Adult Sleep Scoring

Harry Whitmore

Rita Brooks

The scoring of sleep and its events represents at least half of what the technologist must do on a nightly basis, whether the technologist is doing hookups at night, running multiple sleep latency tests (MSLTs) during the day, or scoring records as a scoring technologist. Although the scoring technologist might spend most of the day scoring polysomnograms (PSGs), the technologist working at night must understand and implement the scoring rules to run a PSG well. Similarly, the technologist working during the day running MSLTs and maintenance of wakefulness tests must also understand and implement the scoring rules to execute the respective procedure properly. In short, all technologists should understand the scoring rules.

Scoring is usually performed using a two-pass method, meaning that the entire recording is looked through at least twice. In the first pass, the recording is viewed in 30-second epochs and a sleep stage is assigned to each epoch between lights out and lights on. Arousals are also generally scored during sleep staging. In the second pass, the resolution of the display is changed to 60 seconds, 120 seconds, or even 5 minutes and events are marked. The summary data that result from this process are then tabulated and reported according to the current version of the AASM Manual for the Scoring of Sleep and Associated Events (1).

HISTORICAL PERSPECTIVE

The history of how we describe and quantify sleep is the story of how technology has helped us observe the rhythms of the human in a relatively nonintrusive manner. The first person to report recording the electrical biopotentials of the brain in mammals was Richard Caton, who recorded brain waves from the exposed brains of dogs and rabbits in 1875 (2). It was not until brain waves could be recorded through the skin and scalp that the field of electroencephalogram (EEG) was born. Hans Berger reported the presence of what he called the alpha rhythm using electrodes attached to the scalp of himself and his son in 1929 (Figs. 38-1 and 38-2).

Loomis and others at Harvard adopted this technology and were the first to describe sleep stages using features of the EEG in 1937. They classified five different stages of what we would now call nonrapid eye movement (NREM) sleep (stages A through E). Several other investigators added to this early work in EEG, but it was not until the landmark paper in 1953 by Eugene Aserinsky and Nathaniel Kleitman that described regularly occurring bouts of rapid eye movements (REMs) during sleep that the standard of the recording montage including more than simply the EEG was recognized.

Eugene Aserinsky adopted Jacobson’s technique of monitoring eye movements using EEG electrodes to study infants. It should be noted that he spent many hours collecting the data himself, much like the sleep technologist does today. He noticed regularly occurring periods of “jerky eye movements” in more than 50
recordings by 1952. This added the electrooculogram (EOG) to what we now know as the PSG.

Figure 38-1 German psychiatrist Hans Berger (1873 to 1942) is best known as the inventor of electroencephalography and is the discoverer of the alpha wave rhythm.

Figure 38-2 Hans Berger reported the presence of what he called the alpha rhythm using electrodes attached to the scalp of himself and his son in 1929.

A few years later, a young medical doctor named William Dement joined the Kleitman group at the University of Chicago. In 1957, Dement and Kleitman published a landmark article entitled Cyclic Variations in EEG During and Their Relation to Eye Movements, Body Motility, and Dreaming. This article outlined four stages of what we now call NREM sleep as well as REM sleep.

In 1968, A Manual of Standardized Terminology, Techniques and Scoring for Sleep Stages of Human Subjects was published (2). This manual was the result of the efforts of a committee led by Allan Rechtschaffen and Anthony Kales and represents the first consensus-based set of rules for the scoring of sleep in humans. Its intended application was for the sleep research community to aid in the standardization and compatibility of research data in adults. This R&K manual, as it has become known, was the most widely used set of rules in scoring sleep until 2007, when the American Academy of Sleep Medicine (AASM) put together a task force to compose a new set of rules that would address its shortcomings.

Although the R&K manual was the most widely used scoring system for 39 years, it was not the only scoring system in use. Several years later, in 1971, a similar set of scoring rules was published by a committee led by Anders, Emde, and Parmelee. A Manual for Standardized Terminology, Techniques, and Criteria for Scoring of States of Sleep and Wakefulness in Newborn Infants was intended to serve much the same purpose as the R&K manual: a set of rules by which increased comparability of research results might be achieved, but for PSG data from newborns and infants instead of adults.

A group from the University of Florida also published a scoring system at about the same time as the R&K manual was being developed. The scoring rules put forth by Williams, Karacan, and Hursch used 1-minute epochs instead of the 20- or 30-second epochs suggested in the R&K manual, as well as slightly different frequency bands and amplitude requirements. Although not widely used, the group did publish an impressive summary of normative PSG data using their guidelines in 1974 (3).

The R&K manual has several significant drawbacks that have been described in detail elsewhere (4, 5). Many of these drawbacks stem from the fact that the manual ended up being used for purposes other than its intended original purpose: to aid in the standardization of sleep research data between laboratories. The clarity and economy of its rules made it easy to use and thus, with no other clear alternative in existence, clinicians began to use the scoring rules for sleep studies in the clinical setting. In the clinical setting, several additional channels were added: channels to monitor respiratory airflow and effort, an electrocardiogram (ECG) channel, and electromyogram (EMG) channels to monitor limb movements (LMs). There are no definitions for respiratory events, arousals, or rules for the scoring of limb movements in the R&K manual. To address these other features of sleep, additional guidelines were developed (6, 7, 8).

A second drawback of the R&K manual is the fact that it was intended only for adults and thus does not address the effect of age on sleep. Although a separate manual was developed for the scoring of sleep in newborns and infants (9), the changes in sleep due to advancing age are not addressed in the R&K manual. For example, the 75 µV criteria for the scoring of slow-wave sleep has been questioned (5).

A third issue not addressed in the R&K manual is the absence of digital specifications for the acquisition and scoring of PSG data. The R&K manual was developed in the late 1960s before the use of computers became the norm for recording sleep. Thus, R&K only provides guidelines for the use of paper machines and many of the choices for the number of channels to be recorded were determined by the use of eight-channel recorders. As such, only one central EEG signal was required as a minimum EEG channel because all of the waveforms of sleep can be seen from this derivation. Two EOG leads were suggested, as well as a single chin EMG channel. This allowed two participants to be recorded on a single machine. The advancement of digital technologies now allows for as many as 256 channels to be recorded and analyzed with relative ease.

Taken together, these many drawbacks led to the re-evaluation of the R&K manual. To this end, the board of the AASM approved the development of a new scoring manual. Over the course of 2 years, a steering committee and eight task forces were brought together with members of the board of the AASM and content experts in various aspects of sleep. Six of the eight task forces were developed to investigate and report on major content areas of the new manual: visual scoring, digital scoring, arousal, movement, respiratory issues, and cardiac issues. The issue of age effects was addressed by the two remaining task forces, one for pediatrics and one for geriatrics.

Each task force followed the same process for the development of the new manual. They performed a literature search and review, from which they produced evidence tables and a review paper, which served to provide transparency about the decisions made in the new manual. Each of these papers was published in a single issue of the Journal of Clinical Sleep Medicine in 2007 and provides a thorough review for the curious reader.

Since 2007, there have been eight versions of the AASM Manual for the Scoring of Sleep and Associated Events, with the most recent ninth version 2.5 being released in early 2018. The reader is encouraged to read page 75 of version 2.5 for a brief review of the committee-recommended changes made since the 2007 manual was written. Some of the changes introduced in these eight updates to the manual have been minor to the terms to be used, while other changes were more substantive, including the addition of home sleep apnea testing (HSAT) and infant rules for scoring sleep and associated events. These substantive changes have been examined in detail in other chapters.

TECHNICAL SPECIFICATIONS

The AASM Manual for the Scoring of Sleep and Associated Events contains important technical specifications for the collection of PSG data. These rules represent the first consensus-based set of rules for the use of digital technology in polysomnography.

The maximum impedance allowed for EEG and EOG is 5 kΩ. The minimum digital resolution is 12 bits per sample. The sampling rates and filter settings for each channel are detailed in Table 38-1 (1).

The minimal sampling rate for the EEG, EOG, EMG, and ECG channels is 200 samples per second; the preferred sampling rate is 500 samples per second for these parameters. The sampling rate is chosen on the basis of what the frequency of interest is in the signal being viewed. The AASM Scoring Manual suggests that the sampling rate and high filter settings may be increased if needed to more accurately display EEG spikes. In this case, the sampling rate should be “at least three times the high frequency of interest” (1).

Usually, a sleep study is scored in two passes. The scorer will first view the sleep study in 30-second epochs to determine the appropriate sleep stage. After determining sleep and wakefulness, the scorer will do a second pass in a more compressed view, usually a 60- to 120-second epoch, to determine sleep-related events.

RECOGNIZING WAVEFORMS

There are four features of the EEG signal that are informative to the categorization of the signal as a distinct waveform (Table 38-2). They are the amplitude, frequency, morphology, and distribution of the waveform. We use these four features of the EEG signal to identify the waveforms that are a part of the definition of sleep stages. The amplitude of the signal refers to its vertical distance from the isoelectric baseline, measured in microvolts (µV). The frequency refers to the number of waves per second, measured in Hertz (Hz) or the number of cycles per second. The morphology of the waveform refers to its shape. The distribution of the waveform is the source of the signal, or the region in which the amplitude is the highest.

Alpha Rhythm

The first waveform that we must recognize is the alpha rhythm or Berger’s rhythm. As with several waveforms, the alpha rhythm is defined by its frequency, which is
between 8 and 13 Hz (Fig. 38-3). Different people have slightly different alpha frequencies and some people have no alpha frequency at all. Some people display alpha that cycles at about 11 Hz, whereas others may have an alpha rhythm that cycles at around 9 Hz. Although they are different frequencies, both are considered alpha since they fall between 8 and 13 Hz.

Table 38-1 Recommended Digital Specifications and Filter Settings from the AASM Manual for the Scoring of Sleep and Associated Events

Channel	Desirable Sampling Rate in Hz	Minimal Sampling Rate in Hz	Low-Frequency Rate in Hz	High-Frequency Rate in Hz
Electroencephalogram	500	200	0.3	35
Electrooculogram	500	200	0.3	35
Electromyogram	500	200	10	100
Electrocardiogram	500	200	0.3	70
Airflow	100	25	0.1	15
Oximetry	25	10	(not filtered)	(not filtered)
Nasal pressure	100	25	Direct current (DC) or ≤0.03	15
Esophageal pressure	100	25	0.1	15
Body position	1	1	(not filtered)	(not filtered)
Snoring sounds	500	200	10	100
Rib cage and abdominal movements	100	25	0.1	15
Positive airway pressure device flow	100	25	DC	DC
Reprinted with permission from Berry, R. B., Albertario, C. L., & Harding, S. M. (2018). The AASM manual for the scoring of sleep and associated events: Rules, terminology and technical specifications (Version 2.5). Darien, IL: American Academy of Sleep Medicine.

Another helpful feature of alpha is its occipital distribution (Fig. 38-4). Although alpha can usually be seen in each of the three EEG channels, its amplitude is the highest in the occipital channel.

Table 38-2 Summary of Waveforms Used in Scoring Sleep and Wakefulness

	Alpha Activity	Low-Amplitude Mixed Frequency	Vertex Sharp Wave	Spindle	K Complex	Slow Wave
Amplitude	Not defined	“Low”	Not defined	Not defined	Not defined	75 µV
Frequency	8-13 Hz	4-7 Hz	≥2 Hz	11-16 Hz	<2 Hz	0.5-2 Hz
Waveform	Sinusoidal	Mixed	Sharp	“Distinct”	Negative followed by positive	Not defined
Distribution	Occipital	Not defined	Central	Central	Frontal	Frontal

Low-Amplitude Mixed-Frequency Activity (Theta)

The second waveform to recognize is theta activity, also known as low-amplitude mixed-frequency (LAMF) activity (Fig. 38-5). The frequency of LAMF activity is 4 to 7 Hz.
The amplitude is quite loosely defined as “low.” In the R&K manual, the term “low-voltage mixed frequency” was used and the frequency range was from 2 to 7 Hz, a slightly larger frequency range. This overlapped with what is referred to as theta activity by the International Federation of Clinical Neurophysiology (10). The amplitude is quite loosely defined as “low.” The distribution of the activity is diffuse, meaning it occurs in all channels.

Figure 38-3 Varying amplitudes of the alpha rhythm.

Figure 38-4 Maximal alpha amplitude in the occipital region of a 27-year-old female (horizontal dotted line is an “eyes closed” instruction).

Vertex Sharp Waves

The third waveform to recognize is the vertex sharp wave (Fig. 38-6). The AASM Scoring Manual defines vertex sharp waves as “sharply contoured waves with duration <0.5 seconds maximal over the central region and distinguished from the background activity” (1). There is no amplitude criterion for a vertex sharp wave. The distribution is central, meaning that the vertex sharp wave

will usually have its highest amplitude in the central channels. The most recognizable feature of the vertex sharp wave is its characteristic morphology.

Only gold members can continue reading. Log In or Register to continue