The activity of individual neurons generates electrical patterns due to the intracellular/extracellular charge differential and the action potential. Certain laboratory preparations allow the recording of single neurons or the activity of small regions of the brain. In rather limited circumstances, electrodes placed on the surface (electrocorticography) or within the human cerebral cortex (depth electrodes) allow precisely localized recordings. However, in human research and clinical studies, the EEG is a convenient method that offers a representation of the composite electrical activity of millions of neurons below a particular EEG lead. Soon after Hans Berger’s development of the EEG in the late 1920s, it was used to examine brain activity during sleep. It was evident immediately that sleep was not a homogeneous state but rather was associated with a dramatic range of brain electrical activity. Different EEG patterns generally could be correlated with the depth of sleep, as reflected by observations and experimental arousal thresholds. Loomis, Harvey, and Hobart (1936) developed a system of sleep stages with criteria based on particular EEG wave patterns and the amount of slow-frequency electrical activity present. They suggested that greater slow-wave activity was associated with deeper sleep.
At different times and from different general brain regions, the EEG may record amplified neural electrical activity across a rather wide frequency spectrum and with a significant amplitude range. The frequency distribution typically is divided into bins, which are familiar because of their wide application. They are described in terms of the equivalent units of cycles per second (cps) or hertz (Hz). Although the precise boundaries of frequency bins may vary slightly with different systems, the typical standard demarcations of the EEG frequency spectrum, beginning with the slowest, are shown in table 2.1. The demarcations are somewhat arbitrary; however, they do represent useful clusters in describing brain activity in different stages of sleep and waking.
Standard EEG recordings typically incorporate filters that block out frequencies at the high and low ends of the frequency spectrum (Carskadon and Rechtschaffen, 2000). However, investigators may focus specifically on other ranges. For instance, there has been recent interest in very low frequency brain activity (below 1 cps) and in the spectrum from about 20 to 50 cps, which is termed the gamma range. In fact, the frequency bins can be abandoned entirely in certain research applications that use signal-processing formulas made possible on a large scale by digital technology. With these techniques it is possible to measure the degree of activity of any defined frequency range over a specified time period. For example, the mathematically calculated power of 1 to 2 cps activity per minute can be examined over an entire eight-hour sleep period. The EEG signal amplitude may vary considerably with states of consciousness and among individuals. Children tend to generate signals of greater amplitude than do older adults. For an individual, the signal amplitude typically is greater at the lower end of the standard frequency spectrum. A primary reason is synchronization, wherein a pacemaker organizes simultaneous electrical patterns leading to the accumulation of the signal power. Recall that the amplified activity from a single EEG lead always represents a composite of the local neurons. During wakefulness, the neurons are doing many different things at the same time. During certain phases of sleep, numerous neurons are firing simultaneously. This is a normal phenomenon and is evident especially in what is termed slow-wave sleep. Pathological high-amplitude EEG patterns due to abnormal synchronous discharge commonly are observed in individuals with seizure disorders.
The current standardized system of sleep stages incorporates criteria from one or more EEG leads as well as activity recorded from electrooculogram (EOG) and electromyogram (EMG) leads (Carskadon and Rechtschaffen, 2000). Together, these leads represent the most basic polysomnogram (PSG) for scoring sleep stages, although other parameters (e.g., electrocardiogram, respiratory effort, oxygen saturation) typically also are monitored. The system was codified in 1968 (Rechtschaffen and Kales) and evolved from earlier reports of the EEG and sleep depth but incorporated new understandings of sleep physiology related to the phase of sleep that is termed rapid eye movement (REM) sleep. Polysomnographic records, typically representing an entire night, are evaluated in 30second epochs, each of which is assigned a score of awake, REM sleep, or non-REM (NREM) sleep, the last of which must be specified further as stage 1, 2, 3, or 4. Movement artifact may be coded if signal interference makes the epoch impossible to designate. The duration of the PSG study, most typically a normal nighttime sleep period, can be described in terms of the amounts and percentages of waking and the individual sleep stages, and it can be graphed over time to create a visual representation of the progression of sleep stages throughout the recording time. The PSG is vital in the evaluation of several different sleep disorders, and it is an important tool for research regarding sleep.
The normal waking EEG typically will exhibit a mixed-frequency pattern with relatively low voltage, assuming the subjects have their eyes open or are attending to some mental activity. However, if they are relaxed with their eyes closed, then alpha activity often will predominate on the EEG, particularly in the occipital region. Typically, the sleep-onset transition is identified as relaxed wakefulness becomes light sleep, which is indicated by a disappearance of the alpha pattern and a general slowing of the background rhythm. There continues to be a low-voltage mixed-frequency pattern, but there also may be theta activity and sharp waves at the vertex. This light sleep represents NREM stage 1, and normally it lasts no more than a few minutes at the beginning of the sleep period. It may reappear for brief periods during a normal night. In certain circumstances, including some sleep disorders, however, there may be large amounts of NREM stage 1 sleep. Excessive light sleep measured on the PSG often corresponds with a subjective report of unrefreshing sleep and sometimes the sense of not having slept at all.
Aside from the changes in frequency and amplitude associated with levels of waking and sleep, other EEG complexes normally are encountered while monitoring sleep. Several of them have become integrated into criteria for sleep stages. These include the K-complex and spindles that are found during NREM stage 2 sleep. The K-complex is a distinctive, sharp, negative deflection followed by a positive component, together lasting up to 0.5 second. The spindles are EEG events lasting 0.5 to 1.5 seconds and having a frequency between 12 and 14 cps. As the name suggests, spindles have a characteristic shape with a gradual increase and then decrease in amplitude. Both K-complexes and spindles occur in the context of the typical NREM stage 2 background rhythm of mixed-frequency low-amplitude activity. Most people normally will spend at least one-half of their total nighttime sleep in NREM stage 2.
The EEG pattern evolves a dramatically different appearance when the rhythm slows and the amplitude increases due to thalamic pacemaker activity. Delta waves are identified when the frequency is below 4 cps and they meet the amplitude scoring criteria of at least 75 microvolts. This slow-wave activity occurs mostly during the first few hours of normal nighttime sleep, primarily before the first REM period and to a lesser extent between the first and second REM periods. The degree to which slow-wave activity is present varies considerably among individuals and is influenced strongly by age. Children tend to have high-amplitude and longer-duration slow-wave activity, which may account for 30 percent or more of their total sleep. With aging there is a gradual decline in the amplitude and duration of the EEG activity that would meet the criteria for NREM stages 3 and 4. It is not unusual for older individuals to have little or no slow-wave sleep. The designations of NREM stage 3 and stage 4 sleep are continuous and are based on the percentage of qualifying delta wave activity in each epoch. If more than 20 percent of the epoch consists of delta activity, then it is coded as NREM stage 3 unless the delta waves constitute over half of the epoch, in which case it is coded as NREM stage 4.
The slow-wave or delta EEG pattern is the key feature of NREM stages 3 and 4, and sometimes these stages together are termed slow-wave sleep (SWS) or delta sleep. In contrast to the highly activated brain states of wakefulness and REM sleep, the portions of NREM sleep where intense slow-wave activity predominates (i.e., relatively early during a normal night) reflect a different level of dynamic brain organization. In the terminology of chaos theory, this very regular SWS would be ascribed a rather low correlation dimension. Accordingly, the brain in SWS can be seen as lacking flexibility. Some people have difficulty quickly returning to agile alertness on awakening from sleep, especially after prolonged and intense slow-wave activity. This phenomenon is termed sleep inertia. Memory processing can be impaired, and there may be frank confusion. There is a common anecdote of people being awakened by a ringing telephone about one hour after sleep onset and then not recalling the conversation the following morning. Abundant slow-wave activity seems to play a role in the pathology of certain parasomnia sleep disorders, such as sleep terrors. The problem with feeling worse (e.g., mentally foggy, sluggish) after prolonged daytime napping is this sleep inertia, which becomes more likely as people sleep longer. Since SWS at nighttime is mostly early in the night, there still are several hours to recover from the effects. Compared with SWS, the only global brain activity with a lower level of organization, other than a completely flat line on the EEG, would be the highly regular pathological activity during a seizure. To a certain extent, there is a similarity between the postictal and sleep inertia states.
The NREM sleep stages are punctuated by REM sleep that occurs in distinct episodes totaling about 15 to 25 percent of normal nighttime sleep. The first REM episode typically emerges roughly 90 minutes after sleep onset, and REM sleep recurs about every 90 minutes. Usually, there are four to five episodes during a night. The REM episodes have a tendency to lengthen in duration as nighttime sleep progresses. Thus, the majority of REM sleep typically is during the latter part of the night. This aspect of the timing of REM sleep is influenced by the circadian system. The shift from NREM to REM and back is determined in the pons region of the brainstem, and it is regulated by the balance of cholinergic activity promoting REM and monoaminergic activity (norepinephrine and serotonin) promoting NREM (Siegel, 2000).
REM sleep is vastly different from NREM sleep. During REM sleep, the brain is in an activated state, somewhat similar to being awake but without the same input and output functions. Cerebral blood flow and certain metabolic activities are at the waking level. Although mentation may occur during any sleep stage, REM sleep is associated with active and sometimes intense dreaming. Other physiological changes include greater heart rate and blood pressure lability, markedly decreased thermoregulation, and the presence of penile tumescence. There is a remarkable skeletal muscle atonia during REM sleep. This results from a medullary process of active inhibition that prevents stimulation of the cerebral cortex motor area from being transmitted to the muscles. Accordingly, EMG leads register a decreased muscle tone associated with REM sleep. The other key feature of REM sleep is the presence of distinctive rapid eye movements, which are evident on the EOG tracing. The EEG activity during REM sleep is not particularly distinctive, in that it is a mixed-frequency, low-amplitude tracing, although sometimes there is a characteristic sawtooth pattern. This helps explain why it was two decades after the initial EEG sleep-scoring system before REM sleep was discovered. The data from the EOG and EMG are necessary for the identification and coding of REM sleep.
The hypnogram offers a convenient visual representation of a sleep period, typically a full-night PSG recording. Figure 2.3 shows a relatively normal night of sleep lasting about eight hours and incorporating the generally normal progression and amounts of sleep stages. There is considerable night-to-night and individual variation in the pattern, however, and the amounts of sleep stages are influenced strongly by age. Nevertheless, the hypnogram is quite useful in emphasizing the characteristic tendency for most SWS (NREM stages 3 and 4) to occur during the early hours of the night and for the recurrent REM episodes to lengthen in duration later in the night. A few brief awakenings identified by the PSG generally would be considered normal. Some abnormal patterns can be seen immediately on a hypnogram, such as a prolonged sleep-onset latency, early-morning awakening, frequent nighttime awakenings, and excessive NREM stage 1 sleep. In clinical sleep laboratory recordings, the resulting hypnogram also will show sleep-disordered breathing events and certain body movements, which can be correlated with the changes in the sleep stage.
Source: David N. Neubauer, “Understanding Sleeplessness: Perspectives on Insomnia,” The Johns Hopkins University Press, Baltimore 2003
Republished by Health Care Programs