Basic Features and Regulation of Sleep

The profound biological importance of sleep is supported not only by the fact that humans spend a significant time of their lives sleeping, but also by the fact that sleep is present in virtually all animals as well, some of which spend even more time sleeping than humans(Cirelli and Tononi, 2008), and sleep deprivation generally leads to serious impairment in cognitive abilities and other biological functions. Despite these facts, our current knowledge of the functions of sleep is far from complete (Rosen, 2006). Some features and characteristics of sleep, however, may help highlight its significance for physiological and cognitive functioning.

Sleep is characterized by changes in hormone levels and it affects the functioning of the immune system, thus contributing to ’regeneration’ in a broad sense. After sleep deprivation, immune responses are attenuated due to a lower white blood cell count (Zager et al., 2007). On the other hand, slow-wave sleep increases growth hormone level(Van Cauter et al., 2000), which enables regeneration, wound healing and physical restorative processes of the body. Reduced restorative capacity was found in sleep-deprived rats(Gümüstekin et al., 2004), albeit this effect appears to originate rather from NREM sleep deprivation and it is not present in case of selective REM deprivation (Mostaghimi et al., 2005). A higher amount and better quality of sleep is correlated with higher levels of melatonin in diurnal species, a hormone heavily involved in restorative processes (Bubenik, 2002; Odaci and Kaplan, 2009), suggesting that better sleep quality may be both a cause and an index of the increased ability of the body to heal itself.

The effects of sleep deprivation are certainly more immediate and perhaps even more dramatic in the cognitive domain. The most common and immediate effects of sleep deprivation are sleepiness, the slowing of mental processes as well as the lack of the ability to concentrate. These effects can partially reversed voluntarily, such as by being motivated by rewards (Horne and Pettitt, 1985; Monk, 1991), but they never completely

6

disappear. Sleep deprivation reduces performance in working memory tasks to a particularly striking degree(Turner et al., 2007), in line with increased hemodynamic responses inthe prefrontal cortex, a sign of compensatory recruitment (Drummond et al., 2000; Drummond et al., 2005). However, similarly to the more basic physiological effects of sleep deprivation, alterations in the cognitive domain also appear to rather stem from NREM than REM sleep. REM sleep causes disturbances in emotional regulation(Ellman et al., 1978; Rosales-Lagarde et al., 2012), but it appears to be less involved in sleep-related cognitive processing (Siegel, 2001) and individuals with chronic pharmacological or traumatic REM deprivation are able to live without serious cognitive impairments (Vertes and Eastman, 2000). Thus, sleep cannot be treated as a monolithic process in terms of its effects and functions.

Sleep can be broadly defined as two very distinct states: rapid eye movement (REM) sleep and non-rapid eye movement (NREM) sleep (Rechtschaffen et al., 1968; Iber et al., 2007). At the same time NREM and REM sleep are two alternating phases of the ultradian oscillation serving the basis of the cyclical nature of sleep. Furthermore, subdivisions of these states can be made, reflecting the different depths or electrophysiological states. The following basic description of the most important features of normal sleep – when no other sources are noted – arepresented based on these two classification system (Rechtschaffen et al., 1968; Iber et al., 2007) and one book chapter (Billiard, 2008). Typical EEG features of different sleep stages are illustrated on Figure 1, whileFigure 2 shows a typical hypnogram of all-night sleep..

Figure 1. “EEG patterns of human sleep states and stages”. Figure and caption from

7

“EEG patterns of human sleep states and stages”. Figure and caption from 2008)

“EEG patterns of human sleep states and stages”. Figure and caption from(Billiard,

8

The onset of sleep is characterized by the disappearance of alpha wave trains which are prominent in the resting wakeful EEG signal. There is an increase in theta power, as well as vertex waves and occipital sharp transient waves. This intermediate state is Stage 1 sleep. Stage 1 sleep rarely lasts for more than a few minutes, and instead gives way to either deeper NREM sleep or, if sleep pressure is low (typically in the last periods of night sleep) an awakening.

Stage 2 sleep is characterized by an increased power int he delta band (<4 Hz) and the appearance of its main features, K-complexes and sleep spindles. K-complexes are transient, low-frequency waves which appear spontaneously but can also be elicited by stimulation. Sleep spindles are waxing and waning sinusoidal waveforms which appear all over the scalp but mainly in central and frontal midline derivations, reflecting specific neuronal firing patterns in thalamocortical circuits, mediated by reticular thalamic interference (Steriade, 2003; Lüthi, 2013). Sleep spindles are heavily implicated in the effect sleep exerts on cognition, which is why they will be described in greater detail in later chapters of this thesis.

Stage 3 sleep – together with what is called Stage 4 sleep in an earlier classification system (Rechtschaffen et al., 1968) – is also called slow wave sleep (SWS).

Consequently, this sleep stage is characterized by the proliferation of low-frequency, high-amplitude slow waves, generated by synchronous firing (and silence) in cortical assemblies (Csercsa et al., 2010).

Importantly, slow wave activity in Stage 2 sleep and SWS does not appear in a symmetrically distributed manner, but they are instead organized into cyclic alternating patterns (CAPs) (Terzano et al., 1985; Terzano et al., 2001). Sometimes, slow waves are uniformly distributed for several minutes (Non-CAP), but at other times they appear in sudden, high-amplitude burst series (CAP A1), preceded and followed by a flattened EEG signal devoid of prominent low-frequency, high-amplitude activity (CAP B).

Apart from the CAP A1 subtype, consisting of a transient burst of slow waves, other types of CAP activity are known. CAP A3 is characterized by arousal, reflected by a transient increase in alpha or beta activity and/or muscle tone, while the CAP A2 subtype is characterized by mixed (slow and fast) transient activity. While a detailed description of cyclic alternating patterns is beyond the scope of this thesis, they

9

certainly deserve mention due to many results (Aricò et al., 2010; Esposito and Carotenuto, 2010; Drago et al., 2011) linking them to individual differences in waking cognitive ability. Figure 3 shows EEG recordings from NCAP sleep as well as CAP sequences.

The stages of NREM sleep are typically organized into 90-120 minute long sleep cycles with alternating sleep depth, which continue until awakening. Typically, the deepest stage (reflected by the amount of slow wave activity) of sleep is shallower in each successive sleep cycle.

Rapid eye movement sleep (REM) typically occurs between sleep stages, with increased prominence towards the end of the night. Regarding its appearance and physiological characteristics, REM sleep is radically different from NREM sleep. While reduced muscle tone is typical in all sleep stages, physiological REM sleep is characterized by complete atonia in the skeletal muscles, except for the facial muscles responsible for eye movements. However, REM stage is characterized by increased activity in every other regard, reflected by increased EEG activity in the beta and gamma band (with the complete disappearance of slow waves and sleep spindles), eye movements and prominent – albeit very chaotic – mental activity, which is evident from the fact that dreams are more frequently reported after awakenings from REM sleep.

Importantly, however, dreams also occur in NREM sleep.

Figure 2. Night sleep hypnogram of a healthy young male subject. Note the decreased depth and increased REM prominence in later parts of the night.

The existence and alternation of NREM and REM phases has been explained in numerous ways. One theory (Rial et al., 1993; Rial et al., 2010) proposes that sleep in mammals evolved from reptilian waking states, while mammalian waking is a phylogenically new phenomenon related to the development of a greater and more

10

specialized telencephalon. Based on similarities in EEG patterns and reactivity, these authors proposed that human NREM sleep is analogous to reptilian basking behavior, while REM sleep is analogous to the post-basking behavior of reptiles which is characterized by an observation of the environment and the initiation of new goals.

Another – not necessarily contradicting – theory suggests that sleep phases are tools of energy conservation (Schmidt, 2014). Since thermoregulation is suspended in sleep – especially REM sleep – longer sleep periods are adaptive since they contribute to energy efficiency. This, however, comes at the price of less time available to achieve goals and also a longer time of exposure to predation and other potential dangers. In line with this theory, longer REM phases are observed in larger animals (which have greater thermal inertia) and extremely long periods of continuous waking are observed in niche exploiting animals, such as artic birds with very short mating periods.

Figure 3. “CAP time and non-CAP time in stage 2 NREM sleep. CAP time: alternance of arousal-related phasic events (A) and of the background EEG activities (B). EMG, electromyogram; PNG, pneumogram; EKG, electrocardiogram; CNP, Clinica Neurologica

Parma”. Figure and caption from (Terzano and Parrino, 1992) and (Billiard, 2008) Sleep is frequently investigated using polysomnography for both clinical and research purposes. Polysomnography is the use of multiple electrophysiological exploration methods to accurately determine physiological activity (including and beyond neural activity) during sleep.