Research Article: Functions and Mechanisms of Sleep

Date Published: February 23, 2017


Author(s): Mark R. Zielinski, James T. McKenna, Robert W. McCarley.


Sleep is a complex physiological process that is regulated globally, regionally, and locally by both cellular and molecular mechanisms. It occurs to some extent in all animals, although sleep expression in lower animals may be co-extensive with rest. Sleep regulation plays an intrinsic part in many behavioral and physiological functions. Currently, all researchers agree there is no single physiological role sleep serves. Nevertheless, it is quite evident that sleep is essential for many vital functions including development, energy conservation, brain waste clearance, modulation of immune responses, cognition, performance, vigilance, disease, and psychological state. This review details the physiological processes involved in sleep regulation and the possible functions that sleep may serve. This description of the brain circuitry, cell types, and molecules involved in sleep regulation is intended to further the reader’s understanding of the functions of sleep.

Partial Text

Sleep occurs in every organism to some extent, indicating its physiological importance [1]. Most sleep researchers agree that a single function of sleep is not a realistic view, as will become evident as this review documents sleep’s essential role in many vital physiologic functions including development, energy conservation, brain waste clearance, and modulation of immune responses, cognition, performance, disease, vigilance, and psychological conditions [2,3]. Sleep has been characterized in many species from humans, birds, fish and flies (e.g., Drosophila Melanogaster) to simpler organisms such as worms (e.g., C. Elegans) [4]. Sleep may be disadvantageous, for the animal is less vigilant to potential predators, and sleep prohibits consumption of food and procreation. However, animals are not constantly under such pressure. Perhaps, sleep exists, in part, due to the necessity in maintenance of the aforementioned physiological processes at enhanced levels, in order to increase the animal’s ability to survive and propagate. Abundant breakthroughs in our understanding of the basic mechanisms of sleep regulation have occurred within the past 150 years. Nevertheless, sleep is regulated by molecules and pathways that are redundant and also serve other physiological functions, which has made our understanding of the function(s) of sleep and sleep-related pathologies arduous. Herein, we describe how sleep is regulated globally, regionally, and locally, by means of cellular and molecular mechanisms. This description may then serve to suggest how these processes dictate the many functions of sleep (Figure 1).

The classic definition of sleep is generally based upon physiological characteristics observed in mammals including reduced body movement and electromyographic activity, reduced responsiveness to external stimuli, closed eyes, reduced breathing rates, and altered body position and brain wave architecture assessed by polysomnography (Table 1). However, non-mammalian wakefulness and sleep are often measured by simpler parameters such as decreased relative movement activity, and by rest. Overall, the definition of mammalian sleep relies on activity and metabolism in relation to the electrical brain signals obtained in the electroencephalogram (EEG). Sleep states are typically determined in animals by both the level of muscle activity in the electromyogram (EMG) and EEG characteristics. In humans, sleep states are usually more discernable than in lower animals, including more specific characteristics of EEG architecture. For example, human sleep is classified with 3 defined sub-states of non-rapid eye movement sleep (NREMS) including N1, N2, and N3, which are associated with increasing depth of sleep slower EEG waves, and rapid eye movement sleep (REMS) [5]. Human sleep cycles between NREMS and REMS for approximately 90 minutes for about four to five times during the night in an ultradian cycle [6]. Typically, human sleep is deeper in the beginning of sleep and REMS encompasses a greater proportion of the sleep cycle as sleep persists.

Pioneering work demonstrating neural electrical activity provided the foundation of our understanding of sleep states, sleep phenomena, and the molecular mechanisms that regulate sleep [7]. In 1877, Canton recorded electrical impulses from the surface of the cerebral cortex of both rabbits and monkeys using a galvanometer [8]. This finding was independently confirmed by Danilewsky in 1877 [9]. In the 1880s, Ferrier and Yeo examined cortical localization in apes by means of electrical brain stimulation [10,11]. Thereafter in 1890, the physiologist Beck demonstrated that spontaneous rhythmic cortical oscillations in rabbits and dogs were modulated by light, the circadian entrainment signal [12]. Later in 1912, Pravdich-Neminsky reported the first EEG pictorial recordings using a string galvanometer in dogs [13]. He also observed evoked potentials in the cortex of dogs, again suggesting a crucial role of the cerebral cortex in the generation of the EEG. Consequently, this led to detection of EEG recordings in human sleep by Berger in 1924 [14].

Electrical brain activity occurs, in part, from ionic current changes within neurons and, to a lesser extent, some types of glia [18–20]. EEG signals are largely the product of synchronized synaptic currents generated by the apical dendrites of pyramidal neurons [21]. Nevertheless, intrinsic membrane properties, neuronal firing, and glial activity likely also contribute to the EEG signals. EEG frequency bands provide essential information of how brain regions, cells, and molecules regulate wakefulness, sleep states, and display dysfunction due to related pathologies (Table 2).

Certain physiological functions, such as immune response, development, and cognition, also play a role in regulation of NREMS and REMS, by means of influencing vigilance state-regulating brain regions and networks. For example, research indicates that sleep, especially REMS sleep, is necessary for proper neural development in neonate, termed the ontogenetic REM sleep hypothesis [51]. Sleep loss during development can reduce brain mass, induce neuronal cell death, and increase risk of eventual behavioral problems [52]. One hypothesis of sleep and memory suggests that NREMS and REMS contribute to different types of memory [53]. NREMS appears to be more important for declarative memory whereas REMS is involved more in procedural memory. Others have hypothesized via the sequential hypothesis that both NREMS and REMS work together to enable proper memory functioning. Regardless, animal studies of sleep and sleep loss indicate that proper cognition is dependent upon a complex interaction of brain regions, neuronal and glial interactions, and related molecular mechanisms involving neuroplasticity- and inflammation-related molecules [54]. Recently, Tononi and colleagues proposed the synaptic homeostasis hypothesis (SHY), in which sleep is proposed to selectively attenuate synaptic strength between neurons [55]. Molecular, electrophysiological, and behavioral findings from this group suggest that synaptic strengthening during wake demands energy, and sleep may promote synaptic weakening, ridding the brain of unimportant information, as well as allowing re-establishment of energy reserves and attenuation of cellular stress. Other sleep researchers, though, have argued that there are numerous examples of synaptic plasticity/strengthening occurring during both sleep/wake states and that the SHY hypothesis oversimplifies complex cellular and molecular processes. [56]. Regardless of this dispute, sleep appears to serve a function of re-sculpting the synaptic landscape and energy reserve restoration.

Changes in brain electrical activity, reflected in altered EEG frequency bands, are mediated by molecules that influence ion concentrations, such as potassium (K+) and Ca2+. Ions possess electrical charges that determine the ability of neurons to signal themselves and other cells [98]. The opening and closing of ion channels allow ions to pass through the cell’s plasma membrane, which in turn modulate the membrane potential and, consequentially, may produce action potentials. Ion channel kinetics are regulated by several mechanisms including phosphorylation and dephosphorylation of the protein, the coupling to a messenger molecule such as a neurotransmitter (ionotropic receptors), and by a conformational change in the membrane potential (i.e., voltage-sensitive channels) [99]. Excitatory post-synaptic potentials (EPSPs) occur due to the temporary depolarization of postsynaptic membrane potentials induced by positively-charged ions entering the postsynaptic cell from an opening of their ionic channel. Inhibitory postsynaptic potentials (IPSPs) occur due to negative ionic cellular influx, including chloride, or positively-charged ion efflux. When the postsynaptic potential evokes an action potential it is excitatory, which is typically induced by presynaptic neurons releasing neurotransmitters. The summation of the EPSP and IPSPs of groups of neurons results in hyperpolarized (i.e., down-states) and more depolarized (i.e., up-states) neurons, in turn affecting the EEG and sleep [100].

Cajal first revealed the structure of synapses in his histological work over 100 years ago, and speculated that neurons release signaling molecules that are involved in physiological functions [113]. Neurotransmitters that tend to be excitatory, such as acetylcholine, dopamine, norepinephrine, histamine, serotonin, hypocretins (also known as orexin), neuropeptide S, and glutamate, typically enhance arousal or wakefulness [21]. Arousal-related cholinergic and monoaminergic neurons have been extensively investigated though pharmacologic, genetic, and lesion studies, as well as the recent use of transgenic models.

A wide literature indicates that enhanced waking activity occurring during prolonged wakefulness modulates the immune system [2]. In fact, the sleep status of an animal can affect the ability of an animal to respond to infection and wound healing. Furthermore, impaired sleep is evident in diseases that involve enhanced inflammation, including cancer and type 2 diabetes. Pioneering work done by Ishimori and Pieron at the turn of the 20th century led to an understanding of how humoral factors regulate sleep [127,128]. These researchers independently discovered that dogs injected with cerebral spinal fluid from sleep-deprived dogs exhibited enhanced sleep amounts. Their findings led to a search for a single molecule regulating sleep that was termed sleep promoting factor S. At Harvard University, Pappenheimer, Krueger, and colleagues identified factor S in the brain and urine of rabbits and cats as muramyl peptide—a peptidoglycan component of bacteria [129,130]. This led to the identification of numerous humoral factors including cytokines and hormones, which are activated by waking activity and pathogens, regulating sleep. These sleep-promoting humoral factors are activated, in part, through their respective pattern recognition receptors, thereby providing a link between the immune system and sleep regulation.

Clear evidence indicates that neurons, particularly pyramidal neurons, are important regulators of sleep and EEG activity due to the role they play as the primary electrical conducting cells in the brain [164]. Recent evidence, however, indicates that glia, the most prevalent cell type in the brain, including microglia, astrocytes, and oligodendrocytes, may also play a substantial role in sleep regulation. Glia can express sleep-regulatory neurotransmitters and their respective receptors including GABA and glutamate [18]. Furthermore, evidence indicates that astrocytes and oligodendrocytes can produce action potentials mediated by voltage-gated outward K+ currents. Moreover, both neurons and glia are key producers of sleep regulatory humoral factors including cytokines and energy-related molecules, including adenosine tri-phosphate (ATP) and adenosine (see Metabolism and Sleep section below).

In 1995, Benington and Heller hypothesized that metabolism is involved in sleep regulation, and sleep in turn serves a restorative function [175]. Many genes involved in regulating metabolism are enhanced during sleep compared to wakefulness, in agreement with this hypothesis [176]. Protein synthesis is generally enhanced during sleep vs. wakefulness, supporting the notion that sleep plays a restorative role. Furthermore, sleep deprivation enhances brain mRNAs that modulate glycogen metabolism and glycogen synthesis [177,178]. Additionally, investigations of energy substrates and their derivatives provide further convincing evidence that metabolic molecules and pathways are involved in the initiation and regulation of vigilance states [2].

Various cellular pathways are activated or inhibited by sleep regulatory molecules, and many of these molecules are reciprocally activated or inhibited by these same pathways [2]. Cellular pathways that regulate sleep also modulate inflammation, neurotransmission, and ionic changes within neurons and glia.

Sleep occurs not just within an entire organism (globally) but within localized areas and networks of cells (locally) [2]. Evidence for “local sleep” is found in cases of individuals with parasomnias, who can walk, eat, or have sex but are not entirely conscious [241–243]. An intriguing investigation in epileptic patients with cortical indwelling electrodes demonstrated “local” sleep, in which, during sleep, NREMS characteristics such as SWA and spindles were localized to select regions of the cortex [244]. It was therefore suggested that, as wake is extended by such means as sleep deprivation, sleep-like local activity may intrude into the wake EEG as select neural circuitry goes “offline”. Indeed, Vyazovskiy, Tononi, and colleagues then reported that in the rat, during extended periods of wakefulness, select cortical neurons and related circuitry enter an offline sleep-like state accompanied by SWA in localized cortical regions [245]. This “local sleep” activity increases in incidence as wakefulness was extended. Therefore, although the animal was described by global EEG recordings as awake, localized cortical EEG and neuronal activity recordings indicated that local neuronal populations in the cortex exhibited sleep-like activity. Other evidence of “local sleep” includes studies in which stimulating rodent vibrissae (i.e., whiskers) increases SWA in respective cortical columns, as well as enhances the activity of IL-1β and TNF-α immunoreactive cells [246–248]. Application of TNF-α locally onto one cortical hemisphere enhances SWA, and activates c-Fos and IL-1β in the respective hemisphere, compared to vehicle control treatment of the contralateral hemisphere [249,250]. Recently, Krueger and colleagues developed a novel methodology demonstrating a sleep-like state in a mixed glia and neuron cell culture network [251,252], which was more extensively developed by Tafti and colleagues [253]. Of particular importance, this method was able to document cellular signaling changes after the addition of neurotransmitters, resembling responses observed in mammals during wakefulness.

In summary, we now understand that sleep is regulated globally, regionally, and locally, by means of both molecular and cellular mechanisms. Redundancy in the molecules, cells, and neural circuitry regulating sleep suggest that sleep plays a crucial and protective physiological role, yet such redundancy also renders elusive exact determination of the function(s) of sleep and the regulatory mechanisms involved. Nevertheless, recent basic science technologies are rapidly advancing, further expanding our understanding of the function(s)of sleep.