Sleep function: current questions and new approaches

European Journal of Neuroscience European Journal of Neuroscience, Vol. 29, pp. 1830–1841, 2009 doi:10.1111/j.1460-9568.2009.06767.x REVIEW Sleep function: current questions and new approaches Anne Vassalli1 and Derk-Jan Dijk2 1 Center for Integrative Genomics, Génopode Building, University of Lausanne, CH-1015 Lausanne, Switzerland Surrey Sleep Research Centre, Faculty of Health and Medical Sciences, University of Surrey, GU2 7XP Guildford, UK 2 Keywords: EEG, memory, non-rapid eye movement sleep, plasticity, sleep homeostasis, slow wave, spindle Abstract The mammalian brain oscillates through three distinct global activity states: wakefulness, non-rapid eye movement (NREM) sleep and REM sleep. The regulation and function of these ‘vigilance’ or ‘behavioural’ states can be investigated over a broad range of temporal and spatial scales and at different levels of functional organization, i.e. from gene expression to memory, in single neurons, cortical columns or the whole brain and organism. We summarize some basic questions that have arisen from recent approaches in the quest for the functions of sleep. Whereas traditionally sleep was viewed to be regulated through top-down control mechanisms, recent approaches have emphasized that sleep is emerging locally and regulated in a use-dependent (homeostatic) manner. Traditional markers of sleep homeostasis, such as the electroencephalogram slow-wave activity, have been linked to changes in connectivity and plasticity in local neuronal networks. Thus waking experience-induced local network changes may be sensed by the sleep homeostatic process and used to mediate sleep-dependent events, benefiting network stabilization and memory consolidation. Although many questions remain unanswered, the available data suggest that sleep function will best be understood by an analysis which integrates sleep’s many functional levels with its local homeostatic regulation. Introduction Sleep’s contributions to mental and physical health have been established through large-scale epidemiological studies of normal sleep and clinical studies documenting the adverse health consequences of disordered sleep (Drake et al., 2003; Reid et al., 2006; Bliwise & Young, 2007). Understanding the mechanisms which mediate these associations holds great promise for improving both physical and mental health. This, however, will require a better insight into the basic functions of sleep for the body and the brain. A shared interest in sleep’s contribution to brain function and the underlying neuronal and molecular mechanisms brought together about 140 sleep scientists and clinicians, including 18 speakers, in Lausanne (see Editorial). The focus of this meeting was on research into the neural functions of sleep with an emphasis on the putative role of sleep in neuronal plasticity. The researchers agreed that sleep serves important functions for the brain but, when asked what these functions are, the simple consensus that it was ‘all for plasticity’ did not emerge and lively discussions broke out instead. For the purpose of this review we discuss broad questions of sleep regulation and function in the context of questions that arose during our discussion, during the presentations of new research findings at the meeting, and in selected recent publications relevant to the role of sleep in neuronal plasticity. Background and traditional questions about sleep function Sleep can be defined on the basis of behavioural criteria (Table 1) and on the basis of electrophysiological criteria (Table 2). Traditionally, Correspondence: Dr A. Vassalli, as above. E-mail: [email protected] Received 1 March 2009, revised 29 March 2009, accepted 30 March 2009 questions about sleep and sleep’s function are asked at the level of the whole organism: 1. Do all animals sleep and is sleep function invariable across species? Where in the phylogenetic tree does sleep emerge? 2. Should sleep be viewed as a recovery process, i.e. does sleep contribute to brain function by reversing some of the consequences of wakefulness? Alternatively, is sleep a distinct state, just as hibernation is a different state, not thought to directly contribute to waking brain function? 3. Irrespective of the nature of its function, is sleep an indispensable state, endowed with unique properties that directly mediate functions which could not be executed during wakefulness? Alternatively, is sleep merely a permissive state, during which for example recovery events optimally occur because of reduced interference by the sensory and processing activities of wakefulness? Many of these traditional sleep questions have not yet been answered definitely, but this should not be taken as evidence for lack of progress in the quest for sleep’s functions (Siegel, 2005; Allada & Siegel, 2008). The easiest way to tackle sleep is from its less elusive end: sleep need. Indeed, the study of sleep’s function has relied heavily on the use of the sleep deprivation paradigm. This paradigm has established that wakefulness is accompanied by an increase in sleep propensity (Table 2) and sleep loss is compensated for by a subsequent increase in sleep intensity and ⁄ or duration. The term ‘sleep propensity’ refers to the likelihood of sleep, at any level of organization, global or local. At the behavioural level, the latency to fall asleep can be measured and human subjects can report on their level of subjective sleepiness, and both measures can be used as indicators of sleep propensity. At the neuronal and circuitry levels, bioelectrical events associated with sleep can be measured. Thus, both ª The Authors (2009). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd Sleep function and plasticity 1831 Table 1. Behavioural criteria of sleep Species-specific posture Behavioural quiescence Reversible upon stimulation Elevated arousal threshold Rebound after deprivation global and local correlates of sleep propensity can be studied and will be discussed in this review. The past decade saw an impressive expansion of the spatial and temporal domains within which sleep and the effects of sleep deprivation are analyzed. Sleep and sleep loss have correlates and consequences over a wide temporal range, e.g. seconds (Destexhe et al., 2007), days (Belenky et al., 2003) and months (Gais et al., 2007). Furthermore, sleep and sleep loss are accompanied by changes at many levels of functional organization, from gene expression to network connectivity, from memory consolidation to emotions. Within a given level of organization, sleep and sleep loss have correlates in many different variables, e.g. alterations in cytokine production and in the dynamics of electric field potentials in neuronal networks. We will focus on some of the local aspects of sleep regulation that have recently emerged. These local aspects have consequences for our view on what constitutes the minimal unit in which sleep can be observed and studied. Local aspects of sleep regulation also have implications for the discussion on whether sleep is regulated in a ‘topdown’ or ‘bottom-up’ manner. One of the most robust markers of change in sleep propensity is a change in the slow or delta Table 2. Definitions of terminology and concepts used in sleep research Electrophysiologically defined sleep phenomena Electroencephalogram (EEG) Measure of electrical potentials at the surface of the head Vigilance (or behavioural) state Corresponds to a set or constellation of behavioural and physiological variables. Experimentally assessed by the coordinated recording and analysis of the EEG and the electromyogram (EMG) of a face or neck muscle, as well as by the recording of eye movement, i.e. the electrooculogram. Three vigilance states are distinguished in mammals and birds: wakefulness, non-rapid eye movement (NREM) sleep and REM sleep Wakefulness Vigilance state characterized by a low amplitude, high frequency, mixed EEG pattern NREM sleep Vigilance state characterized by high amplitude, low frequency oscillations, dominated by the slow ⁄ delta and spindle oscillations (see below) and a relaxed muscle tone. In animal studies that use the term paradoxical sleep (see below), NREM sleep is referred to as slow wave sleep REM sleep Vigilance state characterized by an EEG resembling wake or stage-1 sleep in humans, in association with muscle atonia. In rodents, REM sleep is dominated by theta oscillations (see below). In some animal studies REM sleep is referred to as paradoxical sleep Slow wave sleep (SWS) The deepest stages of NREM sleep (stages 3 and 4 in humans) during which slow ⁄ delta waves are especially prevalent and arousal thresholds highest Slow waves (or delta oscillations) EEG oscillations within the 0.75–4.5 Hz frequency range. High amplitude slow ⁄ delta waves are a defining feature of (in this review referred to NREM sleep stages-3 and 4, i.e. SWS as ‘slow ⁄ delta waves’) Slow wave activity (SWA) (or delta power) Mathematical variable extracted from the EEG which quantifies the amplitude and prevalence of slow ⁄ delta waves. The relative value of SWA at the onset of NREM sleep can be accurately predicted on the basis of prior wake duration. Because of this, SWA is widely acknowledged to reflect sleep need and is thought to reflect a sleep homeostatic process The slow (< 1 Hz) oscillation Rhythmic alternation in the activity of cortical neurons during NREM sleep. Intervals of high activity during which bursts of action potential occur (‘up states’) alternate with intervals of almost complete silence and membrane hyperpolarization (‘down states’). The slow (< 1 Hz) oscillation is assessed by intra- or juxtacellular recordings Sleep spindles Transient (0.5–2 s), spindle-shaped 10–15 Hz EEG features which prevail during early sleep (stage-2 in humans), and also herald NREM-REM sleep transitions. Sigma activity refers to the prevalence and amplitude of spindle oscillations in the 10–15 Hz range Theta oscillations EEG oscillations in the 5–10 Hz range. In rodents coherent theta oscillations are characteristic of REM sleep and exploratory waking behaviour and are thought to be of hippocampal origin. In humans, the origin of the EEG activity in this frequency range is less clear Definitions of concepts Sleep propensity Synaptic strength (or weight) Probability or tendency to be in or to transition into a sleep state. At the global level sleep propensity can be assessed behaviourally. At the local level, it can be assessed using electrophysiological markers of the sleep state Input–output relationship of a synapse Synaptic plasticity Ability to change synaptic strength and ⁄ or number Hebbian (experience-dependent) Model by which correlated presynaptic and postsynaptic activity strengthens a synapse whereas uncorrelated activity plasticity weakens it. Long-term potentiation (LTP) and long-term depression (LTD) are electrophysiological assays thought to measure Hebbian plasticity Homeostatic plasticity Non-Hebbian regulatory mechanisms thought to counteract the destabilizing effects of experience-dependent, Hebbian events and maintain postsynaptic excitability within a functional range. This term encompasses a wide variety of processes, including synaptic scaling Synaptic scaling Mechanism thought to adjust all of a neuron’s synapses up or down in strength proportionally, so that while average synaptic strength is regulated homeostatically, the relative strengths of individual synapses remain constant ª The Authors (2009). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience, 29, 1830–1841 1832 A. Vassalli and D.-J. Dijk (slow ⁄ delta) waves measured by electroencephalographic (EEG) recordings (see Table 2 for definitions). We will discuss the proposed relationships between sleep slow ⁄ delta waves and changes in neuronal network connectivity. From considerations about sleep-dependent synaptic strength regulation and the role of sleep-regulatory substances, we will move on to models of the role of sleep in memory consolidation. In this context, the two types of sleep, rapid eye movement (REM) and non-REM (NREM) sleep, must be distinguished. Throughout, we will continue to discuss the functional significance of the markers of sleep and sleep homeostasis we have been using. It has been established that sleep–wake regulation is closely linked to circadian rhythmicity and some of the molecular mechanisms underlying circadian processes have been elucidated (Takahashi et al., 2008). This, together with the recognition that circadian rhythmicity is also about temporal organization of processes that contribute to higher brain functions such as memory and cognition (Barnard & Nolan, 2008), and the recognition that some of the proteins involved in circadian rhythm generation also play a role in the homeostatic regulation of sleep (Franken & Dijk, this issue), has led to a renewed cross-fertilization between these two research areas. This review, however, will not discuss the mechanisms of circadian rhythmicity in any detail. Furthermore, following early work by Tobler (1983) and Kaiser & Steiner-Kaiser (1983), recent years have witnessed the introduction of simple model organisms such as Drosophila melanogaster (Shaw et al., 2000) and Caenorhabditis elegans (Zimmerman et al., 2008) to study the genetic and molecular aspects of sleep regulation as well as sleep’s contribution to central nervous system function and development. Recently, birds have appeared as major players in the sleep and memory consolidation field (Jackson et al., 2008; Shank & Margoliash, 2009). In this meeting, however, the emphasis was on new approaches in the study of sleep in mammals. Finally, the role of sleep in glucose homeostasis, metabolism and other physiological processes is being increasingly studied (e.g. Dijk, 2008; Tasali et al., 2008), but this review will focus on the brain. We hope that an overview of the current approaches and concepts may be useful at a time when scientists from many disciplines enter the sleep field and new models, new methods and new reduced preparations for the study of sleep function are introduced at a high rate. New approaches to the study of sleep function: is sleep a global or local phenomenon? The traditional definitions of sleep are based on observed behavioural and physiological phenomena in the whole organism. Vigilance states (Table 2) are defined as constellations of variables associated with these phenomena, and in mammals and birds three vigilance states are distinguished: wakefulness, NREM sleep and REM sleep (Table 2). The animal is in one, and only one, of these vigilance states at a time. In other words, vigilance states have been considered temporally discrete and spatially global states, but recent observations indicate that these concepts may require some refinement. The discrete nature of sleep and wake states is challenged by the following observations: 1. In the diseased brain elements of different vigilance states can occur simultaneously. For example, in Parkinson’s disease, and in general in REM sleep behaviour disorder, motor actions occur while the patient is in REM sleep (Arnulf et al., 2008). Clinical recognition of mixed states in sleep disorder patients led to the term ‘status dissociatus’ as early as 1991 (Mahowald & Schenck, 1991). 2. Below the surface of observable behaviour, when analyzed with a temporal resolution in the minutes to hour range, vigilance states are not invariable. A defining feature of NREM sleep is the occurrence of high-amplitude slow ⁄ delta waves. Within each NREM sleep episode, slow-wave activity (SWA), also called delta power, an EEG measure of power in the 0.75–4.5 Hz frequency range (Table 2), gradually rises and abruptly declines prior to the onset of REM sleep (Fig. 1). Furthermore, in the course of the resting period (e.g. a night for humans) SWA declines over consecutive NREM–REM sleep cycles. In contrast, wakefulness is characterized by a low-amplitude, high-frequency EEG pattern but the contribution of slow ⁄ delta and theta oscillations (5–10 Hz, Table 2) increases as the duration of wakefulness increases. Thus, as wakefulness progresses the EEG becomes more ‘sleep-like’ (Franken et al., 1991; Cajochen et al., 2002) and as NREM sleep progresses, the EEG becomes more ‘wake-like’ (Dijk et al., 1997). 3. When analyzed with a higher temporal resolution, of the order of seconds and less, the behaviour of cortical neurons during NREM sleep is highly dynamic, yielding bouts of widespread highamplitude synchronized fluctuations in activity. These fluctuations, known as the ‘slow (< 1 Hz) oscillation’ (Table 2) with a period of 2 s (Steriade et al., 1993), consist of intervals with high activity during which bursts of action potential occur (‘up states’) alternating with intervals of almost complete silence and membrane hyperpolarization (‘down states’). These synchronized alternations of up and down states appear in the EEG as slow waves and spread coherently throughout the cortex as a travelling wave (Massimini et al., 2004). It has been suggested that the ‘up states’ of slow-wave sleep (SWS; Table 2) are dynamically similar to wakefulness and provide a context for the ‘replay’ of firing sequences that occurred in the waking animal, for example during execution of a specific task. These ‘fragments of wakefulness’ may contribute to memory consolidation (Destexhe et al., 2007). According to this interpretation, and at this temporal resolution, we are partly awake when we are asleep but are not aware of this! The global nature of sleep and wake states has also been challenged: 1. Local activation of specific brain regions during wakefulness, through unilateral vibrations of hands (Kattler et al., 1994) or twitching of whiskers (Vyazovskiy et al., 2000), leads to local increases in EEG SWA during sleep in those brain areas that were more active during wakefulness. One brain area can be more asleep than another. 2. In dolphins and other aquatic mammals, EEG-assessed NREM sleep can occur in one hemisphere while the other is awake (Mukhametov et al., 1977; Lyamin et al., 2008). This unequivocally demonstrates that vigilance states are not necessarily invariable in ‘brain space’. 3. Sensory stimulation by whisker twitches and auditory clicks evokes local field potentials in the cortex with amplitudes which fluctuate between high and low values, be the animal awake or asleep. It has been hypothesized that these fluctuations reflect local functional state differences within individual groups of highly connected neurons called cortical columns (Rector et al., 2005; Rector et al., 2009, this issue). The high-amplitude responses are most prevalent during NREM sleep and, it is hypothesized, represent a sleep-like state, whereas the low-amplitude responses would indicate a wakelike state. The amplitude of these potentials fluctuates to some extent independently of overall vigilance state and can be different in the two hemispheres and even in adjacent cortical columns. This suggests that some cortical columns may be ‘awake’ while others are ‘asleep’, and this during either sleep or wakefulness. ª The Authors (2009). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience, 29, 1830–1841 Sleep function and plasticity 1833 Fig. 1. Time course of SWA (power in the 0.75–4.5 Hz band; lower curves) and activity in the spindle frequency range (13.25–15.0-Hz band; upper curves) recorded under baseline conditions and after sleep deprivation (36 h of wakefulness). NREM sleep episodes were subdivided into 20 equal intervals and REM sleep episodes into five equal intervals. Mean values per interval were calculated prior to averaging across subjects (n = 8, except for cycle 8 of recovery sleep, where n = 6) and are expressed relative to the mean level in baseline NREM sleep (100%). The mean timing of REM sleep episodes is delimited by vertical lines and horizontal bars above the abscissa [Reanalysis by D. Aeschbach of the data from Dijk et al. (1990)]. Adapted with permission from Elsevier, ª 2009 (Principles and Practice of Sleep Medicine). These phenomena indicate that, although sleep appears in general to be a global phenomenon in both time and space, at higher temporal and spatial resolution, heterogeneity emerges. This leads to a set of new questions concerning sleep function and the level of organization considered. Sleep function and the minimal sleep unit Can we define a sleep unit, a minimal entity that would recapitulate the essence of sleep? Identifying the minimal sleep unit may be useful because this will help to identify protagonists, e.g. neurons, neuronal groups or bioelectrical events, which display causal relationships with sleep functions at the level of the whole organism. Do we need to define this minimal entity in both space and time? Is the minimal sleep unit a whole organism, a brain, a group of closely connected neurons, a single neuron, a synapse? Is the minimal sleep unit a minute, a second, one NREM–REM sleep cycle? The traditional criteria for defining sleep include a species-specific posture, a reduced but reversible responsiveness to stimulation, i.e. elevated arousal threshold, and homeostatic regulation. There is some consensus that the minimal sleep unit may be smaller than the whole organism, even though a brain cannot adopt a species-specific posture. This is, however, where the consensus ends and the discussions become lively. This is in part because our preferred definition of a sleep unit may very well be influenced by our tacit or explicit assumptions about the function of sleep. To Jan Born, who studies the effects of sleep on human memory, ‘sleep is a complex, systems process’ (Born et al., 2006) and the minimal sleep unit is the entire brain. For James Krueger and David Rector, individual cortical columns, consisting of approximately a thousand neurons in the rat, alternate between two states, wake-like or sleep-like (Krueger et al., 2008; Roy et al., 2008). A proponent of the hypothesis that sleep plays a key role in the regulation of neuronal connectivity and plasticity is unlikely to accept the isolated neuron as the minimal sleep unit whereas the proponents of the ‘sleep replenishes energy stores of the brain’ may very well accept the neuron as the minimal unit. The intermingling of descriptive and functional elements is not uncommon and even at the level of the whole organism functional elements are included in the definition of sleep. For example, current definitions of sleep include the requirement that it is ‘homeostatically regulated’, i.e. that loss of it must be compensated for by a subsequent increase in either its intensity or duration, or both (Tobler, 1983; Zimmerman et al., 2008). The consensus which emerged was that it is permissible that definitions of sleep contain function-oriented elements because, after all, sleep researchers search for the contribution of sleep to brain function. It is, however, questionable that we can develop a definition of the minimal sleep unit in the absence of a clear understanding of sleep’s function at the level of the whole organism. Because we are in this catch-22, spending much time on the definition of the minimal sleep unit may not be productive. It may be more fruitful to accept that sleep- and wake-like phenomena can be observed in units that are smaller than an organism, a brain, a hemisphere, a network, etc. Studying sleep-like phenomena and their function in sleep units of all sizes, and relating these phenomena at the ‘smaller unit level’ to sleep and its functions as observed at the level of the ultimate sleep unit, i.e. the whole organism living in its 24-h environment, is the most promising avenue to further progress. During these discussions about sleep function and the minimal sleep unit it also emerged that another contentious issue was whether we ª The Authors (2009). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience, 29, 1830–1841 1834 A. Vassalli and D.-J. Dijk According to some (Krueger & Obal, 1993), sleep is a self-organizing property of any viable neuronal assembly. According to this view, alternating active (‘wake’) and silent (‘sleep’) phases will emerge in any neuronal network. Sleep is thus a local network process that emerges from the bottom up and that would proceed in the absence of top-down regulation by specialized neuroanatomical circuits or nuclei. It is the local activity within a network or minimal sleep unit that will determine the likelihood of sleep to occur and it is at this local level that clues about the function of sleep are to be found. A bottom-up view is consistent with the local differences in the intensity of NREM sleep SWA as induced by local activation during wakefulness and the local regulation of ‘sleep states’ in cortical columns, as well as the remarkable persistence of sleep-like phenomena after widespread lesions across the brain. The bottom-up view is, however, to a first approximation inconsistent with the highly orchestrated nature of the occurrence of vigilance states and the accompanying changes in multiple state variables across the brain and body. In mammals, the complex phenomena of NREM and REM sleep emerge through interaction of local networks and top-down control by neuroanatomical circuitry (Saper et al., 2005; Luppi et al., 2006; Fort et al., 2009, this issue), disruption of which leads to disruption of the normal characteristics of sleep and wakefulness. A prominent example of loss of top-down control is narcolepsy, in which disruption of the orexinergic systems leads to severe disruption in the regulation of behavioural state transitions, including pathological transitions between wakefulness and REM sleep and co-occurrence of consciousness and muscle atonia (Nishino, 2007). Another prominent example of top-down control is the circadian control of sleep, in the absence of which the temporal organization of sleep and wakefulness into long episodes of wakefulness (e.g. 16 h) and sleep (e.g. 8 h) is lost, but sleep and its homeostatic regulation persist (Trachsel et al., 1992; Edgar et al., 1993). The existence of top-down orchestration, however, does not imply that essentials of sleep function could not be related to some of the ‘bottom processes’ that determine local variation in the probability of being in a sleep-like state, i.e. variation in sleep propensity. Understanding local variation in sleep propensity may provide important clues to sleep function and it seems reasonable to assume that bottom-up phenomena emerged first in evolution and later became controlled by specialized sleep-regulatory centres, to optimize spatial and temporal organization of vigilance states which in turn should contribute to behavioural adaptation (Krueger et al., 2008). established beyond doubt that sleep deprivation leads to sleepiness and changes in subsequent sleep duration and structure. These responses to sleep deprivation are one manifestation of sleep-homeostatic regulation, which in essence implies that sleep propensity and the contents of sleep depend on vigilance state history. The phenomenon of sleep homeostasis is very well established and understanding the mechanisms underlying sleep homeostasis has almost become a proxy for understanding sleep function. It may be thought that measuring sleep propensity is a good approach to monitoring the time course of the sleep homeostat during a normal waking episode, but it is not. This is because at the global level sleep propensity is regulated through interactions of sleep homeostasis and circadian rhythmicity (Borbély, 1982; Daan et al., 1984; Dijk & Czeisler, 1995). In this conceptualisation, time spent awake results in physical changes in brain neural networks. During the day (in the case of diurnal species such as humans) these changes do not, however, lead to an overt and proportional increase in sleepiness (sleep propensity) (Dijk et al., 1987) or to performance deterioration (Dijk et al., 1992). It is thought that the mechanism underlying the absence of an increase in sleep propensity and performance deterioration is related to wake-promoting (activating) inputs at the level of either the hypothalamus or the cortex (Edgar et al., 1993; Saint-Mleux et al., 2007). Whereas during the day this activating input ‘masks’ the effects of sustained wakefulness on sleep propensity and performance, during the night these activating inputs diminish and the effects of sustained wakefulness on the brain can now manifest themselves as increased sleep propensity or, in the case of attempting to stay awake, deteriorating performance. The circadian pacemaker, in mammals located in the suprachiasmatic nucleus of the hypothalamus, is thought to drive these activating inputs and, in a top-down manner, organise the temporal pattern of sleep and wakefulness (Dijk & von Schantz, 2005). Thus sleep propensity and performance are not really directly reflecting time awake and its effects on neural networks. Shadows of the covert effects of sustained wakefulness on neural networks can, however, be observed at the global level in the EEG while the brain is awake: as mentioned above, slow ⁄ delta and theta wave activity increase with increasing duration of wakefulness (Franken et al., 1991; Cajochen et al., 2002). These changes in neural networks manifest themselves most prominently during NREM sleep: SWA in the EEG during repeated nap sleep increases monotonically with increasing duration of wakefulness (Dijk et al., 1987). When we override the circadian control on sleep propensity and sleep is initiated during the day, SWA during NREM sleep is very little affected by the circadian pacemaker. Thus even though many aspects of sleep, i.e. sleep propensity, sleep spindles (which are oscillations in the 10–15 Hz frequency range; Table 2) and REM sleep, are under circadian control, SWA and its decline during sleep are nearly independent of circadian phase (Dijk & Czeisler, 1995). This is strong evidence in support of the hypothesis that wakefulness leads to unidentified changes in the brain and that these changes are reversed during sleep, i.e. sleep homeostasis. Because of its response to sleep deprivation and its near independence of the circadian process, SWA is considered a good marker of sleep homeostasis. The regulation of global and local sleep propensity SWA vs. the duration of NREM sleep: how good is SWA as a marker of sleep homeostasis? should assume that sleep has one primary function or many. If we accept that sleep may have multiple (very important) functions, some of these functions may be fulfilled at the cellular level, e.g. energy balance-related functions, whereas other functions of sleep, e.g. plasticity or memory consolidation, may only emerge at the network or systems level. Abolishing the quest for the ultimate function of sleep may be liberating and may open our eyes to a multitude of sleep functions, but the desire to identify a primordial function of sleep remains strong because this function would have been the driver for the emergence of sleep during evolution. ‘Top-down’ or ‘bottom-up’ control of sleep–wake states? Global organization of sleep propensity: circadian and homeostatic components Searching for alterations in brain function following sleep deprivation has been a popular paradigm in the study of sleep function. It has been SWA is widely considered to be a good marker of sleep homeostasis by researchers who use it as their proxy for sleep need. However, although the variation in SWA upon sleep onset, and its decline during NREM sleep, can be accurately predicted on the basis of prior sleep–wake ª The Authors (2009). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience, 29, 1830–1841 Sleep function and plasticity 1835 history, it has been known for many years that SWA fails to predict sleep duration. Thus, rats that were sleep-deprived for 24 h showed a marked increase in SWA above baseline values during the initial 4 h of recovery sleep (‘positive sleep rebound’). SWA then quickly declined to values below baseline and remained below baseline until 36 h into recovery (‘negative rebound’). Yet throughout this period rats slept more than baseline (Franken et al., 1991). SWA does not, therefore, reflect the human or animal’s apparent sleep need and this has led to the notion that sleep intensity and sleep duration are regulated separately (Borbély, 1982). The take-home message is that only one aspect of sleep homeostasis is accounted for by SWA and understanding SWA may not necessarily hold clues to all functions of sleep. Local correlates of sleep propensity: sleep-regulatory substances (SRSs) and synaptic strength Although sleep homeostasis is well established, a key unanswered question is in what form, i.e. metabolic, molecular, structural or bioelectrical, is past waking and sleep activity recorded? In other words, what does the sleep homeostat sense? To answer this question, total sleep deprivation is the obvious paradigm to use but other paradigms in which the contents of wakefulness are changed, through specific (learning) tasks or stimulation of specific sensory pathways, can also be used and may be more informative. Biochemical correlates of the effects of sleep–wake history have long been sought. Originally it was shown that sleepiness could be transferred from one animal to another, using material extracted from the cerebrospinal fluid or brain of sleep-deprived animals (Pappenheimer et al., 1975). More recently, sleepiness, or at least some aspects of the consequences of sleep deprivation, were successfully transferred to a dish: the responsiveness of orexin neurons to noradrenaline remarkably changes from excitatory to inhibitory when the hypothalamic slices that contain them are taken from animals having undergone 2 h of sleep deprivation rather than from animals allowed to sleep. These effects appear mediated by postsynaptic mechanisms (Grivel et al., 2005). Sleep deprivation has also been shown to increase the strength of glutamatergic synapses on orexin neurons (Rao et al., 2007). Thus at the local level, sleep homeostasis is reflected by changes in neuronal responsiveness and synaptic strengths of, in this case, orexin neurons. These changes in the orexinergic system provide us with mechanisms by which wakefulness may lead to an increase in the propensity to fall asleep, yet they do not necessarily provide us with clues about sleep’s contribution to neural function. One molecule that has received much attention as a potential biochemical mediator of sleep homeostasis is adenosine. Some cells release adenosine triphosphate (ATP) from synaptic vesicles together with neurotransmitters, possibly providing an index of prior synaptic use. This ATP is rapidly hydrolyzed to adenosine (Krueger et al., 2008). During prolonged wakefulness extracellular concentrations of adenosine increase, in particular in basal forebrain areas (PorkkaHeiskanen et al., 2000). Adenosine’s effects on A1 receptors and associated activation of K+ channels and inhibition of Ca2+ entry leads to inhibition of neural activity, an increase in sleep propensity and ultimately to SWS (Basheer et al., 2000). Adenosine may only be a member of the group of SRSs, which include interleukin-1b, tumor necrosis factor a (TNFa), brain-derived neurotrophic factor (BDNF) and epidermal growth factor, all of which have been shown to induce sleep and ⁄ or to be induced by sleep loss (Krueger et al., 2008). It is thought that electrical activity patterns and associated postsynaptic events that prevail during wakefulness, and maybe in particular during specific behaviours such as exploration, lead to the production of these ‘somnogenic growth factors’. Such effects can be observed in the intact animal. For example, expression of BDNF is correlated with time spent in exploratory behaviour (Huber et al., 2007) and BDNF expression is thought to depend on neuronal firing rates. This growth factor and others such as TNFa are on the one hand somnogenic and on the other hand thought to activate the molecular machinery required for synaptic plasticity (see below). Sleep deprivation studies in which appropriate controls for circadian processes were implemented have also been used to identify changes in gene expression profiles in association with vigilance state and circadian rhythmicity. Early rat studies (Cirelli et al., 2004) and more recent mouse studies indicate that there is specific variation around the clock in brain RNA expression, driven to a large extent by the alternation between sleep and wakefulness and to a smaller extent by circadian rhythmicity (Maret et al., 2007). In a recent study in which it was attempted to identify those transcripts that were specifically up- or down-regulated, in the brain and not the liver, by 6 h of enforced wakefulness rather than by circadian rhythmicity, and in three different mouse inbred strains, one cDNA stood out: Homer1a (Maret et al., 2007). This Homer1 gene alternate transcript is induced by neuronal activity. This finding may be interpreted within the context of the function of Homer1a in plasticity or protection and recovery from glutamate-induced neuronal activity imposed by wakefulness. Gene expression microarray studies are informative with respect to brainspecific response to sleep loss but they are blind to translational and post-translational regulations, and no causality can be ascribed: genes which are functionally implicated in sleep regulation are not distinguished from genes whose expression merely follows sleep– wake history. Local correlates of sleep propensity: slow ⁄ delta waves and synaptic strength Another approach in the search for sleep’s contribution to neural function has been to follow the EEG slow ⁄ delta waves, as a marker of sleep homeostasis, to their origin. What are the underlying mechanisms of slow ⁄ delta wave generation? Intracellular recordings during SWS have revealed that many cortical and thalamic neurons are triggered to enter a characteristic burst–pause mode of firing when excitatory inputs from diffuse neuromodulator systems are low and their resting membrane potentials become hyperpolarized. This hyperpolarization allows T-type Ca2+ channels (also referred to as low-voltage-activated Ca2+ channels) to de-inactivate, mediate Ca2+ entry and initiate a burst of action potentials (Steriade et al., 1993). Evidence that T-type Ca2+ channels are critical players in the rhythmic synchronized burst discharges characteristic of SWS was provided through analysis of mice bearing gene-targeted mutations in T-type Ca2+ channel pathway components. Thalamocortical relay neurons from mice deficient for one of the three T-type Ca2+ channels (the a1G-subunit, also called Cav3.1) strikingly fail to enter the burst mode of firing action potentials, whereas the tonic firing mode is unimpaired (Kim et al., 2001). These mutant mice exhibit disrupted sleep continuity, with a higher incidence of awakenings lasting > 16 s, but not of brief awakenings lasting < 16 s, when compared to wildtype mice (Lee et al., 2004; Anderson et al., 2005). Oscillatory bursting in neurons of the nucleus reticularis of the thalamus (nRT) also appears important for slow ⁄ delta wave generation. Mice lacking SK2, a K+ channel selectively expressed on dendrites of nRT neurons and coupled to T-type Ca2+ channels, show a NREM sleep EEG ª The Authors (2009). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience, 29, 1830–1841 1836 A. Vassalli and D.-J. Dijk characterized by weakened slow ⁄ delta and spindle oscillations, and fragmented sleep (Cueni et al., 2008). Thus T-type Ca2+ channel de-inactivation is thought to constitute a triggering event in slow ⁄ delta wave generation and associated processes (Destexhe et al., 2007). However, what underlies the gradual increase in slow ⁄ delta waves in the EEG as observed after sustained wakefulness and their gradual dissipation during sleep? In one scenario it reflects the variation in levels of the wake- or arousal-promoting neuromodulators mentioned above, namely monoamines, acetylcholine and orexins, and ⁄ or variations in adenosine levels (McCarley, 2007). In a different scenario, most relevant to the role of sleep in plasticity, variations in slow ⁄ delta waves reflect changes in connectivity within cortical neuronal networks. Krueger & Obal (1993, 2003) proposed that changes in sleep propensity are related to local and use-dependent changes in synaptic strength and hypothesized that sleep serves the maintenance of the ‘synaptic super structure’. Tononi & Cirelli (2003, 2006) proposed that neuronal connections would become on average stronger during wakefulness due to synaptic potentiation associated with learning, and would weaken during sleep due to sleep-dependent synaptic depression or homeostatic downscaling. Here it may be useful to note that synaptic depression and synaptic downscaling, although they may share molecular steps, were defined as distinct phenomena. ‘Homeostatic plasticity’, which encompasses ‘synaptic scaling’ as well as a wide variety of other processes, refers to mechanisms thought to counteract the destabilizing effects of experience-dependent Hebbian events, such as long-term potentiation (LTP) and long-term depression (LTD) (Table 2) (Turrigiano & Nelson, 2004; Nelson & Turrigiano, 2008). It is now still too early to conclusively identify the specific synaptic plasticity processes that occur during sleep and in which slow ⁄ delta waves are thought to be involved, and the field remains equivocal in the use of these terms. In any event, local synaptic strength (or weight, i.e. the input–output relationship of a synapse) within cortical networks is thought to vary over the sleep–wake history. Mechanistically, stronger synaptic connections would increase network synchronization whereas weaker connections would reduce network synchronization. According to this view, which is supported by computer simulations and models of the thalamocortical system (Hill & Tononi, 2005; Esser et al., 2007), these variations in synaptic strength result in variations in slow ⁄ delta waves within cortical networks, which appear as changes in SWA at the level of the EEG. Across the sleep–wake cycle, changes in synaptic strength are thus driving changes in SWA. Furthermore, SWS-associated neuronal firing patterns, including the slow ⁄ delta waves, are in turn assumed to alter synaptic strength. As discussed below, the burst-firing mode of activity can induce LTD (Birtoli & Ulrich, 2004; Czarnecki et al., 2007). Thus slow ⁄ delta waves may both reflect and drive changes in synaptic strength. Synaptic depression and ⁄ or downscaling are required because otherwise all the experiences incurred during wakefulness would lead to unconstrained synaptic weight growth. In this view, and at this level of organization, the function of NREM sleep is synaptic weight homeostasis, a process initially described in the developing nervous system (Turrigiano & Nelson, 2004). Electrophysiological and molecular evidence for sleep-related changes in synaptic strength Mathematical models of neuronal networks have shown that SWA and other EEG variables, e.g. the slope of individual slow ⁄ delta waves (in voltage over time, nV ⁄ s), as well as the slope and amplitude of evoked potentials during sleep and wakefulness, can indeed be a measure of network connectivity (Esser et al., 2007). These parameters have now been shown in rat and humans to increase with wake and decrease with sleep (Riedner et al., 2007; Vyazovskiy et al., 2007, 2008). To what extent these variables are a simple monotonic function of time awake and time asleep or are also modulated by the circadian phase remains, however, to be established. Nevertheless, there is electrophysiological evidence for changes in synaptic strength in synchrony with the sleep–wake cycle. There are, however, contradictory reports on the direction of synaptic plasticity occurring during sleep using molecular markers of plasticity. Vyazovskiy et al. (2008) compared phosphorylation levels of GluR1-containing AMPA receptors and other markers of LTP and LTD after 6 h of predominant sleep vs. wakefulness in rats, and provided evidence for net potentiation after wakefulness and depression after sleep. In contrast, other authors emphasize the role of sleep in the consolidation of cortical plasticity associated with synaptic potentiation. Using the classical plasticity paradigm of remodelling of ocular dominance in the visual cortex of kittens submitted to monocular deprivation, Frank and colleagues recently reported observations seemingly at odds with the synaptic depression or downscaling hypothesis (Aton et al., 2009). These authors initially showed that sleep enhances ocular dominance plasticity and that this effect depends on cortical activity during sleep through unknown mechanisms (Jha et al., 2005). They now report, using some of the same molecular markers of potentiation or depression of glutamatergic synapses as those used in the Vyazovskiy et al. (2008) study described above, that sleep-dependent enhancement of visual cortex remodelling is associated with synaptic potentiation (Aton et al., 2009). They also show that sleep-dependent ocular dominance plasticity is inhibited by antagonists of NMDA receptors or cAMP-dependent protein kinase, known to inhibit LTP. Altogether, it is proposed that ‘off-line’ secondary waves of synaptic potentiation events occur in the primary visual cortex during sleep that follows initial plasticity, and act to reinforce cortical map remodelling. Slow ⁄ delta waves and sleep spindles: effects on measures of plasticity EEG phenomena such as slow ⁄ delta waves may not just reflect changes in synaptic strength but may actually play a causal role in bringing about these changes. A systematic electrophysiological study of how specific patterns of sleep-related neuronal firing affect synaptic plasticity in cortical pyramidal neurons of rat brain slices was performed by Daniel Ulrich and co-workers, as described below. The two most conspicuous neuronal firing patterns of NREM sleep, slow ⁄ delta waves and sleep spindles, are not only different in their frequency range but are also very different in their regulation. Slow ⁄ delta waves decline during sleep but sleep spindle activity remains constant or even increases during sleep (Fig. 1). Whereas slow ⁄ delta waves are not affected by circadian phase, sleep spindles are under circadian control and are not markedly affected by sleep deprivation (Dijk et al., 1997). At the behavioural level, variation in both delta oscillations and sleep spindles have been reported to be associated with consolidation of procedural and declarative memory (Huber et al., 2004; Schabus et al., 2004; Schmidt et al., 2006; Aeschbach et al., 2008). Is there electrophysiological evidence that these phenomena may contribute to plasticity and, if yes, in a similar or dissimilar way? One approach to investigating how NREM sleep may contribute to plasticity has been to investigate the effects of these specific firing ª The Authors (2009). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience, 29, 1830–1841 Sleep function and plasticity 1837 patterns on phenomena such as LTP and LTD, which have been widely implicated in learning and memory. In these studies, cortical pyramidal neurons were stimulated by bursts of action potentials, which are the characteristic of slow ⁄ delta waves, paired with application of an excitatory postsynaptic potential (EPSP). This electrophysiological profile led to LTD (Birtoli & Ulrich, 2004; Czarnecki et al., 2007). In contrast, if the EPSP was paired with a firing pattern mimicking a spindle oscillation, LTP was generated (Rosanova & Ulrich, 2005). It is noteworthy that, in the same experimental setup, application of a profile characteristic of the waking state, i.e. individual action potentials mimicking tonic rather than burst firing, generated LTP. In summary, these studies suggest that the two hallmarks of NREM sleep both contribute to synaptic plasticity, albeit in opposite directions. A role for SWS beyond synaptic depression and downscaling? Synaptic depression or downscaling during SWS may recalibrate the strength of synapses within a functional responsive range, but could it enhance the signal-to-noise ratio of stored (learned) information? Could it actively contribute to memory consolidation? In a first modelling attempt it was shown that an SWS-mediated downscaling could indeed keep synaptic weight in check and maintain selectivity of information that was stored (learned) during the waking day as well (Sullivan & De Sa, 2008). A parallel question is whether the well known negative effects of extended periods of wakefulness and the positive effects of sleep on performance (which are particularly clear in the absence of the circadian confound; Wyatt et al., 1999), may be associated with these changes in synaptic weight over the sleep–wake cycle. Currently, no direct evidence for this link is available. However, irrespectively of the mechanisms involved, there is some experimental evidence that SWS is beneficial for performance and memory consolidation. Consolidation can be defined as the progressive post-acquisition stabilization, and sometimes enhancement, of memory that occur even in the absence of training. Researchers such as Jan Born, Robert Stickgold, Pierre Maquet and their co-workers believe indeed that sleep does more for memory than synaptic scaling. For a variety of memory tasks that are improved by sleep, brain activity was monitored by brain imaging using positron emission tomography (PET) or functional magnetic resonance imaging (fMRI), before or at initial training, during post-training sleep or at retesting. During post-training SWS they observed (re)activation of activity patterns evoked during learning (Peigneux et al., 2004; Rasch et al., 2007). The reactivations were reported to correlate positively with the improvement in performance on the subsequent day. Much earlier, using multi-unit electrode recordings in rats during SWS, McNaughton and colleagues reported ‘replay’ in hippocampal ‘place cells’ of similar spatial and temporal firing patterns as were evoked in those cells when the rat was engaged in spatial exploration and navigation across an arena (Wilson & McNaughton, 1994). More recently it was reported that post-training replay also occurs in cortical neurons, and this takes place specifically during up-states of SWS (Ji & Wilson, 2007). Cortical and hippocampal replays were found to be coupled temporally. It could not be determined whether these replays contribute to, or only reflect, memory consolidation. Nevertheless, the concept that memory consolidation evolves from repeated covert reactivation of newly encoded memory traces during sleep has gained ground. In addition, memory traces can undergo slow and major structural changes in a sleep-dependent manner. The retesting 48 h (Fischer et al., 2005) or 12 h (Walker et al., 2005) after initial training of a nondeclarative procedural motor task (finger tapping) in humans revealed profound alterations in task-concomitant patterns of brain activation as detected by fMRI. These alterations were sleep-specific and associated with gains in performance. They were interpreted as a reorganization of the memory representations of the skill. The scale of reorganizations of memory representations may be such that retrieval of hippocampus-dependent memories becomes over time hippocampus-independent, presumably due to a gradual transfer to cortical networks, in particular to the prefrontal cortex (Maviel et al., 2004; Euston et al., 2007). Well beyond ‘synaptic consolidation’, sleep is thus thought to contribute to memory by promoting a process of ‘system consolidation’ that leads to a redistribution of the memory representation involving new neuronal networks not involved at initial encoding (Dudai, 2004; Born et al., 2006). This may imply a ‘qualitative’ change in the memory, which can be manifested for example as the memory may become less dependent on context, but richer in schematic, abstract structure (discussed in Hoffman et al., 2007). Memory consolidation can be thought to consist of synaptic and system consolidation and, although much remains to be understood, there is evidence for contribution of SWS to both processes. The next near irresistible question is: are the slow ⁄ delta waves themselves instrumental in SWS-dependent memory consolidation? Local changes in SWA during the first half-hour of post-training sleep are correlated with the overnight improvement in performance on a procedural visuomotor learning task (Huber et al., 2004). Slow waves can be manipulated to shed light on their causal links with memory formation. Application of transcranial direct current stimulation in humans can lead to an increase in < 3 Hz activity and improvement in declarative memory (Marshall et al., 2004). In contrast, brain stimulation with oscillations at 5 Hz (theta) left declarative memory unchanged (Marshall et al., 2006). It is to be noted, however, that the protocol used to enhance slow ⁄ delta waves also enhanced spindle activity. Furthermore, it has been reported that acoustic disruption of slow ⁄ delta waves prevents the sleep-dependent improvement in performance on a perceptual learning task (visual texture discrimination). In this study, the improvement in performance correlated with SWA, when the control and experimental conditions were combined in those 16 individuals (out of 20) who met a specific initial performance criterion (Aeschbach et al., 2008). Additional evidence for a role of SWS in memory consolidation was derived by re-exposing sleeping subjects to an odor that had been presented as context during the learning of a declarative memory task (Rasch et al., 2007). This odor exposure led to a boost in post-sleep memory retrieval if the odor was presented during SWS, but not if it was presented during REM sleep. fMRI imaging revealed significant hippocampal activation during SWS upon odor re-exposure. This suggests that reactivation during sleep of a part of a complex memory ‘engram’ can have access to, and functionally alter, the wider memory traces associated with it. The evidence for an association between slow ⁄ delta waves and plasticity is growing, but important unanswered questions and issues remain. The quantitative relation between SWA and measures of plasticity, and even, as mentioned above, the direction of this association, as well as sleepiness levels and performance, have not always been firmly established across the circadian cycle. Early, wellcontrolled studies have failed to identify a specific role for SWS in the regulation of daytime performance (Bonnet, 1986) and results of recent studies have not always been supportive of a role for SWS in memory consolidation (Genzel et al., 2009; Schabus, 2009). ª The Authors (2009). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience, 29, 1830–1841 1838 A. Vassalli and D.-J. Dijk Furthermore, whether the positive effects of sleep on declarative memory is mediated by the slow ⁄ delta waves or the sleep spindles remains unresolved (Schmidt et al., 2006). For some procedural memory tasks, the case for a role of stage-2 sleep (i.e. NREM sleep with few slow ⁄ delta waves and many sleep spindles) in consolidation is much stronger than the case for SWS (Nishida & Walker, 2007). In summary, there is a growing consensus that NREM sleep contributes to plasticity and memory consolidation, although the relative contribution of specific NREM phenomena such as spindles and slow ⁄ delta waves remains to be established. The data in support of a role for sleep in memory consolidation should not detract from the observation that these effects may be learning-task-specific, and that consolidation also occurs during wakefulness (Doyon et al., 2009). A role for REM sleep in memory consolidation? What about REM sleep? Is it also associated with memory trace reactivation? Does it also contribute to memory consolidation? Evidence that brain structures activated during learning are reactivated selectively during subsequent REM sleep was obtained in humans using PET imaging during and following visuomotor skill and other implicit memory tasks (Maquet et al., 2000; Peigneux et al., 2003). From these and other studies the notion emerged that, while declarative, hippocampus-dependent memories would be strengthened particularly during SWS, non-declarative, procedural memories (i.e., perceptual and motor skills) would benefit to a greater extent from REM sleep (Born et al., 2006). Amygdala-dependent emotional memories were also shown to benefit specifically from late REM sleep-rich sleep (Wagner et al., 2001). More recently, Born and co-workers cast doubt on the concept that procedural skills benefit from REM sleep. They reported that suppression of REM sleep by selective serotonin and noradrenaline reuptake inhibitors did not impair consolidation of two different motor skill tasks. In fact, accuracy in finger sequence tapping was significantly enhanced by drug administration and the gain in accuracy was positively correlated with the increase in number and density of fast spindle oscillations (> 13 Hz) during stage-2 sleep and SWS (Rasch et al., 2009). Therefore EEG-defined REM sleep appears not to be required for memory consolidation of this particular procedural task and in fact may be detrimental. A compilation of many recent studies indicates that the hypothesis that REM sleep benefits the consolidation of procedural tasks and SWS is for the benefit of declarative tasks is too simple (Schabus, 2009). early life in mammals and birds is a major contribution to a more general view of the interrelationships between plasticity and sleep throughout the lifespan and across two classes of the animal kingdom. Gender- and age-related and other individual differences in SWA and other EEG variables: implications for differences in plasticity? It is intriguing that aspects of sleep such as SWA and sleep spindles which are thought to be implicated in plasticity and learning can display marked interindividual differences under baseline (i.e. not sleep-deprived) conditions. Gender and ageing are major factors affecting sleep parameters. Thus, women have more SWS and higher SWA than men (Dijk et al., 1989; Carrier et al., 2001) and ageing is associated with reductions in SWS and SWA (Landolt et al., 1996) and sleep duration (Klerman & Dijk, 2008). Twin studies indicate that the interindividual differences in certain aspects of the sleep EEG such as SWA have a heritability of well over 90% (De Gennaro et al., 2008). The identity of some of the genes contributing to interindividual differences in sleep has been revealed. Polymorphisms in genes involved in the adenosinergic system (Retey et al., 2005) and the circadian gene PER3 (Viola et al., 2007) have now been shown to affect SWS and SWA. Although sleep deprivation leads to an increase in SWA in all these individuals, the magnitude of some of these interindividual differences in the baseline EEG, including SWA, dwarf the effects related to sleep deprivation (Tucker et al., 2007). An important question to be addressed in future studies is whether these interindividual differences in sleep measures are indicators of differences in plasticity or, alternatively, reflect individual differences in sleep characteristics, which are independent of the modulation and involvement of the sleep process in plasticity, learning and memory. In this context it will be important to distinguish between individual differences in baseline values of sleep characteristics and individual differences in the response to challenges. In other words, are these differences reflective of differences in EEG-generating mechanisms or in sleep regulatory processes which may be related to differences in plasticity? Studies in mice demonstrate that genetic factors that contribute to inter-strain differences in SWA (analogous to interindividual differences for humans) (Maret et al., 2005) are distinct from those underlying the inter-strain differences in the sleep deprivation-induced changes in SWA (Franken et al., 2001). Concluding remarks A role for sleep in plasticity during development? A role of sleep in neuronal network plasticity is in accordance with the predominance of NREM and in particular REM sleep in the early postnatal animal life. Sleep amount is maximal at the time when plasticity is maximal. As mentioned above, ocular dominance plasticity during the critical period in kittens is enhanced by sleep and this effect depends on postsynaptic cortical activity during sleep (Jha et al., 2005). However, a related but distinct paradigm of plasticity during development, recovery from monocular deprivation, is not enhanced by sleep (Dadvand et al., 2006), showing that different forms of plasticity in vivo are regulated by distinct mechanisms whose dependence on sleep may differ. Recently, birds have emerged as models for studying the role of sleep in two classical developmental learning paradigms: imprinting in the domestic chick (Jackson et al., 2008) and song development in juvenile zebra finches (Shank & Margoliash, 2009). Establishing a contribution of sleep to plasticity in Understanding sleep’s function and contribution to neuronal function requires convergence of evidence from a variety of independent assays, models, levels of analysis, etc. The research that we have reviewed demonstrates that progress has been made in the understanding of the interrelations between classical EEG markers for NREM sleep and sleep regulation and the associated molecular, cellular and network events that mediate plasticity. Many questions remain unanswered. Continuation and expansion of this multidisciplinary sleep research effort holds great promise for furthering our understanding of mechanisms involved in the brain’s ability to interact efficiently with its social and physical environment and adapt its responses based on waking experience. Acknowledgements D.J.D. acknowledges support from the Biotechnology and Biological Sciences Research Council, Air Force Office of Scientific Research, Wellcome Trust- ª The Authors (2009). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience, 29, 1830–1841 Sleep function and plasticity 1839 UK, and the Higher Education Funding Council for England. A.V. was supported by a Marie Heim-Vögtlin award from the Swiss National Science Foundation. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the funding organizations. A.V. thanks Mehdi Tafti for support, Paul Franken for many discussions and invaluable mentoring help, and Jan Born and Daniel Ulrich for insightful comments. We thank Paul Franken, Reto Huber, James Krueger and Anita Luthi for helpful comments on the manuscript. Abbreviations AMPA, a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; EEG, electroencephalogram or electroencephalographic; fMRI, functional magnetic resonance imaging; GluR1, glutamate receptor 1; LTD, long-term depression; LTP, long-term potentiation; NMDA, N-methyl D-aspartate; NREM, non-REM; REM, rapid eye movement; SRS, sleep-regulatory substance; SWA, slow-wave activity; SWS, slow-wave sleep. References Aeschbach, D., Cutler, A.J. & Ronda, J.M. (2008) A role for non-rapid-eyemovement sleep homeostasis in perceptual learning. J. Neurosci., 28, 2766– 2772. Allada, R. & Siegel, J.M. (2008) Unearthing the phylogenetic roots of sleep. Curr. Biol., 18, R670–R679. Anderson, M.P., Mochizuki, T., Xie, J., Fischler, W., Manger, J.P., Talley, E.M., Scammell, T.E. & Tonegawa, S. (2005) Thalamic Cav3.1 T-type Ca2+ channel plays a crucial role in stabilizing sleep. Proc. Natl Acad. Sci. USA, 102, 1743–1748. Arnulf, I., Leu, S. & Oudiette, D. (2008) Abnormal sleep and sleepiness in Parkinson’s disease. Curr. Opin. Neurol., 21, 472–477. Aton, S.J., Seibt, J., Dumoulin, M., Jha, S.K., Steinmetz, N., Coleman, T., Naidoo, N. & Frank, M.G. (2009) Mechanisms of sleep-dependent consolidation of cortical plasticity. Neuron, 61, 454–466. Barnard, A.R. & Nolan, P.M. (2008) When clocks go bad: neurobehavioural consequences of disrupted circadian timing. PLoS Genet., 4, e1000040. Basheer, R., Porkka-Heiskanen, T., Strecker, R.E., Thakkar, M.M. & McCarley, R.W. (2000) Adenosine as a biological signal mediating sleepiness following prolonged wakefulness. Biol. Signals Recept., 9, 319–327. Belenky, G., Wesensten, N.J., Thorne, D.R., Thomas, M.L., Sing, H.C., Redmond, D.P., Russo, M.B. & Balkin, T.J. (2003) Patterns of performance degradation and restoration during sleep restriction and subsequent recovery: a sleep dose-response study. J. Sleep Res., 12, 1–12. Birtoli, B. & Ulrich, D. (2004) Firing mode-dependent synaptic plasticity in rat neocortical pyramidal neurons. J. Neurosci., 24, 4935–4940. Bliwise, D.L. & Young, T.B. (2007) The parable of parabola: what the U-shaped curve can and cannot tell us about sleep. Sleep, 30, 1614–1615. Bonnet, M.H. (1986) Performance and sleepiness following moderate sleep disruption and slow wave sleep deprivation. Physiol. Behav., 37, 915–918. Borbély, A.A. (1982) A two process model of sleep regulation. Hum. Neurobiol., 1, 195–204. Born, J., Rasch, B. & Gais, S. (2006) Sleep to remember. Neuroscientist, 12, 410–424. Cajochen, C., Wyatt, J.K., Czeisler, C.A. & Dijk, D.J. (2002) Separation of circadian and wake duration-dependent modulation of EEG activation during wakefulness. Neuroscience, 114, 1047–1060. Carrier, J., Land, S., Buysse, D.J., Kupfer, D.J. & Monk, T.H. (2001) The effects of age and gender on sleep EEG power spectral density in the middle years of life (ages 20–60 years old). Psychophysiology, 38, 232–242. Cirelli, C., Gutierrez, C.M. & Tononi, G. (2004) Extensive and divergent effects of sleep and wakefulness on brain gene expression. Neuron, 41, 35– 43. Cueni, L., Canepari, M., Lujan, R., Emmenegger, Y., Watanabe, M., Bond, C.T., Franken, P., Adelman, J.P. & Luthi, A. (2008) T-type Ca2+ channels, SK2 channels and SERCAs gate sleep-related oscillations in thalamic dendrites. Nat.Neurosci., 11, 683–692. Czarnecki, A., Birtoli, B. & Ulrich, D. (2007) Cellular mechanisms of burst firing-mediated long-term depression in rat neocortical pyramidal cells. J. Physiol., 578, 471–479. Daan, S., Beersma, D.G.M. & Borbély, A.A. (1984) Timing of human sleep: Recovery process gated by a circadian pacemaker. Am. J. Physiol., 246, R161–R178. Dadvand, L., Stryker, M.P. & Frank, M.G. (2006) Sleep does not enhance the recovery of deprived eye responses in developing visual cortex. Neuroscience, 143, 815–826. De Gennaro, L., Marzano, C., Fratello, F., Moroni, F., Pellicciari, M.C., Ferlazzo, F., Costa, S., Couyoumdjian, A., Curcio, G., Sforza, E., Malafosse, A., Finelli, L.A., Pasqualetti, P., Ferrara, M., Bertini, M. & Rossini, P.M. (2008) The electroencephalographic fingerprint of sleep is genetically determined: a twin study. Ann. Neurol., 64, 455–460. Destexhe, A., Hughes, S.W., Rudolph, M. & Crunelli, V. (2007) Are corticothalamic ‘up’ states fragments of wakefulness? Trends Neurosci., 30, 334–342. Dijk, D.J. (2008) Slow-wave sleep, diabetes, and the sympathetic nervous system. Proc. Natl Acad. Sci. USA, 105, 1107–1108. Dijk, D.J. & Czeisler, C.A. (1995) Contribution of the circadian pacemaker and the sleep homeostat to sleep propensity, sleep structure, electroencephalographic slow waves and sleep spindle activity in humans. J. Neurosci., 15, 3526–3538. Dijk, D.J. & von Schantz, M. (2005) Timing and consolidation of human sleep, wakefulness, and performance by a symphony of oscillators. J. Biol. Rhythms, 20, 279–290. Dijk, D.J., Beersma, D.G.M. & Daan, S. (1987) EEG power density during nap sleep: reflection of an hourglass measuring the duration of prior wakefulness. J. Biol. Rhythms, 2, 207–219. Dijk, D.J., Beersma, D.G.M. & Bloem, G. (1989) Sex differences in the sleep EEG of young adults: visual scoring and spectral analysis. Sleep, 12, 500–507. Dijk, D.J., Brunner, D.P. & Borbély, A.A. (1990) Time course of EEG power density during long sleep in humans. Am. J. Physiol., 258, R650–R651. Dijk, D.J., Duffy, J.F. & Czeisler, C.A. (1992) Circadian and sleep ⁄ wake dependent aspects of subjective alertness and cognitive performance. J. Sleep Res., 1, 112–117. Dijk, D.J., Shanahan, T.L., Duffy, J.F., Ronda, J.M. & Czeisler, C.A. (1997) Variation of electroencephalographic activity during non-rapid eye movement and rapid eye movement sleep with phase of circadian melatonin rhythm in humans. J. Physiol. (Lond.), 505, 3. Doyon, J., Korman, M., Morin, A., Dostie, V., Tahar, A.H., Benali, H., Karni, A., Ungerleider, L.G. & Carrier, J. (2009) Contribution of night and day sleep vs. simple passage of time to the consolidation of motor sequence and visuomotor adaptation learning. Exp. Brain Res. [DOI: 10:1007/s0021-009-1748-y] Drake, C.L., Roehrs, T. & Roth, T. (2003) Insomnia causes, consequences, and therapeutics: an overview. Depress. Anxiety, 18, 163–176. Dudai, Y. (2004) The neurobiology of consolidations, or, how stable is the engram? Annu. Rev. Psychol., 55, 51–86. Edgar, D.M., Dement, W.C. & Fuller, C.A. (1993) Effect of SCN lesions on sleep in squirrel monkeys: evidence for opponent processes in sleep–wake regulation. J. Neurosci., 13, 1065–1079. Esser, S.K., Hill, S.L. & Tononi, G. (2007) Sleep homeostasis and cortical synchronization: I. Modeling the effects of synaptic strength on sleep slow waves. Sleep, 30, 1617–1630. Euston, D.R., Tatsuno, M. & McNaughton, B.L. (2007) Fast-forward playback of recent memory sequences in prefrontal cortex during sleep. Science, 318, 1147–1150. Fischer, S., Nitschke, M.F., Melchert, U.H., Erdmann, C. & Born, J. (2005) Motor memory consolidation in sleep shapes more effective neuronal representations. J. Neurosci., 25, 11248–11255. Franken, P., Dijk, D.J., Tobler, I. & Borbély, A.A. (1991) Sleep deprivation in rats: effects on EEG power spectra, vigilance states, and cortical temperature. Am. J. Physiol., 261, R198–R208. Franken, P., Chollet, D. & Tafti, M. (2001) The homeostatic regulation of sleep need is under genetic control. J. Neurosci., 21, 2610–2621. Gais, S., Albouy, G., Boly, M., ng-Vu, T.T., Darsaud, A., Desseilles, M., Rauchs, G., Schabus, M., Sterpenich, V., Vandewalle, G., Maquet, P. & Peigneux, P. (2007) Sleep transforms the cerebral trace of declarative memories. Proc. Natl Acad. Sci. USA, 104, 18778–18783. Genzel, L., Dresler, M., Wehrle, R., Grozinger, M. & Steiger, A. (2009) Slow wave sleep and REM sleep awakenings do not affect sleep dependent memory consolidation. Sleep, 32, 302–310. Grivel, J., Cvetkovic, V., Bayer, L., Machard, D., Tobler, I., Muhlethaler, M. & Serafin, M. (2005) The wake-promoting hypocretin ⁄ orexin neurons change their response to noradrenaline after sleep deprivation. J. Neurosci., 25, 4127–4130. Hill, S. & Tononi, G. (2005) Modeling sleep and wakefulness in the thalamocortical system. J. Neurophysiol., 93, 1671–1698. Hoffman, K.L., Battaglia, F.P., Harris, K., MacLean, J.N., Marshall, L. & Mehta, M.R. (2007) The upshot of up states in the neocortex: from slow oscillations to memory formation. J. Neurosci., 27, 11838–11841. ª The Authors (2009). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience, 29, 1830–1841 1840 A. Vassalli and D.-J. Dijk Huber, R., Ghilardi, M.F., Massimini, M. & Tononi, G. (2004) Local sleep and learning. Nature, 430, 78–81. Huber, R., Tononi, G. & Cirelli, C. (2007) Exploratory behavior, cortical BDNF expression, and sleep homeostasis. Sleep, 30, 129–139. Jackson, C., McCabe, B.J., Nicol, A.U., Grout, A.S., Brown, M.W. & Horn, G. (2008) Dynamics of a memory trace: effects of sleep on consolidation. Curr. Biol., 18, 393–400. Jha, S.K., Jones, B.E., Coleman, T., Steinmetz, N., Law, C.T., Griffin, G., Hawk, J., Dabbish, N., Kalatsky, V.A. & Frank, M.G. (2005) Sleepdependent plasticity requires cortical activity. J. Neurosci., 25, 9266–9274. Ji, D. & Wilson, M.A. (2007) Coordinated memory replay in the visual cortex and hippocampus during sleep. Nat. Neurosci., 10, 100–107. Kaiser, W. & Steiner-Kaiser, J. (1983) Neuronal correlates of sleep, wakefulness and arousal in a diurnal insect. Nature, 301, 707–709. Kattler, H., Dijk, D.J. & Borbély, A.A. (1994) Effect of unilateral somatosensory stimulation prior to sleep on the sleep EEG in humans. J. Sleep Res., 3, 159–164. Kim, D., Song, I., Keum, S., Lee, T., Jeong, M.J., Kim, S.S., McEnery, M.W. & Shin, H.S. (2001) Lack of the burst firing of thalamocortical relay neurons and resistance to absence seizures in mice lacking alpha(1G) T-type Ca(2+) channels. Neuron, 31, 35–45. Klerman, E.B. & Dijk, D.J. (2008) Age-related reduction in the maximal capacity for sleep – implications for insomnia. Curr. Biol., 18, 1118–1123. Krueger, J.M. & Obal, F. (1993) A neuronal group theory of sleep function. J. Sleep Res., 2, 63–69. Krueger, J.M. & Obal, F. Jr (2003) Sleep function. Front Biosci., 8, d511–d519. Krueger, J.M., Rector, D.M., Roy, S., Van Dongen, H.P., Belenky, G. & Panksepp, J. (2008) Sleep as a fundamental property of neuronal assemblies. Nat. Rev. Neurosci., 9, 910–919. Landolt, H.-P., Dijk, D.J., Achermann, P. & Borbély, A.A. (1996) Effect of age on the sleep EEG: slow-wave activity and spindle frequency activity in young and middle-aged men. Brain Res., 738, 205–212. Lee, J., Kim, D. & Shin, H.S. (2004) Lack of delta waves and sleep disturbances during non-rapid eye movement sleep in mice lacking alpha1Gsubunit of T-type calcium channels. Proc. Natl Acad. Sci. USA, 101, 18195– 18199. Luppi, P.H., Gervasoni, D., Verret, L., Goutagny, R., Peyron, C., Salvert, D., Leger, L. & Fort, P. (2006) Paradoxical (REM) sleep genesis: the switch from an aminergic–cholinergic to a GABAergic-glutamatergic hypothesis. J. Physiol Paris, 100, 271–283. Lyamin, O.I., Manger, P.R., Ridgway, S.H., Mukhametov, L.M. & Siegel, J.M. (2008) Cetacean sleep: an unusual form of mammalian sleep. Neurosci. Biobehav. Rev., 32, 1451–1484. Mahowald, M.W. & Schenck, C.H. (1991) Status dissociatus – a perspective on states of being. Sleep, 14, 69–79. Maquet, P., Laureys, S., Peigneux, P., Fuchs, S., Petiau, C., Phillips, C., Aerts, J., Del, F.G., Degueldre, C., Meulemans, T., Luxen, A., Franck, G., Van Der, L.M., Smith, C. & Cleeremans, A. (2000) Experience-dependent changes in cerebral activation during human REM sleep. Nat. Neurosci., 3, 831–836. Maret, S., Franken, P., Dauvilliers, Y., Ghyselinck, N.B., Chambon, P. & Tafti, M. (2005) Retinoic acid signaling affects cortical synchrony during sleep. Science, 310, 111–113. Maret, S., Dorsaz, S., Gurcel, L., Pradervand, S., Petit, B., Pfister, C., Hagenbuchle, O., O’Hara, B.F., Franken, P. & Tafti, M. (2007) Homer1a is a core brain molecular correlate of sleep loss. Proc. Natl Acad. Sci. USA, 104, 20090–20095. Marshall, L., Molle, M., Hallschmid, M. & Born, J. (2004) Transcranial direct current stimulation during sleep improves declarative memory. J. Neurosci., 24, 9985–9992. Marshall, L., Helgadottir, H., Molle, M. & Born, J. (2006) Boosting slow oscillations during sleep potentiates memory. Nature, 444, 610–613. Massimini, M., Huber, R., Ferrarelli, F., Hill, S. & Tononi, G. (2004) The sleep slow oscillation as a traveling wave. J. Neurosci., 24, 6862–6870. Maviel, T., Durkin, T.P., Menzaghi, F. & Bontempi, B. (2004) Sites of neocortical reorganization critical for remote spatial memory. Science, 305, 96–99. McCarley, R.W. (2007) Neurobiology of REM and NREM sleep. Sleep Med., 8, 302–330. Mukhametov, L.M., Supin, A.Y. & Polyakova, I.G. (1977) Interhemispheric asymmetry of the electroencephalographic sleep patterns in dolphins. Brain Res., 134, 581–584. Nelson, S.B. & Turrigiano, G.G. (2008) Strength through diversity. Neuron, 60, 477–482. Nishida, M. & Walker, M.P. (2007) Daytime naps, motor memory consolidation and regionally specific sleep spindles. PLoS ONE, 2, e341. Nishino, S. (2007) Narcolepsy: pathophysiology and pharmacology. J. Clin. Psychiatry, 68(Suppl. 13), 9–15. Pappenheimer, J.R., Koski, G., Fencl, V., Karnovsky, M.L. & Krueger, J. (1975) Extraction of sleep-promoting factor S from cerebrospinal fluid and from brains of sleep-deprived animals. J. Neurophysiol., 38, 1299–1311. Peigneux, P., Laureys, S., Fuchs, S., Destrebecqz, A., Collette, F., Delbeuck, X., Phillips, C., Aerts, J., Del, F.G., Degueldre, C., Luxen, A., Cleeremans, A. & Maquet, P. (2003) Learned material content and acquisition level modulate cerebral reactivation during posttraining rapid-eye-movements sleep. Neuroimage, 20, 125–134. Peigneux, P., Laureys, S., Fuchs, S., Collette, F., Perrin, F., Reggers, J., Phillips, C., Degueldre, C., Del, F.G., Aerts, J., Luxen, A. & Maquet, P. (2004) Are spatial memories strengthened in the human hippocampus during slow wave sleep? Neuron, 44, 535–545. Porkka-Heiskanen, T., Strecker, R.E. & McCarley, R.W. (2000) Brain sitespecificity of extracellular adenosine concentration changes during sleep deprivation and spontaneous sleep: an in vivo microdialysis study. Neuroscience, 99, 507–517. Rao, Y., Liu, Z.W., Borok, E., Rabenstein, R.L., Shanabrough, M., Lu, M., Picciotto, M.R., Horvath, T.L. & Gao, X.B. (2007) Prolonged wakefulness induces experience-dependent synaptic plasticity in mouse hypocretin ⁄ orexin neurons. J. Clin. Invest., 117, 4022–4033. Rasch, B., Buchel, C., Gais, S. & Born, J. (2007) Odor cues during slow-wave sleep prompt declarative memory consolidation. Science, 315, 1426–1429. Rasch, B., Pommer, J., Diekelmann, S. & Born, J. (2009) Pharmacological REM sleep suppression paradoxically improves rather than impairs skill memory. Nat. Neurosci., 12, 396–397. Rector, D.M., Topchiy, I.A., Carter, K.M. & Rojas, M.J. (2005) Local functional state differences between rat cortical columns. Brain Res., 1047, 45–55. Reid, K.J., Martinovich, Z., Finkel, S., Statsinger, J., Golden, R., Harter, K. & Zee, P.C. (2006) Sleep: a marker of physical and mental health in the elderly. Am. J. Geriatr. Psychiatry, 14, 860–866. Retey, J.V., Adam, M., Honegger, E., Khatami, R., Luhmann, U.F., Jung, H.H., Berger, W. & Landolt, H.P. (2005) A functional genetic variation of adenosine deaminase affects the duration and intensity of deep sleep in humans. Proc. Natl Acad. Sci. USA, 102, 15676–15681. Riedner, B.A., Vyazovskiy, V.V., Huber, R., Massimini, M., Esser, S., Murphy, M. & Tononi, G. (2007) Sleep homeostasis and cortical synchronization: III. A high-density EEG study of sleep slow waves in humans. Sleep, 30, 1643–1657. Rosanova, M. & Ulrich, D. (2005) Pattern-specific associative long-term potentiation induced by a sleep spindle-related spike train. J. Neurosci., 25, 9398–9405. Roy, S., Krueger, J.M., Rector, D.M. & Wan, Y. (2008) A network model for activity-dependent sleep regulation. J. Theor. Biol., 253, 462–468. Saint-Mleux, B., Bayer, L., Eggermann, E., Jones, B.E., Muhlethaler, M. & Serafin, M. (2007) Suprachiasmatic modulation of noradrenaline release in the ventrolateral preoptic nucleus. J. Neurosci., 27, 6412–6416. Saper, C.B., Scammell, T.E. & Lu, J. (2005) Hypothalamic regulation of sleep and circadian rhythms. Nature, 437, 1257–1263. Schabus, M. (2009) Still missing some significant ingredients. Sleep, 32, 291– 293. Schabus, M., Gruber, G., Parapatics, S., Sauter, C., Klosch, G., Anderer, P., Klimesch, W., Saletu, B. & Zeitlhofer, J. (2004) Sleep spindles and their significance for declarative memory consolidation. Sleep, 27, 1479–1485. Schmidt, C., Peigneux, P., Muto, V., Schenkel, M., Knoblauch, V., Munch, M., de Quervain, D.J., Wirz-Justice, A. & Cajochen, C. (2006) Encoding difficulty promotes postlearning changes in sleep spindle activity during napping. J. Neurosci., 26, 8976–8982. Shank, S.S. & Margoliash, D. (2009) Sleep and sensorimotor integration during early vocal learning in a songbird. Nature, 458, 73–77. Shaw, P.J., Cirelli, C., Greenspan, R.J. & Tononi, G. (2000) Correlates of sleep and waking in Drosophila melanogaster. Science, 287, 1834–1837. Siegel, J.M. (2005) Clues to the functions of mammalian sleep. Nature, 437, 1264–1271. Steriade, M., McCormick, D.A. & Sejnowski, T.J. (1993) Thalamocortical oscillations in the sleeping and aroused brain. Science, 262, 679–685. Sullivan, T.J. & De Sa, V. (2008) Sleeping our way to weight normalization and stable learning. Neural Comput., 20, 3111–3130. Takahashi, J.S., Hong, H.K., Ko, C.H. & McDearmon, E.L. (2008) The genetics of mammalian circadian order and disorder: implications for physiology and disease. Nat. Rev. Genet., 9, 764–775. Tasali, E., Leproult, R., Ehrmann, D.A. & Van Cauter, E. (2008) Slow wave sleep and the risk of type 2 diabetes in humans. Proc. Natl Acad. Sci. USA, 105, 1044–1049. ª The Authors (2009). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience, 29, 1830–1841 Sleep function and plasticity 1841 Tobler, I. (1983) Effect of forced locomotion on the rest-activity cycle of the cockroach. Behav. Brain Res., 8, 351–360. Tononi, G. & Cirelli, C. (2003) Sleep and synaptic homeostasis: a hypothesis. Brain Res. Bull., 62, 143–150. Tononi, G. & Cirelli, C. (2006) Sleep function and synaptic homeostasis. Sleep Med. Rev., 10, 49–62. Trachsel, L., Edgar, D.M., Seidel, W.F., Heller, H.C. & Dement, W.C. (1992) Sleep homeostasis in suprachiasmatic nuclei-lesioned rats: effects of sleep deprivation and triazolam administration. Brain Res., 589, 253–261. Tucker, A.M., Dinges, D.F. & Van Dongen, H.P. (2007) Trait interindividual differences in the sleep physiology of healthy young adults. J. Sleep Res., 16, 170–180. Turrigiano, G.G. & Nelson, S.B. (2004) Homeostatic plasticity in the developing nervous system. Nat. Rev. Neurosci., 5, 97–107. Viola, A.U., Archer, S.N., James, L.M., Groeger, J.A., Lo, J.C., Skene, D.J., von, S.M. & Dijk, D.J. (2007) PER3 polymorphism predicts sleep structure and waking performance. Curr. Biol., 17, 613–618. Vyazovskiy, V., Borbely, A.A. & Tobler, I. (2000) Unilateral vibrissae stimulation during waking induces interhemispheric EEG asymmetry during subsequent sleep in the rat. J. Sleep Res., 9, 367–371. Vyazovskiy, V.V., Riedner, B.A., Cirelli, C. & Tononi, G. (2007) Sleep homeostasis and cortical synchronization: II. A local field potential study of sleep slow waves in the rat. Sleep, 30, 1631–1642. Vyazovskiy, V.V., Cirelli, C., Pfister-Genskow, M., Faraguna, U. & Tononi, G. (2008) Molecular and electrophysiological evidence for net synaptic potentiation in wake and depression in sleep. Nat. Neurosci., 11, 200–208. Wagner, U., Gais, S. & Born, J. (2001) Emotional memory formation is enhanced across sleep intervals with high amounts of rapid eye movement sleep. Learn. Mem., 8, 112–119. Walker, M.P., Stickgold, R., Alsop, D., Gaab, N. & Schlaug, G. (2005) Sleepdependent motor memory plasticity in the human brain. Neuroscience, 133, 911–917. Wilson, M.A. & McNaughton, B.L. (1994) Reactivation of hippocampal ensemble memories during sleep. Science, 265, 676–679. Wyatt, J.K., Ritz-De Cecco, A., Czeisler, C.A. & Dijk, D.J. (1999) Circadian temperature and melatonin rhythms, sleep, and neurobehavioral function in humans living on a 20-h day. Am. J. Physiol., 277, R1152–R1163. Zimmerman, J.E., Naidoo, N., Raizen, D.M. & Pack, A.I. (2008) Conservation of sleep: insights from non-mammalian model systems. Trends Neurosci., 31, 371–376. ª The Authors (2009). Journal Compilation ª Federation of European Neuroscience Societies and Blackwell Publishing Ltd European Journal of Neuroscience, 29, 1830–1841