Disrupted Sleep: From Molecules to Cognition

Abstract

Although the functions of sleep remain to be fully elucidated, it is clear that there are far-reaching effects of its disruption, whether by curtailment for a single night, by a few hours each night over a long period, or by disruption in sleep continuity. Epidemiological and experimental studies of these different forms of sleep disruption show deranged physiology from subcellular levels to complex affective behavior. In keeping with the multifaceted influence of sleep on health and well-being, we illustrate how the duration of sleep, its timing, and continuity can affect cellular ultrastructure, gene expression, metabolic and hormone regulation, mood, and vigilance. Recent brain imaging studies provide some clues on mechanisms underlying the most common cause of disrupted sleep (insomnia). These insights should ultimately result in adequate interventions to prevent and treat sleep disruption because of their high relevance to our most prevalent health problems.

SIGNIFICANCE STATEMENT Disruption of the duration, timing, and continuity of sleep affects cellular ultrastructure, gene expression, appetite regulation, hormone production, vigilance, and reward functions.

Introduction

Many insights into why sleep is a highly conserved biological phenomenon can be gleaned from observing the consequences of its disruption due to experimental manipulations or as a result of natural causes such as aging or disease. Given the propensity for modern life to disrupt sleep, whether through behavioral sleep curtailment or sleep fragmentation, and given the mounting epidemiological evidence linking sleep disturbances and disease, a better understanding of the mechanisms underlying the detrimental effects of sleep disruption is clearly needed. In search for answers to fundamental questions on the “essential” (Cirelli and Tononi, 2008), cellular (Cirelli, 2009), multiple (Van Someren, 2010), or nonspecific (Siegel, 2012) functions of sleep, an unabatedly growing number of studies have applied the common methodology of interfering with sleep and observing the consequences. Deprivation can be applied continuously, intermittently, or in a stage-dependent way to either completely or partially deprive organisms of sleep. Consequences differ depending on the modality of experimental sleep disruption. Many studies in the past have described the consequences of acute sleep loss, but the focus has recently been expanded to include chronic sleep disruption, a more common condition in everyday life. Although superficially more subtle than total sleep deprivation (TSD), chronic sleep disruption has far-reaching consequences that will be summarized in the present mini-review, starting from the effects on brain cells and ending with recent insights in the mechanisms involved in the chronically disrupted sleep experienced by people suffering from insomnia, one of the most prevalent disorders. In some cases, negative consequences result from the fragmentation of the normal sleep pattern into short sleep bouts frequently interrupted by brief awakenings, even if the total daily amount of sleep is not decreased. The relevance of findings from experimental studies is supported by observational studies on the consequences of naturally occurring sleep disruption, whether due to environmental and societal demands or pathological conditions such as sleep-disordered breathing or insomnia. The resulting insights lay ground for a mechanistic understanding of the epidemiological finding that disrupted sleep contributes to the major health challenges facing our aging society, including type 2 diabetes, cardiovascular disease, neurodegeneration, and depression.

Effect of disrupted sleep on the ultrastructure of brain cells

One approach to understanding the consequences of chronic sleep restriction is to look at how it affects the brain at the cellular level. For instance, in a recent study, mice were subjected to chronic sleep fragmentation for 14 weeks to mimic the frequent awakenings experienced by patients with sleep apnea. Chronic sleep disruption led to mitochondrial stress, neuronal degeneration, and cell loss in specific groups of neurons important to maintain arousal, including the orexinergic neurons in the hypothalamus and the noradrenergic neurons of the locus ceruleus in the pons (Zhu et al., 2015). Even a month after the end of the experiment, mice were less able to stay awake at night when they are normally more active, suggesting that damage to arousal-promoting cells may be long-lasting (Zhu et al., 2015). Other experiments focused on adolescent mice because of a long-held concern that young brains are more sensitive to the negative effects of sleep loss, even though there is little direct evidence to support this view (Astill et al., 2012; Owens et al., 2014). Recent experiments used electron microscopy to gain insight into the mechanisms by which sleep loss affects brain cells, and specifically pyramidal neurons in the frontal cortex (L. de Vivo, A.B. Nelson, M. Bellesi, J. Noguti, G. Tononi, C. Cirelli, unpublished data). In the cell bodies of these glutamatergic cells, which represent ∼80% of all cortical neurons, the analysis focused on mitochondria and several components of the endocytotic pathway, from early endosomes to lysosomes. These organelles carry out basic cellular functions ranging from energy production and nutrient intake to membrane turnover and prevention of metabolic and oxidative stress. Adolescent mice were studied under four conditions: ad libitum sleep, TSD of 6–8 h, chronic sleep restriction to 1/3 of their habitual sleep for ∼4 d, or recovery sleep for ∼2 d after 4 d of chronic sleep restriction. It was found that there are many quantitative ultrastructural differences in the adolescent mouse cortex that reflect, not only behavioral state (sleep vs wake), but also “quality” of sleep (baseline sleep vs recovery sleep after sleep loss) and duration of wake (acute sleep loss vs chronic sleep restriction). In fact, using a combination of 11 ultrastructural parameters, it was possible to assign individual pyramidal neurons to the correct experimental group with an accuracy of ∼70%. Some of the ultrastructural changes, such as signs of mitochondrial activation, were already present after short sleep deprivation and became more pronounced with chronic sleep restriction. Other parameters, such as signs of lysosomal activation, were only present after prolonged sleep disruption, suggesting that chronic sleep loss imposes a greater burden to brain cells than acute sleep loss. For all of these parameters, the between-mice variance tended to increase after acute and chronic sleep loss relative to sleep. This trend toward higher interindividual variability may reflect true biological heterogeneity unmasked by sleep loss, which is consistent with human studies showing that there are stable, trait-like differences in the susceptibility to cognitive impairment caused by sleep loss (Van Dongen et al., 2004; Rupp et al., 2012).

A time to sleep

Both the amount of sleep and its timing with respect to the internal circadian clock are important. In recent years, the extent and direction of this interrelatedness has been investigated at the behavioral and molecular levels. The circadian phase at which we sleep influences many aspects of sleep. Sleep duration and continuity, the amount of REM sleep, and sleep spindle activity are all modulated by the circadian phase at which we sleep; for example, night or day (Dijk and Czeisler, 1995). Slow waves, putative markers of synaptic strength (Tononi and Cirelli, 2014), have long been thought to be regulated by the duration and intensity of wakefulness and subsequent sleep (Borbély and Achermann, 1999). Some characteristics of slow waves, such as their slope and amplitude, which vary along the anterior–posterior brain axis, are affected by the circadian phase at which we sleep (Lazar et al., 2015). The effects of circadian phase on sleep structure also imply that it matters when you sleep because REM sleep, sleep spindles, and slow waves have all been implicated in functions of sleep such as memory consolidation (for review, see Rasch and Born, 2013). In addition to the timing of sleep, accumulating evidence suggests that there is also a minimum requirement of continuous unperturbed sleep to ensure optimal sleep-supported memory processes (van der Werf et al., 2009; Djonlagic et al., 2012; Varga et al., 2014).

Effect of insufficient or mistimed sleep on rhythmic gene expression

Circadian rhythms are generated by a set of core clock genes and this molecular machinery is present in the suprachiasmatic nucleus (SCN), but also in cells throughout the body (Partch et al., 2014). This machinery drives circadian rhythmicity in an organ- and tissue-specific manner. Forty-three percent of all protein-coding genes show circadian rhythms in transcription (Zhang et al., 2014). Recent animal and human data provide strong evidence that the timing of sleep and sleep deprivation can have a profound influence on this rhythmicity in the peripheral transcriptome (for a recent review, see Archer and Oster, 2015). In mice, sleep deprivation can lead to an 80% reduction in rhythmic transcripts in the brain (Maret et al., 2007). When sleep occurs at night, when temperature is low and melatonin is high, ∼6.4% of the human blood transcriptome is rhythmic. When sleep is delayed by 4 h every day, the percentage of genes that are transcribed rhythmically is reduced to 1% for an overall 84% reduction (Archer et al., 2014). Furthermore, 1 week of insufficient sleep reduces rhythmicity of the blood transcriptome transcriptome by 20% (Möller-Levet et al., 2013). Therefore, sleep timing and sleep debt can influence circadian rhythmicity in the periphery (outside of the SCN). Short or mistimed sleep disrupts the rhythmic expression of core clock genes associated with chromatin modification, transcription, and translation, as well as genes implicated in inflammatory, immune, and stress responses. These alterations could be mechanisms that account for why short and mistimed sleep affects health.

Effect of disrupted sleep on hormones and metabolism

Landmark studies conducted in rats in the late eighties and early nineties demonstrated that extended periods of near TSD produced a reliable syndrome including debilitated appearance, impaired immune function, increased food intake, increased energy expenditure, weight loss, impaired temperature regulation, elevated norepinephrine levels, decreased plasma thyroxine, and eventual death (Rechtschaffen et al., 1989; Everson, 1993; Rechtschaffen and Bergmann, 2002). These studies provided unequivocal evidence that the “function” of sleep extends beyond the brain. Examination of hormonal and metabolic consequences of sleep loss and sleep disruption in these rodent models was largely limited to assays in single blood samples. Whether the main finding from the rat model that sleep deprivation has major adverse effects on health could be extrapolated to conditions of sleep loss in human remained very much in doubt. During the following two decades, sleep disturbances, including insufficient sleep, poor sleep quality, insomnia, obstructive sleep apnea, and mistimed sleep (as occurs in shift workers) became endemic chronic conditions in humans living in industrialized societies. These trends for shorter and poorer sleep developed over the same time period as the dramatic increase in the prevalence of obesity and diabetes. Evidence has accumulated to indicate that sleep loss has important negative health consequences. Prospective epidemiological studies in both children and adults suggest a causative role of short sleep in the increased risk of weight gain. Sleep curtailment is associated with dysregulated neuroendocrine control of appetite that includes a reduction of the satiety factor leptin, an increase in the hunger-promoting hormone ghrelin, and an increase in daytime levels of endocannabinoids (Reutrakul and Van Cauter, 2014). Sleep loss may alter the ability of leptin and ghrelin to signal caloric need accurately, producing an internal misperception of insufficient energy availability that could promote hedonic feeding. The increased hunger induced by sleep restriction leads to increased energy intake, particularly in the form of snacking (Nedeltcheva and Scheer, 2014). Recent studies using whole-room indirect calorimetry have shown that sleep restriction increases energy need, but only modestly (Markwald et al., 2013). In comparison, the increased appetite reported by sleep-deprived subjects and their increased energy intake in the presence of ad libitum feeding appear to exceed the energy demands of extended wakefulness under sedentary conditions. Sleep disturbances, including experimental sleep fragmentation (Tasali et al., 2008; Stamatakis and Punjabi, 2010), insomnia with short sleep duration, sleep apnea, and mistimed sleep, have been linked to abnormal glucose metabolism and increased diabetes and cardiovascular risk. Several large prospective studies suggest that these sleep disturbances are associated with an increased risk of incident diabetes. Obstructive sleep apnea, which combines sleep fragmentation and hypoxemia, is a major risk factor for insulin resistance and possibly diabetes. Whether glycemic control in type 2 diabetes patients can be improved by treating sleep apnea remains controversial. In addition, evidence from animal models and experimental studies in humans have identified disruption of the circadian system as a novel putative risk factor for adverse metabolic outcomes. Short or fragmented sleep decreases insulin sensitivity and increases the risk of type 2 diabetes and obesity due to a decrease in energy expenditure, a dysregulation of the hormonal control of hunger, excessive snacking, and an increased hedonic drive for food intake. The current evidence strongly suggests that these sleep and circadian disturbances may be involved in the epidemic of obesity and type 2 diabetes.

Disrupted sleep affects reward networks in the brain

Although sleep disturbances have adverse neuroendocrine and metabolic effects (see previous section), it is also well known that chronic sleep disruption and sleep deprivation may be associated with severe cognitive performance deficits (see next section). However, impairments in stress coping and emotional regulation often prevail. For example, when healthy human participants engage in risky choices, a night of TSD can heighten the response of brain areas sensitive to reward while dulling responses in areas sensitive to losses (Venkatraman et al., 2011), with major consequences on decision making. How can we explain these deleterious effects of sleep disruption? Neuroimaging, neurophysiological, and clinical studies suggest that emotional and reward networks are activated during sleep. Such activation may promote the reprocessing of emotional or rewarded information during sleep and dreaming and optimize affective regulation and behavioral responses during wakefulness (Perogamvros and Schwartz, 2012, 2054). Supporting this hypothesis, studies in rodents reported that neuronal firing patterns recorded during a reward-searching behavior spontaneously reemerge across hippocampal and reward circuits during subsequent periods of slow-wave sleep. This offline replay is orchestrated by hippocampal ripples and was proposed to strengthen the association between a memory trace and its motivational value (Lansink et al., 2009; Singer and Frank, 2009). In humans, too, the consolidation of motivationally relevant information seems to benefit from oscillatory activity predominating during sleep (sleep spindles and slow-waves) (Wilhelm et al., 2011; Oudiette et al., 2013) and may implicate the activation of the dopamine system during sleep (Feld et al., 2014). By impeding these neural mechanisms, sleep curtailment alters the integration and regulation of reward-based learning and decision making.

In a recent study, K. Igloi, G. Gaggioni, V. Sterpenich, and S. Schwartz (unpublished data) demonstrated that postlearning sleep favors the selectivity of long-term consolidation by retaining the most important (i.e., rewarded) memories. These effects were mediated by an increased functional interplay between the dopaminergic reward regions, the prefrontal cortex and hippocampus. The data further showed that sleep spindles strengthened memory representations based on reward values, suggesting a privileged replay of information yielding positive outcomes. These findings demonstrate that postlearning sleep determines the neural fate of motivationally relevant memories and promotes a value-based stratification of long-term memory stores.

An important, unanswered question is whether reward influences the types of neural patterns replayed during sleep in humans. V. Sterpenich, H.D. Yang, M. van Schie, M. Catsiyannis, A. Ramyead, S. Perrig, D. Van De Ville, and S. Schwartz (unpublished data) used simultaneous EEG–fMRI and neural decoding methods (Horikawa et al., 2013) and showed that spatiotemporal patterns of brain activity associated with an experience that led to a successful outcome (winning a game) had a higher probability of reemerging during slow-wave sleep. Collectively, these new data indicate that sleep disruption has deleterious consequences on a healthy emotional balance by hindering a form of neural reverberation that predominates during sleep and helps our brains learn from our daily failures and accomplishments.

Predicting individual differences in cognitive performance after sleep loss

Sustained attention is a cognitive domain that is robustly affected by sleep loss (Lim and Dinges, 2010; Lo et al., 2012). Degraded neurobehavioral performance shows marked trait-like individual differences (Van Dongen et al., 2004; Rupp et al., 2012), so identifying persons who are more susceptible to vigilance decline without having to expose them to sleep loss could be useful in reducing the negative consequences of sleep restriction (e.g., accidents).

Drift diffusion modeling (DDM) of psychomotor vigilance test response times collected after a night of habitual sleep can predict the rate of behavioral lapsing induced by a night of TSD (Patanaik et al., 2014). DDM models perceptual decision-making by assuming that the brain extracts evidence for the presence of a stimulus at a constant rate over time (drift) that is perturbed by noise (diffusion) (Ratcliff and Van Dongen, 2011). This accumulation stops once enough evidence has been gathered to reach a decision. Vulnerable participants have a slower drift rate, but this is masked by their slightly faster nondrift contribution to response time. Therefore, greater vulnerability is not apparent from merely observing response times. DDM has been shown to predict TSD-induced vigilance decline with data collected at different times of day and the method has reproducibility across different samples (Patanaik et al., 2015).

Autonomic and fMRI markers for individual differences in sensitivity to sleep loss

Three other physiological predictors of vulnerability to vigilance decline after TSD have been identified. Heart rate variability (HRV) is an indicator of autonomic function that reflects the balance between time-varying parasympathetic and sympathetic influences on heart rate. HRV at the low-frequency band (0.02–0.08 Hz) closely tracks the time course of the rate of behavioral lapsing over a 24 h period of sustained wakefulness (Chua et al., 2014a). Persons more vulnerable to vigilance decline after TSD have a slower resting heart rate and a higher HRV relative to those who are less vulnerable (Chua et al., 2014b), which could reflect higher vagal tone in such individuals. Because HRV is also lower in individuals reporting a greater frequency and duration of daily worry, it is possible that persons more resistant to vigilance decline after TSD tolerate high homeostatic sleep pressure because of hyperarousal.

With task-based fMRI, behavioral decline across several attention-demanding tasks is reliably correlated with the attenuation of task-related frontoparietal activation after TSD. This suggests that brain areas mediating attention control go offline to a greater extent when vulnerable persons are sleep deprived (Chee, 2014). Critically, activation appears to be more attenuated at all levels of task difficulty after TSD in these individuals, even with correct trials. Such individuals may thus have lower redundancy of neural activation when in a well-rested state and, as a result, become less well able to buffer the effects of local sleep when sleep deprived.

These findings might lead one to expect that greater task activation in the well-rested state will predict less vulnerability. However, the notion that less vulnerable persons have greater cognitive reserve as reflected by higher task activation appears to hold only for experiments involving working memory (Caldwell et al., 2005; Mu et al., 2005; Chee et al., 2006). Why state-related shifts in activation are a more reliable marker of vulnerability remains to be determined.

A limitation of task-related fMRI is that inferences are restricted to the areas recruited during task performance. In contrast, at least in theory, task-free (or resting-state) fMRI can assess the functional integrity of multiple brain regions and networks with minimal participant engagement. Sleep-deprived persons show functional connectivity changes resembling those observed during light sleep (Sämann et al., 2011), consistent with the occurrence of microsleeps and slow wave activity intrusions. Two sets of network changes characterize the altered functional connectivity that accompanies TSD: (1) reduced within-network connectivity in functionally related cortical regions and (2) reduced segregation of normally functionally segregated (or antagonistic) cortical networks (Sämann et al., 2010; De Havas et al., 2012). Reduced functional connectivity of anticorrelated networks in the rested state appears to be useful for predicting vulnerability to vigilance decline after TSD, although these predictive regions differ slightly from those affected by sleep deprivation (Yeo et al., 2015).

Individual differences in experiencing curtailed and disrupted sleep

Much of the work discussed above concerns the responses to experimentally induced sleep disruption. However, in the absence of any perturbation originating in the environment, sleep can also be disrupted spontaneously. Individual differences occur, not only in the sensitivity to sleep loss, but also in the probability to experience such endogenously curtailed or fragmented sleep. Recent population-based studies suggest that these individual differences are not without consequences for brain function. In particular, people who have fragmented sleep—rather than people with short sleep duration—have increased risks of cerebral small vessel disease and mortality (Zuurbier et al., 2015a; Zuurbier et al., 2015b), as well as of poor cognitive and emotional functioning (Luik et al., 2015a; Luik et al., 2015b). The risks may in part be counteracted by interventions to promote adequate sleep–wake rhythms, especially at advanced age, when the endogenous circadian regulation of sleep weakens (Swaab et al., 1996; Van Someren and Riemersma, 2007).

However, behavioral genetics studies suggest that endogenous sleep problems may not be completely reversible. Heritability estimates of sleep variables range from 20–40% for habitual sleep duration to >90% for sleep EEG-derived parameters (Landolt, 2008). The strength of genetic predispositions is underwritten by the surprising finding of a twin study demonstrating that, despite night-to-night variability, one's subjective evaluation of problems encountered during a single night is in part genetically determined (Boomsma et al., 2008). Molecular genetics studies have started to elucidate the genes involved. Polymorphisms in genes encoding clock and GABA receptor proteins have been proposed to provide the genetic underpinnings of the heritability of chronic insomnia, the most common of all endogenous sleep problems (for review, see Palagini et al., 2014).

For a number of reasons, it is relevant to the present review to examine chronic insomnia in some more detail. First, despite its name, insomnia is not so much a matter of sleep deprivation as it is fragmentation and variability in timing of sleep that appeared so relevant in the topics reviewed above (Vallières et al., 2005; Van Someren, 2007; Buysse et al., 2010; Feige et al., 2013). Second, insomnia is not only the most prevalent sleep disorder, but also the second-most prevalent mental disorder (Wittchen et al., 2011), and, because of its high prevalence, the associated societal costs are extremely high (Gustavsson et al., 2011). Third, insomnia contributes significantly to the risk or severity of cardiovascular, metabolic, mood, and neurodegenerative disorders (for review, see Bassetti et al., 2015). Fourth, a closer look at insomnia will illustrate the complexity of disentangling causes versus consequences of disturbed sleep, as well as their endogenous versus exogenous components.

In addition to the genes encoding for proteins with key roles in the sleep-regulatory mechanisms mentioned above, the heritability of insomnia has been proposed to involve a polymorphism in the gene encoding the 5HTTLPR serotonin transporter (Harvey et al., 2014). Carriers of the s/s allele are more likely to experience disturbed sleep in response to a stressful experience. It is important to note that this predisposition to “endogenous” insomnia involves a pathway that does not primarily serve a sleep-regulating role and is not specific to insomnia; it is also implicated in individual differences in stress responses, depression, anxiety, and posttraumatic stress disorder. Even though insomnia is commonly associated with these conditions, not all people suffering from insomnia are anxious or depressed.

We considered that the neural correlates of insomnia occurring comorbidly with depression or anxiety are not necessarily the same as the neural correlates of insomnia in those without affective symptoms. Preliminary findings of a voxel-based morphometry MRI study within a group of depressed patients suggest that those with a low density of gray matter in the thalamic pulvinar suffer the most severe insomnia (Van Someren et al., 2014). The association was specific to depressive patients and not present in healthy controls or in two patient control groups. Conversely, MRI scans in volunteers who were rigorously screened to exclude those with even slightly elevated symptoms of depression or anxiety pinpointed no involvement of the pulvinar, but rather a role for white matter integrity in individual differences in the strength of spindles and slow oscillations (Piantoni et al., 2013) and a role of orbitofrontal cortical (OFC) gray matter density in the severity of complaints about early morning awakening and insomnia (Altena et al., 2010; Stoffers et al., 2012). Supported by psychometric findings (Raymann and Van Someren, 2008; Diaz et al., 2013), these findings suggest that individuals with a low hedonic capacity are more likely to experience disrupted sleep, possibly because of a reduction in OFC output signaling contentment, a prerequisite for the brain to disengage from environmental monitoring and action preparation, reduce cortical excitability, and give in safely to the unconscious state of sleep (van der Werf et al., 2010; Stoffers et al., 2014).

Two major conclusions can be derived from these examples. First, disrupted sleep does not necessarily involve only primarily sleep regulatory circuits. Second, disrupted sleep is best regarded as a final common path of which the neural correlates differ depending on nonsleep phenotypes such as depression. Multivariate phenotyping appears essential to disentangle the different circuits that can be involved in insomnia (Van Someren et al., 2014).

Discussion and perspective

The presented findings show how research on the consequences of sleep disruption has matured over the last decades. The topics addressed in this mini-review are not exhaustive, not in the least because novel insights in adverse effects of keep emerging. A notable recent example concerns a possible role of disrupted sleep in neurodegeneration. Using real-time two-photon imaging in mice brains, Xie et al. (2013) showed a sleep-related increase in the interstitial space that markedly facilitated convection of interstitial and CSF, which could enhance removal of possibly neurotoxic waste including beta-amyloid. We expect the next decade to bring a wealth of highly relevant neuroscientific and clinical findings on this newly discovered role of sleep.

An emerging consensus from the findings reviewed here is that fragmented sleep may be at least as detrimental as curtailed sleep. Therefore, an outstanding question to be addressed in the next decade is why equal amounts of continuous versus interrupted sleep have such different outcomes. The wealth of studies on the favorable effects of sleep for brain function during the last decade may have led us to presume that any sleep, even fragmented, is better than no sleep. However, it might be that some processes that occur during sleep are so slow or require so many steps that premature interference results in unfinished states that adversely affect brain function. The possibility that sleep can also have adverse effects is already suggested by the amazing improvement of mood in some severely depressed people if they are deprived from their fragmented, incoherent sleep; at the same time, disturbed sleep may render individuals more vulnerable to stress and precipitate depressive episodes. Adverse effects may also occur with mistimed sleep. For example, a recent rodent model of posttraumatic stress disorder suggested that fear extinction is impaired if REM sleep occurs while the locus ceruleus is still activated by the recent fearful experience (Vanderheyden et al., 2014).

Another major question pertains to similarities and differences across the spectrum of a single acute sleep disruption to repeated disruption or even lifetime disruption. What happens on the trajectory from acute to chronic sleep disruption? Are compensatory/rescue mechanisms activated? What are the processes underlying adaptation and why do some processes fail to adapt? Equivocal results have been reported from studies that sought to determine whether allostatic adaptation to restricted sleep occurs (Kim et al., 2007; Leemburg et al., 2010; Van Someren, 2010; Deurveilher et al., 2015). If allostatic adaptation can truly be accomplished, how does it build up, what are the underlying mechanisms, and where are its limits?

In summary, it is clear that sleep disruption interferes profoundly with normal function, from the subcellular level to complex behavior, and with strong impacts on health and well-being.