performance despite
Tadic, V., Pellicano, A., Born, J., Binkofski, F., & Binkofski, F. (2012). acute sleep deprivation? An exploratory study. International Journal of
Skill memory escaping from distraction by sleep—Evidence from dual- Psychophysiology, 99, 96–102. https://doi.org/10.1016/j.ijpsycho.2015
task performance. PLOS ONE, 7(12), Article e50983. https://doi.org/10 .11.008
.1371/journal.pone.0050983 Giuditta, A. (2014). Sleep memory processing: The sequential hypothesis.
*2Esposito, M. J., Occhionero, M., & Cicogna, P. (2015). Sleep deprivation Frontiers in Systems Neuroscience, 8, Article 219. https://doi.org/10
and time-based prospective memory. Sleep, 38(11), 1823–1826. https:// .3389/fnsys.2014.00219
doi.org/10.5665/sleep.5172 *2Gorissen, M., Tielemans, M., & Coenen, A. (1997). Alertness and memory
*2Falleti, M. G., Maruff, P., Collie, A., Darby, D. G., & McStephen, M. after sleep deprivation and diazepam intake. Journal of Psychopharma-
(2003). Qualitative similarities in cognitive impairment associated with 24 cology, 11(3), 233–239. https://doi.org/10.1177/026988119701100306
h of sustained wakefulness and a blood alcohol concentration of 0.05%. *1Griessenberger, H., Hödlmoser, K., Heib, D. P. J., Lechinger, J., Klimesch,
Journal of Sleep Research, 12(4), 265–274. https://doi.org/10.1111/j W., & Schabus, M. (2012). Consolidation of temporal order in episodic
.1365-2869.2003.00363.x memories. Biological Psychology, 91(1), 150–155. https://doi.org/10
Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G*Power 3: A .1016/j.biopsycho.2012.05.012
flexible statistical power analysis program for the social, behavioral, and Groch, S., Zinke, K., Wilhelm, I., & Born, J. (2015). Dissociating the
biomedical sciences. Behavior Research Methods, 39(2), 175–191. https:// contributions of slow-wave sleep and rapid eye movement sleep to
doi.org/10.3758/BF03193146 emotional item and source memory. Neurobiology of Learning and
*1Feld, G. B., Weis, P. P., & Born, J. (2016). The limited capacity of sleep- Memory, 122, 122–130. https://doi.org/10.1016/j.nlm.2014.08.013
dependent memory consolidation. Frontiers in Psychology, 7, Article *2Grundgeiger, T., Bayen, U. J., & Horn, S. S. (2014). Effects of sleep
1368. https://doi.org/10.3389/fpsyg.2016.01368 deprivation on prospective memory. Memory, 22(6), 679–686. https://
Fenn, K. M., Gallo, D. A., Margoliash, D., Roediger, H. L., III, & Nusbaum, doi.org/10.1016/j.biopsycho.2012.05.012
H. C. (2009). Reduced false memory after sleep. Learning & Memory, 16, Guyette, F. X., Morley, J. L., Weaver, M. D., Patterson, P. D., & Hostler, D.
509–513. https://doi.org/10.1101/lm.1500808 (2013). The effect of shift length on fatigue and cognitive performance in
*1Ferrara, M., Iaria, G., De Gennaro, L., Guariglia, C., Curcio, G., Tempesta, air medical providers. Prehospital Emergency Care, 17(1), 23–28. https://
D., & Bertini, M. (2006). The role of sleep in the consolidation of route doi.org/10.3109/10903127.2012.710719
learning in humans: A behavioural study. Brain Research Bulletin, 71(1–3), *1Harrington, M. O., Nedberge, K. M., & Durrant, S. J. (2018). The effect of
4–9. https://doi.org/10.1177/1745691612459059 sleep deprivation on emotional memory consolidation in participants
*1Ferrara, M., Iaria, G., Tempesta, D., Curcio, G., Moroni, F., Marzano, C., reporting depressive symptoms. Neurobiology of Learning and Memory,
De Gennaro, L., & Pacitti, C. (2008). Sleep to find your way: The role of 152, 10–19. https://doi.org/10.1016/j.nlm.2018.04.013
sleep in the consolidation of memory for navigation in humans. Hippo- *2Harrison, Y., & Horne, J. A. (2000a). Sleep loss and temporal memory.
campus, 18(8), 844–851. https://doi.org/10.1002/hipo.20444 The Quarterly Journal of Experimental Psychology: Section A, 53(1),
*1*2Fischer, S., Hallschmid, M., Elsner, A. L., & Born, J. (2002). Sleep 271–279. https://doi.org/10.1080/713755870
forms memory for finger skills. Proceedings of the National Academy Harrison, Y., & Horne, J. A. (2000b). The impact of sleep deprivation on
of Sciences, 99(18), 11987–11991. https://doi.org/10.1073/pnas decision making: A review. Journal of Experimental Psychology: Applied,
.182178199 6(3), 236–249. https://doi.org/10.1037/1076-898X.6.3.236
*1Fischer, S., Nitschke, M. F., Melchert, U. H., Erdmann, C., & Born, J. *2Harrison, Y., Horne, J. A., & Rothwell, A. (2000). Prefrontal neuropsy-
(2005). Motor memory consolidation in sleep shapes more effective chological effects of sleep deprivation in young adults—A model for
neuronal representations. Journal of Neuroscience, 25(49), 11248– healthy aging? Sleep, 23(8), 1067–1073. https://doi.org/10.1093/sleep/23
11255. https://doi.org/10.1523/JNEUROSCI.1743-05.2005 .8.1f
*2Forest, G., & Godbout, R. (2000). Effects of sleep deprivation on perfor- Hempel, S., Miles, J. N., Booth, M. J., Wang, Z., Morton, S. C., & Shekelle,
mance and EEG spectral analysis in young adults. Brain and Cognition, P. G. (2013). Risk of bias: A simulation study of power to detect study-
43(1–3), 195–200. level moderator effects in meta-analysis. Systematic Reviews, 2(1), 1–10.
Fraley, R. C., & Vazire, S. (2014). The N-pact factor: Evaluating the quality https://doi.org/10.1186/2046-4053-2-107
of empirical journals with respect to sample size and statistical power. *1Herzog, N., Friedrich, A., Fujita, N., Gais, S., Jauch-Chara, K., Oltmanns,
PLOS ONE, 9(10), Article e109019. https://doi.org/10.1371/journal.pone K. M., & Benedict, C. (2012). Effects of daytime food intake on memory
.0109019 consolidation during sleep or sleep deprivation. PLOS ONE, 7(6), Article
Funder, D. C., & Ozer, D. J. (2019). Evaluating effect size in psychological e40298. https://doi.org/10.1371/journal.pone.0040298
research: Sense and nonsense. Advances in Methods and Practices in Higgins, J. P., Altman, D. G., Gøtzsche, P. C., Jüni, P., Moher, D., Oxman,
Psychological Science, 2(2), 156–168. https://doi.org/10.1177/2515245 A. D., Savović, J., Schulz, K. F., Weeks, L., & Sterne, J. A. C., (2011). The
919847202 cochrane collaboration’s tool for assessing risk of bias in randomised
Gais, S., Albouy, G., Boly, M., Dang-Vu, T. T., Darsaud, A., Desseilles, M., trials. British Medical Journal, 343, Article d5928. https://doi.org/10
Rauchs, G., Schabus, M., Sterpenich, V., Vandewalle, G., Maquet, P., & .1136/bmj.d5928
Peigneux, P. (2007). Sleep transforms the cerebral trace of declarative Higgins, J. P., & Thompson, S. G. (2002). Quantifying heterogeneity in a
memories. Proceedings of the National Academy of Sciences, 104(47), meta-analysis. Statistics in Medicine, 21(11), 1539–1558. https://doi.org/
18778–18783. https://doi.org/10.1073/pnas.0705454104 10.1002/sim.1186
Gais, S., & Born, J. (2004). Declarative memory consolidation: Mechanisms Hoedlmoser, K. (2020). Sleep and memory in children. Current Sleep
acting during human sleep. Learning and Memory, 11(6), 679–685. Medicine Reports, 6(4), 280–289. https://doi.org/10.1007/s40675-020-
https://doi.org/10.1101/lm.80504. 00194-8
*1Gais, S., Lucas, B., & Born, J. (2006). Sleep after learning aids memory Hu, P., Stylos-Allan, M., & Walker, M. P. (2006). Sleep facilitates consoli-
recall. Learning and Memory, 13(3), 259–262. https://doi.org/10.1101/lm dation of emotional declarative memory. Psychological Science, 17(10),
.132106 891–898. https://doi.org/10.1111/j.1467-9280.2006.01799.x
Hunter, J. E., & Schmidt, F. L. (2004). Methods of meta-analysis: Correcting Lewis, P. A., Knoblich, G., & Poe, G. (2018). How memory replay in sleep
error and bias in research findings. Sage Publications. boosts creative problem-solving. Trends in Cognitive Sciences, 22,
*1Idzikowski, C. (1984). Sleep and memory. British Journal of Psychology, 491–503. https://doi.org/10.1016/j.tics.2018.03.009
75(4), 439–449. https://doi.org/10.1111/j.2044-8295.1984.tb01914.x Lim, J., & Dinges, D. (2008). Sleep deprivation and vigilant attention.
Inostroza, M., & Born, J. (2013). Sleep for preserving and transforming Annals of the New York Academy of Sciences, 1129(1), Article 305. https://
episodic memory. Annual Review of Neuroscience, 36, 79–102. https:// doi.org/10.1196/annals.1417.002
doi.org/10.1146/annurev-neuro-062012-170429 Lim, J., & Dinges, D. F. (2010). A meta-analysis of the impact of short-term
Ioannidis, J. P. (2005). Why most published research findings are false. PLoS sleep deprivation on cognitive variables. Psychological Bulletin, 136(3),
Medicine, 2(8), Article e124. https://doi.org/10.1371/journal.pmed.0020124 375–389. https://doi.org/10.1037/a0018883
Jacoby, L. L. (1991). A process dissociation framework: Separating auto- *1Lin, C. C., & Yang, C. M. (2014). Evidence of sleep-facilitating effect on
matic from intentional uses of memory. Journal of Memory and Language, formation of novel semantic associations: An event-related potential
30(5), 513–541. https://doi.org/10.1016/0749-596X(91)90025-F (ERP) study. Neurobiology of Learning and Memory, 116, 69–78.
Johnson, B. T. (2021). Toward a more transparent, rigorous, and generative https://doi.org/10.1016/j.nlm.2014.08.011
psychology. Psychological Bulletin, 147(1), 1–15. https://doi.org/10 Lipinska, G., Stuart, B., Thomas, K. G., Baldwin, D. S., & Bolinger, E.
.1037/bul0000317 (2019). Preferential consolidation of emotional memory during sleep: a
*2Kaida, K., Niki, K., & Born, J. (2015). Role of sleep for encoding of meta-analysis. Frontiers in Psychology, 10, Article 1014. https://doi.org/
emotional memory. Neurobiology of Learning and Memory, 121, 72–79. 10.3389/fpsyg.2019.01014
https://doi.org/10.1016/j.nlm.2015.04.002 *2Lo, J. C., Chong, P. L., Ganesan, S., Leong, R. L., & Chee, M. W. (2016).
Kilpeläinen, A. A., Huttunen, K. H., Lohi, J. J., & Lyytinen, H. (2010). Effect Sleep deprivation increases formation of false memory. Journal of Sleep
of caffeine on vigilance and cognitive performance during extended Research, 25(6), 673–682. https://doi.org/10.1111/jsr.12436
wakefulness. The International Journal of Aviation Psychology, 20(2), Lo, J. C., Groeger, J. A., Cheng, G. H., Dijk, D. J., & Chee, M. W. (2016).
144–159. https://doi.org/10.1080/10508411003617847 Self-reported sleep duration and cognitive performance in older adults: A
Kim, S. Y., & Payne, J. D. (2020). Neural correlates of sleep, stress, and systematic review and meta-analysis. Sleep Medicine, 17, 87–98. https://
selective memory consolidation. Current Opinion in Behavioral Sciences, doi.org/10.1016/j.sleep.2015.08.021
33, 57–64. https://doi.org/10.1016/j.cobeha.2019.12.009 Lo, J. C., Sim, S. K. Y., Chee, M. W. L. (2014). Sleep reduces false memory
King, B. R., Hoedlmoser, K., Hirschauer, F., Dolfen, N., & Albouy, G. in healthy older adults. Sleep, 37, 665–671. https://doi.org/10.5665/
(2017). Sleeping on the motor engram: The multifaceted nature of sleep- sleep.3564
related motor memory consolidation. Neuroscience and Biobehavioral Lowe, C. J., Safati, A., & Hall, P. A. (2017). The neurocognitive con-
Reviews, 80, 1–22. https://doi.org/10.1016/j.neubiorev.2017.04.026 sequences of sleep restriction: A meta-analytic review. Neuroscience &
Klinzing, J. G., Niethard N., & Born, J. (2019). Mechanisms of systems Biobehavioral Reviews, 80, 586–604. https://doi.org/10.1016/j.neubiorev
memory consolidation during sleep. Nature Neuroscience, 22, 1598– .2017.07.010
1610. https://doi.org/10.1038/s41593-019-0467-3 *1Maddox, W. T., Glass, B. D., Wolosin, S. M., Savarie, Z. R., Bowen, C.,
*1Kolibius, L. D., Born, J., & Feld, G. B. (2021). Vast amounts of encoded Matthews, M. D., & Schnyer, D. M. (2009). The effects of sleep
items nullify but do not reverse the effect of sleep on declarative memory. deprivation on information-integration categorization performance. Sleep,
Frontiers in Psychology, 11, Article 3685. https://doi.org/10.3389/fpsyg 32(11), 1439–1448. https://doi.org/10.1093/sleep/32.11.1439
.2020.607070 *2Maddox, W. T., Glass, B. D., Zeithamova, D., Savarie, Z. R., Bowen, C.,
Korman, M., Doyon, J., Doljansky, J., Carrier, J., Dagan, Y., & Karni, A. Matthews, M. D., & Schnyer, D. M. (2011). The effects of sleep
(2007). Daytime sleep condenses the time course of motor memory deprivation on dissociable prototype learning systems. Sleep, 34(3),
consolidation. Nature Neuroscience, 10(9), 1206–1213. https://doi.org/ 253–260. https://doi.org/10.1093/sleep/34.3.253
10.1038/nn1959 Mander, B. A., Santhanam, S., Saletin, J. M., & Walker, M. P. (2011). Wake
Kumaran, D., Hassabis, D., & McClelland, J. L. (2016). What learning deterioration and sleep restoration of human learning. Current Biology,
systems do intelligent agents need? Complementary learning systems 21(5), R183–R184. https://doi.org/10.1016/j.cub.2011.01.019
theory updated. Trends in Cognitive Sciences, 20, 512–534. https:// Manoach, D. S., & Stickgold, R. (2019). Abnormal sleep spindles, memory
doi.org/10.1016/j.tics.2016.05.004 consolidation, and schizophrenia. Annual Review of Clinical Psychology,
*1Kuriyama, K., Honma, M., Yoshiike, T., & Kim, Y. (2013). Memory 15, 451–479. https://doi.org/10.1146/annurev-clinpsy-050718-095754
suppression trades prolonged fear and sleep-dependent fear plasticity for McClelland, J. L., McNaughton, B. L., & O’Reilly, R. C. (1995). Why there
the avoidance of current fear. Scientific Reports, 3, Article 2227. https:// are complementary learning systems in the hippocampus and neocortex:
doi.org/10.1038/srep02227 Insights from the successes and failures of connectionist models of
*1Kuriyama, K., Soshi, T., & Kim, Y. (2010). Sleep deprivation facilitates learning and memory. Psychological Review, 102(3), 419–457. https://
extinction of implicit fear generalization and physiological response to doi.org/10.1037/0033-295X.102.3.419
fear. Biological Psychiatry, 68(11), 991–998. https://doi.org/10.1016/j McKeon, S., Pace-Schott, E. F., Spencer, R. M. C. (2012). Interaction of
.biopsych.2010.08.015 sleep and emotional content on the production of false memories. PLOS
*1Lahl, O., & Pietrowsky, R. (2006). Does the “sleep effect” on memory ONE, 7, Article e49353. https://doi.org/10.1371/journal.pone.0049353
depend on sleep or on night time? Sleep and Hypnosis, 8(2), 61–70. *2McWhirter, K. K., Morrow, A. S., Lee, B. A., Bishu, S., Zametkin, A. J.,
Lakens, D. (2013). Calculating and reporting effect sizes to facilitate Balkin, T. J., Smith, C. B., & Picchioni, D. (2015). A pilot study on the
cumulative science: A practical primer for t-tests and ANOVAs. encoding of a perceptual learning task following sleep deprivation.
Frontiers in Psychology, 4, Article 863. https://doi.org/10.3389/fpsyg Perceptual and Motor Skills, 121(1), 80–93. https://doi.org/10.2466/23
.2013.00863 .PMS.121c11x9
Lakens, D., & Caldwell, A. R. (2021). Simulation-based power analysis for *1Mograss, M. A., Guillem, F., Brazzini-Poisson, V., & Godbout, R. (2009).
factorial analysis of variance designs. Advances in Methods and Practices in The effects of total sleep deprivation on recognition memory processes: A
Psychological Science, 4, 1–14. https://doi.org/10.1177/2515245920951503 study of event-related potential. Neurobiology of Learning and Memory,
Lewis, P. A., & Durrant, S. J. (2011). Overlapping memory replay during 91(4), 343–352. https://doi.org/10.1016/j.nlm.2009.01.008
sleep builds cognitive schemata. Trends in Cognitive Sciences, 15, 343– *1Montemayor, B. A. (2014). Musical task: Sleep deprivation and its effects
351. https://doi.org/10.1016/j.tics.2011.06.004 on motor learning [Masters dissertation]. University of Houston-Clear
Lake. Proquest. https://www.proquest.com/docview/1532140044?pq- Parallel neuropsychological and biochemical investigations. Neuropsy-

origsite=gscholar&fromopenview=true chobiology, 49(4), 210–217. https://doi.org/10.1159/000077369
Morgan, D. P., Tamminen, J., Seale-Carlisle, T. M., & Mickes, L. (2019). Pilcher, J. J., & Huffcutt, A. I. (1996). Effects of sleep deprivation on
The impact of sleep on eyewitness identifications. Royal Society Open performance: A meta-analysis. Sleep, 19(4), 318–326. https://doi.org/10
Science, 6(12), Article 170501. https://doi.org/10.1098/rsos.170501 .1093/sleep/19.4.318
*1Nesca, M., & Koulack, D. (1994). Recognition memory, sleep and Plihal, W., & Born, J. (1997). Effects of early and late nocturnal sleep on
circadian rhythms. Canadian Journal of Experimental Psychology, declarative and procedural memory. Journal of Cognitive Neuroscience,
48(3), 359–379. https://doi.org/10.1037/1196-1961.48.3.359 9(4), 534–547. https://doi.org/10.1162/jocn.1997.9.4.534
Newbury, C. R., Monaghan, P. (2019). When does sleep affect veridical and *2Poh, J. H., & Chee, M. W. (2017). Degradation of cortical representations
false memory consolidation? A meta-analysis. Psychonomic Bulletin and during encoding following sleep deprivation. Neuroimage, 153, 131–138.
Review, 26, 387–400. https://doi.org/10.3758/s13423-018-1528-4 https://doi.org/10.1016/j.neuroimage.2017.01.080
*2Nilsson, L. G., Bäckman, L., & Karlsson, T. (1989). Priming and cued Pollard, P., & Richardson, J. T. (1987). On the probability of making Type I
recall in elderly, alcohol intoxicated and sleep deprived subjects: A case errors. Psychological Bulletin, 102(1), 159–163. https://doi.org/10.1037/
of functionally similar memory deficits. Psychological Medicine, 19(2), 0033-2909.102.1.159
423–433. https://doi.org/10.1017/s0033291700012460 *1Porcheret, K., Holmes, E. A., Goodwin, G. M., Foster, R. G., & Wulff, K.
Nosek, B. A., Alter, G., Banks, G. C., Borsboom, D., Bowman, S. D., (2015). Psychological effect of an analogue traumatic event reduced by
Breckler, S. J., Buck, S., Chambers, C. D., Chin, G., Christensen, G., sleep deprivation. Sleep, 38(7), 1017–1025. https://doi.org/10.5665/slee
Contestabile, M., Dafoe, A., Eich, E., Freese, J., Glennerster, R., Goroff, p.4802
D., Green, D. P., Hesse, B., Humphreys, M., : : : Yarkoni, T. (2015). Rauchs, G., Bertran, F., Guillery-Girard, B., Desgranges, B., Kerrouche, N.,
Promoting an open research culture. Science, 348(6242), 1422–1425. Denise, P., Foret, J., & Eustache, F. (2004). Consolidation of strictly
https://doi.org/10.1126/science.aab2374 episodic memories mainly requires rapid eye movement sleep. Sleep,
Nosek, B. A., & Lindsay, D. S. (2018, February 28). Preregistration 27(3), 395–401. https://doi.org/10.1093/sleep/27.3.395
becoming the norm in psychological science. APS Observer. https:// *1Rauchs, G., Orban, P., Schmidt, C., Albouy, G., Balteau, E., Degueldre, C.,
www.psychologicalscience.org/observer/preregistration-becoming-the-norm- Schnackers, C., Sterpenich, V., Tinguely, G., Luxen, A., Maquet, P., &
in-psychological-science Peigneux, P. (2008). Sleep modulates the neural substrates of both spatial
*2Occhionero, M., Cicogna, P., & Esposito, M. J. (2017). The effect of sleep and contextual memory consolidation. PLOS ONE, 3(8), Article e2949.
loss on dual time-based prospective memory tasks. American Journal of https://doi.org/10.1371/journal.pone.0002949
Psychology, 130(1), 93–103. https://doi.org/10.5406/amerjpsyc.130 Robertson, E. M., Pascual-Leone, A., & Press, D. Z. (2004). Awareness
.1.0093 modifies the skill-learning benefits of sleep. Current Biology, 14(3), 208–
*1Orban, P., Rauchs, G., Balteau, E., Degueldre, C., Luxen, A., Maquet, P., 212. https://doi.org/10.1016/j.cub.2004.01.027
& Peigneux, P. (2006). Sleep after spatial learning promotes covert *2Roehrs, T., Burduvali, E., Bonahoom, A., Drake, C., & Roth, T. (2003).
reorganization of brain activity. Proceedings of the National Academy Ethanol and sleep loss: A “dose” comparison of impairing effects. Sleep,
of Sciences, 103(18), 7124–7129. https://doi.org/10.1073/pnas 26(8), 981–985. https://doi.org/10.1093/sleep/26.8.981
.0510198103 Rohatgi, A. (2019). WebPlotDigitizer (Version 4.2). https://automeris.io/We
*2Pace-Schott, E. F., Hutcherson, C. A., Bemporad, B., Morgan, A., Kumar, bPlotDigitizer
A., Hobson, J. A., & Stickgold, R. (2009). Failure to find executive Rosenthal, R. (1979). The file drawer problem and tolerance for null results.
function deficits following one night’s total sleep deprivation in university Psychological Bulletin, 86(3), 638–641. https://doi.org/10.1037/0033-
students under naturalistic conditions. Behavioral Sleep Medicine, 7(3), 2909.86.3.638
136–163. https://doi.org/10.1080/15402000902976671 Rothstein, H. R., Sutton, A. J., & Borenstein, M. (Eds.). (2005). Publication
Pardilla-Delgado, E., & Payne, J. D. (2017). The impact of sleep on true and bias in meta-analysis: Prevention, assessment and adjustments. Wiley.
false memory across long delays. Neurobiology of Learning and Memory, Saletin, J. M., & Walker, M. P. (2012). Nocturnal mnemonics: Sleep and
137, 123–133. https://doi.org/10.1016/j.nlm.2016.11.016 hippocampal memory processing. Frontiers in Neurology, 3, Article 59.
Payne, J. D., & Kensinger, E. A. (2010). Sleep’s role in the consolidation https://doi.org/10.3389/fneur.2012.00059
of emotional episodic memories. Current Directions in Psychological Sawangjit, A., Oyanedel, C. N., Niethard, N., Salazar, C., Born, J., &
Science, 19, 290–295. https://doi.org/10.1177/0963721410383978 Inostroza, M. (2018). The hippocampus is crucial for forming non-
Payne, J. D., Kensinger, E. A., Wamsley, E. J., Spreng, R. N., Alger, S. E., hippocampal long-term memory during sleep. Nature 564, 109–113.
Gibler, K., Schacter, D. L., & Stickgold, R. (2015). Napping and the https://doi.org/10.1038/s41586-018-0716-8
selective consolidation of negative aspects of scenes. Emotion, 15, 176– Schäfer, S. K., Wirth, B. E., Staginnus, M., Becker, N., Michael, T., &
186. https://doi.org/10.1037/a0038683 Sopp, M. R. (2020). Sleep’s impact on emotional recognition memory: A
Payne, J. D., Schacter, D. L., Propper, R. E., Huang, L. W., Wamsley, E. J., meta-analysis of whole-night, nap, and REM sleep effects. Sleep Medi-
Tucker, M. A., Walker, M. P., & Stickgold, R. (2009). The role of sleep cine Reviews, 51, Article 101280. https://doi.org/10.1016/j.smrv.2020
in false memory formation. Neurobiology of Learning and Memory, 92, .101280
327–334. https://doi.org/10.1016/j.nlm.2009.03.007 Schapiro, A. C., Reid, A. G., Morgan, A., Manoach, D. S., Verfaellie, M.,
Payne, J. D., Stickgold, R., Swanberg, K., & Kensinger, E. A. (2008). Sleep & Stickgold, R. (2019). The hippocampus is necessary for the consoli-
preferentially enhances memory for emotional components of scenes. dation of a task that does not require the hippocampus for initial
Psychological Science, 19(8), 781–788. https://doi.org/10.1111/j.1467- learning. Hippocampus, 29, 1091–1100. https://doi.org/10.1002/
9280.2008.02157.x hipo.23101
*1Peigneux, P., Schmitz, R., & Willems, S. (2007). Cerebral asymmetries Schönauer, M., Geisler, T., & Gais, S. (2014). Strengthening procedural
in sleep-dependent processes of memory consolidation. Learning and memories by reactivation in sleep. Journal of Cognitive Neuroscience,
Memory, 14(6), 400–406. https://doi.org/10.1101/lm.551207 26(1), 143–153. https://doi.org/10.1162/jocn_a_00471
Piérard, C., Béracochéa, D., Pérès, M., Jouanin, J. C., Liscia, P., Satabin, P., *1Schönauer, M., Grätsch, M., & Gais, S. (2015). Evidence for two distinct
Martin, S., Testylier, G., Guézennec, C. Y., & Beaumont, M. (2004). sleep-related long-term memory consolidation processes. Cortex, 63, 68–
Declarative memory impairments following a military combat course: 78. https://doi.org/10.1016/j.cortex.2014.08.005
Siddaway, A. P., Wood, A. M., & Hedges, L. V. (2019). How to do a systematic Neurobiology of Learning and Memory, 131, 9–17. https://doi.org/10
review: A best practice guide for conducting and reporting narrative reviews, .1016/j.nlm.2016.03.007
meta-analyses, and meta-syntheses. Annual Review of Psychology, 70, 747– *1Tempesta, D., Socci, V., Ioio, G. D., De Gennaro, L., & Ferrara, M. (2017).
770. https://doi.org/10.1146/annurev-psych-010418-102803 The effect of sleep deprivation on retrieval of emotional memory: A
Stepan, M. E., Dehnke, T. M., & Fenn, K. M. (2017). Sleep and eyewitness behavioural study using film stimuli. Experimental Brain Research,
memory: Fewer false identifications after sleep when the target is absent 235(10), 3059–3067. https://doi.org/10.1007/s00221-01705043-z
from the lineup. PLOS ONE, 12, Article e0182907. https://doi.org/10 Tononi, G., & Cirelli, C. (2003). Sleep and synaptic homeostasis: A
.1371/journal.pone.0182907 hypothesis. Brain Research Bulletin, 62(2), 143–150. https://doi.org/10
Sterne, J. A., Sutton, A. J., Ioannidis, J. P., Terrin, N., Jones, D. R., Lau, J., .1016/j.brainresbull.2003.09.004
Carpenter, J., Rücker, G., Harbord, R. M., Schmid, C. H., Tetzlaff, J., Tononi, G., & Cirelli, C. (2012). Time to be SHY? Some comments on sleep
Deeks, J. J., Peters, J., Macaskill, P., Schwarzer, G., Duval, S., Altman, and synaptic homeostasis. Neural Plasticity, 12, 1–12. https://doi.org/10
D. G., Moher, D., & Higgins, J. P. T. (2011). Recommendations for .1155/2012/415250
examining and interpreting funnel plot asymmetry in meta-analyses of Twenge, J. M., Krizan, Z., & Hisler, G. (2017). Decreases in self-reported
randomised controlled trials. British Medical Journal, 343, Article d4002. sleep duration among US adolescents 2009–2015 and association with
https://doi.org/10.1136/bmj.d4002 new media screen time. Sleep Medicine, 39, 47–53. https://doi.org/10
*1Sterpenich, V., Albouy, G., Boly, M., Vandewalle, G., Darsaud, A., .1155/2012/415250
Balteau, E., Dang-Vu, T. T., Desseilles, M., D’Argembeau, A., Gais, Valentine, J. C. (2009). Judging the quality of primary research. In The
S., Rauchs, G., Schabus, M., Degueldre, C., Luxen, A., Collette, F., & handbook of research synthesis and meta-analysis (2nd ed., pp. 129–146).
Maquet, P. (2007). Sleep-related hippocampo-cortical interplay during Sage Publications.
emotional memory recollection. PLoS Biology, 5(11), Article e282. Van Der Werf, Y. D., Altena, E., Schoonheim, M. M., Sanz-Arigita, E. J.,
https://doi.org/10.1371/journal.pbio.0050282 Vis, J. C., De Rijke, W., & Van Someren, E. J. (2009). Sleep benefits
*1Sterpenich, V., Albouy, G., Darsaud, A., Schmidt, C., Vandewalle, G., Vu, subsequent hippocampal functioning. Nature Neuroscience, 12(2), 122–
T. T. D., Desseilles, M., Phillips, C., Degueldre, C., Balteau, E., Collette, 123. https://doi.org/10.1038/nn.2253
F., Luxen, A., & Maquet, P. (2009). Sleep promotes the neural reorgani- *1van Heugten–van der Kloet, D., Giesbrecht, T., & Merckelbach, H. (2015).
zation of remote emotional memory. Journal of Neuroscience, 29(16), Sleep loss increases dissociation and affects memory for emotional
5143–5152. https://doi.org/10.1523/JNEUROSCI.0561-09.2009 stimuli. Journal of Behavior Therapy and Experimental Psychiatry, 47,
*1Sterpenich, V., Ceravolo, L., & Schwartz, S. (2017). Sleep deprivation 9–17. https://doi.org/10.1016/j.jbtep.2014.11.002
disrupts the contribution of the hippocampus to the formation of novel *1Vargas, I., Payne, J. D., Muench, A., Kuhlman, K. R., & Lopez-Duran,
lexical associations. Brain and Language, 167, 61–71. https://doi.org/10 N. L. (2019). Acute sleep deprivation and the selective consolidation of
.1016/j.bandl.2016.12.007 emotional memories. Learning and Memory, 26(6), 176–181. https://
*1Stickgold, R., James, L., & Hobson, J. A. (2000). Visual discrimination doi.org/10.1101/lm.049312.119
learning requires sleep after training. Nature Neuroscience, 3(12), 1237– Vargas, I. P., Aguiar, S. A., & Barela, J. A. (2017). Effects of sleep
1238. https://doi.org/10.1038/81756. deprivation on sustained attention in young adults. Brazilian Journal of
Stickgold, R., & Walker, M. P. (2013). Sleep-dependent memory triage: Motor Behavior, 11(1), 1–9.https://doi.org/10.20338/bjmb.v11i1.96
Evolving generalization through selective processing. Nature Neurosci- Vidya, S., Patlolla, V. R. P., Kamuju, N. R., Ampalam, P., & Kalyan,
ence, 16, 139–145. https://doi.org/10.1038/nn.3303 V. K. S. N. (2019). Impact of shift work on sleep and quality of life in
Stone, J. C., Glass, K., Munn, Z., Tugwell, P., & Doi, S. A. (2020). industrial workers: A cross sectional study. Archives of Mental Health,
Comparison of bias adjustment methods in meta-analysis suggests that 20(2), Article 45. https://doi.org/10.4103/AMH.AMH_3_19
quality effects modeling may have less limitations than other approaches. Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor
Journal of Clinical Epidemiology, 117, 36–45. https://doi.org/10.1016/j package. Journal of Statistical Software, 36(3), 1–48. https://doi.org/10
.jclinepi.2019.09.010 .18637/jss.v036.i03
*2Stricker, J. L., Brown, G. G., Wetherell, L. A., & Drummond, S. P. (2006). Viechtbauer, W., & Cheung, M. W. L. (2010). Outlier and influence
The impact of sleep deprivation and task difficulty on networks of fMRI diagnostics for meta-analysis. Research Synthesis Methods, 1(2), 112–
brain response. Journal of the International Neuropsychological Society, 125. https://doi.org/10.1002/jrsm.11
12(5), 591–597. https://doi.org/10.1017/S1355617706060851 Voderholzer, U., Piosczyk, H., Holz, J., Landmann, N., Feige, B., Loessl, B.,
Szucs, D., & Ioannidis, J. P. (2017). Empirical assessment of published effect Kpasz, M., Doerr, J. P., Riemann, D., & Nissen, C. (2011). Sleep
sizes and power in the recent cognitive neuroscience and psychology restriction over several days does not affect long–term recall of declarative
literature. PLoS Biology, 15(3), Article e2000797. https://doi.org/10.1371/ and procedural memories in adolescents. Sleep Medicine, 12(2), 170–178.
journal.pbio.2000797 https://doi.org/10.1016/j.sleep.2010.07.017
Talamini, L. M., Nieuwenhuis, I. L., Takashima, A., & Jensen, O. (2008). Wagner, U., Gais, S., & Born, J. (2001). Emotional memory formation is
Sleep directly following learning benefits consolidation of spatial asso- enhanced across sleep intervals with high amounts of rapid eye movement
ciative memory. Learning and Memory, 15(4), 233–237. https://doi.org/ sleep. Learning & Memory, 8(2), 112–119. https://doi.org/10.1101/
10.1101/lm.771608 lm.36801
*1*2Tamminen, J., Newbury, C. R., Crowley, R., Vinals, L., Cevoli, B., & Wagner, U., Hallschmid, M., Rasch, B., & Born, J. (2006). Brief sleep after
Rastle, K. (2020). Generalisation in language learning can withstand total learning keeps emotional memories alive for years. Biological Psychiatry,
sleep deprivation. Neurobiology of Learning and Memory, 173, Article 60(7), 788–790. https://doi.org/10.1016/j.biopsych.2006.03.061
107274. https://doi.org/10.1016/j.nlm.2020.107274 Walker, M. P., Brakefield, T., Morgan, A., Hobson, J. A., & Stickgold, R.
*1Tempesta, D., De Gennaro, L., Natale, V., & Ferrara, M. (2015). Emo- (2002). Practice with sleep makes perfect: Sleep-dependent motor skill
tional memory processing is influenced by sleep quality. Sleep Medicine, learning. Neuron, 35(1), 205–211. https://doi.org/10.1016/S0896-
16(7), 862–870. https://doi.org/10.1016/j.sleep.2015.01.024 6273(02)00746-8
*2Tempesta, D., Socci, V., Coppo, M., Ioio, G. D., Nepa, V., De Gennaro, L., Walker, M. P., & Stickgold, R. (2006). Sleep, memory, and plasticity.
& Ferrara, M. (2016). The effect of sleep deprivation on the encoding Annual Review of Psychology, 57, 139–166. https://doi.org/10.1146/
of contextual and non-contextual aspects of emotional memory. annurev.psych.56.091103.070307
Walker, M. P., Stickgold, R., Alsop, D., Gaab, N., & Schlaug, G. (2005). *2Yoo, S. S., Hu, P. T., Gujar, N., Jolesz, F. A., & Walker, M. P. (2007). A
Sleep-dependent motor memory plasticity in the human brain. Neurosci- deficit in the ability to form new human memories without sleep. Nature
ence, 133(4), 911–917. https://doi.org/10.1016/j.neuroscience.2005.04.007 Neuroscience, 10(3), 385–392. https://doi.org/10.1038/nn1851
Wilson, B. M., Harris, C. R., & Wixted, J. T. (2020). Science is not a signal
detection problem. Proceedings of the National Academy of Sciences,
117(11), 5559–5567. https://doi.org/10.1073/pnas.1914237117
Yonelinas, A. P. (2002). The nature of recollection and familiarity: A review Received September 8, 2020
of 30 years of research. Journal of Memory and Language, 46(3), 441– Revision received October 24, 2021
517. https://doi.org/10.1006/jmla.2002.2864 Accepted December 11, 2021 ▪
Members of Underrepresented Groups:

Reviewers for Journal Manuscripts Wanted
If you are interested in reviewing manuscripts for APA journals, the APA Publications and
Communications Board would like to invite your participation. Manuscript reviewers are vital to the
publications process. As a reviewer, you would gain valuable experience in publishing. The P&C
Board is particularly interested in encouraging members of underrepresented groups to participate
more in this process.
If you are interested in reviewing manuscripts, please write APA Journals at Reviewers@apa.org.
Please note the following important points:
To be selected as a reviewer, you must have published articles in peer-reviewed journals. The
experience of publishing provides a reviewer with the basis for preparing a thorough, objective
review.
To be selected, it is critical to be a regular reader of the five to six empirical journals that are most
central to the area or journal for which you would like to review. Current knowledge of recently
published research provides a reviewer with the knowledge base to evaluate a new submission
within the context of existing research.
To select the appropriate reviewers for each manuscript, the editor needs detailed information.
Please include with your letter your vita. In the letter, please identify which APA journal(s) you
are interested in, and describe your area of expertise. Be as specific as possible. For example,
“social psychology” is not sufficient—you would need to specify “social cognition” or “attitude
change” as well.
Reviewing a manuscript takes time (1–4 hours per manuscript reviewed). If you are selected to
review a manuscript, be prepared to invest the necessary time to evaluate the manuscript
thoroughly.
APA now has an online video course that provides guidance in reviewing manuscripts. To learn
more about the course and to access the video, visit http://www.apa.org/pubs/journals/resources/
review-manuscript-ce-video.aspx.

Sleep Deprivation and Memory Meta-Analytic Reviews of Studies On Sleep Deprivation Before and After Learning

Uploaded by

Copyright:

Available Formats

Sleep Deprivation and Memory Meta-Analytic Reviews of Studies On Sleep Deprivation Before and After Learning

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sleep Deprivation and Memory Meta-Analytic Reviews of Studies On Sleep Deprivation Before and After Learning

Uploaded by

Copyright:

Available Formats

Psychological Bulletin

© 2021 The Author(s) 2021, Vol. 147, No. 11, 1215–1240

Sleep Deprivation and Memory: Meta-Analytic Reviews of Studies

Public Signiﬁcance Statement

Keywords: sleep, sleep deprivation, learning, memory, meta-analysis

Supplemental materials: https://doi.org/10.1037/bul0000348.supp

Effect Size Calculation between-study variance. This allowed us to determine whether it

Study feature k M ± SD Mdn Mode Range

Publication date 130 NA 2012 2015 1994–2020

Meta-Analysis 2: Sleep Deprivation Before Learning

Moderator Variable type df Heterogeneity (Q) p

Recovery sleep (yes vs. no) Categorical 1 10.496 <.001*

Study feature k M Mdn Mode Range

Publication date 55 NA 2010 2000/2020 1989–2020

Figure 7 meta-analysis (M = 54.77%, SD = 20.81%, range = 21.22%–

Figure 9 consolidation theory, hippocampal-dependent declarative memories

Lake. Proquest. https://www.proquest.com/docview/1532140044?pq- Parallel neuropsychological and biochemical investigations. Neuropsy-

Members of Underrepresented Groups:

You might also like