Development of Sleep Spindle Bursts During the First Year of Life

Transcription

1 Sleep, 5(1): Raven Press, New York Development of Sleep Spindle Bursts During the First Year of Life Robert J. Ellingson University of Nebraska Medical Center. Omaha, Nebraska Summary: Sleep spindles, as defined in the international electroencephalography (EEG) glossary, are distinguished from spindle bursts seen in the EEGs of premature infants. Classical sleep spindles do not occur in prematures. They first appear clearly in the EEG during slow wave sleep from the 4th week postterm (44 weeks conceptional age) and are normally present in all infants' EEGs by 9 weeks postterm. During the first year of life they may be of high voltage relative to their appearance in older children and adults and are commonly characterized by variable degrees of interhemispheric asymmetry and asynchrony. Individual differences are great. Some features of clinical significance are discussed. Key Words: Human infants-sleep-sleep spindles Development. Before discussing the development of sleep spindle bursts in human infants, it is essential to distinguish clearly between spindles and sleep spindles. The international EEG glossary (Chatrian et al., 1974) defines spindle as a "group ofrhythmic waves characterized by a progressively increasing, then gradually decreasing, amplitude. Cj. sleep spindle." Sleep spindle is defined as a "burst at Hz, but mostly at Hz, generally diffuse but of higher voltage over the central regions of the head, occurring during sleep. Amplitude is variable but is mostly below 50 f.l V in the adult." Therefore, sleep spindles constitute a specific instance of a larger class of graphoelements called spindles. Conversely, not all spindles are sleep spindles. The reason this distinction is important in considering sleep spindle development is that spindle bursts occur abundantly in the EEGs of premature infants (Ellingson, 1979). These spindle bursts are sometimes called brushes (Lombroso, 1975). They have frequently in the past been confused with sleep spindles, resulting in numerous reports of sleep spindles in the EEGs of prematures and the repetition of such erroneous reports in reviews. Explicit distinctions between spindles of prematurity (Fig. 1) and sleep spindles (Fig. 2) are as follows. (1) The frequency range of the spindles of prematurity is roughly 8-20 Hz; that of sleep spindles is Hz. (2) Spindles of prematurity Accepted for publication October Address correspondence and reprint requests to Dr. Ellingson at Nebraska Psychiatric Institute, 602 South 45th Street, Omaha, Nebraska

2 40 R. J. ELLINGSON AGE: 3SW CA Vv-~r-----"",.~.",,~-~~ F8-T4 ~\~~('~J'~~/"~~ T4-T6 ~~~JlrAN<",~~~~"'- T6-02 NII-172 FIG. 1. EEG of a normally developing infant born at 29 weeks gestation (1000 g birth weight) recorded at 35 weeks conceptional age during quiet sleep. The recording was made at a paper speed of 30 mm/s. The non-eeg channels have been omitted from this and the following figure. Spindle bursts of prematurity are prominently displayed. Note the predominant occipital-posterior temporal distribution. where they are superimposed on large slow «1 Hz) delta waves forming spindle-delta complexes. The waves in the spindle bursts are of somewhat variable frequency in the range of Hz. occur during all stages of the wake-sleep cycle, although from about weeks conceptional age (CA) they tend to be more prominent during sleep than during wakefulness. Sleep spindles normally occur only during slow wave sleep (SWS). (3) Spindles of prematurity are usually associated with, and often superimposed Sleep. Vol. 5. No

3 SLEEP SPINDLES DURING INFANCY 41 AGE: 37W ~~~_-",- ~,----/~-----~--J'----v"'---'~~ FP1-F3 ~.~~r-"""'~--"i..--,\.-~...-.\/\./..r""""-.,~,,;'\(\'\'\1rtf''''\i~\i\'v\yn''11\1'\\\'v\'{ 1\/\"V\r\"Ii"W(\I."\'\fV'v',..r'vJ'r'- ~ F3-C3 ~/'-_/--'V'---~-v_r-~"r4iY'~ C3-P3 P3-01 FP2-F4 FP1-F7 I 100"v ~~~~~ F7-13 SEC FP2-FS FS ~'/v-'---j'~rw'""rtfl'y4rjvf'rn'{rlfrtvr'[r(iirlrrfr[frrjfl\ rllfrr[jr!wf~/v#'~~~ FZ-CZI ~--.r'\mr-/'vvv''v.x~ 'NJ\" 'vj'v'\i<lv",jvif'~vii~ CZ-PZ /' N FIG. 2. EEG of a normal full-term infant (2700 g birth weight) recorded at 37 weeks of age. A sleep spindle burst is prominently displayed. Note the rolandic distribution. Compare the 8-s burst seen at the vertex (bottom two tracings) with the apparently independent right and left rolandic bursts seen in channels 6-8 and 2-4, respectively. It is seen that the latter are really asymmetrical extensions of the field of the single peri vertex burst, first to the right and then to the left. Analysis of phase reversals shows that in the midline the field is centered at the vertex (CZ electrode) on the right between the C4 and P4 electrodes and on the left at the P3 electrode, in the anteroposterior dimension. The center of the field in the transverse dimension cannot be determined definitely in the absence of a linked transverse bipolar electrode montage; judging on the basis of relative voltage only, the center of the field was probably to the right ofthe midline at the beginning of the burst and then moved to, or slightly to the left of, the midline. Note some extension of the field first into the right temporal area (channel 14) and then into the left (channel 10). The frequency of the spindle burst waves is consistent at exactly 13 Hz. Sleep. Vol, 5. No. I, 1982

4 42 R. J. ELLINGSON VI CC... Z V CII: A AGE IN WEEKS FIG. 3. Percentages of EEGs of normal full-term infants, 1-13 weeks old, showing clearly identifiable rolandic sleep spindle bursts. The data base was 201 EEG-polygraph recordings of 1-4 h duration obtained approximately weekly on 17 subjects. (Reprinted, with permission of the publisher, from Ellingson and Peters, 1980a.) on, slow «1 Hz) delta waves, forming what are often called spindle-delta bursts or complexes. Sleep spindles are often associated with moderate- to high-voltage vertex sharp transients (Chatrian et ai., 1974) of negative polarity, sometimes called V waves or sleep transients, and are associated with high-voltage diphasic slow waves at the vertex, forming K complexes (Chatrian et ai., 1974).1 Both V waves and K complexes have topographical distributions similar to those of sleep spindles themselves. (4) Spindles of prematurity are topographically widespread, are usually most prominent in the temporal and occipital areas, and often occur independently over the two hemispheres. Sleep spindles are of highest voltage in the rolandic area bilaterally and are often confined to that area. Brazier (1949) demonstrated that the focus of sleep spindles is in the midline just anterior to the vertex. (5) Spindles of prematurity rapidly diminish in voltage and incidence during the ninth month of gestation and have already disappeared in the majority of full-term newborns. In those in whom they persist, they are weak and infrequent. Normally they are not seen beyond 2 weeks postterm (42 weeks CA). Clearly identifiable sleep spindles do not appear in the full term newborn until 4 weeks postterm (44 weeks CA)2 and sometimes not until 9 weeks (Fig. 3). When they first appear they are weak and infrequent, sometimes not more than 3 or 4/hr of SWS. In the following 2-4 weeks, they become more prominent and frequent (Fig. 4), reaching a mean frequency of about 4/min of SWS by 3 months of age. It is quite clear from these distinctions that spindles of prematurity and sleep spindles are different phenomena. The only things they have in common are the 1 V waves and K complexes are not, however, present at 4-9 weeks of age When the first sleep spindle bursts appear. They develop later during the first year of life. 2 They may appear a week or more earlier in the premature (Metcalf 1969, 1970; Ellingson and Peters, 1980b). Sleep. Vol. 5, No. I, 1982

5 SLEEP SPINDLES DURING INFANCY 43 AGE DAYS 48 ss (' If\AJVJ\rJ "\;J~(V~\r"r,''vJr!~~/\vr~j'i'J\ SEC 83,\('J\rJ1IV'V~hvWjrf\.~V~/J~\l~rN'\I(.fiW"V#tj l'4~")~r'f_, Z-ll FIG. 4. Isolated rolandic EEG derivations (C3-Al or C4-A2, depending on which showed the most prominent sleep spindle bursts in a given recording) from weekly EEGs in the same normal full-term infant (birth weight, 3490 g). No spindles were seen at 7 weeks (48 days). Rare weak spindle bursts were seen for the first time at 8 weeks. No change at 9 weeks. Some increase was seen in incidence and duration but not in voltage of bursts at to and 11 weeks. Clear increase in voltage and a further increase in incidence and duration of bursts occurred at 12 weeks, when the rate of bursts was 12.7/min of slow wave sleep. The subject was one of those contributing to the data summarized in Fig. 3. The date of onset of sleep spindle activity for the subject was taken as 52 days (71-2 weeks). spindle-shaped outline of the bursts and the overlap in frequency range. Sleep spindles as defined in the international glossary do not occur in prematures nor in the immediate neonatal period. When sleep spindles first appear at 4-9 weeks postterm in the full-term infant (Lenard, 1970; Metcalf, 1970; Hagne, 1972; Metcalf and Jordon, 1972; Curzi Dascalova, 1977; Ellingson and Peters, 198Oc), only a few weak (generally less than 20 /-tv peak to peak) bursts are seen during an SWS period. A majority of the bursts occur unilaterally, randomly alternating between hemispheres (Fig. 2), and amplitude asymmetry is common. By 9-10 weeks postterm, they should be present in all normal infants, subsequently constituting the premier EEG sign of SWS. Sleep, Vol, 5, No. I, 1982

6 44 R. J. ELLINGSON TABLE 1. Percent of sleep spindle burst synchrony and synunctry as a junction of age in full-term infants" Age postterm (months) Mean SD Range b " Modified, with permission of the publisher, from Ellingson and Peters (l980b). b The difference between this value and both of the two others is significant (p < 0.01). Their absence in infants after this time is considered to be a nonspecific EEG abnormality (Monod and Ducas, 1967), provided that a substantial sample of SWS has been obtained as determined by other criteria, such as regular respiration, tonic electromyographic activity, and absence of rapid eye movements. Distinguishing stages of the wake-sleep cycle on the basis of the frequency and abundance of slow wave activity in the EEG is still too difficult to be reliable at these ages. The onset of sleep spindle activity during the period of 4-9 weeks postterm has been objectively confirmed by quantitative frequency analyses of infants' EEGs (Schulte and Bell, 1973; Sterman et ai., in press, especially Fig. 1). By 13 weeks postterm sleep spindle activity during SWS is much stronger than in the adult and remains so for the rest of the first year of life and often beyond. Still, at 3 months postterm, only half of sleep spindle bursts are bilaterally synchronous and symmetrical, a situation that persists at 6 months (Table 1). By 12 months there has been a significant improvement in bilateral synchrony, continuing into the second year (Kellaway, 1979). At all of these ages, sleep spindle bursts are usually of highest voltage at the vertex (CZ electrode; Fig. 2). Bursts may also occur alone at the vertex or in the left or right rolandic areas (electrodes C3, C4), or they may, infrequently, be of highest voltage at one of the lateral locations (C3 or C4), of lower voltage at CZ, and of lowest voltage contralaterally. Abnormalities of sleep spindle development have been described in several disorders. In phenylketonuria, spindle activity is exaggerated and may already be abnormal at 45 weeks CA (Gross and Schulte, 1969; Schulte et ai., 1973). A paucity of sleep spindle activity has been reported in hypothyroid infants (Schultz et ai., 1968; Lenard and Bell, 1973). In trisomy-21 infants, the appearance of sleep spindles is delayed by 3 weeks on the average, the number of sleep spindle bursts is decreased by about one-fourth, and interhemispheric synchrony of spindle bursts does not improve by 12 months postterm (Ellingson and Peters, 1980c). It must be emphasized that there are large individual differences in sleep spindle manifestation during infancy as well as later. Despite intergroup differences cited, at present the only deviations in sleep spindle activity during the first year of life that can be safely classified as abnormal in evaluating individual patients' EEGs are (1) complete absence of sleep spindles after 8 weeks postterm (at 8-12 Sleep. Vol. 5, No. I, 1982

7 SLEEP SPINDLES DURING INFANCY 45 weeks more or less such absence may indicate only maturational retardation; after that it has more dire pathological implications); (2) virtually total interhemispheric asynchrony of bursts; (3) virtually total interhemispheric asymmetry with a greater than 2: 1 ratio of spindle voltages between the hemispheres during an entire SWS recording and using more than one montage, preferably confirmed by a second recording (this suggests a unilateral brain disorder or greater disorder on one side, which is usually, but not always, the side with lower voltage; when this sign is associated with other focal EEG abnormality, the side of the disorder will be clear). Suspicious sleep spindle deviations include: (1) more than 75% asynchronous spindle bursts after 3 months postterm and more than 50% asynchrony at 12 months, and (2) consistent asymmetry with voltage ratio between 2: 1 and 3:2. REFERENCES Brazier MAB. The electrical fields at the smface of the head during sleep. Electroencephalogr Clin Neurophysiol 1: , Chatrian GE, Bergamini L, Dondey M, Klass OW, Lennox-Buchthal M, and Petersen I. A glossary of terms most commonly used by clinical electroencephalographers. Electroencephalogr Clin Neurophysiol 37: , Curzi-Dascalova L. E.E.G. de veille et de sommeil du nourrisson normal avant 6 mois d'age. Rev EEG NeurophysioI7: , Ellingson RJ. EEGs of premature and full-term newborns. In: OW Klass and DD Daly (eds), Current Practice of Clinical Electroencephalography, Raven Press, New York, 1979, pp Ellingson RJ and Peters JF. Development of EEG and daytime sleep patterns in normal full-term infants during the first 3 months of life: Longitudinal observations. Electroencephalogr Clin Neurophysiol 49: , 1980a. Ellingson RJ and Peters JF. Development of EEG and daytime sleep patterns in low risk premature infants during the first year of life: Longitudinal observations. Electroencephalogr Clin Neurophysiol 50: , 1980b. Ellingson RJ and Peters JF. Development of EEG and daytime sleep patterns in Trisomy-21 infants during the first year of life: Longitudinal observations. Electroencephalogr Clin Neurophysiol 50: , 198Oc. Gross RP und Schulte FJ. Uber vermehrte Spindelaktivitat im Schlaf-EEG bei Kindem mit Phenylketonurie. Z Kinderheilk 105: , Ragne I. Development of the sleep EEG in normal infants during the first year of life. Acta Paediatr Scand Suppl 232:25-53, Kellaway P. An orderly approach to visual analysis: Parameters of the normal EEG in adults and children. In: OW Klass and DO Daly (eds), Current Practice of Clinical Electroencephalography, Raven Press, New York, 1979, pp Lenard RG. The development of sleep spindles in the EEG during the first two years of life. Neuropaediatrie 1: , Lenard RG and Bell EF. Bioelectric brain development in hypothyroidism. Electroencephalogr Clin Neurophysiol 35: , Lombroso CT. Neurophysiological observations in diseased newborns. Bioi Psychiatry \0: , Metcalf DR. The effect of extrauterine experience on the ontogenesis of EEG sleep spindles. Psychosom Med 31: ,1969. Metcalf DR. EEG sleep spindle ontogensis. Neuropaediatrie 1: , Metcalf DR and Jordan K. EEG ontogenesis in normal children. In: WL Smith (ed), Drugs, Development, and Cerebral Function. Thomas, Springfield, Ill., 1972, pp Monod Nand Ducas P. The prognostic value ofthe electroencephalogram in the first two years oflife. In: P Kellaway and I Petersen (Eds), Clinical Electroencephalography in Children, Grune & Stratton, New York, 1967, pp Schulte FJ and Bell EF. Bioelectric brain development. An atlas of EEG power spectra in infants and young children. Neuropaediatrie 4:30-45, Sleep, Vol. 5, No. I, 1982

8 46 R. 1. ELLINGSON Schulte FJ, Kaiser HJ, Engelbart S, Bell EF, Castell R, and Lenard HG. Sleep patterns in hyperphenylalaninemia: A lesson on serotonin to be learned from phenylketonuria. Pediatr Res 7: , Schultz MA, Schulte, FI. Akiyama Y, and Parmelee lr AH. Development of electroencephalographic sleep phenomena in hypothyroid infants. Electroencephalogr Clin NeurophysioI25: , Sterman MB, McGinty DJ, Harper RM, Hoppenbrouwers T, and Hodgman IE. Developmental comparison of sleep EEG power spectral patterns in infants at low and high risk for sudden death. Electroencephalogr C/in Neurophys;o/ (in press). Sleep. Vol. 5, No. I. 1982