Introduction
This article is based on the rules of the definitive sleep scoring manual (Rechtschaffen and Kales). For situations not defined by Rechtschaffen and Kales, other expert polysomnographers have developed logical rules, which are presented in this text as supplementary information to facilitate scoring. The contents of this article reflect only the opinions of the author and do not constitute official policy of the United States Department of Defense.
Sleep-stage scoring is a rule-based art requiring an understanding of the basic mechanisms underlying the generation of cephalic electric potentials. Signals of interest are generated from the brain (ie, cortex and deeper structures) and the facial muscles (ie, signals picked up by periorbital and electromyographic [EMG] leads).
Interference with the signals of interest is encountered through many mechanisms, including physiological attenuation of the cerebral electric potentials by scalp muscle and bone; intrusion into the signal by slow cyclic respiration, movement, ECG signal, or external electric fields; and impaired contacts between the recording electrodes and the skin surface. Discriminating true signal from artifact can be one of the most challenging aspects of scoring the stages of sleep.
At a paper speed of 10 mm/s, 1 page equates to 30 seconds and is defined as 1 epoch. Computerized polysomnography usually displays one video screen as one 30-second epoch.
Polysomnographic Interpretation
Cortical signals
The electroencephalography (EEG) signal is of primary importance in interpreting polysomnographic studies. It records electric potentials generated by the interaction between the cortex and the deeper brain structures, especially the thalamus. Two centrocephalic and 2 occipital cortical channels are recorded. Measurement of EEG signals is possible because of the relative difference in potential between the 2 recording electrodes; 1 electrode is considered negative with respect to the other. Negative discharges, by convention, are represented by an upwardly deflecting wave.
The polysomnograph references the left and right centrocephalic electrodes (C3, C4) or the left and right occipital electrodes (O1, O4) to electrodes on the opposite right and left ears (A2, A1). The general rule is to read only from the left cortical channel. However, when the left channel develops artifact or the validity of the signalissuspected, comparison is made with the right cortical channel. By convention in the United States, the left channels are odd numbers and the right channels are even numbers. By convention in the United Kingdom, the opposite is true, with the left channels described by even numbers.
Cortical signals are defined slightly differently according to the reference used. The following convention is used here:
- Delta is the slowest activity at less than 4 counts per second (cps).
- Theta is between 4 and 8 cps.
- Alpha is between 8 and 14 cps.
- Beta is greater than 14 cps.
- Another range occasionally mentioned is gamma, which is part of the high end of the beta frequency and has been described to range from 30-45 cps.
High frequency signals above 50 cps are finding increased mention in the literature. High frequency bands (HFB) are described in the ranges 51-100 Hz (HFB1), 101-200 Hz (HFB2), and 201-500 Hz (HFB3) for analysis purposes. Frequencies in these bandwidths are reported as being associated with cognitive processing and alertness.
Muscle signals
The EMG signals are muscle twitch potentials that are of secondary importance in polysomnography. Their utilization is based on the finding that, during sleep, muscle activity decreases. During rapid eye movement (REM) sleep, muscle activity is at its nadir. However, in many cases appreciating the decreasing tone is difficult. The relative silence during REM sleep may not be of help in distinguishing REM sleep from the preceding or subsequent sleep stages.
Compounding the problem of interpreting EMG channels is intrusion of artifact into the signal. This has many etiologies. Some examples include cyclic chewing movements, irregular teeth grinding, steady high-amplitude noise generated by increased pressure on the electrode (eg, as caused by lying on the chin). Additionally, muscle artifact may spill over into the cortical leads. ECG signal is a specific type of cardiac artifact that can appear in all or several channels; it can be recognized by the QRS complexes in the cortical or other channels.
Eye movements
The electro-oculographic (EOG) signals measure changes in the electric potential of the positive anterior aspect of the eye relative to the negative posterior aspect. Horizontal axis electrodes are placed near the outer canthi and vertical axis electrodes below and above the eye to measure transient changes in potential during the actual eye movement. During any eye movement, the cornea (positive) moves toward 1 electrode, while the fundus (negative) moves away. When the eye is not moving, the change in relative position is zero, and the eye leads do not record a signal.
Slow, rolling eye movements are recorded as long gentle waves, while rapid jerking movements are represented by sharply contoured fast waves. Blinking of the eyes produces rapid vertical movements. Eye movements during drowsiness and stage I sleep may be jerky, irregular, or gently rolling. In deeper stages of sleep, macro eye movements cease altogether. During REM sleep, eye movements again become active and jerky. The intensity of the bursts of activity is used to describe the density of REM sleep.
Sleep Stages
Initially, the clinician should scroll through the entire record quickly to evaluate the quality of the recording and the usefulness of specific channels. Observe sections that represent the major stages to learn the specific shape of the features that represent the stages in that particular individual and to gain an overall picture of the cycles for that record. Specifically observe for sleep spindles, K complexes, slow waves, and REMs.
Wake stage
The first several epochs of the record will be the wake stage. The EEG will show mixed beta and alpha activities as the eyes open and close and predominantly alpha activity when the eyes remain closed. The EMG will reflect the high-amplitude muscle contractions and movement artifacts. The EOG will show eye blinking and rapid movement. The record will slow in frequency and amplitude as the subject stops moving and becomes drowsy.
Drowsiness
This stage is defined as sleepy but awake with eyes closed. The EEG will show predominant alpha activity, while the EMG activity becomes less prominent. The EOG may show slow, rolling eye movements. If, at any point, the subject rolls over, the record will reflect this as paroxysmal sustained increased artifact and high-amplitude activity. The subject may enter stage I of sleep for 1 or 2 epochs and then reawaken. Transitions may be difficult to score. From wake, sleepers normally proceed to stage I, but infrequently they may enter REM sleep or stage II sleep directly.
Stage I
Stage I sleep is scored when the alpha activity in the EEG drops to less than 50%. A transition is observed from alpha activity to a lower frequency activity, such as theta, possibly intermixed with low-amplitude delta activity. Amplitudes are less than 50-75 µV. Paroxysms of 2-7 cps activity up to 75 µV may occur.
Stage I is usually brief, lasting for 1-7 minutes. Vertex sharp waves may occur, but no sleep spindles or K complexes are recorded. The EOG may show slow, rolling eye movements, especially early in the stage. No REMs are observed. The EMG shows less activity than in wake stage, but the transition is gradual and of little assistance in scoring.
Arousals are paroxysms of activity lasting 3 seconds or longer. The minimum arousal is simply a paroxysmal burst in the EEG channel, usually to alpha or theta activity. Arousal from stage I is common and usually is represented by a burst of activity on the EEG, EOG, and EMG. If the burst results in alpha activity for greater than 50% of the record, then the epoch is scored as wake.
Stage II
The EEG shows predominant theta activity with minimal alpha activity. Delta is permitted for less than 20% of the record. Amplitude may increase from that seen in stage I.
K complexes appear for the first time. K complexes are sharply negative (ie, up-deflecting) monophasic or polyphasic waves followed by a slower, positive (ie, down-deflecting) wave. The complex must persist for at least 0.5 seconds. No minimum amplitude is defined, but characteristically the waves stand out clearly from the background. K complexes can occur in response to a sudden sound and were so named because they were appreciated as following the knocking sound produced by knuckle rapping. In this respect, they may represent a form of cortical evoked potential in a brain still minimally responsive to external stimuli.
Sleep spindles may appear. These are paroxysms of 12-14 cps activity persisting for at least 0.5 seconds (that is, 6-7 small waves in 0.5 seconds). Although classically described as spindle shaped, they vary in morphology and may attach as a tail to a K complex.
No specific criteria exist for EOG and EMG in this stage.
Arousal from stage II may be into stage I or into wakefulness. If the arousal is 3 seconds or longer in duration, and the resulting alpha activity persists for less than 50% of the record, the epoch is scored as stage I. If the alpha persists for greater than 50% of the record, the epoch is scored as stage wake. If the first half of the following epoch demonstrates stage II characteristics (ie, spindles, K complexes, high-amplitude theta/delta activity), that epoch is scored as stage II.
Once in stage II, that score is maintained unless a reason to exit presents. One such reason to exit is described as the 3-minute rule. If no specific stage II indicators appear, and in the absence of arousals and muscle tone changes that would alter the staging, continue to score all epochs as stage II for up to 3 minutes. At 3 minutes, if no specific indicators for stage II have occurred, scroll back 3 minutes and score those epochs as stage I.
Stage III
The EEG shows 4 cps or slower activity, with peak-to-peak amplitudes greater than 75 µV for between 20% and 50% of the epoch. Both K complexes and sleep spindles may be seen in stage III sleep. No specific criteria exist for EOG and EMG. The transition to stage III from stage II may be gradual, and stage III may alternate with stage I.
Stage IV
The EEG shows 4 cps or slower activity, with peak-to-peak amplitudes greater than 75 µV for at least 50% of the epoch. Both K complexes and sleep spindles may be seen in stage IV sleep. No specific criteria for EOG and EMG exist. In general, muscle tone decreases gradually from stage II to stage IV. The transition from stage III to stage IV may be gradual, and stage III may alternate with stage IV.
Stage REM
The EEG of REM sleep shows relatively low-voltage and mixed-frequency activities and may resemble the EEG of stage I. Sawtooth-shaped waves may occur before or with REM EOG bursts. Slow alpha activity may occur, resembling that of wake stage.
Sleep spindles and K complexes are not part of the REM EEG; when they occur, they are reason to consider moving from REM to stage II. If 2 K complexes or spindles occur without REM activity between them, the epochs between the complexes are scored as stage II. If REM activity occurs on both sides of the K complexes or spindles, then the epoch is scored as REM and the complex is considered to represent a momentary breakthrough into REM rather than a change of stage. No high-amplitude activity may be counted as REM. Bursts of delta activity are reason to change sleep stage.
The EOG of REM shows paroxysmal, relatively sharply contoured, high-amplitude activity occurring in all eye leads simultaneously. The EOG activity is not needed to mark the start of an REM period. REM epochs may be recognized by EEG activity before EOG movements start. Small REMs on EOG may serve as a harbinger of REM stage and can indicate the actual onset of REM in another area where interscorer concordance is lower.
The EMG of REM shows an appreciable decrease in tone but may differ little from the EMG of stages III or IV. The EMG should show the lowest tone in the record, but no specific amplitude or frequency criteria are in place.
For excellent patient education resources, visit eMedicine's Procedures Center. Also, see eMedicine's patient education articles Sleep: Understanding the Basicsand Electroencephalography (EEG).
Rem Scoring Subtleties
REM can alternate with stage wake, stage I, or stage II. Rules to help differentiate REM-spindle-K complex intermixing include the following: If a single sleep spindle or K complex occurs during an REM period (REM on both sides of the complex), and the EMG remains unchanged (low amplitude), continue to score REM. If the EMG amplitude increases with the spindle or K complex and REM is not clearly apparent following the complex, score stage II from that point forward. If 2 spindles or K complexes occur in the absence of EMG changes (ie, low amplitude continues), score them and the intervening period as stage II.
Beginning scoring REM
REM periods may begin before the characteristic EOG movements. Occasionally, low-voltage, mixed-activity EEG, sometimes with sawtooth waves and low-amplitude EMG, may begin several epochs prior to the onset of the characteristic EOG movements. When the EOG movements are recognized clearly as REMs, scroll back to the point where the record became REM-like on both EEG and EMG and score those epochs as REM. REM may begin after a brief arousal.
Ending scoring REM
REM may end with a brief arousal, after which the EEG looks qualitatively different, with higher amplitude and slower frequency activity. In the absence of arousal, the transition may be subtle.
Because the EEGs of REM, stage I, and stage wake can be similar, several criteria are available to help differentiate the stages. Look for development of bursts of higher amplitude and slower frequency EEG activity, which may be associated with EOG eye movements similar in morphology to those of REM. Score the epoch according to the applicable stage criteria from the onset of the EEG bursts of slow activity. Remember that REM does not have high-amplitude, slow-EEG activity and that eye movements may occur in all stages. For example, if paroxysms of K complexes begin, score stage II.
Refer to the EMG and EOG as adjuncts to help confirm staging. For example, if the EMG increases in amplitude briefly and without warning, the period may be stage I. If higher amplitude EMG discharges persist and are accompanied by increased alpha activity (for >50% of the record), the epoch may be scored as wake.
General Considerations
Once in a stage, remain in that stage until given an unambiguous reason to change.
Arousal of greater than 3 seconds occurring in the first half of the epoch (from stages II, III, IV, or REM) is reason enough to consider changing the stage.
Arousals may be observed only on EEG, characterized by the development of theta or alpha activity.
Arousals may have EEG activity with EMG activity increases, or they may have EEG, EMG, and EOG activity. For example, if an arousal occurs during stage III, score stage II if the arousal is less than half the record, stage I if the arousal is greater than half the record, and stage wake if the arousal is greater than half the record and associated with sustained alpha activity.
In a given epoch, if an arousal occurs from stage II, normally the remaining epoch would be scored as stage I or wake depending upon the characteristics of the arousal.
An exception to this might be if a K complex or a sleep spindle occurs within 15 seconds of the end of the arousal, in which case the epoch is scored as stage II.
Acknowledgments: Thomas Balkin, PhD, Diplomate of the American Board of Sleep Medicine, Chief of the Department of Neurobiology and Behavior - Division of Neuropsychiatry, Walter Reed Army Institute of Research, provided the pearls, the rule interpretations, and overall guidance on this manuscript.
Multimedia
 | Media file 1: Sleep stage I EEG sample. |
 | Media file 2: Sleep stage II EEG sample. |
 | Media file 3: Sleep stage III EEG sample. |
 | Media file 4: Sleep stage IV EEG sample. |
 | Media file 5: Rapid eye movement sleep EEG sample. |
Keywords
polysomnography, electroencephalography, EEG, cortical electroencephalography, sleep stages, wake stage, drowsiness, rapid eye movement, REM, sleep scoring, sleep-stage scoring, cephalic electric potentials
More on Sleep Stage Scoring |
| References |
References
Butkov N. Atlas of Clinical Polysomnography. Ashland, Ore: Synapse Media;1996.
Kumano-go T, Mikami A, Suganuma N, et al. Three components of obstructive sleep apnea/hypopnea syndrome. Psychiatry Clin Neurosci. Apr 2003;57(2):197-203. [Medline].
Marzec ML, Malow BA. Approaches to staging sleep in polysomnographic studies with epileptic activity. Sleep Med. Sep 2003;4(5):409-17. [Medline].
Niedermeyer E, Lopes Da Silva F. Electroencephalography: Basic Principles, Clinical applications, and Related Fields. 3rd ed. Baltimore, Md: Williams and Wilkins;1993.
Rechtschaffen A, Kales A, eds. A Manual of Standardized Terminology, Techniques, and Scoring System for Sleep Stages of Human Subjects. US Department of Health, Education, and Welfare Public Health Service - NIH/NIND. 1968.
Sing HC, Kautz MA, Thorne DR, et al. High-frequency EEG as measure of cognitive function capacity: a preliminary report. Aviat Space Environ Med. Jul 2005;76(7 Suppl):C114-35. [Medline].
Further Reading
Keywords
polysomnography, electroencephalography, EEG, cortical electroencephalography, sleep stages, wake stage, drowsiness, rapid eye movement, REM, sleep scoring, sleep-stage scoring, cephalic electric potentials
