The Inaugural Meeting of the Association for the Psychophysiological Study
of Sleep (APSS) was convened in 1960 for the purpose of adopting a standard
scoring system for stages of sleep. The opportunity to discuss experimental
results diverted these charter members from their original purpose and instead,
they decided to meet annually thereafter to communicate their latest research
findings.
The need for a standard scoring system was reemphasized by Monroe's study
(1967) which revealed serious unreliability in the scoring of certain sleep
stages. Concern over this unreliability led to the inclusion of a special
session on scoring at the Seventh Annual Meeting of APSS held that year.
Subsequently, an ad hoc committee of investigators was formed under the auspices
of the UCLA Brain Information Service to develop a terminology and scoring
system that might be used universally by sleep researchers. The Committee
members met on several separate occasions and corresponded extensively in the
intervening periods. The Committee was composed of the following members, all of
whom had considerable experience in scoring sleep records:
Ralph J. Berger, Ph.D., Department of Psychology, Crown College.
University of California, Santa Cruz, California.
William C. Dement, M.D., Department of Psychiatry, Stanford
University School of Medicine, Palo Alto, California.
Allan Jacobson, M.D., Department of Anatomy, University of
California School of Medicine, Los Angeles, California.
Laverne C. Johnson, Ph.D., Navy Medical Neuropsychiatric
Research Unit, U.S. Naval Hospital, San Diego, California.
Michel Jouvet, M.D., Laboratoire de Medecine Experimentale,
Faculte de Medecine, Lyon, France.
Anthony Kales, M. D., Department of Psychiatry, University of
California School of Medicine, Los Angeles, California.
Lawrence J. Monroe, Ph. D., Department of Psychology, University of
Illinois College of Medicine, Chicago, Illinois.
Ian Oswald, M.D., Department of Psychiatry, University of Edinburgh,
Edinburgh, Scotland.
Allan Rechtschaffen, Ph.D., Departments of Psychiatry and Psychology,
University of Chicago, Chicago, Illinois.
Howard P. Roffwarg, M.D., Albert Einstein Medical Center, Montefiore
Hospital, The Bronx, New York.
Bedrich Roth, M.D., Department of Neurology, Charles University Medical
Faculty, Prague, Czechoslovakia.
Richard D. Walter, M.D., Division of Neurology, Department of Medicine,
University of California School of Medicine, Los Angeles,
California.
This proposal was prepared by the Committee with the expectation that the
standardization of recording techniques and scoring criteria would be widely
used and would markedly increase the comparability of results reported by
different investigators. An evaluation of how much such standardization
contributes to reliability of scoring will have to await the development of
experience with the system and empirical testing.
Allan Rechtschaffen and
Anthony Kales, Co-Chairmen
(Page 1)
TO PROVIDE CONTINUITY within the very extensive research on sleep stages
which has been published in the last decade, it was the guiding principle of the
Committee to retain, as much as possible, the terminology and criteria which
have had the greatest use. Indeed, it is testimony to the accuracy and judgment
of Dement and Kleitman (1957)* that their description of sleep stages has proven
useful in a decade of voluminous research by hundreds of investigators without
requiring major modification. In general, the present proposal represents a
reaffirmation of their original criteria for the partition of sleep into stages,
combined with the revisions and elaborations which must inevitably follow ten
years of experience with their system.
Research utilizing the Dement-Kleitman criteria for scoring sleep stages
has firmly established the fact, first noted by Loomis, Harvey and Hobart
(1937), that sleep is not a steady state and that the sleep stages follow a
fairly orderly cyclic pattern. While knowledge of the significance of each stage
of sleep is incomplete, specific physiological and behavioral correlates of the
various stages have been found. These and other correlates may eventually
provide more meaningful descriptions of sleep than the stages described in the
following section which emphasizes the EEG changes.
The terminology and scoring systems proposed here are not intended as
restrictions upon the researcher who has substantive reasons for using different
terms and criteria. However, it is strongly recommended that departures from
these terms and criteria should be specified. Such specification would aid in
comparing the results in different studies.
Although there is considerable comparability of sleep stage manifestations
among various species, the differences are sufficiently great to require a
separate scoring system for most species. This proposal is designed for adult
humans. Even among human subjects, however, there are some individuals or groups
whose polygraph recordings may require further description or elaboration than
that provided by the stages proposed here. In such instances, the existence of
established categories does not free the investigator from the task of devising
a descriptive system which better communicates the unique features of the
phenomena. For example, it is well known that human infants show combinations of
polygraphic features which defy classification by the criteria proposed here. A
strict adherence to the proposed system would not yield an adequate description
of infant sleep.
EEG Terms
For the designation of specific EEG phenomena, the proposal of the
Terminology Committee of the International Federation for Electroencephalography
and Clinical Neurophysiology (Brazier, Cobb, Fischgold, Gastaut, Gloor, Hess,
Jasper, Loeb, Magnus, Pampiglione, Remond, Storm van Leeuwen and Walter, 1961)
is adopted except where specifically noted. Similarly, the Ten Twenty Electrode
System of the International Federation (Jasper, 1958) is adopted for designation
of electrode placements.
Stages
A more detailed discussion of scoring criteria will be presented later;
for purposes of preliminary identification, the stages and their most prominent
features are:
Stage W (wakefulness) - The EEG contains alpha activity and/or low
voltage, mixed frequency activity.
Movement Time (MT) - Scoring epoch during which the polygraph record is
obscured by movements of the subject.
Stage 1 - A relatively
low voltage, mixed frequency EEG without rapid eye movements (REMs).
Stage 2 - 12-14 cycles per second (cps) sleep spindles and K complexes on
a background of relatively low voltage, mixed frequency EEG activity.
Stage 3 - Moderate amounts of high amplitude, slow wave activity.
Stage 4 - Large amounts of high amplitude, slow wave activity.
Stage NREM (non-REM) - Stages 1, 2, 3, and 4 combined.
Stage REM - A relatively low voltage, mixed frequency EEG in conjunction
with episodic REMs and low amplitude electromyogram (EMG).
*These scoring criteria were
further elaborated by W.C. Dement in a manual presented to the APSS meeting in
1962.
(Page 2)
It is recommended that the above stage designations be used instead of
other terms. Some of the terms which have been previously used and which are no
longer considered desirable are:
For Stage 1:
descending Stage 1, sleep onset 1, pure 1, transitional sleep, drowsy sleep, low
voltage-fast, Stage B.
For Stage 2: spindle
sleep, Stage C, light sleep.
For Stage 3 and/or Stage
4: transitional sleep, delta sleep, slow sleep, slow wave sleep, light
sleep, deep sleep, quiet sleep, orthosleep, telencephalic sleep, Stages D and
E.
For Stage REM: paradoxical sleep, rhombencephalic sleep, Stage 1REM,
activated sleep, fast sleep, low voltage-fast sleep, deep sleep, light sleep,
desynchronized sleep, D-state, archisleep, parasleep, emergent Stage 1,
ascending Stage 1, Stage 5.
Although the physiological correlates of behavioral sleep remain a
continuing research issue, generally, sleep has been polygraphically defined by
Stages 1, 2, 3, 4, and REM. The same working definition is used here.
As the quality and characteristics of a polygraph tracing depend upon the
techniques used, the designation of criteria for scoring stages of sleep and
wakefulness would be useless without some standardization of
technique.
EEG Recording
A minimum paper speed of 10 mm/sec is recommended as the slowest which
will permit clear visual resolution of alpha and sleep spindle frequency. Time
constants shorter than 0.3 see should not be used. Selective filtering below 20
cps should be reported. A minimal pen deflection of 7.5-10 mm for 50 microvolts
(mv) is recommended; otherwise low amplitude sleep
spindles may escape detection. Electrode resistances should not exceed lOK ohms
at the beginning of the recording.
EEG patterns, and therefore the scoring of stages, may vary according to
placement and derivation. Ideally, a standard array might include a large number
of placements which would yield comprehensive regional information. However,
regional differences are not critical for the scoring of sleep stages, except
insofar as certain critical types of activity, i.e., alpha, vertex sharp waves,
sleep spindles, K complexes. and delta waves are adequately registered. Further,
the wide use of eight channel electroencephalographs and a desire to maximize
data acquisition by running two subjects on each machine often limits the
channels of information from each subject to four. For reasons to be elaborated
in a later section, two eye movement channels and one EMG channel are considered
minimal; EEG information would accordingly be limited to one channel. It is
therefore necessary to specify the single optimal EEG derivation which can be
confidently recommended for scoring sleep stages unless there are compelling
reasons for doing otherwise .
When EEG information is limited to one derivation, the recommended
derivation is C4/A1 or C3/A2 (Fig. 1). Either the right or left side may be
used, since the EEG patterns from homologous areas are generally synchronous
.
Sleep spindles, K complexes, and vertex sharp waves are clearly recorded
from the C3 or C4 placements, and high voltage slow waves show maximal or nearly
maximal amplitude if the referent maximizes interelectrode distance. Although
alpha rhythm is better recorded from occipital areas, there is adequate
registration at C3 or C4 to permit a precise evaluation of sleep onset according
to EEG criteria.
The opposite ear or mastoid (Al, A2) is preferred over scalp referents
because the ear or mastoid placement is also used as the recommended reference
for electrodes that record eye movement potentials. In addition, use of the ear
or mastoid reference maximizes interelectrode distance and avoids mixing
activity from two different scalp areas.
The advantages of standardization are obvious. In the collective
experience of the Committee, sleep stages may be adequately scored using EEG
information obtained from the recommended derivation. If multiple channels of
EEG information can be recorded, and special studies make additional derivations
desirable, the results from the additional placements should be compared with
the results from C3/A2 (or C4/A1) and substantial differences reported if they
exist. There is no problem regarding interlaboratory comparability in scoring
when multiple EEG channels are available, since C4/A1 (or C3/A2) will always be
included in the array. However, when only one channel is recorded, presumed
advantages of other placements must be carefully weighed against the obvious
advantages of a standard EEG technique which maximizes comparability and
replicability. * THE EEG CRITERIA FOR SCORING SLEEP STAGES SHOULD ALWAYS BE
BASED ON TRACINGS OBTAINED FROM C4/A1 OR C3/A2. A schematic illustration of
these electrode placements is given in Fig. 1.
To eliminate confusion between eye movement potentials and other signals
which resemble them, at least two channels are necessary for recording eye
movements. The recommended procedure is to record on one channel the potentials
from an electrode approximately 1 cm above and slightly lateral to the outer
canthus of one eye and a reference electrode on either the homolateral ear lobe
or mastoid. On the second eye movement channel are recorded the potentials from
an electrode 1 cm below and slightly lateral to the outer canthus of the other
eye referred to the contralateral ear or mastoid, i.e., both eyes are referred
to the same ear or mastoid electrode. This arrangement is shown in Fig.
1.
Because eye movements in Stage REM and wakefulness are binocularly
synchronous, the suggested arrangement produces out-of-phase deflections on the
two channels for almost all eye movements. Apparatus artifacts usually register
as in-phase deflections or deflections on one channel only; artifacts from one
of the outer canthus electrodes register as deflections on only one channel;
artifacts from the referent electrode register as in-phase deflections on the
two channels. Thus, the out-of-phase deflections produced by eye movements are
easily distinguished from artifacts.
The above array for detecting eye movements has some disadvantages.
Certain oblique eye movements can result in relatively flat tracings, as when a
potential from the horizontal component of the eye movement is canceled by an
opposite potential from the vertical component. Also, convergent eye movements
generate in-phase deflections with this arrangement. However, both of these
situations are of minor importance because abundant out-of phase deflections are
usually present at the same time. Further, the above array does not permit any
differentiation of horizontal and vertical eye movements. Such differentiation
can be accomplished with the use of a supranasion reference, which produces
in-phase deflections on the two recording channels for vertical eye movements.
However, these vertical eye movements could be Confused with artifacts. Also,
the use of a supranasion reference could result in the introduction of
considerable EEG signal in the eye movement tracings. These arrays with AC
recordings do not yield precise information about the direction, size and speed
of eye movements, e.g., large pen deflections could result from large horizontal
eye movements or from a summation of the horizontally and vertically induced
potentials in certain smaller oblique eye movements. The recommended array is
intended to provide an economical arrangement to survey eye movements ~ se for
the scoring of Stage REM. Measurements of the precise characteristics of these
eye movements can only be accomplished with other techniques.
Where specific information about direction and size of eye movements is
required, a four channel arrangement is suggested where electrodes horizontal to
the outer canthi as well as infraorbital and supraorbital electrodes are each
paired with the same ear or mastoid electrode and accorded a separate channel.
Where specific information about eye position is required, DC recording must be
introduced.
A minimum gain of 7.5 mm for 50 mv is recommended
for eye movement recordings. Because the detection of slow eye movements is
helpful in the analysis of a record, time constants faster than 0.3 see should
not be used.
*Because the standard derivation is
so important, it is advisable to prepare the subject with both C3 and C4
electrodes and Al and A2 electrodes, even though only one pair of these may be
recorded. This prior precaution should enable the investigator to change from
one electrode pair to another in the event that artifact develops, and thus
preserve a standard record. This "spare" could also be used as a reference
electrode for EOG recordings. The use of only a single ear Lobe or mastoid
reference electrode is suggested in contrast to one clinical practice in which
reference electrodes are joined in parallel.
The recording of EMG from muscle areas on and beneath the chin (mental,
submental) is strongly recommended because of its importance in the scoring of
Stage REM which will be discussed later.
Because tonic EMG activity during sleep may be at a relatively low level,
high gains should be used, preferably 20 mv/cm or
higher. A minimum of high frequency filtering should be used, although at very
high gains it is occasionally necessary to filter out AC artifacts. Time
constants of 0.1 or faster should be used to eliminate slow potentials from
other sources, which at the recommended gains. may cause amplifier blocking. The
type of electrodes. their position and firm contact with the skin are critical
factors in obtaining good EMG recordings. One method is described by Jacobson,
Kales. Lehmann and Hoedemaker (1964). More recently EMG recordings with very
little AC artifact have been obtained from large (1 in. diameter) flexible lead
or stainless steel mesh electrodes which are curved around the jawbone on both
sides of the chin and secured with large strips of adhesive tape (Rechtschaffen.
1968). See Figs. 1 and 22.
Scoring by Epochs
An epoch-by-epoch approach is strongly recommended in all scoring
procedures. Specifically, a convenient time interval should be selected which is
not so short as to make scoring prohibitively cumbersome, nor so long as to do
violence to the scoring of short-lived stage changes. A convenient interval for
most investigators would be one page of record, which is 300 mm in the most
commonly used paper sizes. This interval would result in epochs of 30 and 20
sec. respectively for the recommended paper speeds of 10 and 15 mm/sec. Because
the use of much longer epochs could produce a bias against the scoring of
short-lived stage changes. it is suggested that when epoch lengths greater than
30 see are used. the investigator report them specifically.
The epoch-by-epoch approach recommended here does not imply that each
epoch is considered in isolation. There are many instances, which will be
discussed in later sections, where the score assigned to a particular epoch
depends in part on the polygraphic features of preceding and succeeding epochs.
The main features of the epoch system recommended here are that the polygraph
record is divided into consecutive segments of equal size and each segment is
assigned a single score. Once an epoch duration is selected it should be
maintained for the duration of the recording. For example, if a 30 see epoch is
selected. the chart is divided into consecutive 30 see periods from the start of
the recording to the end. Each epoch is assigned a single stage score; portions
of two epochs may not be combined to create a new epoch. Where more than one
stage is present in an epoch, the one which takes up the greatest portion of the
epoch should be scored as the stage of that epoch.
Stage W corresponds to the waking state. It is characterized by alpha
activity and/or a low voltage, mixed frequency EEG.
Certain subjects (alpha producers) may have a virtually continuous alpha
record; other subjects may show little or no alpha activity in the waking
record. This stage is usually, but not necessarily, accompanied by a relatively
high tonic EMG, and often REMs and eye blinks are present in the EOG
tracing.
Movement Time (MT), Body Movements, and Movement
Arousal
The score "MT" is assigned to epochs which immediately precede or
follow sleep stages, but in which the EEG and EOG tracings are obscured in more
than half the epoch by muscle tension and or amplifier blocking artifacts
associated with movement of the subject. Where the EEG and EOG record can be
seen in more than half the epoch in spite of the artifacts, the record is scored
according to the prevailing EEG and EOG pattern. MT is not counted with either
sleep or wake time but scored as a separate category unless the investigator
specifies otherwise. Not enough is known about the behavioral correlates of MT
to classify it unambiguously as either sleep or wakefulness. Where an epoch is
obscured by muscle tension and/or amplifier blocking artifacts but is
immediately preceded and followed by Stage W the epoch is scored Stage W rather
than MT.
MT should not be confused with the scoring of discrete body movements
which may be of relatively short duration. Body movements can be detected by so
many devices, and they can be defined in such a large variety of ways according
to the investigator's interest, that it is left to each investigator to
designate his own criteria (e.g., Rechtschaffen, Wolpert, Dement, Mitchell and
Fisher, 1963). Body movements are not to be considered epoch scores in the same
sense as MT. Body movements should be considered specific physiological events
which can occur during MT or during stages.
Both MT and body
movements are to be distinguished from movement arousals. which will assume some
importance in the scoring of stages to be discussed later. A movement arousal
will be defined as any increase in EMG on any channel, which is accompanied by a
change in pattern on any additional channel. For EMG channels, the change in
pattern may consist of either an increase in amplitude of the EMG signal or an
amplifier blocking artifact. For EOG channels, the change of pattern may consist
of either the occurrence of EMG activity, amplifier blocking artifacts, or blink
artifacts. For EEG channels, the change of pattern may consist of either a
decrease in amplitude, an increase in alpha activity, a paroxysmal burst of high
voltage activity, the presence of EMG activity, or amplifier blocking artifacts.
Movement arousals are not used as epoch scores, nor need they be tallied and
reported. The major purpose in defining movement arousal is to aid in the
scoring of stages by signaling the possibility of stage changes. For example,
the interpretation of a short interval of relatively low voltage, mixed
frequency EEG which follows Stage
2 will depend on whether the interval follows a movement arousal. Where
there is a movement arousal, there is greater confidence that the subsequent
interval represents a partial "arousal" to Stage 1 rather than a simple
transient lapse of spindles.
In summary, a score of MT is used to account for
epochs which are mostly obscured by movement artifacts . Body movements are not
considered epoch scores; they are considered discrete physiological events which
occur during epochs. The criteria for defining body movements are left entirely
to the individual investigator. Movement arousal, which is explicitly defined in
this proposal, is not used as an epoch score as is MT, nor is it intended
primarily as a measure of body displacement as is body movement. Movement
arousal is intended primarily as an aid in the scoring of stages inasmuch as an
increase in muscle activity during sleep raises the possibility that the
continuity of the prevailing sleep stage has been disturbed in some way.
Because MT involves more than half an epoch of movement artifact, and
hence relatively large movements in most cases, the criteria for scoring MT will
usually satisfy the criteria for body movements and movement arousals.
Similarly, the criteria for body movements will usually satisfy the rather
minimal requirements for movement arousal. However, the converse of these
statements need not be true. Body movements may be of relatively short duration
and fall far short of the criteria for ANT. Movement arousal need indicate only
some increase in muscle activity and does not necessarily involve the
substantial spatial displacement of the body which is implied by body movement.
Movement arousal could result from an isolated small muscle contraction, a
vigorous blink, or a facial grimace.
Stage 1 is defined by a relatively low voltage, mixed frequency EEG with a
prominence of activity in the 2-7 cps range. * The faster frequencies are mostly
of lower voltage than the 2-7 cps activity. Stage 1 occurs most often in the
transition from wakefulness to the other sleep stages or following body
movements during sleep. During nocturnal sleep, Stage 1 tends to be relatively
short, ranging from about 1 to 7 min. The highest voltage 2-7 cps
activity (about 50-75 µV) tends to occur in irregularly
spaced bursts mostly during the latter portions of the stage. Also during the
latter portions of the stage, vertex sharp waves may appear, often in
conjunction with the high amplitude 2-7 cps activity. The amplitude of the
vertex sharp wave is occasionally as high as 200 µV. Scoring of Stage 1 requires
an absolute absence of clearly defined K complexes and sleep spindles. Traces of
low voltage activity at 12-14 cps may begin to appear as the transition to Stage 2 approaches, but this
activity is not to be defined as sleep spindles until the rhythmic bursts are
clearly visible for at least 0. 5 sec. at which time they define the presence of
Stage 2 (see below). Stage
1, especially following wakefulness, is characterized by the presence of slow
eye movements, each of several seconds duration, which are usually most
prominent during the early portions of the stage. Rapid eye movements are
absent. Tonic EMG levels are usually below those of relaxed wakefulness.1
*The term "relatively low voltage,
mixed frequency EEG" is preferred over other terms previously used to describe
the EEG pattern which is characteristic of Stage 1 and Stage
REM. The term "Stage 1 EEG" is to be avoided, because the pattern also occurs in
Stage REM and it could cause confusion to speak about "a Stage 1 EEG in Stage
REM". The term "low voltage record" should not be used, because it does not
accurately describe the EEG of Stage 1 or REM.
According to the terminology of the International Federation, this term denotes
a record with "No rhythmic activity above
10m
v, no activity above 20
m v". Many of the vertex sharp waves of Stage 1 and many
of the saw-tooth waves of Stage REM exceed 20
m
v. The term "low voltage fast record" is rejected, not only because it is
rejected by the International Federation, but also because the EEG of Stages 1
and REM contains a considerable amount of activity in the 2-7 ups band. The most
popular term heretofore used to describe the EEG of Stages 1 and REM has been
"Low voltage, random". This term is nag rejected because, according to the
terminology of the International Federation, "random" means "recurring at
inconstant time intervals", rather than denoting a mixture of
frequencies.
The transition from a low voltage waking record to Stage 1 is characterized by a
generalized slowing of the EEG. The transition from an alpha record to Stage 1 is characterized by a
decrease in the amount, amplitude, and frequency of alpha activity. When the
amount of record characterized by alpha activity combined with low voltage
activity drops to less than 50% of the epoch and is replaced by relatively low
voltage, mixed frequency activity, the epoch is scored as Stage 1.
This stage is defined by the presence of sleep spindles and/or K complexes
and the absence of sufficient high amplitude, slow activity to define the
presence of Stages 3 and 4 (see below).
The presence of a sleep spindle should not be defined unless it OR of at
least 0.5 see duration, i.e., one should be able to count 6 or 7 distinct waves
within the half-second period.
K complexes are defined as EEG wave forms having a well
delineated negative sharp wave which is immediately followed by a positive
component. The total duration of the complex should exceed 0.5 sec. Waves of
12-14 cps may or may not constitute a part of the complex. 3 The K complex is
generally maximal over vertex regions. K complexes can occur as a response to
sudden stimuli, but they also frequently occur in the absence of any detectable
stimuli (Johnson and Karpam, 1968). Other polyphasic high voltage slow waves
occurring paroxysmally which do not have the precise morphology of the K complex
are also frequently seen during Stage 2.
Because sleep spindles and K complexes are
transient phenomena, relatively long periods may intervene between these events
without the occurrence of a stage change. If less than 3 min of record which
would ordinarily meet to spindles and/or K complexes, these intervening epochs
are to be scored Stage 2 if
there is no indication of movement arousal or pronounced increase in muscle tone
during the interval in question. 4 If the interval without sleep spindles or K
complexes lasts 3 min or longer, the interval is scored as Stage 1, even if it contains no
movement arousal. If movement arousals or increases in muscle tone do occur
during the interval in question, the portion of the record prior to them should
be scored as Stage 2. The
portion of the record which follows should be scored as Stage 1 until the next sleep
spindle or K complex
occurs, provided, of course, that the epoch requirements and criteria
for Stage 1 are otherwise
met.
1 For a more detailed, differentiated description of EEG changes in the
transition from wakefulness to steep, the reader is referred
to ROTH (1961).
2 Because the term "steep spindles" has been widely used in steep
research, this term wild be retained. The term should be used only to describe
activity between 12 and 14 cps.
3 This definition of X complex is at variance with the definition of the
Terminology Committee of the International Federation which is as follows:
"Combination of vertex sharp waves and sigma paroxysm, occurring spontaneously
and especially in response to sudden stimuli during
steep."
Stage 3 is defined by an EEG record in which at least 20% but not more
than 50% of the epoch consists of waves of 2 cps or slower which have amplitudes
greater than 75 mv from peak to peak (the difference
between the most negative and positive points of the wave ) . The figures "20%"
and "50%'t refer specifically to the time occupied by the high amplitude, slow
waves and do not include intervals of higher frequency and lower amplitude
activity between these waves.5 In actual scoring practice, it will be necessary
to make wave by wave measurements only for epochs with borderline amounts of
high amplitude, slow wave activity, i.e., about 20% and 50%. For most epochs,
judgments about Stages 3 and
4 can be made by comparison
with the tracings shown in Figs. 2-5. Figs. 2 and 4 present EEG tracings which
are just below and just above the threshold for differentiating Stages 2 and 3.
Figs. 3 and 5 present EEG tracings which are just below and just above the
threshold for differentiating Stages
3 and 4.
As noted in the original Dement-Kleitman criteria, an attempt should be
made to distinguish between spontaneous K complexes and delta waves, although
this distinction is not always easy. Sleep spindles may or may not be present in
Stage 3.
4 She selection of a 3 min interval was arbitrary. It was based upon our
judgment that inter-spindle intervals of that Length might occur without a stage
change although such occasions would be rare.
5 In determining the percentage of an epoch which contains waves of 0.5
sec or slower and greater than 75 mv, a
certain amount of unreliability will be introduced because it is sometimes
difficult to define a wave by visual inspection. For example, it is not always
clear whether two contiguous potentials of the same polarity represent a single
slow wave with a small wave of opposite polarity superimposed, or whether they
represent two separate faster waves. To estimate the unreliability introduced by
such ambiguities (sic), three members of the Committee scored 28 selected epochs
of 30 sec. duration each for a percentage of the epoch which contained
"acceptable" stow waves. In the test sample used, these percentages ranged from
about 35%, to 80%; about two-thirds of the epochs fell in the narrower range
between 40% and 60%, i.e., the range in which reliable differentiation of Stages 3 and 4 is most difficult. The
rank order correlations between pairs of scorers on amount of high amplitude
stow activity were .94, .94, and .91. Apparently, the ambiguities involved in
the measurement of EEG waves are not so great or prevalent as to interfere with
reliable scoring of the amount of an epoch which contains the specified stow
wave activity. The above reliability coefficients, however, should not be
confused with the reliability of scoring Stages 3 and 4. In actual practice, a
failure to score wave by wave in doubtful cases could result in lower
reliabilities for stage scoring. On the other hand, actual records of continuous
sleep may not be so heavily biased as the test samples with epochs of about 50%
"acceptable" slow wave activity, and the reliability of scoring Stages 3 and 4 could be higher than
the coefficients cited above, depending on the care and precision of tee
scorers.
*There was considerable discussion
in the Committee about the wisdom applying an amplitude criterion in addition to
a frequency criterion the scoring of Stages 3 and 4. Amplitude is
influenced by several variables apart from cerebral activity. These include:
electrode resistance, time constants, electrode placement, and interelectrode
distance. The variance introduced in the amount of Stages 3 and 4 by these factors will
be minimized by standardization.
Other potential determinants of EEG
amplitude such as skin resistance and individual differences in stow wave foci
are difficult to evaluate. However, some implicit amplitude criteria are always
involved in the visual discrimination of specific EEG activity. When the wave
form is complex and highly characteristic, as in the steep spindle, amplitude
per se contributes little to the detection process, i.e., signal to noise ratio.
In the case of stow wave activity where wave form is secondary, the major factor
for easy detection, aside from frequency, is amplitude. Thus, for purposes of
scoring and particularly for reliability of scoring, detection and rating of
stow wave activity must require an explicit amplitude criterion. Therefore, the
question of specifying the precise scoring parameters of Stages 3 and 4 is the major problem,
and a number of factors entered into the decision of the Committee.
With regard to the amplitude
criterion, there was no question that, although the results of measures of slow
wave activity which stress the amplitude parameter and those which minimize it
are highly correlated, there are some differences which made a choice more than
academic.
The following summarizes the
reasoning of the Committee's choice to utilize a relatively high amplitude
criterion for delta activity in the scoring of Stages 3 and 4.
A) Although some extra-cerebral
determinants of amplitude stand as potential sources of variance in measures
which emphasize amplitude, as abundant research has shown, this variance is not
so great as to preclude relationships between such measures and other empirical
variables.
b) A large majority of the
Committee felt that the sleep EEG would not be adequately described by a scoring
system which did not attend to the marked amplitude variations which are one of
its most prominent features.
c) Most published EEG steep
research has utilized an amplitude criterion in the scoring of Stages 3 and 4; retention of such a
criterion provides some continuity with this
research.
Stage 4 is defined by an EEG record in which more than 50% of the epoch
consists of waves of 2 cps or slower which have amplitudes greater than 75mv peak to peak. Although only slightly more than half of an
epoch may actually contain high amplitude, slow waves which meet the above
specifications, most Stage 4 epochs have the appearance of being completely
dominated by this activity. Intervals of lower amplitude, faster activity rarely
persist for more than a few seconds in Stage 4, but are usually prominent in
Stage 3 epochs.
Sleep spindles may or may not be Present in Stare 4. *
Nevertheless, it is recognized that
measures of slow wave activity which minimize amplitude might yield the same
empirical relationships as the measure chosen by the Committee. Also, it is
possible that alternative measures of slow wave activity might have a usefulness
and empirical significance not enjoyed by the measure chosen. Therefore the
selection of the Committee should not deter investigators from using measures of
slow wave activity other than the one suggested here. However, it is recommended
that departures from the criteria specified here be reported in
detail.
*K complexes can be substituted for
steep spindles in all of these illustrations.
Stage REM is defined by the concomitant appearance of relatively low
voltage, mixed frequency EEG activity and episodic REMs. The EEG pattern
resembles the one described for Stage 1, except that vertex
sharp waves are not prominent in Stage REM. Also, in Stage REM distinctive
"sawtooth" waves (Berger, Olley and Oswald, 1962) frequently, but not always,
appear in vertex and frontal regions in conjunction with bursts of REM. Alpha
activity is usually somewhat more prominent during Stage REM than during Stage 1, and the frequencies
generally 1-2 cps slower than during wakefulness (Johnson, Nute, Austin and
Lubin, 1967). As with the EEG of Stage 1, there is an absolute
absence of sleep spindles and K
complexes.
Stage REM should not be scored in the presence of a relatively elevated
tonic mental-submental EMG (Berger, 1961; Jacobson, Kales, Lehmann and
Hoedemaker, 1964). The term "relatively elevated" requires explanation. At the
high gains used for EMG recording during sleep, using surface electrodes it is
difficult to define the absence of tonic EMG activity; even when the EMG tracing
is of very low amplitude, there is almost always some residual fast activity
which could represent either EMG activity or noise. Neither is it possible to
define absolute EMG amplitude in a useful way, because amplitude varies
considerably from subject to subject and with electrode position. However, for
any given recording session with any given subject, tonic EMG will show
considerable variation about which it is useful to make some relative
statements. Specifically, during Stage REM the tonic mental-submental EMG
tracing is not higher than the level during the preceding sleep stage. It almost
always reaches its lowest levels during Stage REM. These low levels may or may
not be reached during the other sleep stages, but they are reached during
unambiguous REM periods. Therefore, a low amplitude EMG contributes little to
the Scoring of sleep stages, but the presence of a "relatively elevated" tonic
EMG contributes to scoring information by precluding the scoring of Stage REM.
There are infrequent occasions during Stage REM, especially in association with
vigorous bursts of REMs, when the EMG tracing may show a slight transient
increase in tonic activity or bursts of phasic activity for several seconds. For
purposes of stage scoring, these transient changes may be disregarded.
Mixtures of Stage REM and sleep spindles
There are occasions, mostly during the first REM period of the night, when
sleep spindles are interspersed with REMs and the EMG remains at the Stage REM
level throughout. The scoring in these situations is based upon the following
two principles.
2. An interval of relatively low voltage, mixed
frequency EEG record between two sleep spindles or K complexes is considered Stage 2 regardless of EMG
level, if there are no REMs or movement arousals during the interval and if the
interval is less than 3 min long. (This is simply a reiteration of the rules for
scoring Stage 2.)
Examples of application of these rules are given below and illustrated
schematically in Fig. 6. The letter designations of the illustrations correspond
to the letter designations of the text.
A. A single sleep spindle* occurs in an epoch during which EMG is at the
Stage REM level. The preceding and succeeding epochs are Stage REM and contain
no sleep spindles. The epoch containing the sleep spindle is scored Stage REM
because all of the epoch preceding and following the sleep spindle is considered
Stage REM under rule #1
above.
B. In a 30 sec. epoch, sleep spindles occur at the 10th and 20th seconds.
The preceding and succeeding epochs were Stage REM. The first and last 10 see of
the epoch containing the sleep spindles are considered Stage REM under rule #1. The middle 10 see is
considered Stage 2 under
rule #2. Because
two-thirds (more than half) of the epoch is Stage REM, the epoch is scored as Stage REM (Fig. 6,
illustration B1). If, in the above situation, the sleep spindles had occurred at
the 5th and 25th seconds, then two-thirds of the epoch would be considered Stage 2, and the epoch would be
scored Stage 2 (Fig. 6,
illustration B2).
C. There are three successive epochs during which the EMG is at the Stage
REM level; except for the sleep spindles to be noted below, the EEG is
relatively low voltage, mixed frequency. REMs occur during the first 5 see of
the first epoch and during the last 5 see of the third epoch. Sleep spindles
occur at the 20th see of the first epoch and during the 20th see of the third
epoch. There are no REMs or sleep spindles in the middle epoch.
The first epoch is scored Stage REM because the first 20 see of the epoch,
i.e., until the sleep spindle, is considered Stage REM under rule #1.
The middle epoch is scored Stage 2, even though it
contains no sleep spindles, because under rule #2 all of the record between the
two sleep spindles is considered Stage 2.
The third epoch is scored Stage 2 because the first 20
see of the epoch is considered Stage
2 under rule #2.
Start and End of Stage REM
The major problem in scoring Stage REM is the determination of the precise
points at which REM periods begin and end. This problem arises primarily from
the fact that three indicators, EEG, EOG, and EMG activity, which are used to
define Stage REM may or may not change simultaneously. To aid in the
determination of Stage REM onset and termination in instances where the three
indicators do not change simultaneously, the following scoring problems are
presented together with the scoring recommendations.
1. Start of Stage REM
Schematic illustrations of the rules for starting Stage REM are presented
in Fig. 7. The letter designations of the illustrations correspond to the letter
designations of the text below.
A. Sleep spindles and K complexes stop and the EEG changes to relatively
low voltage, mixed frequency for one or more epochs before REMs start. Score all
the record from the last sleep spindle or K complex as Stage REM if
the EMG during the epochs before REMs is at the same level as after REMs and if
there has been no intervening movement arousal. Score the record up to the last
sleep spindle or K complex
as Stage 2
irrespective of EMG level. In rare instances where there is a progression from
Stage 3 or 4 to Stage REM, a parallel rule applies, except that slow wave
activity is substituted for sleep spindles and K complexes.
B. Sleep spindles and K complexes stop and the EEG changes to relatively
low voltage, mixed frequency for an interval of one or more epochs before REMs
start; EMG remains at a relatively high level for some portion of this interval
before it drops to the Stage REM level. Score Stage REM from the point where the
EMG amplitude dropped, providing REMs occur before any additional sleep spindles
or K complexes appear. The portion of the record following the last sleep
spindle or K complex and
prior to the EMG drop is considered as part of the preceding stage (usually Stage 2) unless it is 3 min or
longer, in which case it is scored Stage 1.
C. Sleep spindles and K complexes stop and the EEG changes to relatively
low voltage, mixed frequency for one or more epochs before REMs start; during
these epochs there is a movement arousal. Score Stage REM at the point following
the movement arousal at which the EMG tracing is reduced to the Stage REM level
providing there is no indication of a change to Stage 1 (see discussion under
End of Stage REM, section C, page 11). If the interval between the cessation of
sleep spindles and K complexes and the movement arousal is less than 3 min. it
is scored Stage 2 (Fig. 7,
illustration C1). If the interval is 3 min or longer, it is scored Stage REM
(Fig. 7, illustration C2) .
D. Sleep spindles and K complexes stop and the EEG changes to relatively
low voltage, mixed frequency for an interval of one or more epochs before REMs
start. During this interval, there is a movement arousal, and tonic EMG remains
elevated for one or more epochs following the movement arousal before dropping
to Stage REM level. Score Stage
1 for those epochs following the movement arousal during which EMG is
relatively elevated. Score Stage REM when the EMG drops to the Stage REM level
following the movement arousal. The epochs of relatively low voltage, mixed
frequency EEG which precede the movement arousal are scored Stage 2 or Stage REM according
to the 3 min rule cited under "C" (above).
2. End of Stage REM
Schematic illustrations of the rules for ending Stage REM are presented in
Fig. 8. The letter designations of the illustrations correspond to the letter
designations of the text below.
A. A period of relatively low voltage, mixed frequency EEG, but without
eye movements follows contiguously from an unambiguous Stage REM. Score as Stage
REM, regardless of duration, providing the EMG tracing remains at the Stage REM
level and there are no intervening sleep spindles, K complexes, or movement
arousals.
B. An interval of relatively low voltage, mixed frequency EEG, but without
REMs, follows contiguously (no intervening sleep spindles, K complexes, or
movement arousals) from an unambiguous State REM. Tonic EMG is initially at the
Stage REM level but becomes elevated later in the interval. Score Stage REM up
to the point of EMG augmentation. Score Stage 1 from the point of EMG
augmentation until there is some subsequent indication of a stage change, which
is usually either a resumption of REMs (Fig. 8, illustration B1) or the
occurrence of sleep spindles and/or K complexes (Fig. 8, illustration
B2).
C. A special case frequently arises where a movement arousal interrupts
the continuity of Stage REM, the mental-submental EMG quickly reverts to the
Stage REM level following the movement arousal, the EEG remains relatively low
voltage, mixed frequency, and there is a resumption of REMs or change to Stage 2 one or more epochs
following the movement arousal. The problem is whether to score the interval
following the movement arousal and the resumption of REMs or change to Stage 2 as Stage 1 or Stage REM. There is
argument for Stage 1
inasmuch as a change to a relatively low voltage, mixed frequency pattern
regularly follows many movement arousals which interrupt Stages 2, 3, and 4; there is no reason to assume
that interruptions of Stage REM would be immune to such changes to Stage 1. On the other hand, the
general principle remains that a relatively low voltage, mixed frequency EEG
contiguous with Stage REM is to be scored as Stage REM. Although the decision is
difficult, there are differences between Stage 1 and Stage REM, apart
from the presence or absence of REMs, which can be used to make a discrimination
between the two. The Committee thought it best to list general guidelines for
distinguishing between Stage
1 and Stage REM and leave it to the scorer to apply these guidelines
judgmentally.
In general, the scoring of Stage 1 is favored in
proportion to the size and duration of the movement arousal. This is based on
the fact that, when movement arousals interrupt other stages, the duration of
the post-movement Stage 1 is
generally proportional to the magnitude of the movement.
A most important indication of Stage 1 is the presence of slow
eye movements. Although slow eye movements are sometimes present during Stage
REM, they do not approach the abundance, magnitude, and pendular pattern of the
slow eye movements during Stage
1. As long as there are prominent slow eye movements following the
movement arousal, the investigator can feel confident about scoring Stage 1.
Although the EEG of Stage
1 and Stage REM are very similar, there are some differences which may
be usefully noted. Stage 1
scoring is favored in proportion to the amount of alpha activity immediately
following the movement arousal; this guide is in accord with a general
conception of the post-movement Stage 1 as a transition back to
unambiguous sleep following a change in state which approached wakefulness. The
presence of well formed vertex spikes favors the scoring of Stage 1, whereas the appearance
of well formed saw-tooth waves indicates Stage REM.
Incipient signs of sleep spindles and K complexes may also be useful in
making the difficult discrimination between Stage 1 and Stage REM following
movement arousals. The point here is best indicated by considering the typical
transition from Stage 1 to
Stage 2 at the beginning of
sleep. Just before well formed sleep spindles and K complexes appear in the
record, one can usually detect incipient signs of this activity. For example,
there may be traces of very low voltage 12-14 cps activity scattered in the
record which do not persist for the full half second required to meet the
definition of a sleep spindle; there may be wave forms which begin to approach
the morphology of the K complex, but do not fulfill the requirement. lichen such
"incipient" sleep spindles and K complexes appear in a relatively low voltage,
mixed frequency record which follows a Stage REM interrupting movement arousal,
they may be taken as supporting evidence that the interval between the arousal
and the incipient sign has been Stage 1.
If the epoch following the arousal is scored as Stage 1 then the remainder of
the record is scored as Stage
1 until the reappearance of REMs (Fig. 8, illustration C1) or until the
appearance of the first sleep spindle or K complex (Fig. 8,
illustration C2).
D. Given the above situation, but without any or with only very minimal
signs of Stage 1 following
the movement arousal (large movement, slow eye movements, initial abundance of
alpha, vertex spikes, incipient steep spindles and K complexes) the scoring
should be continued as Stage REM
(Fig. 8, illustration D ) until the next indication of a stage change
(Fig. 8, illustration D2 ) .
Publication Style
If an abbreviation for REM is used in a language other than English, it is
recommended that REM be added at the end in parentheses. For discursive portions
of the text, the term "REM sleep" may be used instead of Stage REM and the term
"NREM sleep" may be used instead of Stage NREM. Stage REM periods may be
designated simply as REM periods. The abbreviation "REM", in and of itself,
should be taken to mean "rapid eye movement". For example, "The subject had 100 minutes of Stage
REM, four REM periods with an average of eight REMs per minute."
Some uniformity is needed in the abbreviations used to indicate quantities
of sleep recording. It is suggested that the letter "T" at the beginning of an
abbreviation always represent "total" and that percentage always be represented
by the sign "%" at the beginning of an abbreviation. The letter "s" in the
second position should always designate "Stage, " and the third letter or number
should indicate the specific stage in question. According to this system,
quantities of sleep stages are given in the form indicated by the examples given
below:
TSR - Total Stage REM
%SR - Percentage Stage REM
TSN - Total Stage NREM
%SN - Percentage Stage NREM
TS1 - Total Stage 1
%S1 - Percentage Stage 1
TS2 - Total Stage 2
%S2 - Percentage Stage 2
Where percentages or fractions are used, it is important to indicate
clearly the specific numerator and denominator.
For the designation of psychophysiological measures, the term nomenclature
and abbreviations of the journal Psychophysiology are recommended.
The above listing of abbreviations should not be interpreted as
necessarily indicating that abbreviations are encouraged, but simply suggests a
standardization of abbreviations in those instances where they are used.
Just as there is saving in "decoding" time when the same abbreviations are
used in different articles, so there is some ease of visual recognition to be
gained by using a common format for the most frequent kinds of graphic
presentation. It is suggested that in histograms, white represent wakefulness,
diagonal lines represent sleep, cross-hatching represent Stage NREM, and black
represent Stage REM. Where it is desirable to represent the individual NREM
stages separately, it is left to the investigator's ingenuity to design
different cross-hatchings or stipling effects for each stage.
Where stages are represented by different ordinal levels, as in figures
which show the sequence of sleep stages with time (which is demarcated on the
abscissa), the stages should be accorded the following ordinal positions (listed
from top to bottom): W. REM, 1, 2, 3, 4. The practice of assigning Stage 1 and Stage REM to the
same ordinal level is not recommended because it increases the possibility of
confusion between the two stages. Vertical marks may be used to designate
movements. This ordinal arrangement of stages was selected because it is
convenient, not because it has any additional significance, such as a
representation of sleep depth. It is now well known that sleep stages cannot be
graded unambiguously on a generalized depth of sleep dimension (e.g.,
Rechtschaffen, Hauri and Zeitlin, 1966).
When the polygraph tracings themselves are presented, it is suggested that
tracings be arranged from top to bottom in the following order: eye movements,
EMG, EEG, and other polygraphic tracings such as ECG and respiration
.
This manual for a standard terminology and scoring system reflects the
consensus of a group of investigators each having many years of experience in
the scoring of polygraphic sleep records. Initial discussion revealed a number
of issues that needed to be resolved. The development of the manual required
extensive discussion, correspondence and revision until each of these issues was
resolved to the satisfaction of the entire Committee. This handbook should be
viewed as a working instrument rather than a statute. Many of the decisions made
are based upon an underlying conception of the organization of sleep stages
which future research may prove wrong. Experience with the manual may suggest
possible revisions. When these suggestions accumulate appreciably, it would seem
in order to have a review of the manual.
FIGURE 1. Placement of EEG, EMG and EOG electrodes.
FIGURE 2. Comparison of EEG tracings in Stages 2 and 3.
FIGURE 3. Comparison of EEG tracings in Stages 3 and 4.
FIGURE 4. Comparison of EEG tracings in Stages 2 and 3.
FIGURE 5. Comparison of EEG tracings in Stages 3 and 4.
FIGURE 6. Schematic illustrations of rules for scoring Stage
REM mixed with sleep spindles.
FIGURE 7. Schematic illustrations of rules for scoring the
start of Stage REM.
FIGURE 8. Schematic illustrations of rules for scoring the
end of Stage REM.
FIGURES 9-22. Epochs from various stages of a 19 year old, normal,
male subject.
FIGURES 23-32. Epochs from various stages of a 21 year old, normal,
male subject.
FIGURES 33-40. Epochs from various stages of a 25 year old, normal,
male subject.
Bibliography
Berger. R. J. Tonus of extrinsic laryngeal muscles during sleep and
dreaming. Science, 134: 840, 1961.
Berger, R.J., Olley, P. and Oswald, I. The EEG, eye movements and dreams
of the blind. Quart. J. Exp. Psychol., 14: 183-186, 1962.
Brazier, M . A. B., Cobb, W. A., Fischgold, H., Gastaut, H., Gloor, P.,
Hess, R., Jasper, H., Loeb, C., Magnus, O., Pampiglione, G., Remond, A., Storm
van Leeuwen, W. and Walter, W. G. Preliminary proposal for an EEG terminology by
the Terminology Committee of the International Federation for
Electroencephalography and Clinical Neurophysiology. Electroenceph. clin.
Neurophysiol., 1;3: 646-650, 1961.
Brown, C. C. A proposed standard nomenclature for psychophysiologic
measures. Psychophysiol., 4: 260-264, 1967.
Dement, W., and Kleitman, N. Cyclic variations in EEG during sleep and their relation to eye movements, body motility, and dreaming. Electroenceph. clin. Neurophysiol.. 9: 673-690, 1957.
Jacobson. A., Kales, A., Lehmann, D., and Hoedemaker, F.S. Muscle tonus in
human subjects during sleep and dreaming. Exp. Neurol.,10: 418 -424,
1964.
Jasper. H. H. (Committee Chairman). The ten twenty electrode system of the
International Federation. Electroenceph. clin. Neurophysiol.,10: 371-375,
1958.
Johnson, L. C. and Karpam, W. E. Autonomic correlates of the spontaneous
K-complex. Psychophysiol., 4: 386, 1968. (Abstract. Article in
press.)
Johnson, L. C., Nute, C., Austin, M. T. and Lubin, A. Spectral analysis of
the EEG during waking and sleeping. Electroenceph. clin. Neuro-physiol.,
23: 80, 1967.
Loomis,A. L., Harvey, E. N., and Hobart, G.A. Cerebral states during sleep
as studied by human brain potentials. J. Exp. Psychol., 21: 127-144,
1937.
Monroe, L.J. Inter-rater reliability of scoring EEG sleep records. Paper
read at the Association for Psychophysiological Study of Sleep Meeting, Santa
Monica, California. April, 1967. Abstract in Psychophysiol ., 4: 3
70-371, 1968.
Rechtschaffen, A. Personal communication, 1968.
Rechtschaffen, A., Hauri, P., and Zeitlin, M. Auditory awakening
thresholds in REM and NREM sleep stages. Percept. Motor Skills, 22:
927-942, 1966.
Rechtschaffen, A ., Wolpert, E . A., Dement, W. C., Mitchell, S . A., and
Fisher, C. Nocturnal sleep of narcoleptic Electroenceph. clin.
Neurophysiol., _: 599-609, 1963.
Roth, B. The clinical and theoretical importance of EEG rhythms
corresponding to states of lowered vigilance. Electroenceph. clin.
Neurophysiol., 13: 395-399, 1961.