Sleep is a physiological state that has been conserved across evolution, even noted in invertebrates lacking a centralized brain (Hendricks et al., 2000; Shaw et al., 2000; Zhdanova et al., 2001; Raizen et al., 2008; Singh et al., 2014; Nath et al., 2017). Although sleep's physiological functions remain poorly understood, sleep loss has been associated with deleterious effects on health and cognition (Rechtschaffen and Bergmann, 1995; Dinges et al., 1997; Durmer and Dinges, 2005; Spiegel et al., 2005; Banks and Dinges, 2007; Knutson et al., 2007; Grandner et al., 2010). Sleep varies based on previous waking experience (Ganguly-Fitzgerald et al., 2006; Huber et al., 2007; Hanlon et al., 2009; Keene et al., 2010; Beckwith et al., 2017; Kirszenblat et al., 2019; Milinski et al., 2021) throughout the lifespan (Roffwarg et al., 1966; Kales et al., 1967; Feinberg and Carlson, 1968; Cauter et al., 2000; Backhaus et al., 2007; Dijk et al., 2010; Feinberg and Campbell, 2010; Carrier et al., 2011; Vienne et al., 2016; Mander et al., 2017), and between species (Lyamin et al., 2008, 2017, 2018; Siegel, 2008; Lesku et al., 2012), suggesting that sleep has multiple functions. However, because sleep coincides with broad changes in neurophysiology and necessitates a loss of consciousness with reduced responsiveness to external threats, it is likely that sleep evolved, at least in part, to support brain function (Rasch and Born, 2013; Tononi and Cirelli, 2014). Notably, sleep is often elevated during periods of synaptic reorganization, including early development (Roffwarg et al., 1966; Shaw et al., 2000; Kayser et al., 2014), recovery from neural injury (Singh and Donlea, 2020; Stanhope et al., 2020), and memory consolidation (Walker et al., 2002; Ganguly-Fitzgerald et al., 2006). These findings each suggest that sleep supports plastic remodeling in the brain. Synaptic plasticity allows behavioral flexibility in response to external stimuli, and enables the processing and storage of information (Hughes, 1958; Zucker and Regehr, 2002; Cooke and Bliss, 2006). However, the underlying cellular and molecular mechanisms that support plasticity during sleep remain an area of intense investigation.

The impacts of sleep loss, interestingly, vary widely depending on age, environmental conditions, and genotype. While organisms typically recover from acute sleep disruptions relatively quickly, early-life sleep disruptions can prevent developmental plasticity during critical periods and result in long-lasting changes in circuit connectivity and behavior (Frank et al., 2001; Seugnet et al., 2011; Kayser et al., 2014). Conversely, some individuals withstand sleep loss with few consequences depending on the physiological conditions or genetic factors (Viola et al., 2007, 2012; Lyamin et al., 2008; Keene et al., 2010; Thimgan et al., 2010; Donlea et al., 2012; Lesku et al., 2012). In some cases, sleep disruption even provides an opportunity to weaken maladaptive memories (Poe, 2017). Examining the variables that can influence an individual's sensitivity to sleep loss could provide new insights into the core mechanisms of sleep-dependent plasticity. In this review, we will discuss roles for sleep in the maintenance of neural and behavioral plasticity during development, and learning/memory. Finally, we outline ethologically relevant conditions in which organisms can maintain neural and behavioral plasticity in the face of sleep loss.

Synaptic plasticity plays a crucial role in brain development, especially in the refining of neural connectivity through the process of pruning (Paolicelli et al., 2011). Defects in synaptic pruning during development are thought to contribute to atypical circuit function seen in neurodevelopmental disorders (Paolicelli et al., 2011; Konopaske et al., 2014; Tang et al., 2014; Cossío et al., 2017; Kim et al., 2017; Neniskyte and Gross, 2017). Daily sleep amounts peak in many species early in development, when the brain is undergoing significant plastic changes (Roffwarg et al., 1966; Jouvet-Mounier et al., 1970; Shaw et al., 2000; Kayser et al., 2014). Studies in humans have found that sleep disruption during development is associated with severe and lasting consequences for behavior and cognition (O'Brien et al., 2004; Halbower et al., 2006; Ednick et al., 2009). While these human studies provide a correlational link between impaired sleep and later cognition, several lines of animal studies described below indicate conserved roles for sleep in neurodevelopment of several species and begin to identify possible mechanisms by which sleep might influence brain development.

Rapid eye movement (REM) sleep is thought to play a particularly important role in development. Infants spend as much as 50% of their time asleep in REM, compared to 25% in adults (Roffwarg et al., 1966; Jouvet-Mounier et al., 1970). This period of increased REM sleep coincides with heightened formation and elimination of synapses in the developing mouse brain (Marks et al., 1995). Previous work found that REM deprivation, but not non-REM (NREM) deprivation, prevents the elimination of newly-formed dendritic spines in layer V pyramidal neurons in the developing mouse motor cortex (Li et al., 2017). Further, elimination of recent spines during REM facilitates the development of new spines at nearby sites. While most newly formed spines are eliminated, persistent spines are strengthened by REM sleep. Notably, similar findings were observed in the adult mouse brain following motor learning (Li et al., 2017).

A unique feature of REM sleep is the occurrence of myoclonic twitches, or spontaneous, discrete, spastic movements of the limbs (Tiriac et al., 2012; Blumberg et al., 2013; Sokoloff et al., 2020). These twitches occur throughout the mammalian lifespan, but are particularly abundant in infancy (Tiriac et al., 2012; Blumberg et al., 2013; Sokoloff et al., 2020, 2021). The development of myoclonic twitches depends on sensory feedback; the spatiotemporal organization of twitches is disrupted in newborn ErbB2 muscle-specific knockout mice which lack muscle spindles and exhibit impaired proprioception in adulthood (Blumberg et al., 2015). Muscle spindles are sensory receptors that relay changes in the length of muscles to the central nervous system and are necessary for intact proprioception (Kröger and Watkins, 2021). These findings suggest that twitches during sleep provide the developing brain with opportunities to refine immature sensorimotor maps and better coordinate limb movements. Twitching during early-life REM episodes, therefore, could facilitate the transformation of uncoordinated movements during infancy to the fine-tuned sensorimotor maps of an adult. Sensory feedback from twitching limbs are thought to contribute to motor learning and sensorimotor integration (Blumberg et al., 2013, 2020; Sokoloff et al., 2015; Rio-Bermudez and Blumberg, 2018; Glanz et al., 2021), as reafference from myoclonic twitches selectively activates brain regions such as the thalamus, cortex, hippocampus, and cerebellum in infant rats (Khazipov et al., 2004; Mohns and Blumberg, 2010; Tiriac et al., 2012; Sokoloff et al., 2015). Because reafference signals from self-movement are gated during waking, sleep disruptions that interfere with twitching, and their corresponding neuronal activity may disrupt sensorimotor maturation (Tiriac and Blumberg, 2016). While these studies provide an important and promising link between early-life sleep episodes and the development of mature sensorimotor representations, the underlying synaptic mechanisms and long-term consequences of myoclonic twitch disruptions remain to be characterized in detail.

A vital role for sleep in early life plasticity is shared across sensory circuits. The study of ocular dominance plasticity (ODP) induced by monocular deprivation (MD) in cats, for example, is a canonical model of critical period plasticity during development that is reliant upon sleep. During an early critical period for visual development, occluding one eye leads to enhanced visual cortex responses to inputs from the non-deprived eye (Hubel and Wiesel, 1970). Sleep enhances ODP; NREM sleep deprivation prevents enhancement of cortical plasticity, suggesting that sleep is vital for consolidating experience-dependent changes in ocular dominance following MD (Frank et al., 2001). More recent work has found that REM deprivation disrupts cortical plasticity after MD as well, perhaps by disrupting replay-like patterns of activity in the visual cortex (Bridi et al., 2015). Additionally, REM sleep following MD is sufficient to prevent reversal of ODP following subsequent manipulations such as further SD (Bridi et al., 2015), cortical inactivation (Jha et al., 2005), and inhibition of NMDA receptors (Aton et al., 2009). The dependence of ODP on REM sleep parallels studies of sensorimotor development described above, suggesting a vital role for REM sleep in permitting developmental refinement across sensory systems. The consolidation of ODP is also reminiscent of hippocampal memory consolidation during sleep (Diekelmann and Born, 2010; Rasch and Born, 2013). These studies suggest that sleep during development is necessary for the consolidation of plastic changes induced by waking experience, which likely guide appropriate behavioral adaptations to a changing environment. Since ODP (along with other forms of developmental plasticity) occurs during a tightly restricted critical period of development, sleep disruptions early in life could have long-lasting effects on neurophysiology and behavior.

Ontogenetic changes in sleep are conserved; sleep amount and intensity are increased early in life for invertebrates, such as the fruit fly, just as they are in mammals (Jouvet-Mounier et al., 1970; Shaw et al., 2000). In Drosophila, 24 h of sleep deprivation following eclosion leads to long-term learning deficits, whereas adults recover from the same duration of sleep loss after one night of recovery sleep (Seugnet et al., 2011). These chronic learning impairments are likely connected with altered dopamine signaling, and can be dampened either by blocking D1 receptor activity during early life sleep loss or by elevating dopamine signaling during the days after developmental sleep deprivation (Seugnet et al., 2011). Additionally, young sleep-deprived male flies, but not mature flies, show deficits in courtship behavior as adults (Seugnet et al., 2011; Kayser et al., 2014). These courtship deficits are accompanied by decreased size of an olfactory glomerulus associated with perception of social pheromones, caused by impaired developmental growth (Kayser et al., 2014). Similarly, 1 week of early life sleep disruption impairs later social bonding in adult prairie voles (Jones et al., 2019). In this study, sleep disruption occurred during the third and fourth weeks of life, which likely falls during a critical period for maturation of GABAergic circuits that contribute to sensory integration (Gogolla et al., 2014). Notably, early life sleep deprivation in prairie voles leads to an increase in parvalbumin immunoreactivity in the primary sensory cortex, a brain region relevant to social bonding (Jones et al., 2019). Chronic changes in parvalbumin signaling could disrupt sensory processing and social behavior by altering excitatory/inhibitory balance (Yizhar et al., 2011). Together, these studies demonstrate that early life sleep is vital for developmental growth of rapidly growing brain regions across many species, and that disrupted sleep during development can result in lasting effects on adult circuitry and behavior.

While human studies have not yet revealed a mechanistic understanding of how sleep promotes neural and cognitive development, animal models indicate that sleep's role in neurodevelopment is evolutionarily ancient. Model system studies, such as those in flies and mice discussed above, have begun to examine how sleep modulates synaptic connectivity in a variety of developing sensory circuits. Further studies in these systems may reveal interventions that facilitate healthy development during insufficient sleep (Seugnet et al., 2011; Kayser et al., 2014; Jones et al., 2019).

In a variety of species, sleep is required for several stages of memory formation and processing (Walker et al., 2002; Graves et al., 2003; McDermott et al., 2003; Ganguly-Fitzgerald et al., 2006; Seugnet et al., 2008; Krishnan et al., 2016). Indeed, sleep deprivation leads to impaired encoding (Walker et al., 2002; Yoo et al., 2007; Seugnet et al., 2008), consolidation (Graves et al., 2003; Diekelmann and Born, 2010), and retrieval (Gais et al., 2007; Lo et al., 2016; Montes-Rodríguez et al., 2019; Heckman et al., 2020) of recent associations. While even a brief nap restores memory in some assays (Seugnet et al., 2008; Ong et al., 2020), other learning and memory impairments persist after days of recovery sleep (Havekes et al., 2016; Yamazaki et al., 2020; Wu et al., 2021). While it is not clear why recovery from sleep loss varies between these conditions, studies have detected several types of longer-lasting cellular and molecular changes that persist after recovery sleep, including altered gene expression (Gaine et al., 2021), protein synthesis (Tudor et al., 2016; Lamon et al., 2021), and circuit connectivity (Weiss and Donlea, 2021). Interestingly, some types of memories seem to be more vulnerable to sleep loss than others. For example, procedural memories and memories acquired with a conscious motivation or reward benefit from sleep more than declarative or unmotivated memories (Stickgold and Walker, 2007; Diekelmann and Born, 2010). In Drosophila, sleep deprivation disrupts consolidation of appetitive sugar reward memories in fed flies, but in not starved flies (Chouhan et al., 2021). Together, these studies indicate that sleep deprivation likely does not have a universal effect on learning and memory, but varies based on physiological, environmental, and behavioral factors.

While the negative impacts of sleep loss on memory formation are typically detrimental, it is possible that targeted sleep disruption could be used to prevent the consolidation of maladaptive memories. Some studies, for instance, suggest that sleep deprivation could be used following trauma to degrade fear memories in patients with post-traumatic stress disorder (PTSD). Studies by Vanderheyden et al. (2015) compared sleep patterns of rats that were susceptible to developing PTSD-like symptoms after trauma to those that were resilient. While susceptible rats exhibited an increase in REM sleep in the hours following the traumatic event, resilient rats slept little during this period (Vanderheyden et al., 2015). Heightened REM sleep following trauma could lead to consolidation and reactivation of the trauma memory, preventing fear extinction, and resulting in generalization of the fear memory (Poe, 2017). Traumatic events drive activation of the mammalian locus coeruleus (LC) (Passerin et al., 2000; Naegeli et al., 2018), a collection of noradrenergic cells that promote long-term potentiation (LTP) (Izumi et al., 1992; Thomas et al., 1996; Izumi and Zorumski, 1999) and are generally quiescent during REM sleep (Foote et al., 1980). Elevated LC activity during REM sleep following a traumatic event can contribute to enhancement of recently formed emotional memories as seen in PTSD (Wassing et al., 2019). Therefore, behavioral sleep deprivation or pharmacological REM suppression following a traumatic event could lead to interventions to prevent the development of PTSD (Vanderheyden et al., 2014, 2015; Poe, 2017). Conversely, given the importance of sleep in memory consolidation (Rasch and Born, 2013) and emotional processing (Palmer and Alfano, 2017; Tempesta et al., 2018), sleep loss following a traumatic event could prevent consolidation of fear extinction memory in other conditions (Pace-Schott et al., 2015). Recent human studies have produced mixed results (Porcheret et al., 2015; Kleim et al., 2016; Cohen et al., 2017), indicating that the role for sleep in consolidating and/or maintaining traumatic memories varies with context or time elapsed since trauma. Further studies will be required to examine the therapeutic potential of sleep manipulations more clearly.

In many contexts, sleep is vital for individuals to learn and adapt their behavior to best fit their environmental conditions. Sleep facilitates brain development and circuit refinement, and early life disruptions in sleep can result in long-lasting behavioral changes. Throughout the lifespan, sleep also impacts whether new memories can be effectively acquired and consolidated. While understanding the mechanisms that contribute to sleep-dependent plasticity remain an area of intense interest, many studies have already identified molecular and synaptic connectivity changes that occur during sleep to facilitate memory formation. More clearly identifying these mechanisms and developing strategies to manipulate them could open opportunities to support cognitive processing during sleep loss. Finally, individuals exhibit varying responses to sleep loss due to intrinsic and environmental factors. Understanding the benefits and detriments of variations in sleep, as well as the biological basis for inter-individual differences, will help resolve the function(s) of sleep and elucidate how sleep patterns affect future behavior.

JW and JD wrote sections of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

This work had support from NIH/NINDS R01NS105967 to JD and a Jessamine K. Hilliard UCLA Neurobiology Graduate Student Grant to JW.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.