Ontogenesis of nocturnal organization of sleep spindles: a longitudinal study during the first 6 months of life

Electroencephalography and clinical Neurophysiology, 83 (1992) 289-296

289

© 1992 Elsevier Scientific Publishers Ireland, Ltd. 0013-4649/92/$05.00

EEG91607

Ontogenesis of nocturnal organization of sleep spindles: a longitudinal study during the first 6 months of life * J. Louis, J.X. Zhang a, M. Revol, G.

D e b i l l y a n d M.J. C h a l l a m e l

INSERM U. 52, Service d'Explorations Fonctionnelles Neurologiques, Pay. 4E, Centre Hospitalier Lyon-Sud, 69310 Pierre B~;nite (France), and u Universit~ d'Anhui, Anhui (People's Rep. of China) (Accepted for publication: 17 June 1992)

Summary Ontogenesis of sleep spindles was studied on overnight longitudinal recordings in 12 full-term infants at 1.5-3-4.5 and 6 months of life. Six parameters (density, duration, frequency, amplitude, asymmetry and asynchrony) were analyzed during both slow wave sleep (SII and delta) and during 5 periods of the night. Results show a significant increase of most parameters between 1.5 and 3 months of age. All spindle patterns developed quite rapidly during the first 3 months of infancy, possibly reflecting developmental changes in thalamo-cortical structures and maturation of the physiological system that produces spindles. The density of 12-14 Hz spindle frequency was higher in stage II when compared to stage delta, as in adults. Our data confirm previous reports on spindle ontogenesis and give a more complete aspect of this ontogenesis in relation to sleep development. Three months of age appeared to be a turning point in maturational processes and might reflect changes in central nervous system activity and behavior which take place during that period. Sleep spindle evolution seems to be an accurate reflection of the slow wave sleep (SWS) development, and our results are discussed in terms of the developmental aspect of SWS production and characterization of sleep stages in young infants. Concordance between quantitative aspects and nocturnal organization leads us to consider that the individualization of slow wave sleep (SWS) in infants occurs from 4.5 months of life. Key words: Human infants; Sleep development; Sleep spindles; Ontogenesis

A sleep spindle has been defined as an "11-15 Hz burst" but mostly at 12-14 Hz, occurring during sleep and recorded on the central regions of the scalp (international EEG glossary, Chatrian et al. 1974). However, not all spindles are sleep spindles. The reason for this distinction is important when considering sleep spindle development: "spindle bursts" occur abundantly in the EEGs of premature infants (Dreyfus-Brisac and Curzi-Dascalova 1975; Ellingson 1979); these spindle bursts were called "brushes" (Watanabe and Iwase 1972; Lombroso 1975) and in the past they had frequently been taken for sleep spindles. In fact, infant "real sleep spindles" appear around the age of 8 weeks post term (Metcalf 1969, 1970; Lenard 1970; Hagne 1972; Tanguay et al. 1975; Curzi-Dascalova 1977; Wu et al. 1980; Ellingson 1982; Jankel 1985). Infant spindles were considered to be an indication

Correspondence to: Jacqueline Louis, Service d'Explorations Fonctionnelles Neurologiques, Pay. 4E, Centre Hospitalier Lyon-Sud, 69310 Pierre B6nite (France). T61.: Int + 78.86.17.90. * This work was supported by grants from INSERM U.52, CNRS URA 1195 and DRET 89203.

of normal sleep to such a degree that their absence or abnormality was considered as strongly suggestive of a cerebral dysfunction or pathology (Monod and Ducas 1967; Dreyfus-Brisac and Curzi-Dascalova 1975; Shibagaki et al. 1982). Spindle development during the first year of life might thus serve as an indicator of normal or abnormal central nervous system function, and also for some authors as an electrophysiological measurement of brain maturation (Schultz et al. 1968). However, spindle ontogenesis studies have not often been carried out on longitudinal data, and the recording sessions were often of short duration. Moreover, spindles are one of the most characteristic features of sleep and a comparative analysis of the ontogenesis of sleep and sleep spindles could be relevant. To have a better understanding of the development of sleep spindles during the first 6 months of life, and to define their normality in relation to sleep stages, we studied the nocturnal quantitative organization of sleep spindles during all-night polygraphic recording in normal infants. These polygraphic records were previously used for the study of the development of sleep during the first 6 months of life (Challamel et al. 1988). Thus, our study represents the first overnight longitudinal study of spindle development.

290

J. LOUIS ET AL.

Methods

Subjects Subjects were 12 full-term healthy infants (6 boys, 6 girls) selected according to the following criteria: negative family history of neurological disease; normal pregnancy and birth; mean gestational age (GA) > 37 weeks; birth weight 2800-3800 g; Apgar at 5 rain, 9-10; negative postnatal physical and neurological examination; no medication. Informed consent was obtained from each family for recording.

Recording Forty-four overnight polygraphic records were made longitudinally at around the age of 1 . 5 - 3 - 4 . 5 - 6 months from 21.00 h to 07.00 h. Records include 7 E E G channels, 1 electro-oculogram (EOG), 1 submental electromyogram (EMG), 4 respirograms and 1 electrocardiogram (Challamel et al. 1988). The E E G was recorded with the montage: Fpl-C3, C3-T3, T3-O1,

Fp2-C4, C4-T4 , T4-O2, Cz-Pz, of the international 10-20 system. Paper speed was 1 c m / s e c and amplitude 5 mm = 50 ~V. The time constant for E E G was 0.3 sec.

Data analysis Sleep staging.

Each epoch of 30 sec was visually scored for sleep states and stages. For the 1.5-monthold infants the records were scored according to the criteria of Anders et al. (1971) in indeterminate, quiet and active sleep. The usual modified Rechtschaffen and Kales (1968) criteria were used in older infants, and in this study, records were scored in stage II (SII) and stage delta corresponding to stages III and IV taken together. Sleep spindles. Five consecutive minutes were analyzed during non-consecutive stages II and delta of sleep over 2 h periods (21.00-23.00 h, 23.00-01.00 h, 01.00-03.00 h, 03.00-05.00 h, 05.00-07.00 h) in all

s0 vT_ lsec FPI-C3 c3-o

FP2- C4 C4-O2 Cz- Pz MO ,

EMG ECG

,4.4 4~4 A44~.14

.4

~. ,44.

lllljl

II

'

Lji]]]]

!

p

illllllllll]lI]

RR 1,5 Month

A

~

6 Months

:lO.uV

iOH'z

B 4H'z

Fig. 1. Example of sleep spindles in EEG recording sessions at 1.5 and 6 months of age. Enlarged spindles are shown in A and B.

l 111

SLEEP SPINDLE ONTOGENES1S

291

records at successive ages. For all 2 h periods, spindles were analyzed during the 2 stages of sleep (SII and delta). The 5 consecutive minutes were selected from the middle of the stages in the most stable place. Spindle definition was as follows: duration _>0.5 sec, amplitude > 10 /xV, frequency 10-15 Hz. They were quantified by two observers according to the following criteria: Location: spindles were measured at the loci of their maximum amplitude; however, a precise study of the location was not done. Density: mean number of spindles per minute. Spindle frequency: spindles were classified in 3 ranges: 10-12 Hz, 12-14 Hz and 14-15 Hz. Duration: in seconds. Amplitude: 4 ranges: 10-25 /xV, 25-50 /xV, 50-75 /xV and 75-100/xV. Asymmetry: we considered a burst as asymmetrical when there was a voltage ratio greater than 2:1, between hemispheres. Asynchrony: when there was, on each hemisphere, an interval of more than 2 sec between the middles of the two bursts; or isolated burst. Some of these criteria are illustrated in Figs. 1 and 2. The 6 parameters: density, frequency, duration, amplitude, asymmetry and asynchrony are expressed for the two stages of sleep (II and delta) during the 5

nocturnal periods for all records at the different ages. For each infant the percentages of asymmetry and asynchrony were calculated considering the total spindle burst amount, and the percentages of frequency and amplitude were evaluated by totaling all the same frequency or same amplitude spindles and relating them to the number of all spindles. Data were analyzed in terms of age for the 6 parameters in the following 3 conditions: global evolution, evolution in relation to the two stages of sleep, and evolution in relation to the different hours in the night. Considering the density and duration, the mean value for each infant was calculated and a Student's t test was applied to compare these parameters for each age. In order to analyze percentage of spindles according to age, frequency, amplitude, asymmetry and asynchrony, an ANOVA was performed. A multiple range test was used to study interaction of age with the other parameters. The level of significance corresponds to the plot of interaction for frequency, amplitude, asymmetry or asynchrony by ages.

Results One hundred and twenty periods of 5 min and 12,488 spindles were analyzed. Only 2 infants had no spindles at about 1 month and 15 days of age. These 2

4,5 Months

6 Months 50 IJV] , I s e c FPI- C3

FP2 - C4 I

C4-02

~

i

C z - Pz

ASYMMETRY

ASYNCHRONY

Fig. 2. Example of spindle asymmetryand asynchronybetween the two hemispheres in EEG recording sessions at 4.5 and 6 months of age.

292

J. LOUIS E T AL.

DENSITY

DURATION

second 4

Nb/min 6"

3

2.

1'

1:5

3

4:5

d

0

1.5

3

months

4.5

6

months

Fig. 3. Evolution of sleep spindle density and duration at 1.5-3-4.5 and 6 months of age for each infant. Sleep spindle density and duration are the total mean n u m b e r of spindle bursts per minute for the two sleep stages (SII and delta). * * P = 0.0076 using the Students t test (see text).

infants were aged 1 month and 6 days (GA = 40 weeks) and 1 month and 9 days (GA = 38 weeks). For all infants spindles were prominent in the fronto-central area.

96 lOO

of sleep spindle density for each infant. This density increased rapidly and significantly up to 3 months of age (2.1/min + 1.3 at 1.5 months versus 3.7/rain + 2.2 at 3 months, P = 0.0076) and did not much vary after. However, the inter-individual variations were large. Duration. The longest sleep spindles appeared at 1.5 and 3 months (mean = 3 sec + 0.7; Fig. 3). At this age, bursts lasted often more than 6 sec and sometimes up to 20 sec. After the age of 3 months, a downward trend was observed. Frequency. Fig. 4 shows the age-related evolution. While the percentage of the 10-12 Hz frequencies decreased with age, the 12-14 Hz frequency percentage reached 86.4% of the total spindle burst amount at the age of 6 months. The fast frequencies (14-15 Hz) appeared at 4.5 months of age (1.6%) and increased later. At 1.5 months of age, the 10-12 and 12-14 Hz frequencies displayed an equal distribution (45.6 and 53.9%). The main finding was the significant decrease ( P < 0.03 for interaction for frequency by age) of the 10-12 Hz frequencies between 1.5 months and the other ages. Amplitude (Fig. 4). Sleep spindle amplitude increased with age. The 4 amplitude samples (10-25/zV,

amplitude

100

10 - 25 ~tV

\

8o

12 - 14Hz~

Global evolution of spindles with age Density. Fig. 3 indicates the age-related incidence

96

frequency

/

\

\ \/

*

60

150 I~V 4o 20 10- 12Hz~" o

1.5

96

--<....

14- 15 HZ 3

4.5

i 1.5

6M

asynchrony

96

J 3

i 4.5

J 6M

asymmetry

m . m

8

32

22

ill

/////

0

l 1.5

3

4.5

6M

-

is

3

4'5

~--

Fig. 4. Results of A N O V A . Plot of interaction for frequency and amplitude by age, and single factor analysis for asynchrony and asymmetry by age. * Significantly different between h o m o g e n e o u s groups (3, 4.5 and 6 months) and the first point of the curve (1.5 months). Horizontal bars separated by an asterisk show groups significantly different from others.

293

SLEEP SPINDLE ONTOGENESIS TABLE I Evolution for the two sleep stages (SII and delta) of all parameters at 1.5-3-4.5 and 6 months for all infants. States

Ages 1.5 months SII

Density Duration Frequencies 10-12 Hz 12-14 Hz 14-15 Hz Amplitudes 10-25/~V 25-50/~V 50-75 ~V Asymmetry Asynchrony Infants Spindles

3 months Delta

1.7_+0.5 2.8 -+ 1.5

4.5 months

SII

2.1 _+1.1 2.1 _+0.9

24% 74% * 1.6%

53% 46% 0.7%

92% 7.7% 0% 0% 26.9% n=5 353

94% 5.7% 0% 0% 45.1% n = 10 1026

Delta 3.3_+0.9 3.3 _+0.7

46% 44% 9.8% 2.9% 23.5% n = 11 1540

SII

3.9-+0.7 2.8 _+0.7

14% 86% 0%

6 months Delta

4.1 _+0.6 2.2 _+0.3

SII

4.0-+0.8 2.3 -+0.5

Delta 3.9+ 1.3 2.0 -+0.3

4.2+0.9 2.1 _+0.4

25% 75% 0.1%

15% 55% 2.5%

18% 81% 0.2%

10% 84% 5.3%

12% 85% 2.6%

43% 44% 12% 5.5% 26% n = 11 2054

27% 59% 13% 13.7% 25.2% n = 10 1986

24% 64% 11% 10.7% 30.9% n = 10 1956

24% 61% 15% 11.6% 29% n = 10 1831

30% 57% 13% 11.3% 34.8% n = 10 1742

* P = 0.096 (Student's t test). percentage throughout age.

25-50/xV, 50-75/xV, 75-100/~V) show different evolutions. The percentage of the smaller amplitude (10-25 /xV) s i g n i f i c a n t l y d e c r e a s e d w i t h a g e ( P < 0.01), w h i l e t h a t o f 2 5 - 5 0 /zV a m p l i t u d e s i g n i f i c a n t l y i n c r e a s e d ( P < 0.01). A t t h e a g e o f 1.5 m o n t h s , o n l y s p i n d l e s o f 1 0 - 2 5 / x V a m p l i t u d e w e r e f o u n d (95.9%). It is i n t e r e s t ing to n o t e t h a t l a r g e a m p l i t u d e ( 5 0 - 7 5 /~V) s p i n d l e s a p p e a r e d at t h e a g e o f 3 m o n t h s ( 1 9 . 2 % ) a n d t h e i r p e r c e n t a g e did n o t i n c r e a s e t h e r e a f t e r . T h e l a r g e s t s p i n d l e a m p l i t u d e ( 7 5 - 1 0 0 / ~ V ) r e p r e s e n t e d a v e r y low

Asymmetry (Fig. 4).

A s y m m e t r y c o u l d n o t b e as1.5 m o n t h s o f a g e b e c a u s e o f t h e small o f t h e s p i n d l e bursts. T h e p e r c e n t a g e o f i n c r e a s e d f r o m 4 % at 3 m o n t h s to 11% at 6 months. Asynchrony (Fig. 4). T h e p e r c e n t a g e o f a s y n c h r o n y was h i g h e s t at t h e a g e o f 1.5 m o n t h s ( 3 6 % ) a n d d e c r e a s e d ( P < 0.03) to 2 5 % later. T h e h i g h e s t p e r c e n t s e s s e d at amplitude asymmetry the age of

ASYMMETRY

DENSITY

96

nb'min 4

15"

FREQUENCY

3

10-

%_

2"

2o I

=

10

~ • ........ t-

I-

•

0

O.

sec 3

_-

_

m--------- m ~ . . .

10 -12Hz

~Ir

A.

. .*. .

.....................

O-

;"-•

60 40

2. OJ .. 2'1

IF 21

23

i

3

DURATION

5

"lh

25-50,uV

- ..........

---';---e"" w

23

1

AMPLITUDE

-"

--

,

•

i

_-

w

~

; 1.5M i

1.5M

.... •

"---L-------A

3

=3

"=._.......I-="

l

1

=""•

5-

% ~ ' ~ m

all ages, e v e n at 6 m o n t h s o f

5

7h

20 21

2'3

'i

~1

,5

"ih

ASYNCHRONY

Fig. 5. Evolution of nocturnal sleep spindle organization at 1.5-3-4.5 and 6 months of life for the 6 parameters: density, duration, 10-12 and 12-14 Hz frequencies, 10-25 and 25-50 ~zV amplitude, asymmetry and asynchrony. For all these parameters, calculations are made during 5 periods (21.00-23.00 h, 23.00-1.00 h, 1.00-3.00 h, 3.00-5.00 h, 5.00-7.00 h) for all infants. Black circles = 1.5 months; black squares = 3 months; open circles = 4.5 months; black triangles = 6 months.

294 age at 1.5 months was due to a greater number of isolated bursts at this age.

Spindle et,olution during the two sleep stages Density. Table I shows the comparative density in stages II and delta. The density tended to be greater in stage delta at 1.5 and 3 months of age, but the difference between the two sleep stages is not statistically significant. Duration. Conversely, the burst duration tended to be longer in stage II than in stage delta at the ages of 1.5 and 3 months. At 4.5 and 6 months no difference was noted between the two stages (Table I). Frequency. Table I shows some differences between the low and the high frequencies. While the percentage of spindles with low frequency (10-12 Hz) was higher in stage delta, the percentage of those with frequencies 12-14 and 14-15 Hz was higher in stage II. This discrepancy mainly occurred at 1.5 months of age ( P = 0.096 between SII and delta for the 12-14 Hz frequencies). A larger amount of spindles with higher frequencies (14-15 Hz) persisted at 6 months in SII, but the difference between the two stages was not significant. Amplitude. No difference was observed between stages II and delta whatever the ages (Table I). Asymmetry and asynchrony (Table I). The asymmetry was not dependent on the two sleep stages. On the other hand, the percentage of asynchrony was higher in stage delta than in stage II at all ages and especially at 1.5 months, when a large number of isolated bursts was observed.

Spindle euolution during the night Density. Nocturnal spindle evolution, as shown in Fig. 5, did not show any nycthemeral organization at any age. Only a small upward trend was seen in the middle of the night between 1 and 3 h for all infants at 6 months of age. Duration. The length of sleep spindles appeared greater at the beginning of the night (21-23 h) at 1.5 months of age and remained stable afterwards (Fig. 5). Frequency (Fig. 5). Only the two more abundant frequencies 10-12 and 12-14 Hz were analyzed. Their evolution during the night at 1.5 and 6 months of age was different. The percentage of the lower frequencies did not change at the beginning of the night between 1.5 and 6 months of age, but in the second part of the night this percentage was increased at 1.5 months and decreased at 6 months. Thus, the difference between the two ages increased in the second part of the night. On the contrary, the comparative percentage of 12-14 Hz frequencies between 1.5 and 6 months showed an increase in the first part of the night, with a parallel decrease at the end of the night.

J. LOUIS ET AL.

Amplitude. Only the two more frequent amplitude samples (10-25 and 25-50 txV) were analyzed (Fig. 5). For these two samples, the evolution during the night between the ages of 1.5 and 6 months was stable and no difference between the beginning and the end of the night was noted. Asymmetry and asynchrony. Regarding asymmetry, the evolution during the night was not clear. It did not change at 1.5 and 3 months of age, but increased during the night at 4.5 months. At 6 months, the curve showed an asymmetry: an increase at the beginning of the night and a decrease at the end (Fig. 5). The asynchrony presented (Fig. 5) a higher percentage at 1.5 months. This high percentage was mainly concentrated during the beginning of the night. Later, the asynchrony was stable throughout the night for all ages.

Discussion

The data presented here were based on overnight sleep records and cannot really be compared to similar data based on short daytime records (Metcalf 1970; Wu et al. 1980). Moreover, few data were longitudinal (Ellingson and Peters 1980). While daytime study of spindle ontogenesis was directly applicable to the clinical situation, our study based on longitudinal night-time data might be helpful in reaching a better understanding of the process which regulates the production and organization of sleep in infants. Our sleep spindle analysis was performed visually to study the ontogenesis of sleep (Challamel et al. 1988). For this reason, a bipolar montage was used, which did not allow us to study the precise location of spindle waves. Similarly, our paper st~eed (1 c m / s e c ) did not allow us to analyze the precise form of the spindle wave. However, despite these drawbacks, we found the same frequency and the same amplitude described in ontogenetic studies of sleep spindles, especially in the classification of Metcalf (1970). This work, based on night-time longitudinal data, enabled us to study the evolution with age, in the same infant, of sleep spindles and its relation to the sleep organization rather than an analysis of spindle variability only as a function of age. The relevance of our data concerned the precise quantification of 6 parameters during the two stages of sleep and their evolution during the night. Thus, we analyzed the two components of the spindle ontogenesis: the quantitative development and the nycthemeral organization. In agreement with previous findings, we found that spindles appeared between 6 and 9 weeks post term (Metcalf 1969, 1970; Lenard 1970; Hagne 1972; Tanguay et al. 1975; Curzi-Dascalova 1977; Wu et al. 1980; Ellingson 1982; Jankel 1985).

SLEEP SPINDLE ONTOGENESIS

The greatest and significant differences in parameter values were observed between 1.5 and 3 months of age. At this moment, the density, the duration, the percentage of 12-14 Hz frequencies and 10-25, 25-50 /xV amplitudes, reached their maximum. These results suggest that when the physiological system that produces spindles matures, it acts as a single unit affecting density, duration, frequency and amplitude together. In agreement with Ellingson (1982), spindle bursts often occurred unilaterally. This led to a high percentage of asynchrony at 1.5 months of age. The large modifications between 1.5 and 3 months, with a peak at 3 months of age, had been observed with other physiological parameters such as the rapid increase of a clustering pattern of sleep rapid eye movements (Ktonas et al. 1990a), the emergence of voluntary movements (Prechtl and Hopkins 1986) and the decrease of body movements during REM sleep (Hashimoto and Tomita 1986). Spindle patterns developed quite rapidly during the first 3 months of infancy, possibly reflecting developmental changes in thalamocortical structures (Steriade et al. 1985), and myelination and growth of dendrites (Schad6 and Meeter 1963). Between 3 and 6 months of age, density, duration and percentage of 12-14 Hz frequencies were stable with a slight downward trend. Asymmetry and asynchrony were stable. Only the percentage of large amplitudes (25-50, 50-75 tzV) increased. The stability of parameter values in this period might possibly be related to the stable characteristics of sleep from the fourth month of life (Harper et al. 1981). After 6 months of age there is a gradual evolution of these spindles of infancy (Jankel 1985). It was interesting to note that there was still a slight percentage of 10-12 Hz frequencies up to 6 months of age. This finding might be relevant since the 10 Hz spindle type was found in about 40% of healthy adults (Gibbs and Gibbs 1950) and did not represent a lack of maturation. The differences in our findings between stage II and delta were transient; they were mainly observed before 3 months of age. At this time, duration and frequency were higher in stage II than in stage delta as in adults (Matsubayashi et al. 1981) and infants (Samson-Dollfus et al. 1988). Density of sleep spindles was higher in stage delta as compared to stage II, and the spindle frequency was lower. This last point could be interpreted as an index of "immaturity" of stage delta during this period. The "immaturity" of our stage delta before 4.5 months of age raises the problem of the emergence of this sleep stage from quiet sleep (Schulz et al. 1989). Did this stage in infants correspond to adult SWS? This question brought up the problem of the coherence between EEG variables for the determination of SWS in infants. In our sleep spindle study we could delineate a "true" stage delta of SWS from 4.5 months of life, when no differences between the two

295

sleep stages were noticed for most of the spindle parameters. This is concordant with Ktonas' study which revealed that a significant variation in delta and sigma band power (increase delta and decrease sigma band) occurred between 4 and 12 months of life in the quiet sleep period (Ktonas et al. 1990b). The density of spindles did not show any nycthemeral variation. This corresponds to the results obtained by Gaillard and Blois in 1981 in adults and SamsonDoilfus et al. (1988) in infants. However, for some of the spindle parameters, a few variations during the night were observed in this study. They took place at the beginning or at the end of the night. Higher levels of the duration and asynchrony at the beginning of the night were observed at 1.5 and 3 months of age and disappeared afterwards. Variations at the end of the night were observed later on at the ages of 4.5 and 6 months for frequency, with a downward trend. This result was also reported by Samson-Dollfus, using an automatic analysis system (personal communication), preferentially in 5-month-old infants, whereas this tendency did not clearly appear in younger ones. Infants tended to produce SWS in successive cycles, although the production rate of SWS was always high in the first sleep cycle in adults (Schulz et al. 1989). The mechanism which regulated SWS production at this developmental stage is not well understood, but with regard to our longitudinal results, it seems that some features, characteristic of adults, had already appeared at the age of 4.5 months. Indeed, spindle duration was longer at the beginning of the night, with a higher frequency. Thus, spindle bursts, used as a hallmark of SWS, may also serve to analyze the processes which regulate the production and organization of SWS in infants. Adult spindle characteristics appear rapidly and this fact led us to think that the development of this sleep EEG pattern has to do with its function. However, further investigations are necessary for a better understanding of the precise role of these bursts in the sleep-waking process. We would like to thank Christine Cannard for her technical assistance, and Dr. H61~ne Bastuji and Dr. Luis Garcia-Larrea for their friendly help in discussing this paper. Thanks also to Dr. Joan Cahill for reviewing the English manuscript.

References Anders, T., Emde, R. and Parmelee, A. (Eds.). A Manual of Standardized Terminology, Techniques and Criteria for Scoring of States of Sleep and Wakefulness in Newborn Infants. UCLA Brain Information Service/BRI Publications Office, Los Angeles, CA, 1971. Challamel, M.J., Debilly, G., Leszczynski, M.C. and Revol, M. Sleep state development in near-miss sudden death infants. In: R.M. Harper and H.J. Hoffman (Eds.), Sudden Infant Death Syn-

296 drome: Risk Factors and Basic Mechanisms. PMA Publishing, New York, 1988: 423-434. Chatrian, G.E., Bergamini, L., Dondey, M., Klass, D.W., LennoxBuchthal, M. and Peters~n, 1. A glossary of terms most commonly used by clinical electroencephalographers. Electroenceph. clin. Neurophysiol., 1974, 37: 538-548. Curzi-Dascalova, L. EEG de veille et de sommeil du nourisson normal avant 6 mois d'~ge. Rev. EEG Neurophysiol., 1977, 7: 316-326. Dreyfus-Brisac, C. and Curzi-Dascalova, L. The EEG during the first year of life. In: A. REmond (Ed.), The Evolution of the EEG. Handbook of Electroencephalography, Vol. 6B. Elsevier, Amsterdam, 1975: 24-30. Ellingson, R.J. EEGs of premature and full-term newborns. In: D.W. Klass and D.D. Daly (Eds.), Current Practice of Clinical Electroencephalography. Raven Press, New York, 1979: 149-177. Ellingson, R.J. Development of sleep spindle bursts during the first year of life. Sleep, 1982, 5: 39-46. Ellingson, R.J. and Peters, J.F. Development of EEG and daytime sleep patterns in normal full-term infants during the first 3 months of life: longitudinal observations. Electroenceph. clin. Neurophysiol., 1980, 49:112-124. Galliard, J.M. and Blois, R. Spindle density in sleep of normal subjects. Sleep, 1981, 4: 385-391. Gibbs, F.A. and Gibbs, E.L. Atlas of Electroencephalography, Vol. 1, Addison-Wesley, Reading, MA, 1950. Hagne, I. Development of the sleep EEG in normal infants during the first year of life. Acta Paediat. Scand. (Suppl.), 1972, 219: 25-53. Harper, R.M., Leake, B., Miyahara, L., Mason, J., Hoppenbrouwers, T., Sterman, M.B. and Hodgman, J. Temporal sequencing in sleep and waking states during the first 6 months of life. Exp. Neurol., 1981, 72: 294-307. Hasbimoto, T. and Tomita, T. Body movements during sleep-developmental changes: Third SNCC Symp. on Developmental Neurobiology, Tokyo, 1986: 53-65. Jankel, W.R. and Niedermeyer, E. Sleep spindles. J. Clin. Neurophysiol., 1985, 2: 1-35. Ktonas, P.Y., Bes, F.W., Rigoard, M.T., Wong, C., Mallart, R. and Salzarulo, P. Developmental changes in the clustering pattern of sleep rapid eye movement activity during the first year of life: a Markov-process approach. Electroenceph. clin. Neurophysiol., 1990a, 75: 136-140. Ktonas, P.Y., Fagioli, I. and Salzarulo, P. Delta and sigma power dynamics throughout QS in infants. In: 10th Congress of the Europ. Sleep Res. Society, Strasbourg, 1990b. Lenard, H.G. The development of sleep spindles during the first two years of life. Neurop~idiatrie, 1970, 1: 264-276. Lombroso, C.T. Neurophysiological observations in diseased newborns. Biol. Psychiat., 1975, 10: 527-558.

J. LOUIS ET AL. Matsubayashi, K., Ishiyama, Y., Homma, I. and Ebe, M. Some characteristics of sleep spindles derived from automatic analysis. Sleep, 1981, 4: 392-399. Metcalf, D.R. The effect of extrauterine experience on the ontogenesis of EEG sleep spindles. Psychosom. Med., 1969, 31: 393-399. Metcalf, D.R. EEG sleep spindle ontogenesis. Neurop~idiatrie, 1970, 1: 428-433. Monod, N. and Ducas, P. The prognostic value of the electroencephalogram in the first two years of life. In: P. Kellaway and I. Peters6n (Eds.), Clinical Electroencephalography in Children. Grune and Stratton, New York, 1967: 61-76. Prechtl, H.F.R. and Hopkins, B. Developmental transformations of spontaneous movements in early infancy. Early Hum. Dev., 1986, 14: 233-238. Rechtschaffen, A. and Kales, A. (Eds.). A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. National Institute of Health, Publ. 204, U.S. Government Printing Office, Washington, DC, 1968. Samson-Dollfus, D., Delapierre, G., Senant, J., De Brucq, D., Bertoldi, I. and Dreano, E. Automatic spindle recognition and sleep in infants. In: W.P. Koella, F. Obal, H. Schulz and P. Visser (Eds.), Sleep 86. Karger, Basel, 1988: 264-265. Schad~, J.P. and Meeter, K. Neuronal and dendritic patterns in the uncinate area of the human hippocampus. Prog. Brain Res., 1963, 3: 89-110. Schultz, M.A., Schulte, F.J., Akiyama, Y. and Parmelee, A.H. Development of electroencephalographic sleep phenomena in hypothyroid infants. Electroenceph. clin. Neurophysiol., 1968, 25: 351358. Schulz, H., Bes, E. and Salzarulo, P. Rhythmicity of slow wave sleep developmental aspects. In: A. Wauquier (Ed.), Slow Wave Sleep: Physiological, Pathophysiological and Functional Aspects. Raven Press, New York, 1989: 49-60. Shibagaki, M., Kiyono, S. and Watanabe, K. Spindle evolution in normal and mentally retarded children: a review. Sleep, 1982, 5: 47-57. Steriade, M., Desch~nes, M., Domich, L. and Mulle, C. Abolition of spindle oscillation in thalamic neurons disconnected from nucleus reticularis thalami. J. Neurophysiol., 1985, 54: 1473-1497. Tanguay, P.E., Ornitz, E.M., Kapplan, A. and Bozzo, E.S. Evolution of sleep spindles in childhood. Electroenceph. clin. Neurophysiol., 1975, 38: 175-181. Watanabe, K. and Iwase, K. Spindle-like fast rhythms in the EEGs of low-birthweight infants. Dev. Med. Child Neurol., 1972, 14: 373381. Wu, H.S., Gould, J.B., Lee, A.F.S. and Fineberg, N. Factors affecting sleep spindle activity during infancy. Dev. Med. Child Neurol., 1980, 22: 344-351.