~I
Sleep, 7(4):289-303
© 1984 Raven Press, New York
EEG Sleep of Normal Healthy Children.
Part I: Findings Using Standard
Measurement Methods
*tPatricia A. Coble, *David 1. Kupfer, *Lynn S. Taska, and *Judith Kane
*Western Psychiatric Institute and Clinic, and tDepartment of Psychiatry, University of
Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, US.A.
Summary: Despite the increasing application of all-night electroencephalographic (EEG) sleep studies to children for clinical as well as for research purposes,
readily available normal EEG sleep standards for the period of childhood have
remained sparse and, at present, reflect data on only approximately 100 children
6 to 16 years of age. As part of a large scale study examining various aspects of
EEG sleep among children, findings derived using standard recording and scoring
methods are reported for a new sample of nearly 100 normal, healthy children
and are compared with existing standards. Data obtained add substantially to the
existing database and generally confirm findings of previous normative reports
on children in this age range. Key Words: EEG sleep--Standard measurement
methods-Normal sleep, children.
The objective measurement of human sleep using all-night recording methods represents
an area of ongoing interest and activity, most notably~ince the discovery of REM sleep in
1953 (1,2). Although precise mechanisms have yet to be determined, work in this field has
been both rich and diverse, contributing greatly to our present understanding of the complex
processes involved in human sleep-wake cycles (3-6). Increasing interest in human sleep
has been even more apparent in the last decade with the emergence of growing numbers
of highly specialized clinical sleep facilities and the establishment of a working classification
system for the sleep and arousal disorders (7). Thus, in addition to their continued application
in research efforts to realize more fully the neurophysiological bases of sleep and the various
sleep states, all-night electroencephalographic (EEG) sleep studies are being applied in the
clinical assessment of a broad range of sleep-wake disturbances among individuals of all
ages. These more recent activities have not only renewed interest in various aspects of
"normal" sleep-wake behavior across the life span but have also highlighted the relative
paucity of EEG sleep standards for certain age groups. Among children, for example, the
most frequently referenced normative data include the preliminary observations of Roffwarg
Accepted for publication July 1984.
Address correspondence and reprint requests to Dr. Coble at Western Psychiatric Institute and Clinic, 3811
O'Hara Street, Pittsburgh, Pennsylvania 15213, U.S.A.
289
290
P. A. COBLE ET AL.
et al. (8), the work of Feinberg et al. (9) and Ross et al. (10), and the well known "atlas"
of Williams et a1. (11). All of these are based on data obtained well over a decade ago
and, taken together, reflect EEG sleep findings on a total of only approximately 100 children
between the ages of 6 and 16 years old. To our knowledge, with the exception of some
preliminary work from our laboratory (12) and some recently published normal standards
by Carskadon and others for early and older adolescents (13,14), no other systematically
conducted and purely normative studies are readily available for the period of childhood.
In light of increasing applications of all-night EEG sleep studies to this age group for clinical
and for research purposes, efforts to replicate and to add to the existing normal standards
for children would seem both timely and highly desirable. Although the present study was
undertaken primarily to establish normal standards for children using computer-based
automated techniques developed in our laboratory setting for the scoring of REM and Delta
sleep activity (15,16), all-night EEG sleep measures using the standard criteria of Rechtschaffen and Kales (17) were obtained simultaneously as well. Availability of these latter
measures for a recently studied sample of normal, healthy children provided us with a
built-in opportunity to examine our findings with respect to replication of previous work
in this area, and thus, to add to the existing normal standards for children. The focus of
this report will be limited to our findings using the standard criteria for recording and
scoring EEG sleep. Findings derived for this normative sample using automated measurement techniques will be presented in a separate report.
METHODS
Subjects
The subjects included in this study were 87 normal, healthy children, 45 girls and 42
boys, ranging in age from 6 years 0 months to 15 years 11 months. All children were
reported to be of at least normal intelligence and were generally within the normal range
of height and weight centile for their respective age groups by gender. All children were
likewise without personal past or present histories of significant medical, neurological, or
psychiatric illness and suffered from no current or recent (with 12 months) major sleep
disturbances or parasomnias. The sample was predominantly white and from middle and
upper-middle class families based on parental education/occupation. All children lived with
at least one biological parent either within the city of Pittsburgh itself or in nearby suburban
areas. Thirty-seven children or approximately 43% of the sample were biologically unrelated
to any other children in the study; the remaining 50 children included sibling groups as
follows: 15 sibling pairs (one set of identical twins), five sibling trios, and one group of
five siblings. Same and opposite sex siblings were balanced across all age groups.
Screening and consent procedures
Initial contact with interested children and their parents consisted of a general explanation
of the nature and purposes of the study, a systematic review of the eligibility criteria for
inclusion in the study, and establishment of both the willingness and ability of parents to
provide specified information on their child(ren) including: certain demographic data,
information pertaining to their child(ren)'s overall health and development, and information
relating to their child(ren)'s biological family's health history. Standardized forms were
used to collect these data and were mailed to parents for completion at home prior to a
formally scheduled visit to the sleep laboratory with their child(ren). During this subsequent
prestudy visit, all information on these forms was reviewed with parents by one or another
of two of the authors (P. A. C. or 1. K.).
Sleep, Vol. 7, No.4, 1984
EEG SLEEP OF NORMAL CHILDREN
291
While their parents were thus occupied, children were acquainted with the study separately. Familiarization with the laboratory facilities, an age appropriate explanation of the
nature of the study and its purposes, and a demonstration of the procedures involved were
provided by a research staff member who was regularly present in the sleep laboratory on
those nights when children's studies were performed. At the end of this familiarization
session, children were rejoined by their parents and the research investigator. The children
themselves were then asked to give their parents a "tour" of the facilities, to explain to
them in their own words the nature and purposes of the study, to demonstrate the various
procedures involved, and, finally, to discuss their decision as to whether or not they wished
to participate in the study. No child who expressed any reluctance or resistance to participate
at this point was accepted into the study regardless of the desires and/or willingness of
parents. Written informed consent of parents and children was obtained only after these
preliminary steps had been taken and in accordance with guidelines established by the
Institutional Review Board for Psychosocial Research at the University of Pittsburgh.
Children were scheduled to return to the sleep laboratory. for all-night EEG sleep studies,
usually within 2 or 3 weeks of their prestudy visit.
Procedures for recording and scoring EEG sleep
Our laboratory setting and general procedures for recording and scoring EEG sleep have
been reported previously for adults and for children (18,19). Briefly, our sleep laboratory
area, in addition to the instrumentation/control room itself, consists of six private bedrooms
with comfortable furniture comparable to that of a good motel, a denlike TV lounge, and
a small kitchenette. In addition, a computer terminal with access to a variety of computer
games was made available to children during the evening hours and rapidly became a
popular attraction. All children were scheduled to sleep in the laboratory for 3 consecutive
nights and were required to be free of any minor ailments and medications for a period of
at least 2 weeks prior to their sleep study nights. Children's studies were conducted on
weeknights during the school year in an attempt to achieve some reasonable degree of
control with regard to daytime activity and sleep-wake schedules. Thus, on laboratory study
nights children maintained bedtimes and rising times according to what was normal for
them during the school week. Instructions in these regards were provided by parents on
each study night. Children were awakened by laboratory staff at times designated by parents
unless they awakened spontaneously prior to these times. Finally, in nearly all cases children
came to the sleep laboratory in pairs, having chosen overwhelmingly an available option
to be scheduled for studies on the same nights as a close peer or sibling. Children reported
to the sleep laboratory at least 2 h prior to their usual bedtime for electrode application.
Electroencephalographic (C3/A2 or C4/Al, at 50 fLV/cm for 1-30 Hz), referential electrooculognlphic (right and left outer canthi), and bipolar electromyographic (chin muscle)
activity were recorded on a Grass Model 78B polygraph at a paper speed of 10 mmls. In
addition, all sleep data were simultaneously recorded onto 1 inch frequency modulated
tapes using a Honeywell 5600E recorder, a routine procedure in our setting allowing for
analyses using automated techniques (15,16). The visual scoring of all children's records
was based on 1 min epochs and was carried out independently by scorers having demonstrated high interscorer reliability using the established criteria of Rechtschaffen and
Kales (17).
For purposes of presentation, the visually scored sleep variables are summarized in three
main categories: sleep continuity measures, non-REM sleep measures, and REM sleep
measures. Measures of sleep continuity as derived and defined in our setting include: the
Sleep, Vol. 7, No.4, 1984
292
P. A. COBLE ET AL.
total recording period (TRP), the time from the beginning of the recording at "lights out"
until ihe end of the recording at "good morning time"; sleep latency, the time from the
beginning of the recording until the onset of stage 2 sleep for 10 consecutive min interrupted
by not more than 2 min of stage 1 or wakefulness; early morning awakening, the time
spent awake after the last sleep occurrence (Stage 2-4, or REM for at least 10 consecutive
min interrupted by not more than 2 min of stage 1 and/or wakefulness) prior to the end of
the recording; awake, time spent awake that is bounded by sleep; the number of arousals;
time spent asleep (TSA); and sleep efficiency, a ratio of TSA to TRP x 100. The nonREM sleep measures include the amounts of time spent in each of the different non-REM
sleep stages, stages 1-4, and in total Delta sleep (stages 3 and 4 combined), expressed as
percentages of TSA; and Delta ratio, a ratio of the amount of stage 3 sleep to total Delta
sleep. Finally, the REM sleep measures in our setting include: REM latency, the number
of minutes from the onset of sleep until the onset of the first REM period minus any
wakefulness that occurs during that interval; REM sleep time (RT) , expressed both in
minutes and as a percentage of TSA; REM activity (RA), an approximated measure or
estimate of the amount of phasic eye movement activity during REM sleep (based on a
score of 0 to 8 for each minute of REM sleep and reported as "units" of REM activity)
(20); REM intensity (RA/TSA); REM density (RNRT); and the number of REM periods.
Definition/limits of the independent variables
The independent variables of interest in this study were age, sex, and pubertal status.
With regard to chronological age, children were divided into five groups: (a) 6 years 0
months to 7 years 11 months; (b) 8 years 0 months to 9 years 11 months; (c) 10 years 0
months to 11 years 11 months; (d) 12 years 0 months to l3 years 11 months; and (e) 14
years 0 months to 15 years 11 months. These groupings are similar to those used by others
in normative studies among children except for the division of 6 to 9 year olds into two
separate groups. We decided to report values separately for children in this age range
primarily because initial examination of our sleep data suggested that some differences
might exist, particularly in those measures reflecting adaptation to the laboratory. The
variable pubertal status was based on the physical changes as described by Tanner (21),
but due to the inability of many of the parents to obtain a current or recent Tanner score
from their children's pediatricians, it is essentially a relative rating of pubertal change as
ascertained from parents and their children. Pubertal status was, thus, assigned by us to
the subjects in this study as follows: where there was clearly no evidence of change, a
rating of prepubertal was assigned; a rating of postpubertal was assigned to subjects who
were ascertained to be sexually mature for a period of at least 1 year. The remaining
subjects, that is, those exhibiting any of the changes associated with pubertal development,
were assigned a rating of pubertal.
ANALYSES OF THE DATA
The visually scored sleep variables were first examined for the entire sample of children
within-subjects, across nights by means of paired t test comparisons. The effects of age
and sex on night-to-night differences were then examined using a repeated measures analysis
of variance (ANOVA). Finally, sleep pattern differences, based on the mean values for the
second and the third study night, were examined in a three-way ANOVA using age, sex,
and pubertal status. Due to the strong correlation between age and pubertal status in these
children (r = 0.74 for girls; r = 0.65 for boys), interaction effects were assessed separately
using analyses of covariance.
Sleep, Vol. 7, No.4, 1984
EEG SLEEP OF NORMAL CHILDREN
293
RESULTS
Examination of night-to-night differences
Table 1 presents the findings of paired t test comparisons on the standard sleep measures
across the three study nights for the entire sample of children. As can be seen, laboratory
adaptation or "first night" effects are most marked for measures of sleep continuity. Among
these measurements, significant decreases in sleep latency (p < 0.001), early morning
awakening (p < 0.05), awake time (p < 0.01), and the number of arousals (p = 0.050)
were observed to occur from Night 1 to. Night 2 in association with a significant increase
in the TSA (p = 0.010). Accordingly, overall sleep efficiency rose from approximately
91 % on Night 1 to 94% on Night 2 (p < 0.001). From Night 2 to Night 3, only two
significant differences were noted for these measures: TSA showed continued improvement
(p < 0.05) as did sleep efficiency, which rose from 94% on Night 2 to 95% on Night 3
(p < 0.05). In contrast, night-to-night differences were notably absent for the non-REM
sleep measures. A gradual but nonsignificant decline in the percent of stage 1 sleep was
noted, whereas all other non-REM sleep stage percents appeared to be remarkably stable
across the 3 study nights. Finally, among the REM sleep measures, there appeared to be
evidence for both first night and for continued adaptation effects. For example, REM latency
appeared to reflect primarily a first night effect, with a significant decrease observed only
from Night 1 to Night 2 (p < 0.05). Also, REM sleep time, REM percent, and the number
of REM periods appeared to reflect first night effects as evidenced by significant increases
in all three measures from Night 1 to Night 2 (p < 0.001 for each). However, values for
these measures continued to increase from Night 2 to Night 3 (p < 0.01 for REM time;
p < 0.05 for REM percent and for the number of REM periods). With respect to measures
of REM activity during REM sleep, total REM activity was seen to increase significantly
from Night 1 to Night 2 (p = 0.001) with further increases from Night 2 to Night 3
(p < 0.001) as was also the case for REM intensity (p = 0.001 for each comparison).
Only REM density was shown not to differ significantly across the 3 study nig ... s.
The effects of age and sex on night-to-night differences were examined by meilns of a
three-way repeated measures ANOVA. Table 2 summarizes the results of this analysis.
Although there were no significant age-by-sex or sex-by-night interactions, a number of
age-by-night interactions were observed. For measures of sleep continuity, significant effects
were demonstrated for sleep efficiency (F = 2.80, P < 0.01), for TSA (F = 3.85,
P < 0.001), and for early morning awakening (F = 2.46, P < 0.05). In all instances these
differences appeared to be due largely to a marked first night effect, most pronounced in
the younger children. With regard to sleep efficiency, for example, 6 and 7 year aIds showed
an increase from approximately 83% on Night 1 to 94% and 95% on Nights 2 and 3. Time
spent asleep showed a similar although not identical pattern reflecting increases from Night
1 to Nights 2 and 3 of greater than 1 h in 6 and 7 year aIds and of approximately 20-30
min in 8 and 9 year aIds as contrasted with increases of only approximately 10 min among
the older age groups. Early morning awakening, which was noted to be quite variable in
all age groups, nevertheless, was again most pronounced in children 6 to 9 years of age.
These children showed decreases from Night 1 to Nights 2 and 3 of approximately 10-20
min as contrasted with only 1-2 min decreases in the older children. Among the non-REM
sleep measures, the only significant interaction effect was for stage 2 sleep percent (F = 2.16,
P < 0.05), which was seen to decrease to a greater extent among older children as compared
with children 6 to 9 years of age. Among the REM sleep measures, significant age-bynight effects were demonstrated for both total REM activity (F = 2.23, P < 0.05) and
Sleep. Vol. 7. No.4. 1984
tv
~
~
:I:l
~
."
~
.'"
~
TABLE 1. Night-to-night differences in EEG sleep measures for normal healthy children
EEG sleep parameters
Sleep continuity measures
Total recording period (TRP) (min)
Sleep latency (min)
Early morning awakening (min)
Awake (min)
Number of arousals
Time spent asleep (TSA) (min)
Sleep efficiency (TSA/TRP x 100)
Non-REM sleep measures
Stage 1 percent
Stage 2 percent
Stage 3 percent
Stage 4 percent
Delta percent (stage 3 + 4)
Delta ratio (stage 3/Delta)
REM sleep measures
REM latency (minus awake) (min)
REM time (RT) (min)
REM percent (RT/TSA)
REM activity (RA) (units)
REM intensity (RAITSA)
REM density (RA/RT)
Number of REM periods
!:iight 1
x (SD)
501.3
26.5
6.3
13.9
2.9
454.5
90.9
(58.6)
(13.4)
(19.9)
(25.2)
(3.4)
(52.5)
(6.4)
Paired t test comparisons
!:iight 2
x (SD)
!:iight 3
x (SD)
Night 1 vs. 2 (p)
Night 2 vs. 3 (p)
495.6 (59.5)
21.0 (12.0)
1.7 (7.2)
5.4 (7.4)
2.1 (2.3)
467.5 (55.5)
94.4 (2.7)
499.2 (58.6)
20.0 (11.3)
0.6 (3.0)
4.3 (7.2)
1.8 (2.1)
474.2 (53.0)
95.1 (2.5)
<0.05
<0.001
<0.05
<0.01
0.05
0.01
<0.001
NS
NS
NS
NS
NS
<0.05
<0.05
"'\:I
>
("')
a
0:1
8.3
52.0
7.5
13.6
21.1
0.37
(4.0)
(6.5)
(3.1)
(5.5)
(5.3)
(0.17)
7.7
51.0
7.4
13.2
20.6
0.38
(3.4)
(5.3)
(2.9)
(5.1)
(4.6)
(0.18)
7.2
50.1
7.7
13.4
21.1
0.38
(3.5)
(5.5)
(3.3)
(5.3)
(5.1)
(0.18)
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
141.6
82.7
18.1
99.3
0.21
1.17
3.9
(43.5)
(20.9)
(3.6)
(48.5)
(0.09)
(0.43)
(1.0)
126.2
95.9
20.4
114.8
0.24
1.15
4.3
(41.7)
(22.8)
(3.7)
(56.4)
(0.11)
(0.41)
(1.0)
122.3
101.2
21.3
128.2
0.27
1.22
4.4
(40.2)
(24.7)
(4.1)
(65.6)
(0.12)
(0.43)
(1.0)
<0.05
<0.001
<0.001
0.001
0.001
NS
<0.001
NS
<0.01
<0.05
<0.001
0.001
NS
<0.05
~
t:1
'-1
~
Age range, 6-16 years.
Number of children, 87.
~
-~
TABLE 2. Summary of ANOVA results, age x sex x night for normal healthy children
Significant main effects
Sex
Age
p
F
~
~
~
.' l
~
:'"
"-
10
~
Sleep continuity measures
Total recording period (TRP) (min)
Sleep latency (min)
Early morning awakening (min)
Awake (min)
Number of arousals
Time spent asleep (TSA) (min)
Sleep efficiency (TSNTRP x 100)
Non-REM sleep measures
Stage I percent
Stage 2 percent
Stage 3 percent
Stage 4 percent
Delta percent (stage 3 + 4)
Delta ratio (stage 3/Delta)
REM sleep measures
REM latency (minus awake) (min)
REM time (RT) (min)
REM percent (RT/TSA)
REM activity (RA) (units)
REM intensity (RA/TSA)
REM density (RA/RT)
Number of REM periods
F
Significant interactions·
Age x sex x
Age x night
night
Night
P
F
P
F
P
F
P
~
c:;';)
15.84
8.66
15.78
<0.001
<0.001
<0.001
6.40
<0.001
7.86
5.57
7.25
<0.001
0.001
<0.001
2.56
4.45
<0.05
<0.01
3.35
<0.05
6.39
<0.001
Age range, 6-16 years .
Number of children, 87.
aThere were no significant age x sex or sex
X
2.97
4.13
<0.05
16.84
10.68
14.49
3.20
22.83
42.71
<0.001
<0.001
<0.001
<0.05
<0.001
<0.001
3.85
2.80
<0.001
<0.01
4.39
<0.05
2.16
<0.05
2.46
<0.05
<0.01
~
~
a'"lj
"ti
~
~
~
~
0
~
38.48
28.79
26.29
19.68
<0.001
<0.001
<0.001
<0.001
14.80
<0.001
~
2.23
2.06
<0.05
<0.05
night interactions.
N
~
P. A. COBLE ET AL.
296
TABLE 3. Summary of ANOVA results, age x sex x pubertal status for
normal healthy children
Significant main effects
Sex
Age
Sleep continuity measures
Total recording period (TRP) (min)
Sleep latency (min)
Early morning awakening (min)
Awake (min)
Number of arousals
Time spent asleep (TSA) (min)
Sleep efficiency (TSNTRP x 100)
Non-REM sleep measures
Stage 1 percent
Stage 2 percent
Stage 3 percent
Stage 4 percent
Delta percent (stage 3 + 4)
Delta ratio (stage 3/Delta)
REM sleep measures
REM latency (minus awake) (min)
REM time (RT)
REM percent (RT/TSA)
REM activity (RA) (units)
REM intensity (RAITSA)
REM density (RNRT)
Number of REM periods
F
P
7.01
<0.001
4.58
<0.01
8.32
<0.001
2.73
<0.05
5.55
3.45
4.25
0.001
<0.05
<0.01
F
Pubertal status
P
5.86
<0.05
6.91
0.01
F
P
3.70
<0.05
Age range, 6-16 years.
Number of children, 87.
for REM intensity (F = 2.06, P < 0.05). For these measures, both the youngest (6 and
7 year olds) and the oldest (14 and 15 year olds) children in our sample showed dramatic
increases from Night 1 to Night 2 with more modest increases from Night 2 to Night 3
whereas children 8 and 9 years of age showed this pattern in reverse. In contrast, children
from 10 to 13 years of age showed relatively small but steady increases in REM activity
across the 3 study nights. Finally, only one significant three-way interaction was noted,
for TRP (F = 2.97, P < 0.01). In this instance, boys consistently spent more time in bed
than did girls, there was a greater decrease in the TRP from Night 1 to Night 2 among
boys than among girls, and for both boys and girls the TRP was seen to decrease as a
function of increasing age.
Sleep pattern differences as related to age, sex, and pubertal status
Based on the results of analyses examining internight differences, Night 1 was considered
to be an adaptation night and was excluded from all subsequent analyses. Sleep patterns
as reported here reflect, therefore, the mean values for study Nights 2 and 3 only. Sleep
pattern differences were examined by means of a three-way analysis of variance. Interaction
effects were assessed separately using analyses of covariance due to the strong correlation
between age and pubertal status in this sample of children. Significant main effects on the
EEG sleep measures are presented in Table 3. As can be seen, age accounted for nearly
all of the significant differences observed. In light of this and in the absence of any significant
interaction effects mean values for the EEG sleep measures in this normative sample are
Sleep. Vol. 7. No.4. 1984
.~
EEG SLEEP OF NORMAL CHILDREN
r
I
297
presented in Tables 4a and 4b by age groups collapsed across gender with discussion
separately of the two Delta sleep measures for which a significant sex difference was
observed.
For measures of sleep continuity, significant age effects were observed for TRP (F = 7.01,
P < O.OOl),forTSA(F = 8.32,p < O.OOI),andforearlymorningawakening(F = 4.58,
P < 0.01). Total recording period and TSA demonstrated a gradual and linear decline from
approximately 9Y2 and 9 h, respectively in 6 and 7 year olds to 7Y2 and 7 h in 14 and 15
year olds. Early morning awakening, although highly variable in all age groups, was most
clearly evident among the very youngest children 6 and 7 years of age. Among the nonREM sleep measures, the percent of stage 2 sleep showed a gradual and linear increase
across the five age groups (F = 2.73, P < 0.05) as the percent of total Delta sleep declined
(F = 3.45, P < 0.05). This decline in Delta sleep appeared to be largely accounted for
by a significant decrease in the percent of stage 4 sleep from mean values of approximately
18% in 6 and 7 year olds to only approxim~tely 9% in 14 and 15 year olds (F = 5.55,
P = 0.001). Accordingly, Delta ratio was seen to increase from a mean value of 0.26 in
6 and 7 year olds to a mean value of 0.53 in 14 and 15 year olds (F = 4.25, P < 0.01).
With regard to the REM sleep measures, although some changes were noted to occur across
the five age groups, no. significant differences were noted to occur as a function of age,
sex, or pubertal status. Figure 1 graphically summarizes the major sleep pattern differences
observed across the five age groups.
Finally, as previously mentioned, two significant effects were noted for sex: for the
percent of stage 3 sleep (F = 5.86, P < 0.05) and for Delta ratio (F = 6.91, P = 0.010).
In general, these effects were accounted for by higher stage 3 percents and, accordingly,
higher Delta ratios among the boys in this sample (see Fig. 2).
DISCUSSION
Over the past decade, all-night EEG sleep studies have been increasingly applied to
children for a diversity of purposes, although normal standards for these age groups have
remained both sparse and, for the most part, based on studies conducted nearly 2 decades
ago. Indeed, at present, available normal standards for the period of childhood reflect data
on only approximately 100 children 6 to 16 years of age. Yet, the ready availability of such
data is a necessary prerequisite to identifying and, ultimately, to understanding any alterations in sleep that may occur in association with the various developmental processes and
pathological states in childhood. The present study, undertaken to examine various aspects
of EEG sleep in childhood, was, nevertheless, designed in such a way as to allow us to
compare our findings for the standard EEG sleep parameters with previous reports and to
add to existing normal data a substantial number of healthy children. The results of our
study on the standard, visually scored EEG sleep measures among children 6 to 16 years
of age are, overall, consistent with the major findings of previous reports for this age group.
Laboratory adaptation was evident among the children in our sample, affecting primarily
measures of sleep continuity and certain aspects of REM sleep in much the same manner
as has been reported for adults (22) and children (23). That is, children experienced both
sleep onset difficulty and increased wakefulness after sleep onset, resulting in a shorter
sleep period and a significantly lower sleep efficiency on their first night in the laboratory
as compared with subsequent study nights. In our sample of normal, healthy children we
found this effect to be particularly marked in the youngest children 6 to 7 years of age.
Similarly, among the REM sleep measures, first night effects were reflected both in the
Sleep. Vol. 7. No.4. 1984
'"'"
N
~
~
~
TABLE 4a. All-night EEG sleep measures in normal healthy children (mean values for Nights 2 and 3)
:'I
~
....
~
Sleep continuity measures'
Total recording period (TRP) (min)
Sleep latency (min)
Early morning awakening (min)
Awake time (min)
Number of arousals
Time spent asleep (TSA) (min)
Sleep efficiency (TSA/TRP x 100)
Non-REM sleep measures
Stage 1 sleep (min)
(%)
Stage 2 sleep (min)
(%)
Stage 3 sleep (min)
(%)
Stage 4 sleep (min)
(%)
Total Delta sleep (min)
(%)
(Stages 3 and 4 combined)
Delta ratio (stage 3/total Delta)
REM sleep measures
REM latency (minus awake) (min)
REM sleep time (RT) (min)
REM sleep percent
REM activity (RA) (units)
REM intensity (RA/TSA)
REM density (RAlRT)
Number of REM periods
Age
6 and 7 year
olds
(n = 9)
8 and 9 year
olds
(n = 14)
579.6 (64.1)
18.4 (9.0)
6.3 (10.3)
8.2 (9.0)
3.2 (2.2)
546.6 (53.2)
94.5 (2.9)
42.7
7.7
258.5
47.2
34.8
6.3
94.9
17.6
129.7
24.0
0.26
142.3
113.0
20.7
155.2
0.28
1.35
5.2
7.4
10 and 11 year
olds
(n = 20)
12 and 13 year
olds
(n = 27)
14 and 15 year
olds
(n = 17)
543.3 (46.4)
24.6 (15.6)
0.7 (2.6)
4.3 (3.0)
2.7 (1.6)
513.7 (38.7)
94.7 (2.5)
490.5 (40.5)
19.5 (12.4)
0.5 (0.2)
3.7 (5.3)
1.6 (1.9)
467.2 (33.2)
95.4 (2.5)
478.3 (40.0)
20.3 (6.0)
(3.3)
1.1
5.6 (5.4)
2.5 (2.1)
451.1 (37.6)
94.4 (1.6)
448.3 (33.9)
20.9 (9.7)
0.0 (0.0)
4.8 (5.8)
2.4 (1.9)
422.6 (33.2)
94.2 (1.9)
(20.1)
(3.4)
(43.2)
(5.4)
(11.3)
(2.1)
(9.6)
(3.0)
(16.5)
(3.9)
(0.19)
39.7
7.7
247.5
48.3
34.0
6.5
77.7
15.3
111.7
21.8
0.32
(12.7)
(2.2)
(24.0)
(4.6)
(14.0)
(2.5)
(16.9)
(3.9)
(14.8)
(3.1)
(0.12)
32.3
6.9
242.8
52.0
32.5
7.0
65.0
14.0
97.5
21.0
0.33
(14.3)
(3.1)
(26.6)
(4.5)
(16.0)
(3.4)
(18.9)
(4.2)
(21.0)
(4.6)
(0.16)
37.2
8.1
227.6
50.5
37.2
8.3
56.7
12.6
93.9
20.9
0.40
(17.6)
(3.7)
(23.4)
(3.9)
(10.1)
(2.2)
(18.4)
(4.2)
(19.3)
(4.4)
(0.13)
29.6
6.9
225.3
53.7
36.0
8.6
38.8
8.9
73.7
17.4
0.52
(9.3)
(2.5)
(21.0)
(5.3)
(11.4)
(2.6)
(24.1)
(5.4)
(21.2)
(4.4)
(0.23)
(38.7)
(25.7)
(3.9)
(55.1)
(0.08)
(0.28)
(1.0)
(0.6)
129.5
113.9
22.0
157.9
0.30
1.34
4.8
9.4
(31.9)
(26.9)
(4.1)
(64.9)
(0.11)
(0.35)
(1.0)
(0.9)
132.5
93.1
19.8
109.6
0.23
1.14
4.2
11.4
(30.8)
(17.6)
(2.9)
(52.3)
(0.10)
(0.41)
(0.8)
(0.6)
119.3
91.3
20.2
11 o. 7
0.24
1.17
4.1
13.0
(27.9)
(16.8)
(3.1)
(50.1)
(0.11)
(0.42)
(0.7)
(0.6)
106.1
92.3
21.9
99.0
0.23
1.04
4.0
14.8
(34.8)
(19.2)
(3.3)
(58.2)
(0.12)
(0.39)
(0.7)
(0.5)
"tI
~
(]
§5
t'"<
t"rl
t"rl
....,
~
Measures given as mean (SO).
.
~
TABLE 4b. Individual REM period measures in normal healthy children (mean values for Nights 2 and 3)
'"
"
~
~
1st REM period
REM time (RT) (min)
REM activity (RA) (units)
REM density (RA/RT)
2nd REM period
REM time (RT) (min)
REM activity (RA) (units)
REM density (RA/RT)
3rd REM period
REM time (RT) (min)
REM activity (RA) (units)
REM density (RA/RT)
4th REM period
REM time (RT) (min)
REM activity (RA) (units)
REM density (RA/RT)
5th REM period
REM time (RT) (min)
REM activity (RA) (units)
REM density (RA/RT)
REM/NREM cycle lengths
1st REM/NREM cycle (minus awake)
(min)
2nd REM/NREM cycle (minus awake)
(min)
3rd REM/NREM cycle (minus awake)
(min)
4th REM/NREM cycle (minus awake)
(min)
6 and 7 year
olds
(n = 9)
8 and 9 year
olds
(n = 14)
10 and 11 year
olds
(n = 20)
12 and 13 year
olds
(n = 27)
14 and IS year
olds
(n = 17)
n = 9
14.6 (6.3)
10.8 (5.4)
0.72 (0.13)
n = 9
19.5 (4.1)
21.9 (8.0)
1.12 (0.34)
n = 9
22.8 (8.1)
30.3 (11.5)
1.33 (0.42)
n = 8
26.1 (7.1)
40.4 (17.3)
1.51 (0.52)
n = 8
23.0 (11.4)
37.4 (25.7)
1.56 (0.53)
n = 14
15.7 (7.0)
14.0 (11.6)
0.79 (0.37)
n = 14
24.7 (7.5)
31.6 (19.9)
1.19 (0.45)
n = 14
28.0 (10.1)
36.6 (15.3)
1.32 (0.25)
n = 14
24.4 (10.1)
36.5 (25.1)
1.31 (0.57)
n = 10
26.8 (14.7)
49.3 (38.1)
1.64 (0.64)
n = 20
12.5 (6.4)
8.8 (6.9)
0.63 (0.28)
n = 20
24.2 (7.7)
24.6 (12.9)
1.00 (0.40)
n = 20
28.2 (9.3)
35.9 (20.2)
1.20 (0.53)
n = 19
24.8 (8.3)
30.7 (17.1)
1.21 (0.45)
n = 9
26.8 (12.4)
43.7 (29.9)
1.71 (0.55)
n = 27
12.6 (6.6)
9.2 (7.1)
0.67 (0.30)
n = 27
26.2 (10.0)
31.3 (23.0)
1.09 (0.44)
n = 27
25.8 (9.6)
30.7 (20.5)
1.13 (0.47)
n = 22
25.6 (9.0)
37.8 (23.5)
1.39 (0.54)
n = 10
18.0 (12.5)
27.4 (29.1)
1.19 (0.70)
n = 17
14.8 (9.2)
11.9 (11.8)
0.68 (0.33)
n = 17
28.3 (9.9)
28.7 (24.6)
0.91 (0.46)
n = 17
24.7 (12.4)
23.9 (13.0)
0.97 (0.46)
n = 16
28.2 (12.0)
37.9 (30.9)
1.19 (0.64)
n = 5
21. 7 (12.2)
33.7 (29.4)
1.45 (0.67)
~
C':l
~
~
a
'"ti
~
~
~
~
t.::;
\::)
::tI
85.9 (14.2)
97.1 (20.9)
89.7 (14.2)
92.5 (17.0)
95.8 (27.1)
92.7 (17.6)
93.2 (12.7)
98.8 (11.6)
101.9 (14.0)
99.8 (15.0)
(9.7)
91.2 (17.2)
89.5 (16.1)
87.5 (13.3)
77.0 (20.8)
87.1 (25.2)
74.0 (17.3)
69.1 (17.0)
76.2 (13.4)
89.0 (12.2)
84.4
~
:'-'
~
Measures given as mean (SD).
~
......
'C
'l!;
N
\0
\0
..
P. A. COBLE ET AL.
300
500-1
400
rn
Q)
:lto:
i
300
200
~
Legend
~
Stage
CJ Stage
[SI Stage
1Z2I Stage
_ Stage
100
a
6-7 Yrs
8-9 Yrs
10-11 Yrs
12-13 Yrs
14-15 Yrs
REM
4
3
2
1
FIG. 1. Time spent asleep by sleep stage for normal healthy children ages 6 to 16 years.
prolongation of the REM sleep latency and in reduced amounts of REM sleep time and
phasic REM activity. In fact, the latter measures were seen to continue to increase significantly across the subsequent 2 nights among these children. In contrast, significant night0.65
0.60
'"iii'
0.55
~
Q)
0
..,«l
{!.
~
0.50
0.45
<II
..,""
'"
S
.Sl
..,
0.40
0::
0.35
'"
..,'"
Qj
0
0.30
Legend
0.25
o~
0.20-'------r-------.-----,-------.----.-----
6-7 Yrs.
FIG. 2.
Sleep, Vol. 7, No.4, 1984
6-9 Yrs.
10-11 Yrs. 12-13 Yrs. 14-15 Yrs_
Age Group
Sex differences in Delta ratio as a function of age.
.~~-
EEG SLEEP OF NORMAL CHILDREN
"
301
to-night differences were not observed for any of the non-REM sleep stages; indeed, these
measures were noted to be remarkably stable across the 3 study nights.
With regard to sleep pattern differences among children in this age range, in our study
as in others of similarly aged subjects, chronological age was demonstrated to be most
strongly related to the differences observed (9,11). Both the TRP and the TSA showed a
steady and significant decline with increasing age. Specific measures of sleep continuity
or sleep maintenance, however, remained constant yielding similar, high sleep efficiencies
of 94 to 95% for children in all age groups. The non-REM sleep measures were marked
by a gradual decline in slow-wave stage 4 sleep and an associated increase in stage 2 sleep
with increasing age, whereas no significant differences were noted for the REM sleep
measures across age groups. Rapid eye movement latency showed a gradual but nonsignificant decrease with increasing age as did the number of REM periods, and the percent
of REM sleep remained relatively constant reflecting mean values of 20 to 22% for children
in all age groups. As has been reported previously for both children and adults (1), REM
period lengths in our sample tended to increase across the night. Finally, with regard to
measures of REM activity within REM sleep, no significant differences were noted to be
present as a function of any of the independent variables examined, although a gradual
decline was noted across age groups for total REM activity, REM intensity, and REM
density. Furthermore, as the lengths of the REM periods increased across the night, so did
these measures of REM activity. It should be noted that for these measures, as well as for
REM latency, between-subject variability was relatively high, a finding that has been noted
previously in studies among children 00,11,13).
The effects of gender and pubertal status on the EEG sleep measures in our sample of
children were overwhelmingly nonsignificant. The only significant effect noted for pubertal
status was for the TRP, and in this instance the effect of age was clearly much stronger.
One could argue, quite legitimately, that our pubertal rankings were insufficiently sensitive
to register such changes. However, the studies of others including the work of Williams et
al. (24), Karacan et al. (25), and the more recent work of Carskadon and others (13,14)
have similarly failed to show strongly consistent and unique effects of puberty on nocturnal
sleep patterns despite more discrete pubertal groupings and longitudinal study designs. A
problem with our study, as with those that have been conducted by others, is the strong
correlation between age and pubertal status among normal, healthy children in this age
range. Given this apparent relationship, it may be that neither our present measures for
pubertal status nor our measures for the EEG sleep parameters are sensitive enough to
partition these effects. As for the effect of gender on the EEG sleep measures, only two
significant and related differences were demonstrated in our sample: boys 10 years of age
and above showed higher slow-wave stage 3 sleep percents, although not higher stage 4
sleep percents and, accordingly, higher Delta ratios as compared with their female counterparts. This finding was also reported in the work of Williams et al. (11), but only among
10-12 year olds and was coupled in their study with a tendency for boys at this and at
subsequent ages to show an increased number of stage shifts as compared with girls.
In summary, in our study among normal, healthy children chronological age appears to
exert the strongest effect on sleep pattern differences observed in this age range. Like other
investigators, we observed few gender differences and were unable to demonstrate significant
and unique effects attributable to pubertal development. Current knowledge on the neuroendocrinology of puberty would, in fact, suggest that such changes might indeed be quite
subtle as the onset of puberty is no longer viewed as an isolated event but rather as a critical
stage along a continuum of gonadal functioning (26). Although the major restraint on the
Sleep, Vol, 7, No.4, 1984
302
P. A. COBLE ET AL.
onset of puberty during childhood is believed to be exercised by the central nervous system,
the reversai of this restraint is believed to be a gradual process; one heralded by a sleepassociated augmentation in the secretion of leutinizing hormone but spanning a period of
several years thereafter until full sexual maturation and fertility have been achieved. Thus,
although the lack of demonstrable effects for pubertal change on EEG sleep may, in part,
be confounded by the gradual nature of these and related maturational changes, it may also
be that our current measurement techniques are simply insufficiently sensitive to quantify
such changes. In light of advances in our knowledge regarding both sleep and puberty and
in the technological tools with which to measure these parameters, applications of alternative
and more precise measurement methods would seem to offer great promise as a means of
further investigating the effects of these maturational changes in childhood. Hormonal
measurement techniques, for example, are presently available to more precisely determine
phases of pubertal development. Likewise, automated measurement methods such as those
developed in our laboratory for REM and Delta sleep activity offer greater precision in the
measurement of the EEG sleep states among children. Future studies utilizing such techniques might well shed more light on the hormonal/sleep relationships known to accompany
normal growth and pubertal development in childhood.
Acknowledgment: This manuscript was supported in part by the National Institutes of Health
grant numbers 30915 and 24652.
REFERENCES
1. Aserinsky E, Kleitman N. Regularly occurring periods of eye motility and concomitant phenomena during
sleep. Science 1953;118:273-4.
2. Dement W, Kleitman N. Cyclic variations in EEG during sleep and their relation to eye movements, body
motility, and dreaming. Electroencephalogr Clin Neurophysiol 1957;9:673-90.
3. Borbely AA. A two process model of sleep regulation. Hum NeurobioI1982;1:195-204.
4. Kleitman N. Basic rest-activity cycle-22 years later. Sleep 1982;5:311-7.
5. McGinty DJ, Drucker-Colin RR. Sleep mechanisms: biology and control of REM sleep. lnt Rev Neurobiol
1982;23:391-436.
6. Moore-Ede MC, Sulzman FM, Fuller CA. The clocks that time us: physiology of the circadian timing
system. Cambridge, MA: Harvard University Press, 1982:202-19.
7. Association of Sleep Disorders Centers. Diagnostic classification of sleep and arousal disorders. 1st ed.
Prepared by the Sleep Disorders Classification Committee, HP Roffwarg, Chairman. ~leep 1979;2: 1-137.
8. Roffwarg HP, Dement WC, Fisher C. Preliminary observations of the sleep-dream pattern in neonates,
infants, children and adults. In: Harms E, ed. Problems of sleep and dreams in children. New York:
Pergamon Press, 1964:60-72 (Monographs on child psychiatry, No 2).
9. Feinberg I, Koresko RL, Heller N. EEG sleep patterns as a function of normal and pathological aging in
man. J Psychiatr Res 1967;5:107-44.
10. Ross n, Agnew HW, Williams RL, Webb WB. Sleep patterns in preadolescent children: an EEG-EOG
study. Pediatrics 1968;42:324--35.
11. Williams RL, Karacan I, Hursch CJ. Electroencephalography (EEG) of human sleep: clinical applications.
New York: John Wiley and Sons, 1974.
12. Coble PA, Taska LS, Shaw DH, Kupfer DJ. The sleep of normal children: chronological age vs maturation.
Sleep Res 1981 ;10: 110.
13. Carskadon MA. The second decade. In: Guilleminault C, ed. Sleeping and waking disorders: indications
and techniques. Menlo Park, CA: Addison-Wesley Publishing, 1982:99-125.
14. Caraskadon MA, Orav JE, Dement WC. Evolution of sleep and daytime sleepiness in adolescents. In:
Guilleminault C and Lugaresi E, eds. Sleep/wake disorders: natural history, epidemiology, and longterm evolution. New York: Raven Press, 1983:201-16.
15. McPartland RJ, Kupfer DJ. Computerized measures of electrooculographic activity during sleep. lnt J
Biomed Comput 1978;9:409-19.
16. Kupfer DJ, Shaw DH, Ulrich RF, Coble PA, Spiker DG. Application of automated delta analysis. Sleep
Res 1981;10:173.
17. Rechtschaffen A, Kales A, eds. A manual of standardized terminology, techniques, and scoring system
for sleep stages of human subjects. Washington, DC: US Government Printing Office (Public Health
Service), 1968.
Sleep, Vol. 7, No.4, 1984
...
EEG SLEEP OF NORMAL CHILDREN
303
18. Kupfer DJ, Foster FG, Coble PA, McPartland RJ, Ulrich RF. The application of EEG sleep for the
differential diagnosis of affective disorders. Am J Psychiatry 1978;135:69-74.
19. Kupfer DJ, Coble PA, Kane J, Petti T, Conners KC. Imipramine and EEG sleep in children with depressive
symptoms. Psychopharmacology 1979;60:117-23.
20. Taska LS, Kupfer DJ. A system for the quantification of phasic ocular activity during REM sleep.
Sleepwatchers 1983;6:14-17.
2l. Tanner JM. Growth at adolescence. 2nd ed. Oxford: Blackwell, 1962.
22. Agnew HW, Webb WB, Williams RL. The first night effect: an EEG study of sleep. Psychophysiology
1966;2:263--6.
23. Benoit 0, Goldenberg-Leygonie F, LaCombe J, Marc ME. Sommeil de l'enfant presentant des manifestions
episodiques du sommeil: comparison avec I'enfant normal. Electroencephalogr Clin Neurophysiol 1978;
44:502-12.
24. Williams RL, Karacan I, Hurch C, Davis GE. Sleep patterns of pubertal males. Pediatr Res 1972;6:643.
25. Karacan I, Anch M, Thornby JK, Okawa M, Williams RL. Longitudinal sleep patterns during pubertal
growth: four-year follow-up. Pediatr Res 1975;9:842-6.
26. Grumbach MM. The neuroendocrinology of puberty. In: Krieger DT, Hughes JC, eds. Neuroendocrinology.
Sunderland, MA: Sinauer Associates. 1980:249-58.
Sleep, Vol. 7, No.4, 1984