Electroencephalography and clinical Neurophysiology 94 (1995) 6-11
Dynamics of EEG background activity level within quiet sleep
in successive cycles in infants
I. Fagioli a,* F. Bes b, p. Peirano c, p. Salzarulo d
a Dipartimento di Teoria, Storia e Ricerca Sociale, Universith di Trento, Via Verdi 26, 38100 Trento, Italy
b Laborfiir Klinische Psychophysiologie, Universitiitsklinikum RudolfVirchow, Freie UniversitiitBerlin, Berlin, FRG
c Instituto de Nutricion y Tecnologia de los Alimentos, Universitad de Chile, Santiago, Chile
d Dipartimento di Psicologia Generale, dei Processi di Sviluppo e di Socializzazione, Universitlz di Firenze, Florence, Italy
Accepted for publication: 8 August 1994
We investigated in infants the emergence of the trends of the EEG synchronization throughout quiet sleep (QS) as a function of the QS
rank. The night sleep of 3 groups with 6 subjects each (aged respectively 9-18 weeks, 21-47 weeks, and 16-45 years) was recorded. A
parameter value reflecting the degree of synchronization of the EEG background activity for successive epochs was computed by
automatic analysis. For each QS phase 3 indicators of the dynamics of the time course of the EEG parameter activity were determined:
the range (difference between the EEG parameter value at the beginning of the QS episode and that at the trough), the trough latency
(after QS onset), and the rate of synchronization (range/trough latency). The range and the trough latency increased with age, whereas
the rate of synchronization decreased. The range and the rate of synchronization decreased in the successive cycles, whereas the trough
latency increased. These results provide further support for the hypothesis of the early emergence of the process S mechanisms and
suggest that the framework of the 2-process model could account also for the development of both the EEG background activity dynamics
and the sleep-wake organization.
Keywords: Sleep; Infants; EEG; Automatic analysis; NREM-REM sleep cycle
The non-homogeneity of EEG activity during sleep has
been observed since the thirties, when Loomis et al. (1937)
described 5 stages in human adult sleep, and Blake and
Gerard (1937) reported the predominance of delta activity
at the beginning of sleep, followed by a decline during the
successive hours. More recently, within non-rapid eye
movement (NREM) sleep, 4 stages (stages 1, 2, 3 and 4:
Rechtschaffen and Kales, 1968) have been differentiated
principally on the basis of the amount of slow waves (less
than 2 Hz, more than 75/xV); this classification provided a
rough quantification of slow waves and enabled sleep
researchers to show a predominance of stages 3 and 4 in
the first half of the night. The peak of slow wave activity
takes place within the first hour after sleep onset. The
* Corresponding author. Tel.: (39 461) 881323; Fax; (39 461) 881440.
localization of the greater amount of slow waves at the
beginning of the sleep was confirmed by research carried
out using methods of automatic analysis (Feinberg et al.,
1977, 1978, 1980; Borb~ly et al., 1981; Haustein et al.,
1986; Achermann and Borbrly, 1987).
The dynamics (time course or trend) of slow wave
activity within each NREM period of the first 3 cycles of a
sleep episode have been already investigated in the adult
by Achermann and Borbrly (1987): these authors showed
that several indicators of the EEG synchronization during
NREM sleep (power density, rise rate, maximum rise and
maximum fall rate, but not rise time) decrease progres-
sively in successive cycles. To account for these results
concerning the dynamics of EEG slow wave activity within
NREM sleep periods, the first version of the 2-process
model (Borb~ly, 1982; Daan et al., 1984) was further
elaborated (Achermann and Borbrly, 1990). While the
basic assumptions were maintained (i.e., the sleep propen-
sity is determined by a homeostatic process S - which is
built up during the waking period and decays at a faster
0013-4694/95/$09.50 © 1995 Elsevier Science Ireland Ltd. All rights reserved
SSDI 0013-4694(94)00227-4 EEG 93728
L Fagioli et aL/Electroencephalography and clinical Neurophysiology 94 (1995) 6-11 7
rate during sleep - and by a circadian process C; the
interaction of S and C determines the timing of sleep and
waking), in the new version of the model the relationship
between process S and dynamic aspects of the EEG activ-
ity has been stressed (Achermann and Borb61y, 1990).
In infants, data on the organization of the EEG activity
throughout quiet sleep (QS, the usual denomination for
NREM sleep in infants) are fragmentary. Concerning the
dynamic aspects of the EEG activity within the QS episodes
in premature and full-term infants, Monod and Pajot (1965)
reported that the periods of trac~ lent usually precede the
more lasting periods of tracd alternant. Differences in
delta waves evaluated right at the beginning and at the end
of a QS phase during the first few months of life were
reported by Paul et al. (11973). Intra-state EEG activity
organization in infants ihas recently been investigated
(Salzarulo et al., 1991), using the automatic analysis
method developed by Ha~astein et al. (1986) and adapted
for infant studies by Beset al. (1988); different qualitative
patterns of the time course of the degree of synchroniza-
tion of the EEG background activity were observed at
Elsewhere we have put: forward the hypothesis that the
differences in physiological activities between the first and
the following cycles found at early ages could be a
"sketch" of the Process S described in the adults (Salzarulo
and Fagioli, 1992a; Beset al., 1994). The aim of the
research described here was to substantiate this hypothesis,
and it attempted, as a further step, to quantify the dynam-
ics of the EEG synchronization within the NREM sleep
phases of the first 3 consecutive NREM-REM cycles
within the first uninterrupted nocturnal sleep episode. Three
main indicators of the dynamics of the synchronization of
the EEG activity (the rang,e between the extreme values of
EEG synchronization and desynchronization, the latency
of the maximum of synchronization value, and the rate of
synchronization) were therefore selected and analyzed in
human subjects of different ages.
2. Methods and materials
Twelve healthy, drug-free infants, products of full-term
pregnancies and of uncomplicated deliveries, free from
peri- and post-natal pathology, were continuously moni-
tored by polygraphic recording and behavioral observation
over a 12 h period commencing at 8 p.m. in a dimly lit
sound-attenuated individual room 'of a pediatric hospital
ward (INSERM, Paris, P. Salzarulo), to which they had
been adapted for at least 4 days. They were subdivided in
2 groups of 6 subjects each according to age: young
infants (9-18 weeks old; mean and S.D.: 14.00 + 3.46),
and older infants (21-47 weeks old: 35.67 + 11.15). The
lower limit of age was :set at 9 weeks, given the rare
occurrence of at least 3 consecutive NREM-REM cycles
within an uninterrupted sleep episode at earlier ages. All
the infants had one (or more) nap(s) during the daytime;
since the recordings started after the meal and the toilet of
the evening, a waking period of at least 90 min preceded
the sleep onset; a period of REM sleep preceded the first
QS phase in all the infants' recordings at sleep onset.
Six healthy, drug-free adults (16-45 years old; 28.00 _
11.14), habitual non-nappers and previously adapted to the
laboratory setting, were submitted to polygraphic sleep
recording between 11 p.m. and 7 a.m at the Max-Pank-In-
stitut f'fir Psychiatrie in Munich (recordings provided by
courtesy of Dr. H. Schulz).
The polygraphic data were simultaneously recorded on
paper and on magnetic tape, they included bipolar EEG
(electrode positions being according to the international
10-20 system, leads were Fp1-C3, C3-O1, Fp2-C4, C4-O2;
time constant = 0.3 sec), vertical and horizontal electro-oc-
ulograms, chin electromyogram, electrocardiogram, and
respiratory rhythm. The recording techniques and scoring
procedures for sleep states in infants (see details in Fagioli
and Salzarulo, 1982) allowed us to define 4 states by
combining behavioral and electrophysiological data: QS
(i.e., NREM sleep): eyes closed, no eye movements, di-
minished body movements, EEG with slow waves (and
spindles) or trac~ alternant, regular respiration; paradoxi-
cal sleep (PS, i.e. REM sleep): eyes closed (or alternately
half open and closed), eye movements and, additionally,
body or limb movements, low voltage EEG, chin muscular
atonia, irregular respiration; ambiguous sleep: including
characteristics of both QS and PS; waldng: eyes open, eye
movements, irregular respiration and, in addition, body
movements. The recordings of the adults were visually
scored according to the manual by Rechtschaffen and
The C3-O1 lead was submitted to the method of auto-
matic analysis developed by Haustein et al. (1986), after
pre-filtering with a 30 Hz low pass filter and digitizing at a
sampling rate of 100 Hz. By non-recursive digital filtering,
the signal of every 10 sec epoch was divided into 6
frequency bands with equal relative bandwidths and, for
each frequency band, the peak values, defined as the local
maxima and minima of amplitude, were classified into 8
amplitude ranges: this procedure yielded a 6 × 8 matrix.
The matrix was then reduced to 6 values by calculating the
logarithmic measure of the mean peak value for each
frequency band. The application of a constant weighting
vector (previously obtained by averaging the individual
vectors of 12 subjects, see Haustein et al., 1986) allowed
us to further reduce each 6-value matrix to a unique
parameter value and to optimize the variation between the
values representing the two extremes in EEG activity, i.e.,
maximum of synchronization (high amplitude low fre-
quency) and maximum of desynchronization (low ampli-
tude high frequency). Finally, each 30 see a value of the
8 L Fagioli et al./Electroencephalography and clinical Neurophysiology 94 (1995) 6-11
EEG parameter was computed as the average of at least
two 10 sec parameter values, obtained from artifact-free 10
sec epochs out of 3 consecutive 10 sec epochs. The value
of the EEG parameter varies between 10 (corresponding to
the maximum EEG desynchronization) and 0 (correspond-
ing to the maximum EEG synchronization); its time course
gives a continuous representation of the EEG background
activity level, showing altemance between episodes of
EEG synchronization (low EEG parameter values) and
desynchronization (high EEG parameter values; see Fig.
1). Further details on the method used are reported in
Haustein et al. (1986) and Beset al. (1988); for the
procedure for the elimination of the artifacts see Baroncini
et al. (1986).
Two successive smoothings of the curve were obtained
by computing a 5-point moving average of the original
value on the whole night recording. Then, a separate
analysis of the EEG parameter dynamics was carried out
for each QS phase of each of the first 3 NREM-REM
cycles. To determine the trough (maximum of synchro-
nization) in the time course of the EEG parameter in each
cycle, the same criteria as Achermann and Borb61y (1987)
were applied: whenever there were multiple troughs within
a QS episode, the first trough was selected, unless its
decrease represented less than 90% of the decrease of the
subsequent trough (see Fig. 1). Then, the following main
indicators of the dynamics of the EEG synchronization
were determined: the range (difference between the EEG
parameter value at the beginning of the QS episode and
that at the trough), the trough latency (interval between
QS onset and trough), and the rate of synchronization
(range/trough latency) (Fig. 1).
Data concerning these indicators were submitted to a
2-way ANOVA for 1 factor between subjects (age: young
infants, older infants and adults) and 1 factor within sub-
jects (rank of the cycles: 1st, 2nd and 3rd).
The range increased significantly with age (F2, 15 =
5.372, P < 0.05). The range was significantly different in
the 3 cycles (F2, 30 = 3.541, P < 0.05), without significant
interaction between the two factors (Fa. 30 = 0.820, n.s.). A
posteriori analysis showed significant differences between
the first and the third cycle in young infants and between
the second and the third cycle in adults (see Fig. 2a).
The trough latency increased significantly with age (F2,
15 = 14.675, P < 0.001). Moreover, it was significantly
different in the 3 cycles (Fz. 30 = 10.292, P < 0.001); the
interaction was not significant (F4. 30 = 1.786, n.s.). A
posteriori analysis showed that in young infants the differ-
ence between the first and the second cycle just failed to
reach statistical significance; in older infants it was shorter
in the first cycle than it was in the third; in adults it was
shorter in the first and in the second cycles with respect to
the third one (see Fig. 2b).
The rate of synchronization showed a tendency to
decrease with age, just failing to reach statistical signifi-
cance (F2, 15 = 3.618, P < 0.06, n.s.); moreover, the rate
.,'~./\. "'v,'-d'. :. ,/'w-.
: "" " ~ LL,---./~.'.~'V'" i "i
,.--'r'., r -..'v ..... '- ........
: ~ . . . . . . .
,"' "-' ', ............
2 3 4
I l~t I
Fig. 1. EEG parameter time course (upper panel) and sleep stages (lower panel) during the first 5 h of recording in a 21-week-old infant. X-axis: time of
recording 0onger vertical lines delimit 1 h periods, shorter vertical lines delimit 10 min periods). Upper panel y-axis: arbitrary EEG parameter units;
pr = EEG parameter range; tl = trough latency (both measures are represented by dashed lines). Lower panel y-axis: sleep stages, W= waking;
a = ambiguous sleep; R = REM sleep; 1, 2, 3 and 4 = stages 1, 2, 3 and 4 of NREM sleep. The timing of the 3 NREM-REM cycles is indicated by a
horizontal bar at the bottom; the 3 cycles are delineated by vertical lines.
I. Fagioli et al. / Electroencephalography and clinical Neurophysiology 94 (1995) 6-11 9
• 1st cycle
 2nd cycle
 3rd cycle
• 1st cycle
 2rid cycle
 3rcl cycle
rate of synchronization
I st cycle
 2nd cycle
 3rd cycle
Fig. 2. EEG parameter range (a), trough latency (b) and rate of synchro-
nization (c) in young infants, older infants and adults. The 3 adjacent bars
represent the mean (plus standard deviation) of the first 3 quiet sleep
(NREM) episodes. The lines below the bars indicate significant differ-
ences between cycles (* P < 005; * * P < 0.01).
of synchronization was significantly different in the 3
cycles (F2, 30 = 20.848, P < 0.001), again without signifi-
cant interaction (F4, 30 = 1.903, n.s.). A posteriori analysis
showed that in both young and older infants the rate of
synchronization was faster in the first than both the second
and the third cycles, and in the adults it was slower in the
third cycle than in both the first and the second ones (see
As far as the general trend of the dynamics of the EEG
synchronization with age is concerned, it should be stressed
that the range of the EEG parameter and the trough latency
are subject to clear modifications as a function of age; both
indicators increase with age for the first 3 cycles of the
night sleep. According to the 2-process model (Borb61y,
1982; Daan et al., 1984), the greater synchronization of
EEG activity at sleep onset is the consequence of a higher
level of process S; the increase with age in the EEG
parameter range (i.e., of the EEG synchronization) could
therefore be interpreted as an increase in the level of the
homeostatic process S at sleep onset, possibly due to the
lengthening of the duration of prior waking. Another ex-
planation, not necessarily conflicting with the previous
one, could take into account a maturational factor, the
increase with age of the synaptic connections in the cere-
bral cortex (Feinberg et al., 1977), which could lead to an
increase in slow wave activity. The increase with age of
the trough latency parallels the increase of the duration of
the QS phase, independently of age and rank of the cycle
(as shown in infants by Fagioli et al., 1990; see also
Peirano et al., 1993).
Given that the increase with age of the trough latency
outweighs that of the EEG parameter range, it follows that
the rate of synchronization shows a decreasing trend with
age. Since the model based on data obtained in the adult
predicts an increase in both the level of slow wave activity
and the rate of synchronization when process S increases
(Borb61y and Achermann, 1992), the conflicting evidence
of an increase in the EEG parameter range and of a
decrease in the rate of synchronization suggests that the
age-related development of the intra-state temporal trend
of the EEG activity cannot be interpreted merely as an
increase in process S; this issue should be investigated in
adequately designed studies in the future.
The rank effect has not been observed consistently in
all the age groups for all the indicators. Whereas in the
adult the EEG parameter range and the rate of synchro-
nization show a clear tendency to decrease across the first
3 cycles, and the trough latency shows a trend to increase,
in infants the same trends are observed only between the
first and the second cycle. A clear example of this phe-
nomenon is represented in Fig. 2c for the rate of synchro-
nization, which shows a progressive decrease in adults and
a stabilization from the second cycle onwards in infants; a
similar picture at different ages also clearly emerges for
the EEG parameter range and the trough latency (see Fig.
2a and b). The results of the a posteriori statistical analysis
support the previous statements: in adults, the rate of
synchronization of the third cycle is lower than that of the
preceding cycles, while in both groups of infants only the
differences between the first and the following cycles are
significant. These results suggest that the trends of the
dynamics of the EEG background activity across the cy-
L Fagioli et al. / Electroencephalography and clinical Neurophysiology 94 (1995) 6-11
cles observed in the adults are already sketched in young
infants; however, in infants, the trends are confined to the
first two cycles, confirming those of other physiological
activities, such as respiratory rate (Nogues et al., 1992),
heart rate (Nogues et al., 1990; Salzarulo et al., 1991;
Fagioli et al., 1994): for general comments see also
Salzarulo and Fagioli (1992a). In older infants the trends
of the values of EEG parameter range and trough latency
from the second to the third cycle are reversed compared
with those of young infants; this pattern confirms the
peculiar sleep organization from the fifth month onwards,
first observed by visual analysis (Schulz et al., 1989; Bes
et al., 1991), and consisting in the alternate presence of
slow wave sleep in successive QS phases. All these data
corroborate the hypothesis (Salzarulo and Fagioli, 1992a;
Beset al., 1994) that the development with age of the
dynamics of EEG activity is not straightforward.
In the adult, the amount of slow wave activity in the
first cycle has been shown to be influenced by the dura-
tion of theowaking prior to the sleep episode (Borb61y et
al., 1981; Akerstedt and Gillberg, 1986; Dijk et al., 1987,
1990), while its night trend (time course) is characterized
by an exponential fall across sleep cycles (Dijk et al.,
1991). From the developmental perspective, it is well
known that the waking preceding the sleep episode is
shorter in infants than in both children and adults (cf.,
Fagioli and Salzarulo, 1982; Coons and Guilleminault,
1984; Salzarulo and Fagioli, 1992b); this factor could be
taken into account to explain the developmental trend of
the EEG dynamics. According to recent elaborations of the
2-process model (Achermann and Borb61y, 1990; Acher-
mann et al., 1990; Borb61y and Achermann, 1992) both the
global declining trend of slow wave activity over consecu-
tive cycles and the dynamics of the EEG activity within
NREM episodes depend on the duration of the prior
However, we suggest that it is unlikely that the duration
of prior waking could be the only factor accounting for the
main difference between the pattern of the adults and that
of the infants. A predominant role of prior waking duration
should entail two consequences with growth: (a) an in-
crease in the difference between the first and the second
cycle, due to the exponential decline of process S (and of
the slow wave activity) across sleep cycles, and (b) an
increase in the rate of synchronization. However, these
predictions were not confirmed by the results of this study.
In fact, (a) none of the interactions between the factors age
and cycles for any indicators was statistically significant,
and (b) the rate of synchronization showed an opposite
trend (i.e., to decrease) with age. Therefore, the role of
prior waking would be obvious only "within" the same
group of age, in particular adults (see for humans: Borb61y
et al., 1981; Achermann and Borb61y, 1990; for rats:
Borb61y et al., 1984; Tobler and Borb61y, 1986), but not
"between" different age groups. To ascertain the role of
the duration of the prior waking on the difference between
the first and the second sleep cycle in infants, further
investigations are necessary.
The framework of the 2-process model allows us to put
forward some hypotheses which concern the alternance in
infants of several waking and sleep episodes in the 24 h
period (polyphasic rhythm), and which point to an explana-
tion of the transition to the adult's monophasic rhythm.
The "instantaneous build-up rate" or "rise rate" during
waking and the "instantaneous breakdown rate" or "de-
cay rate" during sleep (respectively in Daan et al., 1984,
and Achermann and Borb61y, 1990) are greater in infants
and decrease with age (as can be argued from Beset al.,
1994); the maximum amount of process S bearable during
waking, without falling asleep (i.e., the threshold H level:
Daan et al., 1984), increases with age.
According to both these hypotheses, in infants, with
respect to adults, the increase in a shorter time of the
process S during waking and the lower level of the thresh-
old H should entail earlier sleep onset, i.e., shorter waking
episodes; moreover, the decrease in a shorter time of the
process S during sleep should entail shorter sleep episodes:
as a consequence, several sleep and waking episodes alter-
nate in the 24 h period. Thus, these hypotheses could
account for why we observe a polyphasic sleep organiza-
tion in infants and a monophasic one in adults. The sleep
fragmentation in condition of continuous bed rest has been
explanined by a decrease of threshold H level by Daan et
al. (1984). Finally, we propose an additional explanation
(again within the framework of the 2-process model),
which is consistent with the previous ones, for the transi-
tion from the infants' polyphasic to the adults' monophasic
sleep-wake organization: the process C fluctuations might
be less ample in infants than in adults, as a consequence of
the lower amplitude of the circadian rhythms, a factor
which has been demonstrated to facilitate the re-establish-
ment of a polyphasic organization in the elderly (Weitz-
man et al., 1982).
The present results encourage further research, aiming
to explain the changes during development of the EEG
dynamics during sleep and taking into account also the
anatomical (synaptic density) and/or physiological
(cerebral metabolic rate) changes with age, in the way
pioneered by Feinberg et al. (1990). In conclusion, these
results provide further support for the hypothesis of the
early emergence of the process S mechanisms and suggest
that the framework of the 2-process model could account
also for the development of both the EEG background
activity dynamics and the sleep-wake organization.
We gratefully acknowledge Dr. Adrian Belton for revis-
ing the English manuscript.
I. Fagioli et al. / Electroencephalography and clinical Neurophysiology 94 (1995) 6-11
Achermann, P. and Borb61y, A.A. Dynamics of EEG slow wave activity
during physiological sleep and after administration of benzodi-
azepines hypnotics. Hum. Neurobiol., 1987, 6: 203-210.
Achermann, P. and Borb61y, A.A. Simulation of human sleep: ultradian
dynamics of electroencephalographic slow-wave activity. J. Biol.
Rhythms, 1990, 5: 141-157.
Achermann, P., Beersma, D.G.M. and Borb61y, A.A. The two process
model: ultradian dynamics of sleep. In: J.A. Home (Ed.), Sleep '90.
Pontenagel Press, Bochum, 1990: 296-300.
erstedt, T. and Gillberg, M. Sleep duration and the power spectral
density of the EEG. Electroenceph. clin. Neurophysiol., 1986, 64:
Baroncini, P., Bes, F., Franc, B. and Salzarulo, P. A semi-automatic
method for elimination of the artifacts from the EEG night recordings
in babies. In: 8th European Congress of Sleep Research, Szeged,
Hungary, September 1-5, 1986: 21.
Bes, F., Baroncini, P., Dugovic, C., Fagioli, I., Schulz, H., Franc, B. and
Salzarulo, P. Time course of EEG activity during night sleep in the
first year of life based on automatic analysis. Electroenceph. clin.
Neurophysiol., 1988, 69: 501-507.
Bes, F., Schulz, H., Navelet, Y. and Salzarulo, P. The distribution of
slow-wave sleep across the night: a comparison for infants, children
and adults. Sleep, 1991, 14: 5-12.
Bes, F., Fagioli, I., Peirano, P., Schulz, H. and Salzarulo, P. Trends in
EEG synchronization across non-REM sleep in infants. Sleep, 1994,
Blake, H. and Gerard, R.W. Brain potentials during sleep. Am. J.
Physiol., 1937, 119: 692-703.
Borb61y, A.A. A two process model of sleep regulation. Hum. Neurobiol.,
1982, 1: 195-204.
Borb61y, A.A. and Achermann, P., Concepts and models of sleep regula-
tion: an overview. J. Sleep Res., 1992, 1: 63-79.
Borb61y, A.A., Baumann, F., Brandeis, D., Strauch, I. and Lehmann, D.
Sleep deprivation: effect on sleep stages and EEG power density in
man. Electroenceph. clin. Neurophysiol., 1981, 51: 483-493.
Borb61y, A.A., Tobler, I. and Hanagasioglu, M. Effect of sleep depriva-
tion on sleep and EEG power spectra in the rat. Behav. Brain Res.,
1984, 14: 171-182.
Coons, S. and Guilleminault, C. Development of consolidated sleep and
wakeful periods in relation to day/night cycle in infancy. Dev. Med.
Child Neurol., 1984, 26: 169-176.
Daan, S., Bersma, D.G.M. and Borb61y, A.A. Timing of human sleep:
recovery process gated by ~t circadian pacemaker. Am. J. Physiol.,
1984, 246: R161-R178.
Dijk, D.J., Beersma, D.G.M. and Daan, S. EEG power density during nap
sleep: reflection of an hourglass measuring the duration of prior
wakefulness. J. Biol. Rhythms, 1987, 2: 207-219.
Dijk, D.J., Brunner, D.P., Beersraa, D.G.M. and Borb61y, A.A. Electroen-
cephalogram power density and slow wave sleep as a function of
prior waking and circadian phase. Sleep, 1990, 13: 430-440.
Dijk, D.J., Cajochen, C., Toblen I. and Borb61y, A.A. Sleep extention in
humans: sleep stages, EEG power spectra and body temperature.
Sleep, 1991, 14: 294-306.
Fagioli, I. and Saizarulo, P. Sleep states development in the first year of
life assessed through 24-hour recordings. Early Hum. Dev., 1982, 6:
Fagioli, I., Bes, F., Peirano, P. and Salzarulo, P. Dynamics of EEG
background activity level within quiet sleep in successive cycles in
infants. In: 10th Congress of the European Sleep Research Society,
Strasbourg, May 20-25, 1990: 224.
Fagioli, I., Peirano, P. and Salzarulo, P. Heart rate and sleep states in
infants. Rev. EEG Neurophysiol. Clin., 1994, 24: 45-50.
Feinberg, I., Hibi, S. and Carlson, V.R. Changes in EEG amplitude
during sleep with age. In: K. Nandy and I. Sherwin (Eds.), The Aging
Brain and Senile Dementia. Plenum Press, New York, 1977: 85-98.
Feinberg, I., March, J.D., Fein, U., Hoyd, T.C., Walker, J.M. and Price,
L. Period and amplitude analysis of 0.5-3 c/sec activity in NREM
sleep of young adults. Electroenceph. clin. Neurophysiol., 1978, 44:
Feinberg, I., Fein, O. and Floyd, T.C. Period and amplitude analysis of
NREM EEG sleep: repeatibility of results in young adults. Electroen-
ceph. clin. Neurophysiol., 1980, 48: 212-221.
Feinberg, I., Thode, H.C., Chugani, H.T. and March, J.D. Gamma
distribution model describes maturational curves for delta wave am-
plitude, cortical metabolic rate and synaptic density. J. Theor. Biol.,
1990, 142: 149-161.
Haustein, W., Pilcher, J., Klink, J. and Schulz, H. Automatic analysis
overcomes limitations of sleep stage scoring. Electroenceph. clin.
Neurophysiol., 1986, 64: 364-374.
Loomis, A.L., Harvey, E.N. and Hobart, G.A. Cerebral states during
sleep, as studied by human brain potentials. J. Exp. Psychol., 1937,
Monod, N. and Pajot, N. Le sommeil du nouveau-n6 et du prematur6.
Biol. Neonat., 1965, 8: 281-307.
Nogues, B., Louvet, S. and Vecchierini-Blineau, M.F. Respiratory and
cardiac rates in 35 normal, 2-month-old, infants during daytime and
nighttime recordings. In: 10th Congress of the European Sleep Re-
search Society, Strasbourg, May 20-25, 1990: 227.
Nogues, B., Vecchierini-Blineau, M.F. and Louvet, S. Variation de la
fr6quence respiratoire au cours du sommeil de jour et de nuit chez 35
nourrissons normaux ~g6s de 2 mois. Rev. EEG Neurophysiol. Clin.,
1992, 22: 167-177.
Paul, K., Dittrichova, J. and Pavlikova, E. The course of quiet sleep in
infants. Biol. Neonat., 1973, 23: 78-89.
Peirano, P., Fagioli, I., Bes, F. and Salzarulo, P. The role of slow wave
sleep on quiet sleep duration in infants. J. Sleep Res., 1993, 2:
Rechtschaffen, A. and Kales, A. A Manual of Standardized Terminology,
Techniques and Scoring System for Sleep Stages of Human Subjects.
U.S. Government Printing Office, Washington, DC, 1968.
Salzarulo, P and Fagioli, I. Post-natal development of sleep organization
in humans: speculations on the emergence of the "S process." Rev.
EEG Neurophysiol. Clin., 1992a, 22: 107-115.
Salzarulo, P. and Fagioli, I. Sleep-wake rhythms and sleep structure in the
first year of life. In: C. Stampi (Ed.), Why We Nap. Birkhauser,
Boston, MA, 1992b: 50-57.
Salzarulo, P., Fagioli, I., Peirano, P., Bes, F. and Schulz, H. Levels of
EEG background activity and sleep states in the first year of life. In:
G.M. Terzano, P.L. Halasz and A.C. Declerek (Eds.), Phasic Events
and Dynamic Organization of Sleep. Raven Press, New York, 1991:
Schulz, H., Bes, F. and Salzarulo, P. Rhythmicity of slow wave sleep:
developmental aspects. In: A. Wauquier, C. Dugovic and M. Radulo-
vacki (Eds.), Slow Wave Sleep: Physiological, Pathophysiological and
Functional Aspects. Raven Press, New York, 1989: 49-60.
Tobler, I. and Borb61y, A.A. Sleep EEG in the rat as a function of prior
waking. Electroenceph. clin. Neurophysiol., 1986, 64: 74-76.
Weitzman, E.D., Moline, M.L., Czeisler, C.A. and Zimmerman, J.C.
Chronobiology of aging: temperature, sleep-wake rhythms and en-
trainment. Neurobiol. Aging, 1982, 3: 299-309.