51: Sleep and Dreaming

• Sleep Consists of Alternating REM and Non-REM Periods

  • Non-REM Sleep Has Four Stages

  • REM and Non-REM Dreams Are Different

• Sleep Obeys Circadian and Ultradian Rhythms

  • The Circadian Rhythm Clock Is Based on a Cyclic Production of Nuclear Transcription Factors

  • The Ultradian Rhythm of Sleep Is Controlled by the Brain Stem

  • Sleep-Related Activity in the EEG Is Generated Through Local and Long-Range Circuits

• Sleep Changes with Age

• The Characteristics of Sleep Vary Greatly Between Species

• Sleep Disorders Have Behavioral, Psychological, and Neurological Causes

  • Insomnia Is the Most Common Form of Sleep Disruption

  • Excessive Daytime Sleepiness Is Indicative of Disrupted Sleep

  • The Disruption of Breathing During Sleep Apnea Results in Fragmentation of Sleep

  • Narcolepsy Is Characterized by Abnormal Activation of Sleep Mechanisms

  • Restless Leg Syndrome and Periodic Leg Movements Disrupt Sleep

  • Parasomnias Include Sleep Walking, Sleep Talking, and Night Terrors

  • Circadian Rhythm Sleep Disorders Are Characterized by an Activity Cycle That Is Out of Phase with the World

• An Overall View

Sleep is a remarkable state. It consumes fully a third of our lives—approximately 25 years in the average lifetime—yet most of us know little about this daily excursion into our inner world. Perhaps even more surprising, we are still hard pressed to give a raison d'etre for sleep.

The exact functions of sleep and of dreaming, one of the more spectacular components of sleep, have been debated over the ages and are still not known. Do dreams reveal some inner psychological functioning of unconscious mental processes, as Sigmund Freud first suggested, or are they merely the consequence of random firing of neurons in the brain? The psychic content of dreams has been a rich subject of speculation throughout history. Plato, anticipating Sigmund Freud, thought that all of us have a "terrible, fierce, and lawless blood of desires, which it seems are revealed in our sleep," whereas Aristotle believed that dreams are merely afterthoughts of the day's activities and experiences.

Although we have only limited understanding of the functions of sleep, our insight into its mechanisms has increased greatly over the past 50 years. We have moved away from the intuitively appealing but incorrect notion that sleep is a period of relative inactivity and rest—one that occurs reflexively in response to reduction of sensory input—to the current view that sleep is a highly organized state generated by the cooperative interplay of many behavioral and neural components. Although in many respects the biological function of sleep remains a mystery, we have begun to understand the cellular and molecular processes underlying sleep.

Sleep affects all of our bodily and mental functions, from the regulation of hormonal levels to muscle tone, from the regulation of respiration rate to the content of our thought processes. Given these important behavioral changes, it is not surprising that the brain's overall electrical activity changes significantly with sleep. During wakefulness the electroencephalogram (EEG) shows relatively low-voltage, high-frequency, fast activity, reflecting an active cerebral cortex busy with perception and cognition. Slow synchronized oscillatory activity is either at a minimum or occurs only transiently in small groups of neurons (Figure 51–1). Relaxing or closing the eyes may produce alpha waves, especially over the visual cortex, indicating relatively synchronous rhythmic activity in the underlying cortical networks.

To describe sleep quantitatively and to distinguish its stages, sleep researchers routinely use three measures: brain activity measured by the EEG (see Chapter 50), eye movements recorded by the electro-oculogram (EOG), and muscle tone measured by the electromyogram (EMG). These three measures are used because of their reliability, ease of recording, and discriminatory power.

Non-REM Sleep Has Four Stages

Based on these measures, sleep is subdivided into five stages. Stages 1 to 4 comprise non-rapid eye movement (REM) sleep, while stage 5 is REM sleep. Stage 1 sleep is the transition between waking and sleep, the period in which sleep is thought to be imminent but not yet developed. During stage 1 the EEG exhibits a decrease in the high-frequency activity that characterizes the waking state. Stage 2, the first true stage of sleep, is marked in the EEG by the onset of spindle waves, 7–15 Hz oscillations over a period of 1–2 seconds that resemble a spindle of thread, and K-complexes (Figure 51–1). Spindle waves and K-complexes reflect slow, synchronized oscillations of neuronal and synaptic activity within the thalamus and cerebral cortex. This state is the result of the relaxation and general hyperpolarization of neurons and neuronal networks that follow gradual inactivation of the brain mechanisms of arousal. During stage 2 sleep muscle tone decreases, and the eyes slowly roll back and forth. Respiration becomes more regular and slows (Figure 51–2), and body temperature begins to fall.

Stage 3 sleep is heralded in the EEG by the appearance of a significant fraction of delta wave oscillations (0.5–4 Hz). These signal a further reduction in arousal processes in the brain and increased synchronization of cortical and thalamic activity. A predominance of delta waves (> 50% of the time in the EEG) indicates that the person is in stage 4 sleep, the deepest stage. During stages 3 and 4 respiration continues to be slow and regular, heart rate slows, muscles relax, and temperature slowly drifts downward (Figure 51–2).

The progression from waking to stage 4 sleep at the beginning of the night typically occurs relatively quickly—over the first 30 minutes (Figure 51–1). After approximately 30 minutes in stage 4 sleep the sleeper ascends quickly through all four stages of sleep. Instead of waking, however, the sleeper now enters a unique stage known as rapid eye movement or REM sleep, a period in which dreams can become vivid, even bizarre.

Rapid eye movement (REM) sleep was discovered in 1953 when Eugene Aserinsky and Nathaniel Kleitman first recorded the EEG and EOG of sleeping adults and observed that during sleep these tests showed changes in activity approximately four to five times per night (changing to a state of higher frequency and lower amplitude similar to that of the waking state). During this "activated" period the eyes dart back and forth, thus giving this stage its name. Sleepers roused from REM sleep report experiencing vivid dreams approximately 80% to 95% of the time.

Interestingly, REM sleep is associated with an almost complete loss of muscle tone, owing to inhibition of the spinal motor neurons by descending pathways. Apparently the motor neurons in the brain stem that control eye movements are not inhibited, because the eyes move during REM sleep. During REM sleep temperature regulation is at a low point, and body temperature begins to fall further still.

Although sleep can be segregated into five stages, the REM stage is so distinctive that sleep is often separated into two phases, referred to simply as REM and non-REM (or NREM) sleep. The cycle of sleep—the non-REM sequence from stage 1 to stage 4 and the reverse sequence followed by a brief period of REM sleep—occurs repeatedly during the night, in a pattern that is best described as a damped oscillation. As the night progresses, the depth of non-REM sleep decreases and the duration spent in REM sleep increases (Figure 51–1). As a result of this pattern, and the fact that it is necessary to wake up to remember your dreams, the dreams that take place in the morning are the ones most often remembered.

REM and Non-REM Dreams Are Different

In adults REM sleep occupies approximately 25% of the total sleep time. Most people would be surprised to find that they had vivid dreams for approximately 2 hours every night. Indeed, the inability to remember dreams (unless awakened) is one of the great mysteries of sleep.

Analysis of dreams and their content have revealed some surprising results. Most people believe that dreams are only occasional and fleeting occurrences of highly bizarre or emotional content, in which the entire scenario is dreamt in an instant. In reality, specific dreams recur with a regular and predictable periodicity, and mostly reflect everyday events. Even the fleeting nature of dreams is a misperception. Events in dreams occur over a period of time that is about as long as they would in real time. Only a small percentage of dreams contain bizarre and fantastic elements.

Dreams occur in both REM and non-REM sleep but the characteristics of REM and non-REM dreams differ. REM dreams are relatively long, primarily visual, somewhat emotional, and usually not connected to the immediate events of the everyday life of the dreamer. These resemble what Freud referred to as the latent content of the dream. Non-REM dreams are shorter, less visual, less emotional, more conceptual, and usually related to the current life of the dreamer. They resemble what Freud referred to as the day's residues, or the manifest content of the dream. This conceptual, "thought-like" mentation can occupy up to 50% of non-REM sleep.

What sensory modalities are experienced during dreams in REM sleep? Vision is preeminent, occurring in all dreams (except, of course, in the congenitally blind), whereas auditory events occur in approximately 65%, vestibular 8%, and temperature 4%; tactile, olfactory, and gustatory experiences are rare (only 1% each). The emotional content of dreams varies from anxiety (14%), to surprise (9%), joy (7%), sadness (5%), and shame (2%). Men experience penile erections, and women experience the physiological counterparts of sexual arousal during REM sleep. The regularity of penile erection during REM sleep (Figure 51–2) can help determine if impotence in a patient has a psychological or physiological origin. The emotional content is organized at several levels of the central nervous system but is, for the most part, unrelated to the specific content of the dream.

Sleep is a circadian behavior that is composed of cyclical or ultradian stages. In all species the period of sleep is dictated by the rotation of the earth and time clues (German, zeitgebers) that indicate the day-night cycle. The principal zeitgeber is the daily rising of the sun. It was once thought that the circadian rhythms of plants and animals are strongly or completely driven by the rising and setting of the sun. However, in 1729 the Swiss astronomer Jean-Jacques d'Ortous de Mairan observed that even in the dark the leaves of plants move according to a diurnal cycle. Since then endogenous rhythms of about 24 hours (ie, circadian rhythms) have been found to be prevalent throughout the plant and animal kingdoms. Circadian rhythms seem to have evolved so as to maximize the use of the day-night cycle by anticipating the rising and setting of the sun, even without clear signals from sunlight or other zeitgebers.

The features of the circadian rhythm of sleep in human beings, in the absence of zeitgebers, has been documented by a few brave individuals willing to live in isolation in specially constructed rooms in labs or caves. These rooms contain no clues as to the time of day in the outside world—no clocks, windows, radios, televisions, e-mail, Internet, or sounds that would tell subjects the time in the outside world. Under these conditions subjects were allowed to sleep and be active whenever their body dictated. Interestingly, the circadian rhythm persists, but its duration varies among subjects and is typically slightly longer (eg, 25 hours) than that of the normal day-night cycle, causing each person in isolation to slowly drift out of phase with the outside environment. Reexposure to the external day-night cycle rapidly resets the sleep-wake cycles (Figure 51–3A).

The sleep-wake cycle is not the only circadian rhythm in the body. Many other physiological processes also exhibit daily maximums and minimums. The level of arousal, the ability to do cognitive tasks such as math problems, body temperature, hormone release, and kidney function all follow endogenous 24-hour rhythms that normally are synchronized to the day-night cycle.

How do sleep and other circadian rhythms become synchronized to the day-night cycle? A small number of retinal ganglion cells (see Chapter 26) respond directly to light and project to the suprachiasmatic nucleus in the hypothalamus (Figure 51–3B). It is this pathway, described in detail later, that synchronizes the circadian rhythmicity with the day-night cycle. Although rare, some people with tumors or lesions of the hypothalamus lose the circadian rhythmicity of their sleep-wake cycle and therefore sleep for short periods throughout the day. Other people, equally rare, have a normal circadian rhythm in their sleep-wake cycle but are unable to synchronize it with the day-night cycle of their environment. The phase relationship between their own endogenous need to sleep, and the time of day, changes continually. Perhaps in these people the neural path conveying light information to the suprachiasmatic nucleus is defective.

Jet lag demonstrates other properties of our circadian rhythm. Travel to a distant time zone disrupts the synchrony of one's circadian rhythm with the day-night cycle. One may temporarily experience excessive sleepiness during the day, hunger at the wrong times, cold despite being in a warm environment, and awakenings in the middle of the night despite sleep deprivation. This feeling of being "out of phase" is common and has a physiological basis. Flying west is generally easier than flying east, presumably because our endogenous rhythm is approximately 25 hours. When flying west we take advantage of this long rhythm by staying up late and sleeping in. Indeed, many individuals practice this on a weekly basis—staying up late and sleeping in on the weekend. The jarring experience of the "Monday morning blues" results from a rapid shift in the circadian cycle and perhaps a bit of sleep deprivation thrown in as well.

Sleepiness—the drive to sleep—depends on several factors. Two of the strongest are the time since the last full period of sleep and the circadian rhythm. We therefore can think of the drive to sleep as two separate processes: a "sleep deficit" that slowly builds up without sleep and dissipates with sleep, and a circadian rhythmicity in arousal. The difference between these processes reflects the sleep drive. After a full night of sleep the sleep deficit should be very low. In the morning our level of arousal slowly increases as does our sleep deficit. Because sleep deficit and alertness increase together, the difference between these two processes—the drive to sleep—is small and therefore weak. Toward evening, however, our sleep deficit continues to increase, and the circadian rhythm of arousal begins to wane, resulting in a larger and larger sleep drive (Figure 51–4A). Sleeping a full night once again removes the sleep debt, and the process starts over.

If a night of sleep is missed, one is likely to find that at approximately 3–4 a.m. it is nearly impossible to stay awake. This is because the sleep deficit is high and continues to increase, while the circadian rhythm in arousal (and other body functions such as temperature) is at a low point. Thus the sleep drive is high (Figure 51–4B). Sleep once again restores the balance, relieving this built-up sleep pressure.

The Circadian Rhythm Clock Is Based on a Cyclic Production of Nuclear Transcription Factors

The search for the neural basis of circadian rhythmicity of sleep has led to one site, the suprachiasmatic nucleus (Figure 51–3B), so named because of its location just above the optic chiasm, the place where the optic nerves cross underneath the hypothalamus.

The firing frequency of neurons in the suprachiasmatic nucleus follows an endogenous circadian rhythm (see Figure 51–6C). Thus the 20,000 neurons in the suprachiasmatic nucleus make up the master clock for circadian rhythms. They are the critical pacemaker for organizing sleep into a circadian pattern. The circadian activity of these neurons is entrained by environmental stimuli such as light, as measured by release of the transmitter vasopressin (see Figure 51–6C). Lesioning this nucleus in animals causes complete loss of the circadian rhythmicity of sleep-wake cycles. Animals with these lesions continue to have a normal daily amount of sleep, but sleep occurs randomly throughout the day and night. Transplantation of the suprachiasmatic nucleus from another animal can restore a circadian sleep rhythm, with the host animal adopting the sleep-wake cycle of the donor.

The suprachiasmatic nucleus is organized into distinct functional groups and acts as the controller for the rhythmic oscillations of clocks located in other organs in the body. These "peripheral" clocks are capable of maintaining their own circadian rhythms for only a few days without input from the suprachiasmatic nucleus.

Like a grandfather clock, the biological clocks that control circadian rhythms have many parts: A complex set of transcription factors, proteins, kinases, phosphatases, and regulatory molecules that have been remarkably conserved throughout the evolution of species. A number of molecules that play key roles in the clock mechanism have been identified over the last 20 years, largely from experiments in the fruit fly Drosophila and the mouse. The essence of the clock mechanism in the suprachiasmatic nucleus, as well as in other organs, is a pair of transcriptional feedback loops; one forms the core circadian mechanism whereas a second forms a modulatory loop that stabilizes the circadian rhythm. Such interlocking feedback loops are quite similar in flies and mammals (see Chapter 3).

At the center of the two loops are two transcriptional activators, CLOCK and BMAL1. These transcription factors bind to each other and form a heterodimer that enhances transcription of the mouse gene per (mPer1–3) and cryptochrome genes (mCry1–2), thereby increasing the cytoplasmic concentrations of the PER and CRY proteins. PER and CRY form heterodimers, reenter the nucleus, and inhibit CLOCK and BMAL1, thus repressing PER and CRY transcription. This sequence creates the core circadian mechanism in which Bmal1 RNA reaches a peak 12 hours out of phase with mPer and mCry RNA. CLOCK and BMAL1 are also at the center of the modulatory loop that modifies the levels of CLOCK/BMAL1 heterodimers. Mutations in genes in the stabilizing loop do not disrupt circadian rhythms to the same degree as those in the core loop.

Once the central clock in the suprachiasmatic nucleus generates the rhythm and the environmental light-dark cycle synchronizes it, this information must be communicated to various systems through humoral or electrical signals. To achieve this, the rhythmic molecular signal is transduced into electrical activity in suprachiasmatic nucleus neurons and then transmitted to other brain regions through action potentials.

A number of areas in the hypothalamus receive input from the suprachiasmatic nucleus and play a role in integrating the output of the suprachiasmatic nucleus. For example, a large number of axons from the suprachiasmatic nucleus terminate in the dorsal and ventral subparaventricular zones of the hypothalamus. The dorsal subparaventricular zone appears to project to areas that are important for circadian rhythms of body temperature, whereas the ventral subparaventricular zone and the downstream dorsomedial nucleus of the hypothalamus are important in the sleep-wake cycle. The dorsomedial nucleus may serve as a common final pathway for a number of circadian rhythms as lesions of this nucleus interrupt circadian rhythms for feeding, locomotion, and corticosteroid secretion as well as the sleep-wake cycle.

The Ultradian Rhythm of Sleep Is Controlled by the Brain Stem

Our level of arousal can best be described by a circadian cycle of sleep and wakefulness, and an ultradian rhythm in the sleep period consisting of a damped oscillation between non-REM and REM periods (Figure 51–1B). Although the suprachiasmatic nucleus is critical to the generation and synchronization of circadian rhythms, it is not the generator of sleep and arousal, for animals without a functional suprachiasmatic nucleus continue to exhibit these states.

The search for brain structures essential for arousal and for the ultradian rhythm of sleep focused on the brain stem and hypothalamus. In 1949 Horace Magoun and Giuseppe Moruzzi demonstrated that electrical stimulation of anterior portions of the brain stem results in arousal of the forebrain (see Chapter 46). They dubbed this afferent pathway the "ascending activating system." This pathway in concert with portions of the hypothalamus appears to be responsible for maintaining the waking state. By cutting the brain at different levels, Michel Jouvet in France and others localized the neuronal groups needed to generate the brain activation, muscle atonia, and rapid eye movements of REM sleep to a restricted region of the pons and medulla.

Putting the two findings together it became clear that the brain stem is a central organizer for the control of arousal as well as the other components of REM sleep. For example, chemical stimulation of particular portions of the brain stem can cause a state similar to REM sleep, complete with activation of the EEG and muscle atonia (because of descending inhibition of spinal motor neurons). Lesions of the same region can result in an animal that lacks the muscle atonia of naturally occurring REM sleep, resulting in an animal that acts out its dreams. Similarly, persons with REM behavior disorder, an unusual sleep disorder in humans, become physically active during REM sleep.

Although the precise neuronal circuits generating the circadian and ultradian rhythms of sleep and arousal are not known, we do know some of the major cellular components. Many cholinergic neurons within the brain stem and basal forebrain fire in anticipation of either waking or REM sleep. These cholinergic neurons in turn project throughout the forebrain, including the thalamus, cerebral cortex, and hippocampus. The release of acetylcholine (ACh) depolarizes, and therefore activates, many of these target cells.

Noradrenergic and serotonergic neurons of the brain stem and histaminergic neurons of the hypothalamus also project widely throughout the nervous system and modulate neuronal excitability during different phases of the sleep-wake cycle. These cells fire in anticipation of the waking state and maintain a regular, almost pacemaker-like, discharge rate throughout waking. The discharge rate of the noradrenergic cells increases in response to novel stimuli, while that of the serotonergic cells increases in response to changes in blood pH. Finally, the ventral lateral preoptic nucleus of the anterior hypothalamus includes cells that inhibit ascending activating networks, thus promoting sleep.

Although the rate of firing of noradrenergic, serotonergic, and histaminergic neurons increases in anticipation of and during the waking state, it decreases ahead of REM sleep and may even become completely silent. These cells are therefore known as REM-OFF cells. Other neurons, some of which are cholinergic, discharge in anticipation of and during the period of REM sleep and are known as REM-ON cells (Figure 51–5). Presumably, the discharge of these REM-ON cells is responsible for the activation of the forebrain during REM sleep. Interestingly, the final step of Otto Loewi's experiment demonstrating that ACh is a chemical transmitter came to him in a dream. Thus ACh participated in its own discovery.

The cyclicity of REM and non-REM sleep can be modeled as an antagonistic interaction between REM-OFF (histaminergic, noradrenergic, and serotonergic) cells and REM-ON (cholinergic) cells.

During REM sleep phasic bursts of action potentials occur throughout the forebrain in association with rapid eye movements. These bursts of activity are strongest within the pontine and geniculate nuclei and the occipital (visual) cortex and therefore are known as ponto-geniculate-occipital (PGO) waves. These waves appear to be initiated by the phasic discharge of brain stem cholinergic neurons; the activation of nicotinic ACh receptors on cells in the forebrain results in a rapid depolarization of the postsynaptic cells. During waking PGO waves may result from activation of brain stem mechanisms mediating shifts in attention associated with eye movements.

Sleep-Related Activity in the EEG Is Generated Through Local and Long-Range Circuits

Sleep is associated with characteristic patterns of activity in the EEG. Non-REM sleep is dominated by sleep spindles, appearing as 1–2 second periods of 7–14 Hz waxing and waning oscillations, and delta (0.5–4 Hz) or slower (< 1 Hz) oscillations. How and where are these EEG rhythms generated?

Slices of forebrain isolated from the rest of the brain can generate activity similar to that which occurs during sleep. For example, thin slices of either the thalamus or the cerebral cortex generate spontaneous activity that resembles the EEG rhythms of non-REM sleep (Figure 51–6). Thus at least some of the spontaneous firing patterns of the forebrain during non-REM sleep result from local neuronal networks alone, without requiring external modulatory input. These firing patterns reflect the basal, or resting, state of the forebrain when it is relieved of the barrage of neurotransmitters from other structures in the waking state and REM sleep.

In fact the progression from waking to stage 1 to stage 4 sleep is the consequence of a gradual withdrawal of modulatory neurotransmitters, allowing forebrain neurons to gradually hyperpolarize and shift into the basal state and the generation of the slow EEG-rhythms of non-REM sleep. Once the activity of these neurons becomes slow and synchronized, cognitive function is reduced or abolished. The observation that waking and REM sleep are active states that must be maintained by ongoing activity in modulatory pathways fits well with our intuitive notion of sleep. We isolate ourselves from stimuli and relax into a hopefully blissful repose.

The purpose, if any, of rhythmic brain activity during sleep is not yet known. It may merely be the expression of the system properties of neural circuits that are idle during sleep. Or rhythmic activity may serve to keep cells active without the brain being sensitive to external stimuli. This activity could tune or recalibrate the brain during sleep—allowing useless information to be forgotten and important facts to be retained.

Sleep shows striking and characteristic changes with age. As every new parent quickly learns, the sleep of a newborn, although plentiful (16–18 hours a day), is distributed almost randomly throughout the day. What the parent may not realize, however, is that more than 50% (8–9 hours) of that sleep is spent in a REM-like state (Figure 51–7).

Sleep recordings from a premature infant exhibit an even higher percentage of REM sleep, indicating that in utero the fetus spends a large fraction of the day in a brain-activated but movement-inhibited state. As neuronal activity influences the development of functional circuits in the brain, it is reasonable to think that the spontaneous activity of the immature brain during sleep facilitates the development of circuits.

By approximately 4 months of age the average baby begins to show diurnal rhythms that are synchronized with day and night, much to the relief of weary parents. The total duration of sleep gradually declines, however, so that by 3 to 5 years of age the child may sleep 10 to 12 hours a day. At these early ages sleep is deep; stages 3 and 4 are prominent, with an abundance of delta waves in the EEG. As a result, children are not easily wakened by environmental stimuli.

Even though children ages 3 to 5 years may spend a large percent of their sleep in REM sleep, only approximately 33% of awakenings result in recall of dreams. These dream reports are brief, and contain virtually no story, self-representation, or interaction between the characters in the dream. However, by the age of 7 to 9 years the dream reports take on more of the qualities of those of adults, including the ability to analyze, abstract, manipulate, and construct visuospatial images or ideas. The content of dreams parallels the development of cognitive skills in the dreamer.

With age, sleep becomes less deep and more fragmented. The percent of time spent in stages 3 and 4 of non-REM sleep may drop dramatically, especially in old age. This results in a significant decrease in the perceived quality of sleep and an increase in daytime sleepiness. Disorders of sleep, such as insomnia and sleep apnea, also become more prevalent with age.

Although all animals have circadian rest periods, the similarity between these and human sleep decreases with evolutionary distance. Nevertheless, some scientists contend that all animals, from fruit flies to man, do in fact sleep. The duration of sleep and the behaviors associated with it vary widely between species. Sleep seems essential to life, but its precise features have been modified to fit the requirements of each species.

The sleep habits of more than 90 mammals have been studied. Some sleep in well-protected burrows (eg, moles and rabbits), whereas others sleep in the open in herds. Giraffes sleep only approximately two hours a day, cats sleep approximately 13 hours, and the brown bat sleeps almost 20 hours. Some mammals (eg, cattle) sleep with their eyes open. Some, such as dolphins and porpoises, sleep with only half of their brain at a time. Presumably this adaptation allows these sea mammals to continually swim to the surface for air. All mammals have clearly recognizable non-REM sleep and nearly all have REM sleep. One notable exception is the spiny anteater, which has only non-REM sleep.

Birds also have REM and non-REM sleep, although during REM sleep they do not lose muscle tone completely, allowing them to remain perched. A state resembling non-REM sleep is recognizable in reptiles, amphibians, and fish, although REM sleep is not evident.

Nocturnal animals are active at night, exploiting an ecological niche, and sleep during the daylight. Other species, such as humans, make use of the abundant sunlight during the day and are diurnal.

More than half of the population experiences significant difficulties with sleep at least on occasion, and as many as one in five persons suffers from chronic sleep problems. Disruption of sleep and waking is the most prevalent health disorder in the United States. Falling asleep while driving, for example, is thought to be responsible for at least 100,000 traffic accidents every year. Lack of sleep was a contributing factor in the disasters at the Three Mile Island and Chernobyl nuclear power plants.

Millions of people struggle through daily life sleepy, and many are irritable or unmotivated owing to disruptions in sleep. As a result, sleep disorders are common in general medical practice. Most sleep problems have mundane causes and may simply require a change in habit. Others involve complicating factors such as shift work, depression, or substance abuse. Some sleep disorders now provide new insights into brain function.

As sleep is organized into several cyclical, roughly 90-minute periods of REM and non-REM sleep, each component can be disrupted. The most common disorders are related to a breakdown in the transition between sleep and waking, such as difficulty falling asleep or early morning wakening. More than half of adults experience insomnia at some point each year. Their sleep is either too short, difficult to obtain, or not refreshing. The lack of adequate or fulfilling sleep often leads to another common complaint: daytime sleepiness.

Other sleep disorders may represent a breakdown in specific neural circuits. For example, sudden activation of the inhibitory descending motor pathways during the waking state can lead to bouts of cataplexy (loss of muscle tone), which is often associated with narcolepsy (described below). The opposite can also happen. In REM behavior disorder the descending inhibitory motor pathways do not function during REM sleep, such that the affected individual may jump up and move about during vivid dreams.

In other sleep disorders circadian rhythmicity may fail, resulting in a phase advance or phase delay in the sleep-wake cycle. Finally, sleep may be associated with inappropriate or unwanted behaviors such as night terrors, sleep walking, tooth grinding, bed wetting, and so on. These unusual behaviors disrupt sleep and are collectively known as parasomnias.

Insomnia Is the Most Common Form of Sleep Disruption

Insomnia, the most common form of sleep disruption, can be prolonged and severe or temporary and mild, as in response to short-term stress. The incidence of insomnia increases with age and is more common in women. Insomnia can manifest as difficulty falling asleep, difficulty staying asleep through the night, or early morning awakening before sufficient sleep is obtained. Insomnia may result from physical or emotional complications or simply from poor sleep habits (eg, consumption of excessive caffeine, alcohol, or food, or exercising vigorously before sleep). Insomnia is often associated with depression where early morning awakening is common.

Benzodiazepines, commonly used as anxiolytics (mild tranquilizers), are also commonly used for short-term treatment of insomnia. Benzodiazepines facilitate the opening of GABA_A (γ-aminobutyric acid type A) receptor-channels through a binding site separate from that for GABA. Although these drugs facilitate sleep, they also suppress stage 4 sleep. Chronic use may be habit-forming and lead to a lightening and fragmentation of sleep. A related compound, zolpidem (sold under the trade name Ambien), binds selectively to a subset of benzodiazepine receptors, does not suppress deep sleep, and is considered relatively selective for facilitating sleep. It is therefore safer than general benzodiazepines for treatment of insomnia.

Over-the-counter sleep aids are often antihistamines (H₁ histamine receptor antagonists). Antihistamines are sedative in some individuals because they antagonize the activation of H₁ receptors, which normally excite thalamic and other central neurons. Muscarinic ACh receptor antagonists also have sedative effects, although these vary between individuals. Anticholinergic actions are common to many psychoactive drugs, such as tricyclic antidepressants, and therefore these drugs are used for insomnia. As discussed earlier, activation of muscarinic ACh receptors in the brain has a general arousal effect by exciting the principal neurons and, in some parts of the brain, by inhibiting GABAergic inhibitory interneurons.

Excessive Daytime Sleepiness Is Indicative of Disrupted Sleep

Sleepiness during the day is the primary complaint of individuals seeking help at hospital sleep centers. The invention of electric lights has allowed people to stay active well into the night. The increasing demands of work, family, and social life along with ever increasing opportunities for entertainment have led to a dramatic shortening in the time spent asleep, from an average of 10 to approximately 7 hours.

This trend has led to a chronic sleep deficit in the average adult that affects not only mood and general feelings of well being, but also work or school performance. Sleep deficit can be dangerous, as when driving or operating machinery. Telltale signs of insufficient sleep include unusually short sleep-onset latencies (less than 15 minutes), the need to sleep in on weekends, overuse of stimulants such as caffeine, and dependence on an alarm clock to wake up in the morning.

Excessive daytime sleepiness is a symptom of several underlying sleep disorders including sleep apnea, narcolepsy, and restless leg syndrome. Sleep apnea and restless leg syndrome are two of the most prevalent disorders that lead to daytime sleepiness, largely through a reduction in the quantity and quality of sleep. Narcolepsy is present in only a small fraction of the population but is devastating in its negative effects.

The Disruption of Breathing During Sleep Apnea Results in Fragmentation of Sleep

Reduction of muscle tone can result in either an annoying disturbance of sleep, as in snoring, or in greatly disturbed sleep, as in sleep apnea (cessation of breathing). Approximately 4% of middle-aged men (2% of middle-aged women) have sleep apnea. At older than age 65 years these percentages increase to more than 28% of men and 24% of women.

Sleep apnea is a serious condition, for it not only disrupts sleep but also can cause excessive sleepiness during the day, early morning headaches, depression, irritability, sexual dysfunction, and learning and memory difficulties. In severe cases sleep apnea can lead to cardiac dysfunction and be life-threatening.

The most common form of sleep apnea, obstructive sleep apnea, results from a physical obstruction of the pharynx. Breathing ceases for 10 or more seconds. A pathological condition involves at least 5 to 10 of these interruptions per hour. In central sleep apnea the disruption of breathing has a neurological origin. In either type of sleep apnea the decrease in blood oxygen tension and increase in CO₂ leads to an arousal response (but not complete awakening). The period of apnea ends with a deep breath, and the sufferer then falls back into a deeper sleep, only to repeat the cycle once again (Figure 51–8).

These frequent disruptions cause fragmentation of sleep, a general decrease in quality of sleep, and therefore daytime sleepiness. Because the affected individual usually does not become fully awake with each apneic period, he may be completely unaware of the problem except for the resultant daytime sleepiness or the concern of a friend or spouse who observes an apneic spell. This pattern is usually easy to diagnose in a sleep laboratory.

REM sleep is especially conducive to sleep apnea owing to the characteristic reduction in muscle tone, which facilitates collapse of the airway. During REM sleep responses to hypoxia are reduced, and the response to increases in carbon dioxide levels (hypercapnia) is completely lost. In addition, the brain stem mechanisms responsible for regulation of blood O₂ and CO₂ levels are weakened, allowing these levels to drop well below those that would be tolerated in the waking state. Obesity can be a precipitating factor for obstructive sleep apnea because airflow through the pharynx is typically reduced.

Obstructive sleep apnea is treated by keeping the pharynx open with positive pressure through a face mask. This treatment can lead to an almost immediate reversal of the symptoms, a more restful sleep, and a reduction in daytime sleepiness. However, most patients eventually reject use of the mask because they find it too uncomfortable or claustrophobic to use consistently.

Narcolepsy Is Characterized by Abnormal Activation of Sleep Mechanisms

Narcolepsy occurs in approximately 0.04% of the population or approximately 120,000 people in the United States. It is characterized by a breakdown in the transition between waking and sleep. Sleep and sleep mechanisms invade daytime periods, often at inappropriate times, and sleep at night is fragmented and disrupted by multiple awakenings. Narcolepsy is characteristically manifested in five symptoms, some or all of which may be seen in any one individual.

The most prevalent symptom is excessive daytime sleepiness and irresistible "sleep attacks" during the waking hours. These sleep attacks (usually less than 20 minutes) can come at any moment and sometimes are a source of embarrassment. For example, falling asleep during conversation or at work may lead to the incorrect assumption that the narcoleptic is uninterested in the conversation or in doing a full day's work. Sleep attacks are precipitated by any behavior that is relatively passive and boring such as watching television, driving a car, or studying for finals. Unfortunately, sleeping does not completely alleviate the tendency to have sleep attacks.

A second symptom, cataplexy, occurs in approximately 70% of narcoleptic patients. Cataplexy is a sudden bilateral loss of muscle tone, typically in the knees and face and neck, leading to a sagging of the jaw and falling to the floor. Consciousness is preserved, however, and the sufferer is typically awake but feels either unable or barely able to move. The onset of a cataplexic episode occurs over a couple of seconds; the episode itself lasts for seconds but in rare instances can last minutes. Emotion, most typically laughter, can provoke an attack, perhaps owing to the fact that during laughter there is a general decrease in muscle tone. Cataplexy is thought to result from abnormal activation of the motor inhibition that normally occurs during periods of REM sleep.

Third, vivid dreamlike experiences may occur during the transition between sleep and waking. These events are known as either hypnagogic (sleep onset) or hypnopompic (sleep offset) hallucinations. What differentiates them from dreams is that the narcoleptic person is not fully asleep and is aware that he is not dreaming—the images seem real. The dreamlike experiences are typically bizarre, frightening, and unpleasant; they are predominantly visual, although auditory and tactile experiences are also frequent. As with dreams, sensations of smell or taste are rare. Hallucinations of this type may also occur in normal people at transitions between sleep and waking.

A fourth symptom, sleep paralysis, also typically occurs in the transition between sleep and waking (either going to sleep at night or waking in the morning). Unlike cataplexy it is not triggered by emotion and episodes last longer, sometimes up to 10 minutes. As sufferers are unable to make even the smallest of movements, such as opening their eyelids or lifting a finger, the experience is frightening and unpleasant. As with cataplexy it is assumed that sleep paralysis results from inappropriate activation of the inhibitory descending motor pathways that are normally responsible for inhibiting movement during REM sleep. Sleep paralysis also occurs in non-narcoleptics (up to 30% of the general population).

The final symptom of narcolepsy is disturbed nocturnal sleep. Although narcoleptics may fall asleep quickly and often immediately fall into REM sleep (Figure 51–9), their sleep is interrupted by frequent arousals. For most narcoleptics these awakenings are brief, but some may stay awake for hours at night.

A diagnosis of narcolepsy requires a minimum of either (1) daytime lapses into sleep and cataplexy or (2) a pattern of excessive daytime sleepiness, sleep paralysis, or hypnagogic or hypnopompic hallucinations, and a polysomnogram demonstrating a short latency to sleep and a period of REM sleep either at the beginning of sleep or shortly thereafter.

The different components of narcolepsy are treated with different pharmacological agents. The excessive sleepiness and sleep spells of narcolepsy are typically treated with central nervous stimulants such as amphetamines. These agents enhance the release of catecholamines and inhibit their reuptake. Cataplexy is treated with tricyclic antidepressants, which inhibit the reuptake of norepinephrine and serotonin and are potent inhibitors of REM sleep.

Both genetic and environmental factors contribute to narcolepsy. In only 1% of cases does narcolepsy run in a family and only approximately one-fourth of identical twins are concordant for narcolepsy. Thus narcolepsy genes appear to confer merely a susceptibility to narcolepsy, whereas environmental factors contribute strongly. By studying narcoleptic dogs, Emmanuel Mignot found that these animals have a defect in the gene coding for the hypocretin-2 receptor. Hypocretin-1 and -2 (also called orexin A and B) are neuropeptides produced in a small cluster of neurons within the hypothalamus that project widely throughout the brain, including the brain stem nuclei responsible for REM sleep (eg, locus ceruleus). Activation of hypocretin receptors in locus ceruleus and other neurons has a slow excitatory effect. Through an unknown mechanism, a loss of this or other modulatory actions of hypocretin results in the abnormal differentiation of sleep and waking typical of narcolepsy.

To study the role of the hypocretin system in feeding behavior, Masashi Yanagisawa and collaborators knocked out the genes for both hypocretin-1 and -2 in mice and found that these mice displayed symptoms typical of narcolepsy. In narcoleptic humans the cerebral spinal fluid levels of hypocretin are remarkably reduced even though the hypocretin genes are not mutated. Immunocytochemical examination of the brains of deceased narcoleptics has revealed a striking loss of hypocretinergic neurons (Figure 51–10).

Restless Leg Syndrome and Periodic Leg Movements Disrupt Sleep

Restless leg syndrome occurs in approximately 8% of the population and is characterized by an irresistible urge to move the legs. This symptom occurs during the day but is usually worse at night while resting in bed. The patient may feel relief only by moving the legs either in bed or by walking about.

Approximately 80% of people suffering from restless leg syndrome also experience periodic leg movements in sleep, when the legs move for a few seconds every 10 to 20 seconds. These leg movements result in a lightening of sleep and therefore an increase in daytime sleepiness. The prevalence of restless leg syndrome and periodic leg movements increases greatly with age, becoming more common in the elderly.

Parasomnias Include Sleep Walking, Sleep Talking, and Night Terrors

Parasomnias are repeated disruptions of sleep that are relatively common and usually mild. They include sleep walking, sleep talking, confusional arousals, bed wetting, night terrors, and REM behavior disorder. Although not normally considered a sleep disorder, sudden infant death syndrome (SIDS) is an example of fatal respiratory failure that occurs during sleep (see Chapter 46).

Sleep walking, sleep talking, and confusional arousals are relatively common in children and typically occur during stages 3 and 4 sleep. These events may be precipitated by disturbing or arousing the child from deep sleep. For short events the EEG is a continued pattern of slow waves; for longer events the pattern changes to an "activated" pattern—low-voltage, desynchronized high-frequency activity—characteristic of waking. Usually sleepwalkers or talkers are unaware of an event and have no memory of it. Speech is largely incoherent during an event; remarkably sleep walkers are often able to avoid colliding with objects.

Night terrors also occur in stages 3 and 4 sleep and are common in children (approximately 2–3%). The child cries as if in great fear, perhaps with eyes wide open and heart rate greatly elevated, for several minutes or longer. Even though they may appear awake, they are actually asleep. During the episode they are inconsolable, and attempts to calm or wake the child may only cause the screams and fearful behavior to worsen. The child usually does not remember the night terror, in contrast to remembering nightmares. Thus night terrors are typically much more difficult for the parent than for the child.

During REM sleep descending inhibition of motor neurons in the spinal cord normally prevents people from acting out their vivid dreams. This descending pathway may be damaged in people suffering from REM behavior disorder (usually men and elderly people). During REM sleep the body may become rigid or extremely tense and may exhibit prominent muscle twitching in association with the phasic eye movements of REM sleep. If the dream has intense physical activity, the patient may even get out of bed and injure himself or his bed partner with his vigorous activity. The mechanisms underlying REM behavior disorder are not yet known.

Circadian Rhythm Sleep Disorders Are Characterized by an Activity Cycle That Is Out of Phase with the World

The most frequent circadian rhythm disorder involves problems with entrainment to the day-night cycle. A patient may find it impossible to fall asleep until very late at night, for example 3 a.m. (delayed sleep phase disorder or DSPD) or impossible to stay awake past the early evening hours, for example 7 p.m. (advanced sleep phase disorder or ASPD).

The most prevalent disruption of normal circadian rhythms occurs in those who work either part-time or full-time during the night. In some individuals forcing a lifestyle that is out of phase with the natural day-night cycle can lead to several medical problems, including unstable mood, vigilance impairments, colds, high blood pressure, stomach problems, and weight gain. In general, shift workers also experience a higher incidence of workplace and automobile accidents because of increased sleepiness.

In rare cases patients may lack an endogenous circadian rhythm or may not be able to synchronize their rhythm with that of the environment. As one can imagine, these disorders cause great difficulties for those afflicted.

For a behavior that involves up to 25 years of our lifetime, it is surprising how much we do not yet fully understand about the functions of sleep. Sleep is essential for life, at least for rats. A rat deprived of all sleep dies after approximately 3 to 4 weeks owing to a breakdown in metabolic processes. In fact, sleep deprivation can cause death even more quickly than food deprivation.

In practice we sleep because we are sleepy, just as we eat or drink because we are hungry or thirsty. But the function of sleep is not to relieve sleepiness. Intuitively we think of sleep as restorative, but exactly what is restored is unknown. In addition to a restorative function, other functions have been proposed, including energy conservation, memory consolidation and sorting, recalibration of neuronal networks, and behavioral suppression during periods of the day for which the animal is not well adapted.

One explanation for this diversity of function is that bodily functions have, over evolution, become segregated into different portions of the day/night cycle so as to make efficient use of resources. Indeed, studies of gene expression reveal that a massive reorganization of the molecular and cellular biology of neurons occurs in the transition from alertness to sleep, suggesting that a similar marked reorganization occurs during sleep. Sleep is therefore likely to have many functions, just as waking does.

In fact, it may be useful to turn the question "Why do we sleep?" on its head and ask instead, "Why do we wake up?" In general, periods of waking are generated to accomplish a goal, for example, the procurement of food and housing, having a family, building a society; otherwise rest is warranted. By keeping the animal inactive during periods in which it is not well-suited for wandering about, sleep may ensure adequate rest and the reservation of energy stores. Waking up and becoming active can cost at least an additional 10% of energy. Given scarce food resources, ceasing activity for a significant fraction of each day can lead to a small but significant increase in survival. Thus sleep may be a light and short version of hibernation, an easily generated and reversible period of rest and energy conservation.

Another idea is called the protection hypothesis. Most animals are adapted for activity during the day or night. Being active at the wrong period brings with it considerable risks both from other animals (eg, predators) or environmental conditions (eg, cold during the night or heat during the day, or unseen objects because of low light levels).

Sleep involves the entire organism and in fact involves the family and society as a whole. Even so, many neuroscientists believe that sleep is of and for the brain. One neural function that is often proposed for sleep is memory consolidation. Disruption of sleep has detrimental effects on some types of learning, particularly procedural memory, although the stressful effects of lost sleep are always a confounding factor.

One idea holds that the spontaneous activity occurring in the sleeping brain is a type of neuronal exercise machine—keeping the brain "in tune" for the requirements of daily activities. By passing spontaneous patterned activity throughout neuronal circuits during the night, the brain can recalibrate the strength of synaptic connections as well as the electrophysiological properties of single neurons, resulting perhaps in the retaining of important information and the forgetting of irrelevant information. The presence of spontaneous patterns of neuronal activity in all animals during sleep, which is energetically more costly than neuronal silence, suggests that this activity plays an important functional role in sleep. Indeed, the spontaneous patterns of activity occurring in the brain during REM sleep are the neural basis of dreams.

The importance of REM sleep is a particularly thorny issue. Patients on some forms of antidepressant may go years with a near complete suppression of REM sleep without any marked deficits in the ability to form new memories or otherwise function normally. REM sleep may allow for a more immediate response to the environment during sleep (which has obvious survival value in the wild) because animals awakened from REM sleep are more responsive and show better sensory and motor function than those awakened from non-REM sleep. In a group of sleeping animals at least one or more of them may be in this state of readiness at any given moment and therefore able to alert the pack of an intruder.

Both REM and non-REM sleep are highly regulated processes that the brain and body appear to crave. Deprivation of REM sleep results in "REM rebound" in which the loss is made up by excessive REM sleep in the following night's sleep. This regulation suggests that REM sleep, and sleep in general, plays one or more critical functions. It remains for us, or rather our brains, to discover the nature of these functions so that we may more fully understand the more ethereal third of our lives.