SEARCH

SEARCH BY CITATION

Keywords:

  • adenosine;
  • nitric oxide;
  • sleep homeostasis

Abstract

  1. Top of page
  2. Abstract
  3. What is sleep and how it is measured
  4. Theories of why we have to sleep
  5. Physiological significance of sleep and sleep need
  6. Regulation of wakefulness and sleep
  7. Modulation of sleep
  8. Problems with sleep
  9. Treatment of common sleeping problems
  10. Conflict of interest
  11. References

The state of sleep consists of different phases that proceed in successive, tightly regulated order through the night forming a physiological program, which for each individual is different but stabile from one night to another. Failure to accomplish this program results in feeling of unrefreshing sleep and tiredness in the morning. The program core is constructed by genetic factors but regulated by circadian rhythm and duration and intensity of day time brain activity. Many environmental factors modulate sleep, including stress, health status and ingestion of vigilance-affecting nutrients or medicines (e.g. caffeine). Acute sleep loss results in compromised cognitive performance, memory deficits, depressive mood and involuntary sleep episodes during the day. Moreover, prolonged sleep curtailment has many adverse health effects, as evidenced by both epidemiological and experimental studies. These effects include increased risk for depression, type II diabetes, obesity and cardiovascular diseases. In addition to voluntary restriction of sleep, shift work, irregular working hours, jet lag and stress are important factors that induce curtailed or bad quality sleep and/or insomnia. This review covers the current theories on the function of normal sleep and describes current knowledge on the physiologic effects of sleep loss. It provides insights into the basic mechanisms of the regulation of wakefulness and sleep creating a theoretical background for understanding different disturbances of sleep.

Disturbances of sleep are among the most common causes to seek medical counselling. The feeling of unrefreshing sleep, difficulty to fall asleep and early morning awakenings are typical complains. Common experience, as well as experimental studies, evidences severely compromised cognitive and social performance after a prolonged period of wakefulness and/or unrefreshing sleep. Recent epidemiological research has identified associations between many common diseases and short/long sleep: among them cardiovascular diseases and type II diabetes. These findings have been largely confirmed in experimental set-ups. Although it has become clear that sleep plays an essential role in our well-being and performance, we still do not know why exactly we have to sleep – what does sleep do to our brain and body? What, however, has become clear is that sleep is a precisely regulated, complicated process consisting of periods of synchronized cortical activity flanked by periods of excessive activity when, while in sleep, cortex reaches the level of waking activity. The events in brain are accompanied and coordinated with physiological changes in the rest of the body, creating a network of programmed physiological activities during each night. Any disturbances in the coordination of this programme will disturb sleep, and thus, it is not surprising that a spectrum of different sleep problems has arisen.

What is sleep and how it is measured

  1. Top of page
  2. Abstract
  3. What is sleep and how it is measured
  4. Theories of why we have to sleep
  5. Physiological significance of sleep and sleep need
  6. Regulation of wakefulness and sleep
  7. Modulation of sleep
  8. Problems with sleep
  9. Treatment of common sleeping problems
  10. Conflict of interest
  11. References

Sleep is a behavioural state that has been identified in all species so far carefully examined (Cirelli & Tononi 2008). It is usually characterized by immobility, typical sleeping posture and reduced sensory threshold, which leads to diminished ability to communicate with the surroundings. As a behavioural state, sleep is ‘global’, meaning that it is a synchronized state enclosing the entire body. In this regard, sleep and wake are also mutually exclusive: a healthy organism is either awake or asleep. However, when we use more sophisticated methods to examine brain activity during sleep, it becomes evident that sleep has different phases, and thus, it is not a single state but rather a process that advances through the night.

A truly reliable definition of vigilance state can be obtained by measuring brain electrical activity with electroencephalography (EEG). This applies mainly to mammals and birds that have fairly developed brains. In EEG, waking is characterized by low-amplitude, high-frequency waves. Sleep consists of two main phases: non-rapid-eye-movement sleep (NREM) and rapid-eye-movement (REM) sleep. During NREM sleep, the amplitude of EEG waves increases and the frequency decreases, while in REM sleep, EEG is indistinguishable from that in waking. These states can be separated based on muscle tone (measured using electromyography, EMG) and saccadic eye movements (measured using electro-oculography, EOG): in waking, muscle tone is high; in NREM sleep, it decreases; and in REM sleep, it practically disappears. During REM sleep, eyes undergo characteristic rapid movements, of which the state has got its name. A finer division of NREM sleep is based on the proportion of the low-frequency, high-amplitude waves (called slow-wave activity, SWA). NREM sleep is divided to three stages (S1, S2 and S3) in increasing order of SWA (Fig. 1). Sleep stages alternate in the course of the night in a regular manner: sleep starts by S1 and deepens via S2 to S3 and then proceeds to REM sleep. After the REM sleep period, the cycle starts from the beginning. The duration of one sleep cycle is about 90 min. The sleeps stages calculated across the night are often presented as a hypnogram, which describes the order and duration of each sleep stage (Fig. 2). During SWA sleep, virtually all cortical neurones are engaged in a slow (<1 Hz) oscillation consisting of alternating ON and OFF periods (also called UP and DOWN states). During ON periods, cortical cells are depolarized and fire action potentials at high rate, while during OFF periods, cells are hyperpolarized and silent.

image

Figure 1. EEG stages of human sleep. Human EEG recording showing the different sleep stages: waking and REM sleep characterized by low-amplitude, high-frequency waves and NREM stage 1 through 3 in the order of decreasing frequency and increasing amplitude of the waves.

Download figure to PowerPoint

image

Figure 2. Hypnogram. Hypnogram shows the distribution of sleep stages across the night and presents the cyclic structure of sleep. Blue colour: NREM (stages 1–3), hatched area: REM sleep.

Download figure to PowerPoint

A more detailed view of the vigilance-state characteristics can be obtained by EEG power spectral analysis in which the broadband EEG signal is divided into specific frequency bands using fast Fourier transformation, and the ‘power’, that is, the proportion or the amount of waves in each band, is calculated. The commonly used spectral bands in sleep research are as follows: the low delta (0.5–4 Hz) band characteristics of NREM sleep, the theta (4–9 Hz) band visible in REM sleep and waking, the sigma (7–15 Hz) band consisting of the sleep spindle oscillations and the fast gamma (30–60 Hz) band pronounced in active waking. The exact boundaries between frequency bands are not absolute but vary between species, method of recording and the study question. Spectral analysis of EEG has proved to be a valuable tool when designing faster and more reliable vigilance-state scoring methods (Borbely et al. 1981, Gath & Bar-On 1983). Moreover, as the neurophysiological mechanisms underlying EEG oscillations are fairly well characterized, the use of spectral analysis in basic research has had a great impact for our understanding of the brain mechanisms and processes that regulate sleep.

Interestingly, a recent study (Vyazovskiy et al. 2011) in rats showed that during extended wakefulness, local ensembles of cortical cells enter a slow oscillation during behavioural wakefulness and that these short periods of slow oscillation correlated with lapses in performance. This could be one of the mechanisms underlying the well-known adverse effects of sleep deprivation on cognitive performance. Similar findings have been made also in humans (Nobili et al. 2012). This phenomenon could be nominated as ‘local sleep’ as opposed to behaviourally defined ‘global sleep’.

Many physiological changes accompany vigilance states: NREM sleep is characterized by decreases in heart rate, blood pressure, breathing rate and core body temperature. In REM sleep, muscle tone in peripheral muscles decreases dramatically – only the muscles innervated by cranial nerves and muscles utilized for breathing stay active, all other skeletal muscles are paralysed. The functions of the autonomic nervous system are lost: temperature regulation is unable to follow ambient temperature, heart rate and blood pressure, as well as breathing is irregular. The cause for these physiological changes is unknown.

Sleep is thus a complex, highly coordinated process, which synchronizes changes in brain activity to changes in the activity of the autonomic nervous system and the muscle tone.

Theories of why we have to sleep

  1. Top of page
  2. Abstract
  3. What is sleep and how it is measured
  4. Theories of why we have to sleep
  5. Physiological significance of sleep and sleep need
  6. Regulation of wakefulness and sleep
  7. Modulation of sleep
  8. Problems with sleep
  9. Treatment of common sleeping problems
  10. Conflict of interest
  11. References

The key question ‘why do we have to sleep’ has remained unanswered, but we have a fairly good idea of the probable explanations. Many ideas are based on the knowledge gained from sleep deprivation experiments: what happens to the brain and body when we do not sleep enough. There is a general agreement that neuronal activity during the waking state drives the ‘sleep pressure’, which is the force that makes us to sleep. However, the exact aspects and mechanisms of this activity have remained unclear.

The main theories concerning the purpose of sleep can be divided into three categories: (i) energy metabolism related, stating that sleep is needed to restore energy resources of the body, (ii) inflammation/threat related, stating that staying awake initiates defence reactions in the body, and (iii) neural plasticity related, stating the restoration of synaptic homeostasis, underlying learning and memory functions, requires sleep. There is more experimental data to support each of these theories, and interestingly, the same results can often be used to support several of them.

Energy metabolism-related theories

All physiological processes need energy in the form of ATP, which can be produced from several sources. The fastest sources are creatine phosphates, then glucose and glycogen and finally fatty acids. In the brain, lactate derived from astrocytes is used as a fast energy source under strong neural activation (Magistretti et al. 1993, Brown & Ransom 2007).

It has been suggested that the lack of energy is a core factor that forces the physiological functions to silence themselves, as in sleep (‘The energy hypothesis of sleep’, Benington & Heller 1995, Scharf et al. 2008). Experimental evidence shows that, indeed, reductions in energy expenditure such as glucose and ATP utilization take place during sleep (Jung et al. 2011, Dash et al. 2013). In addition to saving energy, sleep may also serve to replenish the brain's glycogen stores and to synthesize large molecules (Kong et al. 2002, Scharf et al. 2008). Glycogen and adenosine are the key molecules that couple energy shortage to sleep pressure and sleep (recently reviewed by Scharf et al. 2008, Porkka-Heiskanen & Kalinchuk 2011).

Brain metabolic rate is coupled to neuronal and synaptic activity (Attwell & Gibb 2005), with most of the energy needed to sustain glutamatergic neurotransmission. Studies on cortical neuronal firing rates suggest that the firing, and therefore also energy consumption, is highest during waking, REM sleep and the depolarized up-states of NREM sleep (Madsen et al. 1991a,b, Maquet 1995). Changes in metabolic demand would therefore reflect the homeostatic sleep process with prolonged wakefulness consuming excessively energy. This was demonstrated by studies showing that (i) cortical neurones increase firing in the course of waking (Vyazovskiy et al. 2009) and (ii) that this increase in cortical activity during waking is paralleled by increases in the brain's lactate level (Urrila et al. 2003, Wigren et al. 2009, Dash et al. 2012). Also, many of the genes involved in energy metabolism and synaptic potentiation are upregulated during waking (Cirelli et al. 2006).

The phenomenon of ‘local sleep’ can be easily explained through the energy depletion hypothesis. The idea is that neurones that have been most active during waking need the most sleep to recover – this ‘use-dependent’ sleep would thus be locally induced and form the core of sleep need. Several experiments both in humans and animals give support to this idea, showing that local activation during waking will induce more slow waves and delta power at the same location during subsequent sleep (Vyazovskiy et al. 2000, Huber et al. 2006, Vyazovskiy & Tobler 2008, Hanlon et al. 2009, Landsness et al. 2009, Nir et al. 2011). When waking is prolonged, the brain's energy resources become restricted, and there is a gradual shift in energy balance such that energy demand exceeds the production (Benington & Heller 1995). Adenosine concentration increases locally in BF in the course of wakefulness (Porkka-Heiskanen & Kalinchuk 2011), which has been used as one argument to support the energy depletion theory (see also 'Adenosine').

(Immune) defence-related theories

Accumulating evidence suggests that sleep regulation and immune responses are linked. The pro-inflammatory cytokines interleukin-1 beta (IL-1β) and tumour necrosis factor-alpha (TNF-α) are somnogenic and promote NREM sleep even in the absence of immune challenge, while anti-inflammatory molecules inhibit sleep (Krueger 2008). In humans, the plasma levels of TNF-α have also been shown to covariate with sleep intensity (NREM SWA) (Pollmacher et al. 1995, Mullington et al. 2000). Experimental studies with humans show that adequate sleep helps to fight infection (Cohen et al. 2009) and improves antigen titres after vaccination (Lange et al. 2003, 2011). On the other hand, immune challenges such as influenza and Escherichia coli infections increase the amount and intensity of NREM sleep (Toth & Krueger 1989, Toth et al. 1993, Fang et al. 1995). The increase in sleep following immune challenge is thought to be mediated mainly by the pro-inflammatory cytokines in response to pathogen recognition and subsequent inflammatory cellular pathways (Zielinski & Krueger 2011). Interestingly, evolutionary increases in mammalian sleep durations are strongly associated with enhanced immune defence as measured by the number of immune cells circulating in peripheral blood. Thus, sleep may not only help to recover from infections, but also boosts the immune system's resistance to parasites (Preston et al. 2009).

Many immunological parameters show circadian rhythmicity. Disruption of the circadian rhythm imp-airs the immune responses by desynchronizing the body's circadian clocks (Bollinger et al. 2010) (see also 'The circadian system and melatonin'). It also increases the susceptibility to chronic inflammatory diseases (Narasimamurthy et al. 2012). Sleep loss induces changes in the expression of genes linked with immunity (Moller-Levet et al. 2013) and increases the production and release of pro-inflammatory cytokines, such as IL-1 beta and TNF-alpha (Opp et al. 1992, van Leeuwen et al. 2009). C-reactive protein (CRP) levels increase during experimental sleep deprivation (Meier-Ewert et al. 2004, van Leeuwen et al. 2009), while epidemiological studies evidence association with high CRP levels and short sleep (Meier-Ewert et al. 2004, Haack et al. 2007, van Leeuwen et al. 2009, Martinez-Gomez et al. 2011). The potential activation of the immune system may help to explain the associations between several major diseases, such as cardiovascular diseases and type II diabetes with deviant sleep duration. In extreme conditions, rats totally deprived of sleep die of a systemic invasion of bacteria, suggesting that sleep is crucial in facilitating recovery from bacterial infections (Everson & Toth 2000).

Brain plasticity-related theories

Neuronal plasticity is the foundation of behavioural adaptability, learning and memory. Experimental evidence, as well as personal experiences, witnesses that sleep enhances memory and learning (Diekelmann & Born 2010), while extended waking or poor sleep impairs cognitive performance (Lo et al. 2012). The mechanisms by which sleep causes these effects are less clear but several hypotheses have been put forward for explanation.

During waking, the brain receives enormous amount of information, only a fraction of which is finally selected for long-term storage. As the flow of information to the cortex is stopped by closing the thalamic gate (see also 'Thalamus'), sleep provides an optimal time for off-line processing of memory traces created during the waking period. In addition, the specific sleep oscillations (slow waves, spindles and hippocampal ripples) are well suited to modulate plasticity (Sadowski et al. 2011). Also, the neuromodulatory milieu (see also 'Regulation of wakefulness and sleep') is different from that of waking, potentially contributing to the special communication between brain areas during sleep. Encoding of new information takes place in waking, while sleep is beneficial for memory consolidation: a process that transforms new unstable memory traces acquired during waking into stable form and integrates them with pre-existing memories (reviewed in Diekelmann & Born 2010, Stickgold & Walker 2007).

Sleep-dependent memory consolidation

The relationship between sleep and memory is complex: some forms of memory benefit more from sleep than others and differently from different sleep stages. These findings have been conceptualized into two hypotheses: (i) the dual-process hypothesis stating that NREM sleep benefits declarative (hippocampal-dependent) memory, while REM sleep is important for non-declarative (hippocampus-independent) memory and (ii) the sequential hypothesis assuming that undisturbed sleep cycles, in which NREM sleep is followed by REM sleep, are needed for sleep-dependent memory consolidation for both types of memory (declarative and non-declarative) (Diekelmann & Born 2010). Studies where the actual neuronal representations of memory traces during the acquisition phase in waking have been recorded also during sleep have provided important information on the sleep-dependent memory consolidation. These studies show that rather than sleep stages per se, it is the sleep oscillations that are important in coordinating the re-activation and redistribution of memory traces. An example of this is the replay of freshly acquired hippocampal memory traces and their subsequent import into cortex during coordinated hippocampal ripples and cortical slow oscillations (Sirota et al. 2003, Ji & Wilson 2007).

Synaptic homeostasis hypothesis

The synaptic homeostasis hypothesis (SHY) takes a more fundamental view on the relationship between sleep and plasticity by proposing that the universal function of sleep is synaptic homeostasis, that is, the renormalization of synaptic strength (and number) to a sustainable level after global net increase brought up by waking (Tononi & Cirelli 2006, 2012). In relation to sleep, the hypothesis states that during waking, more synapses are formed, and during sleep (particularly NREM), they are downscaled (Tononi & Cirelli 2006).

According to SHY, waking leads to a net/global increase in synaptic strength/number. As the maintenance of synapses is expensive in terms of space, energy and cellular supplies, and a progressive increase in net synaptic strength would saturate the potential for further strengthening, synaptic strength needs to be returned to sustainable level. This renormalization of synaptic is best achieved during sleep, which provides optimal conditions for unbiased off-line processing. Importantly, global downscaling of all synapses during sleep does not affect the relative synaptic strengths, in which information is coded, but only re-establishes the ability to make for stronger and new synapses in waking. SHY further proposes that sleep homeostasis, the core evidence of the importance of sleep (see also 'Sleep homeostasis and the two-process model of sleep regulation'), can be explained by the increase in synaptic strength in waking and the decrease in sleep. As the maintenance of stronger synapses consumes more energy (Harris et al. 2012), SHY does not contradict with the idea that sleep is needed for the replenishment of brain energy stores.

Another aspect of the hypothesis is that synaptic strength and number should influence the amplitude (and slope) of the SWA, which is an EEG marker of both neuronal synchrony and homeostatic sleep pressure. Supporting this, neuronal synchrony has been shown to be directly related to the number of synaptic connections in a network (Riedner et al. 2007, Vyazo-vskiy et al. 2009). In addition, SHY predicts that SWA is locally regulated in an activity-dependent manner, for which there is experimental evidence (Huber et al. 2004, Hanlon et al. 2009, Lesku et al. 2011). Finally, the neuronal firing pattern of sleep slow waves may be causally involved in downscaling of net synaptic strength, the actual mechanisms by which this is achieved still remains to be elucidated.

SHY has been derived from a more general concept, the synaptic scaling as response to neural activity. This concept states that synaptic strength increases when neuronal activity decreases and vice versa (Turrigiano 2008). Accordingly, a recent publication showed that during NREM sleep, the brain responsiveness was rather upscaled than downscaled, as measured by evoked potentials (Chauvette et al. 2012). An additional aspect was introduced by experiments, which indicated that synaptic downscaling takes place during REM sleep rather than NREM (Grosmark et al. 2012). It is thus clear that a number of details need to be worked out in connection with SHY.

Physiological significance of sleep and sleep need

  1. Top of page
  2. Abstract
  3. What is sleep and how it is measured
  4. Theories of why we have to sleep
  5. Physiological significance of sleep and sleep need
  6. Regulation of wakefulness and sleep
  7. Modulation of sleep
  8. Problems with sleep
  9. Treatment of common sleeping problems
  10. Conflict of interest
  11. References

How much sleep do we need? A simple definition of sufficient sleep is a sleep duration that is followed by a spontaneous awakening and leaves one feeling refreshed and alert for the day. Sleep needed is individual, largely genetically determined (Partinen et al. 1983, Watson et al. 2012), but also age and sex have influence. Children need more sleep than adults; sleep need decreases with age from around 14 h a day at 1 year of age to around 9.5 h a day by age 12, after which the amount of sleep needed remains the same until age 18 when it drops to the adult level of 7–8 h a day (Williams et al. 2013). Older people tend to sleep less than young (Ohayon et al. 2004), which could reflect either reduced sleep need or reduced ability to obtain the sleep that is needed, although the biggest differences in sleep with ageing are found in sleep quality rather than quantity. Gender differences in sleep duration are also significant. Women sleep on average 20 min longer than men (Kronholm et al. 2006). Women also suffer from a lack of sleep more than men, partly because women's sleep tends to be lighter and more easily disturbed than men's (Reyner et al. 1995).

The need for sleep varies dynamically according to duration of previous wakefulness, health status, stress and many other factors. The homeostatic regulation of sleep, compensating for a previous sleep loss, in healthy individuals adjusts the duration and intensity of sleep to match the physiological need for sleep (see also 'Sleep homeostasis and the two-process model of sleep regulation'). It gives us the freedom to stay awake for longer periods when necessary and then regain the lost sleep when possible without detriment. However, if sleep is permanently curtailed, either voluntarily or forced by environmental factors, performance and health consequences become apparent.

What happens if we do not sleep enough?

Sleep duration has been on the decline for the past three decades (Kronholm et al. 2008, Knutson et al. 2010) irrespective of ethnical background. Hectic western life style, shift work and poor sleep habits contribute to a sleep debt, which can have severe consequences to our health and well-being. Recent research has evidenced that in addition to brain functions and performance, the functions of the rest of the body will also suffer from restricted/bad-quality sleep.

Insufficient sleep causes serious attention lapses and performance deficits (Dinges et al. 1997, Van Dongen et al. 2003) and predisposes to mood disorders including depression (Paunio et al. 2009, Utge et al. 2010, 2011). In addition, learning and memory are compromised (Stickgold & Walker 2007, Lim & Dinges 2010). Epidemiological studies have associated short sleep duration with increased risk of mortality (Cappuccio et al. 2010) as well as metabolic and cardiovascular diseases, including type 2 diabetes, hypertension and coronary heart disease (Buxton & Marcelli 2010, Kronholm et al. 2011). These findings are largely supported by experimental sleep deprivation studies where sleep restriction has led to adverse metabolic and cardiovascular events, such as imbalance in the insulin-to-glucose ratio and increased heart rate and blood pressure (Van Cauter et al. 2008, van Leeuwen et al. 2009, 2010, Buxton et al. 2012). Increasingly, people cut their sleep short during weekdays and catch up with sleep during the weekend. However, at least in children such irregular sleep schedules, especially in combination with short sleep duration, associate with adverse metabolic outcomes (Spruyt et al. 2011). Also, there is doubt whether weekend is enough to recover from the sleep debt incurred during the week. A recent sleep restriction study simulating a working week and followed by recovery weekend found that the production of proinflammatory cytokines, including IL-1 beta, IL-6 and IL-17, remained elevated even after the recovery period, accompanied by increased heart rate and serum CRP (van Leeuwen et al. 2009). An association between short sleep duration and obesity has also been reported (Spruyt et al. 2011, Knutson 2012). Sleep deprivation increases appetite, leading to increased food consumption (Spiegel et al. 2004, van Leeuwen et al. 2010). This observation is partially explained by changes in the levels of the hormones leptin and ghrelin, which regulate satiety and appetite respectively. Ghrelin has been reported to increase with sleep deprivation, whereas results concerning leptin are more mixed with some studies reporting an increase in leptin level and others a decrease (Spiegel et al. 2004, van Leeuwen et al. 2010). The increase in leptin level during sleep deprivation suggests the development of leptin resistance, a phenomena often seen in individuals with high BMI. Sleep loss activates the sympathetic nervous system, increasing both heart rate and blood pressure, in part through increased levels of the stress hormones cortisol (Spiegel et al. 1999) and noradrenaline (Dettoni et al. 2012). As stress affects the immune system and reduces its ability to fight antigens, it is no surprise that sleep restriction also modulates the immune functions both directly and through the body's stress systems.

Regulation of wakefulness and sleep

  1. Top of page
  2. Abstract
  3. What is sleep and how it is measured
  4. Theories of why we have to sleep
  5. Physiological significance of sleep and sleep need
  6. Regulation of wakefulness and sleep
  7. Modulation of sleep
  8. Problems with sleep
  9. Treatment of common sleeping problems
  10. Conflict of interest
  11. References

For summary, see Fig. 3.

image

Figure 3. Overview of the sleep regulatory areas and neurotransmitters. The nuclei of the brain stem, hypothalamus and the basal forebrain as well as thalamus send neuronal projections to cortex and, through release of neurotransmitters, regulate the cortical excitability. Red colour indicates excitatory projections/neurotransmitters, blue inhibitory.

Download figure to PowerPoint

Thalamus

Thalamus gates sensory information to the cortex. The basic thalamic processing is handled by a neural circuitry consisting of the thalamic (relay) neurone, the cortical neurone and the thalamic reticular nucleus neurone. This circuit regulates the generation of cortical EEG delta and spindle activity (Steriade et al. 1991). Information through the gate is relayed when the circuit is in tonic activity mode, elicited by depolarization of the thalamocortical cells by different neurotransmitters during waking and REM sleep. When the circuit adopts rhythmic, synchronized activity (0.5–4 Hz), information flow decreases and the typical features of sleep EEG (spindles, SWS) appear (McCormick & Bal 1994). Thus, the rhythmic bursting of thalamic cells creates a state in which the sleeper does not react to external stimuli, or the reactivity is greatly reduced. This reversible disconnection from the environment is regarded as one of the most important features of sleep.

Waking nuclei

Cortical activity is made possible by a complicated, overlapping system that here is called the waking system. The system consists of small neuronal groups (‘the waking nuclei’) in brain stem, pons, hypothalamus and basal forebrain. The cortical projections of these nuclei regulate cortical excitability: the more neurotransmitter is released, the higher excitability. The secretion of the neurotransmitters decreases in NREM sleep. In REM sleep, the secretion of noradrenaline (Aston-Jones & Bloom 1981), serotonin (McGinty & Harper 1976) and histamine (Vanni-Mercier et al. 2003) decreases further, almost stopping, but that of acetylcholine returns to the level of waking secretion (Vazquez & Baghdoyan 2001).

The noradrenergic neurones are located in the locus coeruleus. Novel stimuli and stress activate the cells of this nucleus best (Aston-Jones & Bloom 1981). The noradrenergic neurones have a special role in neural plasticity (Cirelli et al. 2005). The serotoninergic system consists of many groups of neurones located in the brain stem and pons, of which the dorsal raphe group is mostly indicated in regulation of sleep. Sensory stimulation and stress activate these neurones (Heym et al. 1982). Of the many serotonin receptor subtypes, the 5-HT2 receptors have significance in sleep regulation, because their antagonists are able to increase slow-wave sleep (Dugovic & Wauquier 1987). Another serotonin receptor that is involved in sleep regulation is the autoreceptor, 5-HT1A (Bjorvatn & Ursin 1998). The cholinergic system consists of two neuronal groups: one located in the brain stem (LDT/PPT) and the other in the basal forebrain (BF) (Mesulam et al. 1983). Pontomesencephalic nuclei (LDT/PPT) send projections to the thalamus, hypothalamus and basal forebrain (Steriade et al. 1988, Losier & Semba 1993). The basal forebrain (BF) is a heterogeneous region in the ventral region of forebrain. The magnocellular cholinergic neurones provide most of the cholinergic cortical innervation (Jones 2004). The cholinergic cells in this area are mixed with other cells, particularly GABAergic and presumably glutamatergic neurones, which also project to neocortex (Gritti et al. 1998). The BF cholinergic neurones are also wake/REM active: their discharge rate is highest during waking and REM sleep and lowest during slow-wave sleep (Lee et al. 2005) and they have an important role in cortical arousal. Basal forebrain appears to have a special role in the regulation of recovery sleep (see also 'Adenosine'). The histaminergic cells lie in the tuberomammillary nuclei in the posterior hypothalamus projecting diffusely to all brain regions (Panula et al. 1990). The histamine levels during wakefulness remain at constant level and do not accumulate in the course of prolonged wakefulness, indicating that they do not carry information on the duration of the wakefulness (Zant et al. 2012). The orexin/hypocretin cells are situated in the lateral hypothalamus and send projections to other wake-promoting brain areas (Peyron et al. 1998). This system is necessary for the maintenance of consolidated periods of waking (Kilduff & Peyron 2000). The connection between the orexigenic system and sleep regulation became apparent when two groups identified it as target for narcolepsy (Sakurai et al. 1998, Lin et al. 1999).

In addition to the waking nuclei, the ubiquitous neurotransmitter, glutamate, is generally wake-promoting. Glutamate excites thalamocortical relay cells but inhibits thalamic reticular cells (McCormick 1992), thus promoting the tonic activity mode and flow of sensory information through the thalamus to the cortex. Part of the BF cortically projecting cells are presumably glutamatergic, and possibly through these projections, administration of NMDA or AMPA into the BF increases wakefulness (Wigren et al. 2007). Glutamate is also involved in the regulation of REM sleep (Luppi et al. 2012).

Sleep nuclei. The ventrolateral preoptic nucleus (VLPO)

Most neurones in the brain cells decrease their firing rate in sleep, but in the hypothalamus, there are small nuclei that are more active during sleep (NREM) than during waking: ventrolateral preoptic area (VLPO) (Sherin et al. 1996) and the median preoptic nucleus (MnPN) (Szymusiak et al. 2001). These nuclei are called sleep nuclei, and they send projections to all wake-active nuclei (Steininger et al. 2001), which also send projections to the sleep nuclei. This reciprocal inhibition system is the basis for the regulation of transition from sleep to wakefulness and back: when waking-active neurones are on, they inhibit the sleep nuclei and vice versa (McGinty & Szymusiak 2000). It has also been called the ‘flip-flop’ model (Saper et al. 2005). Lesion of the VLPO cells reduced both NREM and REM sleep (Lu et al. 2000).

The circadian system and melatonin

Most physiological functions express rhythms, circadian, diurnal, monthly or annual. Sleep–wake cycle is a typical circadian rhythm. The circadian rhythm is generated by the activity of the suprachiasmatic nucleus (SCN), a group of neurones situated above the optic chiasm in the brain. While there has been rapid advance in understanding of the molecular machinery that produces the base rhythm in the SCN (Siepka et al. 2007), the output signals are less well understood. The main projections from the SCN target hypothalamus, explaining the pathways that regulate the rhythmic secretion of many hormones and temperature rhythms. Whether the circadian regulation of sleep–wake cycle is mediated through hypothalamus is not known, but as hypothalamus hosts many sleep regulatory cell groups (see 'Regulation of wakefulness and sleep'), it is probable. The circadian rhythm, associated with the dark–light cycle, tells the body when it is a proper time to sleep: the body temperature decreases and many metabolic processes slow down. In humans, the sleeping time is the dark period, but in many other species, it is the light period.

The period of the inner rhythm(s) generated by the SCN is not exactly 24 h but usually slightly longer. This means that the rhythm has to be entrained to the solar time every day (Czeisler et al. 1999). The main entrainer of the circadian rhythms is light but also other cues, including feeding time, exercise and melatonin, can be used (Mistlberger & Skene 2005), particularly if light is not available for entrainment. Correct entrainment is essential for the coordination of the inner clock to the surrounding world.

Melatonin is secreted from the pineal gland during the dark phase of the day. The secretion starts upon reduced light in the evening and stops in the morning light, providing brain with information of the duration of the light period (Shaw 1977, Shanahan et al. 1997). This signal is used by annually reproductive species to entrain their reproductive activity to a proper time of the year. During inactive reproduction periods, the gonads shrink, and if the species goes to torpor, metabolic activity in the body decreases (Reiter 1980).

Melatonin production is regulated by SCN, but there is also a feedback mechanism so that melatonin can act as internal synchronizer of circadian rhythms, when light information is not available, for example in blind people (Skene & Arendt 2007).

Sleep homeostasis and the two-process model of sleep regulation

A core feature of sleep is its homeostatic regulation. This means that a prolonged waking period is followed by a prolonged period of sleep, called recovery sleep. Sleep during this recovery period is also deeper, containing more SWS than normal sleep (Borbely 1982). But how does the brain keep count on the duration of wakefulness? One possibility, suggested more than a hundred years ago, is that some molecules, ‘hypnotoxines’, as they were called, during wakefulness accumulate in the brain and induce sleep. During the last 20 years, a number molecules that could act as ‘hypnotoxines’ have been characterized.

Adenosine

Adenosine, the core of adenosine triphosphate (ATP), remains when the three phosphates are removed. As such it has an intimate connection with energy metabolism: increase in extracellular adenosine concentration signals for energy depletion (Van Wylen et al. 1986), but adenosine has also many other roles in physiology. In the central nervous system (CNS), it is an inhibitory neurotransmitter when acting through A1 receptors. The energy hypothesis of sleep (see 'Sleep-dependent memory consolidation') uses adenosine as its core argument: during (prolonged) wakefulness, neuronal activity consumes (excessively) ATP leading to energy depletion, which releases adenosine to extracellular space. There adenosine binds to its receptors, although A1 receptors inhibits neuronal activity. When the activity of wake-promoting neurones (like the cholinergic neurones) is inhibited, cortical activity declines and sleep can be initiated.

Several experiments have provided evidence to support this hypothesis. Adenosine agonists administered either systemically or into the brain increase sleep, while adenosine antagonists, including caffeine, decrease it (Radulovacki et al. 1984). Extracellular adenosine levels are higher during waking than during sleep, and in the basal forebrain (BF), they increase during prolonged wakefulness, increasing sleep (=recovery sleep) (Porkka-Heiskanen et al. 1997). However, if A1 receptors are blocked during the prolonged wakefulness period, sleep does not increase (Gass et al. 2009), showing that the recovery sleep is indeed induced by adenosine through A1 receptors. The waking-active BF cholinergic neurones are tonically inhibited by adenosine (Rainnie et al. 1994), and this adenosine-mediated reduction in cortical activation enables the cortex to enter SWS (reviewed in Porkka-Heiskanen & Kalinchuk 2011, Brown et al. 2012). During recovery sleep, adenosine level decrease back to basal level, thereby restoring cortical activation and the waking state (Porkka-Heiskanen et al. 1997). Interestingly, adenosine levels during prolonged wakefulness increase only very locally in the BF, restricting to the area where the cholinergic cells lie. If the cholinergic cells are specifically destroyed using IgE–saporin, prolonged wakefulness will no more increase adenosine levels, and no recovery sleep is induced (Kalinchuk et al. 2008), indicating the BF cholinergic cells play a key role in the regulation of sleep homeostasis. In humans, genetic variations in the adenosine deaminase enzyme that metabolizes adenosine modulate both the duration and intensity of SWS (Retey et al. 2005, 2007).

Nitric oxide

Nitric oxide, a gaseous neurotransmitter, has similar effects on sleep as adenosine: NO concentration increases in the BF during prolonged wakefulness, experimental increase in NO concentration in the BF induces sleep and blocking its increase during prolonged wakefulness prevents induction of recovery sleep (Kalinchuk et al. 2006a,b). Surprisingly, this increase is induced by inducible nitric oxide synthase (iNOS) (Kalinchuk et al. 2006b), which normally is not present in the brain but is induced by inflammatory challenge or stress (Nathan & Xie 1994, Moro et al. 1998). This implies that prolonged wakefulness is able to trigger defence responses, including immunological defence.

Cytokines

Cytokines act mainly as mediators of immune responses, but also other physiological functions are regulated by them, including stress. Of the cytokines, IL-1β and TNF-α have been thoroughly studied in connection with sleep (Krueger et al. 2011). Their concentration in different brain areas increase during (prolonged) wakefulness and their administration increases sleep (Krueger et al. 2001). It has been suggested that they are involved in the increased sleep in connection with infections (Pollmacher et al. 2000). Prolonged wakefulness induces, in addition to iNOS, also other (immune) defence-related molecules, like heat shock proteins (Terao et al. 2003), CRP (Meier-Ewert et al. 2004) and unfolded protein response (Naidoo 2009). In extreme conditions, rats totally deprived of sleep die of a systemic invasion of bacteria, suggesting that sleep is crucial in facilitating recovery from bacterial infections (Everson & Toth 2000).

Prostaglandins

Of the many prostaglandins, prostaglandin D2 (PGD2) has been indicated in the regulation of sleep homeostasis. Administration of PGD2 into the subarachnoidal space (in close vicinity of the BF/hypothalamic area) increases NREM sleep (Hayaishi 2000). PGD2 releases adenosine, and it has been suggested that its actions on sleep are mediated via adenosine, particularly via adenosine A2 receptors (Scammell et al. 1998).

GABA

GABA is the ubiquitous inhibitory neurotransmitter in the brain. In regulation of sleep, three main sites of action have been discussed: the retinothalamic nucleus, the basal forebrain and the ventrolateral preoptic nucleus (see Fig. 3). GABA may participate also in the regulation of REM sleep (Luppi et al. 2006). The importance of GABA neurones in sleep regulation is suggested by the notion that the most frequently used hypnotics, benzodiazepines, act through the GABA receptor (Mohler 2009).

Modulation of sleep

  1. Top of page
  2. Abstract
  3. What is sleep and how it is measured
  4. Theories of why we have to sleep
  5. Physiological significance of sleep and sleep need
  6. Regulation of wakefulness and sleep
  7. Modulation of sleep
  8. Problems with sleep
  9. Treatment of common sleeping problems
  10. Conflict of interest
  11. References

Ageing

Ageing is associated with poor sleep quality, including increased sleep fragmentation, attenuated amplitude of the diurnal sleep–wake rhythm, and decreased sleep amount and intensity (Cajochen et al. 2006). Age-related changes in the mechanisms that regulate sleep as well as undiagnosed medical conditions or excessive medication may explain this.

During ageing, the circadian pacemaker, SCN, loses progressively its ability to produce precise rhythms, leading to disrupted circadian cycles with reduced amplitude (Hofman & Swaab 2006). The phase relationship between sleep and core body temperature is also altered so that the temperature minimum occurs substantially earlier in the major nocturnal sleep period (Cajochen et al. 2006). This well-documented circadian phase advance, together with a weaker arousal signal in the evening, promotes earlier sleep timing in older individuals. The secretion of melatonin may decreases with ageing, although this view has been challenged (Zeitzer et al. 1999).

One of the prominent changes with ageing is the increased vigilance-state fragmentation (Dijk et al. 2010). Together with increased brain arousal during sleep, it leads to involuntary awakenings, thereby reducing sleep efficiency and also causing inadvertent sleep episodes during waking (Klerman et al. 2004). Both the increase in fragmentation and increase in cortical arousal could be, at least partly, explained by changes in the hypocretin system of the hypothalamus (see 'Waking nuclei'), which regulates vigilance-state transitions and arousal. Animal studies have revealed an age-related decrease in the hypocretinergic tone accompanied by a gradual loss of hypocretin neurones (Porkka-Heiskanen et al. 2004, Sawai et al. 2010). SWA decreases with age in both humans (Landolt et al. 1996) and animals (Mendelson & Bergmann 1999, Rytkonen et al. 2010). Whether this reflects reduced build-up of homeostatic sleep pressure with ageing, reduced ability to sleep or problems producing slow waves, is currently not known (Dijk et al. 2010). Recent studies suggest, however, that the age-related decrease in SWA is most likely caused by reduced homeostatic sleep pressure (Duffy et al. 2009, Wigren et al. 2009, Dijk et al. 2010, Rytkonen et al. 2010), although the underlying molecular mechanisms are not well understood. The reduction could be due to changes in the homeostatic sleep factors (Duffy et al. 2009, Rytkonen et al. 2010) or due to reduced synaptic density and connectivity (Carrier et al. 2011).

Stress

The hormones of the stress axis (CRH secreted in the hypothalamus, ACTH secreted in the pituitary gland and corticosteroids secreted by the adrenal gland) promote wakefulness (Steiger 2002). As long as the levels of these hormones in the body are high, initiation of sleep is not possible. The modern way of life has increased the number of people suffering of chronicle stress, and accordingly, stress is a prevalent self-reported cause of insomnia in working age population (Ohayon 2002, Kim et al. 2011, Kompier et al. 2012). The biological basis of this hyperarousal state is unclear, but the hormones of the stress axis as well as neurotransmitters of the waking system most probably play an important role (Bonnet & Arand 2010).

Problems with sleep

  1. Top of page
  2. Abstract
  3. What is sleep and how it is measured
  4. Theories of why we have to sleep
  5. Physiological significance of sleep and sleep need
  6. Regulation of wakefulness and sleep
  7. Modulation of sleep
  8. Problems with sleep
  9. Treatment of common sleeping problems
  10. Conflict of interest
  11. References

Circadian rhythm-related problems

The inability to entrain the inner clock to the environment (e.g. in jet lag, shift work, blind people) results in desynchronization, or misalignment of the rhythms, which means that the inner, circadian clock shows different time than the solar clock. Thus, a person ends up in sleeping during the daytime, when the temperature and the cortisol secretion are high, while the melatonin secretion is low. Ageing decreases the amplitude and exact timing of the rhythms, inducing the physiological weakening of the circadian signal for sleep (Van Cauter et al. 1996).

Jet Lag

When crossing several time zones, the internal rhythm stays as it was in the departure spot, but the external time is different creating desynchronization of the rhythms. Re-entrainment starts to take place as soon as the person is exposed to daylight at right time point. The limit for re-entrainment is about one hour per day, even when the light pulse is timed correctly.

Symptoms of jet lag include severe problems with sleep: difficulties in falling asleep, waking up prematurely and in wrong time of the day, feeling of unrefreshing sleep and daytime sleepiness. In healthy persons, these symptoms disappear as soon as re-synchronization to the new time zone is complete. Bright light exposure at correct time of the day, as well as correctly timed melatonin ingestion, is helpful in re-installing the circadian rhythms (Czeisler 1995, Arendt & Skene 2005), but light exposure at inappropriate time may slow the re-entrainment (Czeisler 1995).

Shift Work

As shift work is common in many service professions, an increasing number of people work at irregular hours. Desynchronization of the internal and external rhythms is an inevitable consequence of shift work. This will results in compromised performance, decreased vigilance and attention levels, sleep problems and metabolic disturbances. The conflict between the resting state of the body metabolism and the need for energy consumption during the night shift creates a discrepancy that will increase health problems. As a consequence, shift workers have increased risk for coronary disease, gastric ulcers and some cancers (Knutsson et al. 1986, 2013, Knutsson & Boggild 2010).

Seasonal affective disorder (winter depression) (SAD)

This form of depression is tied to day length: when days start to shorten in fall and winter, the depressions starts. Symptoms include increased appetite, increased need for sleep and weight gain. It has been suggested that the decreased amount of light would be insufficient to synchronize the circadian rhythms and thus results in desynchronized state, but there is little evidence to show that the condition would be induced by internal desynchrony (Van Dongen et al. 1998, Reid et al. 2000).

Waking system-related problems

The common problem is over activation: the thalamic gate stays open because of either voluntary or forced prolonged sensory stimulation. Increase in secretion of the neurotransmitters of the waking system, as well as that of the stress hormones, contributes to this hyperarousal state, which is typical for insomnia (Buysse 2013). Low levels of orexin or its receptors are associated with narcolepsy – a serious neurological disease that is characterized by inability to produce stabilized periods of either sleep or wakefulness, resulting in awakenings during night and sleep periods, including intrusion of REM sleep (with atonia) into wakefulness. The dramatic muscle atonia periods during wakefulness are often triggered by positive emotions (Choo & Guilleminault 1998).

Sleep homeostasis-related problems

Too much sleep pressure

A common problem is curtailment of sleep either voluntarily or because of work or social conditions that force the person to reduce sleep. The accumulation of sleep pressure is normal, but because of too short sleep period, its dissipation is insufficient, leaving residual sleep pressure in the brain. This pressure keeps accumulating night after night until normal sleep period topped with sufficient recovery sleep period is allowed (Van Dongen et al. 2003, Banks et al. 2010). As a result, the performance in many tasks suffers, and the worsening of the performance is proportional to the amount of accumulated sleep pressure. In addition to individual suffering and lost opportunities, this type of sleep curtailment forms severe risk for the society. Particularly, professional tasks that include shift work in safety-related positions, like control room work or road traffic, are vulnerable for compromised performance because of shortage of sleep. Several disasters and traffic accidents have been evaluated to have originated from a human error due to lack of sleep (Philip et al. 2010).

Too little sleep pressure

Ageing reduces the accumulation of sleep pressure, presumably by affecting the mechanisms of sleep homeostasis (Rytkonen et al. 2010). The accumulation can also suffer from reduced brain activation during the day (Wigren et al. 2007), resulting in reduced need to sleep. These two conditions may co-exist in old people's homes, if the inhabitants are not provided with activities during the day, resulting in early morning awakenings and unnecessary medication for ‘sleep disturbance’.

Caffeine

Caffeine is an adenosine receptor antagonist and through this action promotes wakefulness (Nehlig et al. 1992). The sensitivity for caffeine varies markedly between individuals. Recently, genetic variations in genes related to adenosine metabolism have introduced at least partial explanation to this variability (Retey et al. 2005, 2007). The desired refreshing effect of caffeine is also its curse on sleep, inducing more sleep problems. It can be noted that not only coffee in its different forms but also many soft drinks contain considerable amounts of caffeine. The main effects of caffeine on sleep are decreased sleep latency, shortened total sleep time, decrease in power in the delta range and sleep fragmentation. Caffeine may also decrease the accumulation of sleep propensity during waking, thus inducing long-term harmful effects on sleep quality (Landolt et al. 2004).

Treatment of common sleeping problems

  1. Top of page
  2. Abstract
  3. What is sleep and how it is measured
  4. Theories of why we have to sleep
  5. Physiological significance of sleep and sleep need
  6. Regulation of wakefulness and sleep
  7. Modulation of sleep
  8. Problems with sleep
  9. Treatment of common sleeping problems
  10. Conflict of interest
  11. References

The treatment of sleep problems should be based on careful evaluation of the cause of the problem. The recommended treatment is behavioural, combined with short medication period, if needed. The medication period should always be short and take place under supervision of medical staff.

Sleep hygiene and cognitive behavioural therapy

Sleep hygiene aims at trying to make circumstances good for sleep, mostly by giving simple advice that will help to restore natural conditions for sleep. The regular daily rhythm is the core condition for sleep success: waking always at the same time of the day and going to bed equally. De-stressing in the evening, before going to bed, forms another key rule for sleep hygiene. Stopping electric media one hour before the aimed sleep time helps to shut the thalamic gate and lowering the stress hormone level, including noradrenaline. The same effect is gained by avoiding hard physical activity shortly before bed time. Silence and pleasant temperature in the bed room further help in calming down the mind (Stepanski & Wyatt 2003). Cognitive behavioural therapy has proved to be effective for particularly insomnia patients but also in treatments of other sleep disorders (Sanchez-Ortuno & Edinger 2012).

Medication

The widely used GABA receptor modulators (typically benzodiazepines) promote falling asleep but the medicine-induced sleep does not have the normal distribution of sleep states, particularly, the amount of SWA and REM sleep are decreased (Borbely et al. 1985). Another classical target neurotransmitter, serotonin, may provide (e.g. the 5-HT2A receptor) new drugs for the treatment of insomnia (Caliendo et al. 2005). Relatively new drugs are melatonin agonists, which act as chronobiotics (Carpentieri et al. 2012). Orexin receptor-based drugs are under development and will most probably be on market soon. Histamine receptor antagonists have been used as sedatives, particularly for children, and development of drugs continues (Krystal et al. 2013). Also many psychosis-indicated drugs (typically olanzapine) have been used also as sleep-inducing drugs with much smaller doses than when used to treat psychosis (Khazaie et al. 2013). All sleep-promoting medication is advised to be used temporarily, because there may be considerable side effects, like addiction and memory lapses (particularly benzodiazepines) (Vermeeren & Coenen 2011).

References

  1. Top of page
  2. Abstract
  3. What is sleep and how it is measured
  4. Theories of why we have to sleep
  5. Physiological significance of sleep and sleep need
  6. Regulation of wakefulness and sleep
  7. Modulation of sleep
  8. Problems with sleep
  9. Treatment of common sleeping problems
  10. Conflict of interest
  11. References