Sleep, 15(5):442-448
© 1992 American Sleep Disorders Association and Sleep Research Society
Comparison of EEG Sleep Measures in Healthy Full-Term and Preterm Infants at Matched Conceptional Ages *tMark S. Scher, *Doris A. Steppe, tRonald E. Dahl, *tShobha Asthana and *Robert D. Guthrie *Departments of Pediatrics and tNeurology, Magee- Womens Hospital and Children's Hospital of Pittsburgh, Pittsburgh, Pennsylvania 15213, U.S.A.; and t.Department of Psychiatry, Western Psychiatric Institute and Clinic, University of Pittsburgh, School of Medicine, Pittsburgh, Pennsylvania 15213, U. S.A.
Summary: Continuous electroencephalogram (EEG) sleep studies were obtained on healthy full-term and preterm infants at matched conceptional ages. Studies were recorded under environmentally controlled conditions. Eighteen healthy preterm infants were matched to 18 full-term infants based on conceptional age, sex, race and socioeconomic class. The initial 3 hours of a 12-hour recording were simultaneously recorded on paper and computer. The visually scored data based on the paper recordings for sleep architecture and continuity measures were studied. Differences in each sleep organization for the preterm infants included the following: a longer ultradian sleep cycle (70 minutes vs. 53 minutes, p = 0.02) was noted. More abundant trace altemant (34% vs. 28%, p = 0.02) and less abundant low-voltage irregular active sleep (13% vs. 17%, p = 0.05) were noted. Although no differences were observed for sleep latency and efficiency, the preterm infants had fewer numbers and shorter durations of arousals, fewer body movements and rapid eye movement (REM) (p < 0.0 I), particularly during quiet sleep. The extrauterine experience or the earlier birth of the preterm infant may influence specific sleep architecture and continuity measures when compared with the sleep of full-term infants who experienced a complete intrauterine gestation. Key Words: EEGsleep- N eonates- Preterm - Fullterm.
For more than three decades, investigators have explored the use of sleep measures as a route to understanding early neonatal development (1-9,17). Studi,es have included longitudinal and cross-sectional designs delineating the maturational patterns of sleep in the preterm infant maturing to term. It has been generally accepted that the sleep cycle in the infant born prematurely is highly variable until at least 36 weeks gestation (11), after which time it will resemble the sleep cycle ofthe appropriate-for-age full-term infant. It has also traditionally been assumed that the ultradian sleep pattern of the full-term infant is the "gold standard" to which the sleep rhythm of the preterm infant at term should be compared (1,17). Even when matched for conceptional age, however, the premature infant has achieved the final aspects of development in an extrauterine environment, with a variety of experiences that may directly impact the development of the brain as expressed by the neonatal sleep cycle.
The present study reports on electroencephalogram (EEG) sleep in two cohorts of infants: a normal group of full-term infants and a group of healthy premature infants at corrected term ages. Groups were matched for sex, race, conceptional age, and socioeconomic class, and EEG sleep measures of architecture and continuity were compared. The primary hypothesis of the study was that despite comparable conceptional ages, preterm infants would show differences in specific sleep architecture and continuity parameters, reflecting the influences of extrauterine development or their premature birth.
METHODS Patient population
Healthy preterm infants of < 32 weeks estimated gestational age (EGA) were recruited from a neonatal population admitted to the neonatal intensive care unit of a large obstetrical center. Selection was based on a Accepted for publication May 1992. Address correspondence and reprint requests to Mark S. Scher, review of maternal and neonatal medical records and M.D., Magee-Womens Hospital, Developmental Neurophysiology Laboratory, Halket and Forbes Avenue, Pittsburgh, Pennsylvania on consultation with the attending neonatologist. Infants with any significant medical illnesses were ex15213, U.S.A.
442
Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
EEG SLEEP MEASURES IN INFANTS cluded, including those with sepsis, respiratory disease, intracranial hemorrhage, or other major organ system illnesses. Infants were enrolled only after 7 days oflife, permitting verification of their healthy medical condition. At least one cranial ultrasound was available to document the absence of significant intracranial hemorrhage or peri ventricular echodensities. When each preterm subject reached a conceptional full-term age (i.e. estimated gestational age at birth plus weeks of life), an appropriate-for-gestational-age fullterm infant was selected from the well-child nurseries. Careful review of the medical records and physical examinations were carried out to verify the healthy status of these full-term infants. Each control subject was matched to a preterm infant at term by conceptional age (within 1 week), sex, race and socioeconomic class (Hollingshead, unpublished). All infants received a neurodevelopmental assessment on the day of each recording (12). Information regarding feeding schedules, sleep behavior and medical care were documented. After completion of the sleep recording portion of the study, all infants were followed clinically for at least 18 months and judged to be healthy, including ageappropriate neurodevelopment milestones. Psychometric testing included Bayley Motor and Mental Performance Scales, Carey Temperament Scales, and Vineland Social Maturity Scales. EEG Sleep Recording Sessions
Twelve-hour EEG sleep Twelve-hour EEG sleep studies were carried out in an environmentally controlled setting in which sound, light, humidity and tactile stimulation were carefully monitored. All infants were studied while they slept in open beds. Preterm infants had received monthly studies from birth until their term conceptional age. The full-term subjects received a single sleep study between 2 and 4 days of life. For analyses comparing preterm subjects with fUll-term control infants as presented in this study, only the term conceptional age data were used.
Recording parameters Continuous recordings began after a feeding and diaper change between 2000 to 2100 hours and ended between 0800 and 0900 hours of the following day. The subject was observed in a prone position, which was the usual sleeping position in the nursery. While the entire 12-hour study was digitized on an Apollo computer workstation (Hewlett-Packard Corp.), the initial 3 hours were simultaneously recorded on paper using a 21-channel EEG machine (Nihon-Kohden,
443
model 4100) for visual scoring. Results in this paper encompass only the visually scored sleep architecture and continuity measures from this 3-hour recording obtained at a term conceptional age. For the 3-hour paper recording, 14 EEG channels were utilized, with 11 cerebral electrodes placed in the standard 10-20 system. N oncerebral channels included one tonic chin electromyogram (EMG) channel, two electrooculographic channels, two respiratory channels (thoracic and abdominal positions), electrocardiogram (EKG), skin and rectal temperature and pulse oximetry. All movements were recorded directly on the paper and into a computer-driven event-marker file operated by the technologist. Sensitivity was selected at 7 m V/ mm, with a time constant of 0.3 and a paper speed of 15 mm/second. The EEG technologist documented all relevant events regarding subject behavior, manipulations (i.e. feedings, diaper changes, etc.) and sources of artifact. A neonatal research nurse was responsible for the clinical care of the infant during each recording session. Sleep and feeding behavior, diaper changes, medication administration and technical comments (i.e. equipment malfunctions, environmental measurements, etc.) supplemented the technologist's comments. All infants were medication-free at the time of the sleep studies.
Data entry and verification procedures Demographic, clinical, technical and EEG sleep data were recorded on hand-written data sheets and then entered into a relational data base (Interbase, Inc.). An electroencephalographer (M.S.S.) assigned a score for EEG sleep state, and arousal number and duration for each minute of the 180 minutes of paper recording. Accurate data entry into the computer was verified. In order to assess scoring reliability, 10% of the records were rescored by one individual (M.S.S.) for state, arousal, motility and REM. There was a 96% agreement with the original scoring of all minutes. Sleep architecture and continuity measures were tabulated for each of the initial 180 minutes of the allnight recording (Table 1). These measures included I-minute numerical values for the following: EEG sleep state; arousal number and duration; total body, facial, head, small and large body movement counts; and rapid eye movement counts (REM). Statistical Analysis
EEG sleep variable definitions Nine measures were defined based on the minuteby-minute tabulated data (Table 1): Sleep, Vol. 15, No.5, 1992
Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
444
M. S. SCHER ET AL. TABLE 1. Sleep architecture and continuity measurements Architecture
Continuity
Phasic events
EEG sleep state (ESS) Mixed-frequency (M) High-voltage slow (HVS) Trace alterant (T A) Low-voltage irregular (LVI) Cycle length (minutes)
Arousals Sleep latency Slleep efficiency
Body movements (BM) Face/suck (F) Head (H) Small (S) Large (L) REMs
EEG sleep state (ESS). Six state categories defined traditional EEG state (I, 10): two active sleep segments (mixed-frequency and low-voltage irregular), two quiet sleep segments (high-voltage and trace alterant), indeterminate sleep and wakefulness. An awake state was defined as 3 minutes of continuous eye-opening associated with an appropriate EEG pattern. Less than 1% of the minutes could not be scored because excessive artifact prevented an electrographic interpretation. Arousals (A). Transient epochs of EEG activity during which a desynchronization of the EEG background activity was noted with more than a 50% reduction illl amplitude and greater than 90% reduction in slow EEG frequency activity. Arousal durations were recorded to the nearest second. Coincidental movements or sources of artifact were also noted. Spontaneous arousals were distinguished from evoked arousals, based on coincident artifact-producing events. Body movements (BM). Total body, facial (including suck), head, small and large body movements were summed for each minute. Rapid eye movements (REMs). The determination of REM was based on synchronized out-of-phase pen deflections in two electrooculographic channels (left and right outer canthi) of at least 20 mv in amplitudl~. Sleep onset. This measure was defined as the first 3 minutes of continuous eye closure after "lights out". Sleep latency (SL). This measure was defined as the time from "lights out" to sleep onset. "Lights out" was determined when the lighting was turned out after the infant was placed in a prone position following a feeding and diaper change. Sleep efficiency (SE). This measure was the ratio of total sleep time (minutes) over the total recording time (minutes). REM latency (RL). This measure was the time from sleep onset to the first REM period. TABLE 2.
Cyclicity (C). This measure was the time (in minutes) between 2 consecutive periods of the same sleep stage (a minimum of 3 consecutive minutes of asleep stage was designated to denote a "period"). For example, cyclicity is the time between quiet to quiet sleep states, trace to trace alterant, etc. Calculations were based on the first cycle after the infant initiated sleep. Statistical analyses
All computer analyses were performed using a statistical package (13) that was configured to our relational data base. Exploratory calculations included tabulation of average numbers and durations of all measures for each subject. Scatterplots and histograms were also used to define outliers and patterns of distribution. Two-tailed, matched-pair t tests were performed on each preterm-full-term pair, matched for conceptional age, sex, race and socioeconomic class. Nonparametric calculations (i.e. Mann-Whitney calculation) were used when distributions were non-normal. Chi-squared calculations were used for categorical data in demographic comparisons.
RESULTS Demographic and clinical data
Eighteen preterm at term (PTT) and full-term (FT) pairs were selected. All neonates were appropriate for gestational age at birth and demonstrated normal growth parameters. Table 2 lists the range of gestational age, conceptional age at the time of the EEG sleep study, birth weight, length and occipital-frontal circumferences. Maternal histories for the preterm group included six women with preeclampsia and 12 with premature
Group comparisons Preterm group
Estimated gestational age (weeks) Conceptional age (weeks) Birth weight (g) Length (cm) Occipital frontal circumference (cm)
Full-term group
Mean
Range
Mean
Range
29.3 40.9 3,698 50.9 36.5
27-32 38-42 2,705-3,950 47-52 34.3-37.5
40.5 40.5 3,591 51.8 34.9
38-43 38-43 2,810-4,020 47.5-53 34.5-37.8
Sleep. Vol. 15. No.5. 1992 Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
EEG SLEEP MEASURES IN INFANTS TABLE 3. Comparisons ofsleep architecture measurements in preterm and full-term infants (using paired t tests) % Preterm group (± SD)
EEG sleep state Mixed (M) High-voltage slow (HVS) Trace alterant (TA) Low-voltage irregular (LVI) Indeterminate sleep (IS) Waking (W) M + LVI HVW + TA Cyclicity (C) (minutes) MtoM TA to TA LVI to LVI
a NS
=
% Full-term Signifgroup (± SD) icance
34.9 (±9.6)
34.2 (± 11.6)
NSa
5.6 (±3.6) 33.9 (±S.7)
4.7 (±2.3) 27.S (±7.1)
NS 0.02
12.7 (±4.7)
17.4 (±S.6)
0.05
13.2 (±6.0) 2.1 (±4.3) 46.S (±9.4) 39.5 (±S.7)
15.5 (±6.S) 3.5 (±7.2) 51.6 (±9.1) 32.5 (±S.O)
NS NS 0.07 0.02
62.1 (± 12.4) 70.2 (± 15.9) 66.9 (±25.2)
52.5 (± IS.9) 52.9 (±19.6) 50.7 (± IS.6)
0.07 0.02 0.05
not significant.
445
Sleep architecture measures EEG sleep state
Significant differences between PTT and FT groups were noted for percentages of trace alterant (TA) and low-voltage irregular sleep segments (LVI), as shown in Table 3. The PTT group showed a significantly greater percentage of T A (34% vs. 28%, p = 0.02) and less LVI than the FT group (13% vs. 17%, p < 0.05). When the percentages of both quiet sleep and active sleep segments were added together, the quiet sleep percentage remained significantly higher in the PTT than in the FT group. It is notable that the state percentages of mixedfrequency active sleep, indeterminate sleep, high-voltage slow quiet sleep and waking segments were not significantly different between the two groups. Cyclicity
rupture of membranes. The average gestational age at birth was 29.5 weeks. Average Apgar scores at 1 and 5 minutes were 5 and 8. Twelve infants were delivered vaginally and six were delivered by Cesarean section, four for placental abruption or prolapsed cord, one for chorioamnionitis, and one for breech presentation. Six infants were black and 12 were white. Nine males and nine females were included in this group. Sixteen infants had an initial oxygen requirement for an average of 3 days. Two infants had grade I or II intraventricular hemorrhage. Pediatric and neurological examinations for this group were all normal. Maternal histories for the full-term group were unremarkable. Six infants were black, 12 were white. Nine males and nine females were included in this group. Five were delivered vaginally and 13 by Cesarean section, eight for previous Cesarean section or failure to progress, three for signs of fetal distress, and two for cephalopelvic disproportion. Average Apgar scores at 1 and 5 minutes were 6 and 9. All control infants were studied between the second and fourth day of life and demonstrated normal examination at the time of the EEG sleep study.
The length of the ultradian EEG sleep cycle was significantly longer in the PTT group compared to the FT group. The PTT group's sleep cycle length between two consecutive T A segments was 70.2 minutes compared to the FT group's length of 52.9 minutes (p < 0.02). Similarly, the sleep cycle length between two consecutive LVI segments was 66.9 minutes for the PTT group, compared to only 50.7 minutes in the FT group ' (p < 0.05). Sleep continuity measures Arousals
The number and durations of spontaneous episodes of transient electrographic reactivity (i.e. arousals) were significantly lower in the PTT group (Table 4). The average number of arousals was decreased throughout the entire ultradian sleep cycle (i.e. 0.13 vs. 0.2l/minute, p = 0.0 I). However, duration of arousals was also significantly decreased in the quiet sleep segments (i.e. 0.8 seconds vs. 3.0 seconds, p = 0.000). No significant differences were noted between the two groups for sleep latency or efficiency.
TABLE 4. Comparisons of sleep continuity measurements in preterm and full-term infants (using paired t tests) Arousals
Preterm group (± SD)
Full-term group (± SD)
Significance
Average no. arousals/study Average no. arousals/quiet sleep Average duration (seconds)/study Average duration /quiet sleep
0.13 (±0.04) 0.06 (±0.05) 3.2 (±I.3) 0.77 (±0.7l)
0.21 (±O.ll) 0.30 (±0.15) 3.5 (±2.1) 3.0 (±2.0)
0.009 0.000 NSa 0.000
12.3 (± 1.23) 0.S5 (±0.09)
16.1 (±IS.I) O.SI (±0.1O)
NS NS
Sleep latency (minutes) Sleep efficiency a
NS
=
not significant. Sleep, Vol. 15, No.5, 1992
Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
M. S. SCHER ET AL.
446
TABLE 5. Comparisons of body movements and REM during sleep in preterm and full-term neonates (using paired t tests)
Body movements (BM) Total BM Total head (H) Active sleep Quiet sleep Total face (F) Active sleep Quiet sleep Total small (S) Active sleep Quiet sleep Total large (L) Active sleep Quiet sleep Rapid eye movements (REMs) Total REMs/study REMs/minute a NS = not significant.
Preterm group
Full-term group
Minutes (± SO}
Minutes (± SD)
Significance
0.64 (±0.33) 0.13 (±0.06) O.IS (±0.09) 0.04 (±0.05) 0.29 (±0.33) 0.34 (±0.19) 0.22 (±0.51) 0.02 (±0.02) 0.03 (±0.03) 0.007 (±0.02) 0.19 (±O.IO) 0.24 (±0.14) 0.03 (±0.02)
0.S4 (±0.34) 0.25 (±0.16) 0.30 (±O.IS) 0.16 (±O.IS) 0.30 (±0.24) 0.33 (±0.22) 0.26 (±0.36) 0.05 (±0.05) 0.05 (±0.07) 0.04 (±0.06) 0.24 (±0.13) 0.23 (±0.12) 0.16 (±0.14)
0.06 0.007 0.06 0.01
S55.S (±352.3) 5.2 (±2.1)
0.003 0.003
447.7 2.7
(±299.9) (± I.S)
Body movements and REM Significantly fewer body movements and REMs were noted during sleep in the PTT group. As shown in Table 5, the PTT group showed significantly fewer head movements during both active and quiet sleep (0.13 vs. 0.25/minute, p = 0.007). The PTT group had fewer small and large body movements during quiet sleep (i.e. 0.007 vs. 0.04/minute, p = 0.02, and 0.03 vs. 0.16/minute, p = 0.001, respectively). The total score for body movements as well as facial movements, however, was not significantly different from Ff infants (0.64 vs. 0.84, p = 0.06). Rapid eye movements (REMs) were significantly lower in the PTT infants. Reductions were seen both in the total REM counts for the entire 3-hour study period (i.e. 0.447 vs. 0.855, p = 0.003) as well as the REM counts per minute for each active sleep segment (i.e. 2.7 vs. 5.2, p = 0.003). DISCUSSION Our study has delineated both differences and similarities in the sleep architecture and continuity measures of PTT and Ff neonates. Although the basic ultradian sleep rhythm is similar for the PTT infant, the length of the first quiet to quiet sleep interval is increased by almost 113 for the PTT infant. There are also changes noted in the duration of two segments within the sleep cycle: a longer trace alterant quiet sle,~p and a shorter low voltage irregular active sleep segment. The PTT infant also demonstrates fewer and shorter arousals, fewer body movements and fewer REMs. These findings have important implications for an understanding of normal sleep ontogeny for an inSleep, Vol. 15, No.5, 1992 Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
NSa NS
NS 0.06 NS
0.02 NS NS
0.001
fant deprived of the intrauterine environment during the last months of fetal development. Differences in EEG features between PTT and Ff neonates have been previously demonstrated. Longer bursts during trace alterant, more immature patterns for the stated conceptional age, better phase stability and earlier sleep spindle development at older ages have been described for PTT infants (14). Differences in polysomnographic features of sleep for PTT and Ff infants include longer quiet sleep segments and lower correlation among sleep state measures for the PTT group (15-19). The basic order of the ultradian sleep cycle and most of the sleep state percentages are comparable between PTT and Ff groups. Our findings reaffirm earlier reports that quiet sleep is longer in the PTT infant. However, we have now identified that the trace alterant segment, rather than the high-voltage slow segment, contributes to this increase in quiet sleep percentage. In fact, the high-voltage slow segment is brieffor both the Ff and PTT infant. No significant changes were noted for the combined active sleep percentage (i.e. mixed-frequency plus low-voltage irregular), but a shorter low-voltage irregular active sleep segment was noted for the PTT infant. We also reported that the overall EEG sleep cycle length is longer for the PTT and Ff infants as measured by specific EEG sleep patterns. Unlike previous reports, which defined this cycle length on behavioral or polygraphic criteria (20), we defined this biologic rhythm using both EEG and sleep criteria. A number of confounding variables may contribute to the specific differences noted in EEG sleep measures between PTT and Ff infants. Comparatively higher
EEG SLEEP MEASURES IN INFANTS
numbers of arousals and body movements in the FT group may have occurred in part because of the stress of a more recent delivery. These children were born within days of the recording. The experience of delivery may have contributed to greater instability in the child's sleep behavior, resulting in more arousals and body movements. However, this explanation is not supported by our findings of similar state percentages for active sleep, transitional sleep and wakefulness between PTT and FT groups. We should have expected increases in these state percentages in the FT neonate, whose sleep cycle was disrupted by the recent experience of parturition. Later recordings during the second week of life might have helped answer this question, but such studies were not performed. Adaptation effect may have contributed to group differences. That is, the PTT infants had undergone previous recordings with EEG and, in general, were more accustomed to a wide variety of stimuli and manipulations because of their extrauterine experiences in the preceding weeks. Thus, the PTT group may have shown smaller adaptation effects to the process of recording EEG sleep in our laboratory. The entire issue of adaptation is complex and difficult to address methodologically. One way to address the possible confound of adaptation effects would be to use noninvasive monitoring techniques, such as infra-red video recordings. However, these techniques do not permit the more quantitative and state-scoring measurements described in this paper. A third variable concerns the circadian effect on the neonatal EEG sleep cycle. We have not yet analyzed the other 9-hour segments throughout each recording period. Sleep architecture and continuity measures may be affected by a circadian effect (21). Despite a significant decrease in phasic activities (arousal, body movements and REMs) in PTT infants, these differences do not affect the overall ability of the PTT infant to efficiently achieve and maintain sleep similar to the FT infant. There is a predictable ontogeny of body motility during development (22). Myoclonic and large body movements predominate in preterm infants, but small, localized movements occur more frequently in term infants. It has not been previously reported, however, that specific body movements, principally small and large body movements, are fewer in PTT infants during quiet sleep than in FT infants. Only head movements are decreased in all sleep states. These differences in motility between the two groups exist by a term age despite the early occurrence of body movements within active rather than quiet sleep segments in preterm infants as young as 31 weeks conceptional age (23). Although neuronal systems responsible for movements determine when during the sleep cycle such activity will occur (i.e. as early
447
as 12 weeks premature), other neurophysiological mechanisms may determine the quantity expressed in each specific movement category. Few reports have addressed transient electrographic arousal periods on neonatal EEG sleep recordings. Although there is an increase in reactivity to external stimuli during maturation for the preterm infant (24), no comparative studies of PTT and FT infants exist with respect to this particular physiological measure. Our findings of decreased arousal numbers and durations, particularly during quiet sleep in PTT infants, have not been previously reported. Transient electrographic arousals may have an important function by which infants can shift between segments of the sleep cycle. The longer T A and shorter LVI segments may, in part, result from an underdeveloped arousal system in the PTT infant. The PTT infant cannot shift as promptly from T A to LVI segments as the FT infant. Consequently, the T A segment is longer and the LVI segment is shorter for the PTT infant. By contrast, behavioral arousal to wakefulness abruptly ends all sleep activity. However, this physiological response may involve different alerting mechanisms from transient electrographic micro arousals, which only shift between sleep stage segments. Rapid eye movements represent the principal physiological feature identified with the active sleep state, even in preterm neonates as early as 31 weeks conceptional age (25). Although reports document increasing numbers ofREMs during ontogenesis (26), no comparative studies at term between PTT and FT infants have yet been reported. Our finding ofdecreased REMs during active sleep in PTT infants reflects the same decrease in phasic activities, as with body movements and arousals. Neuronal aggregates responsible for these phasic activities may develop differently for the PTT infants. Sleep is the most frequent state of consciousness of the neonate, interrupted by brief arousal and waking periods. Consequently, EEG sleep patterns are the dominant neurophysiological expressions of cerebral function at this early stage of brain development (27,28). Neurophysiological maturation in the brain will occur either in an intrauterine or extrauterine environment. However, differences in sleep development between the PTT and FT infants may reflect different congenital or environmental influences on developmental neurobiological processes. Behavioral differences in sleep architecture are largely resolved by one year corrected age in healthy PTT infants (29). However, the differences noted between PTT and FT infants in sleep architecture and continuity measures suggest the potential for continued divergent adaptation during the first year of postnatal brain development. If specific medical illnesses, toxin exposures or enviSleep. Vol. 15. No.5. 1992
Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
448
M. S. SCHER ET AL.
ronmental conditions continue beyond birth into infancy, deviations from expected patterns of sleep maturation may persist. Acknowledgements: This study was supported in part by grants NS 01110 and NS 26793 to Mark S. Scher from the Scaife Family Foundation and The Twenty Five Club of Magee-Womens Hospital. This paper was presented in part at the June 1991 meeting of the Association of Professional Sleep Societies in Toronto, Canada. Elaine Artinger and Lisa Bonnaure prepared the manuscript.
REFERENCES I. Anders T, Parmelee RA. A manual ofstandardized terminology, technique and criteria for scoring states ofsleep and wakefulness in newborn infants. Los Angeles: UCLA Grant Information Service, NINDS, Neurological Information Network, 1971. 2. Dreyfus-Brisac C. Neonatal electroencephalography. In: Scarpelli EM, Cosmi EV, eds. Review in perinatal medicine, Vol. 3. New York: Raven Press, 1979:397-430. 3. Ellingson RJ. Studies of electrical activity of the developing human brain. In: Himwich WA, EH, eds. The developing brainprogress in brain research. Amsterdam: Elsevier Press, 1964:26-· 53. 4. Harper RN, Leake B, Miyahara L, Mason J, Hoppenbrouwers T, Sterman MB, Hodgman J. Temporal sequencing in sleep and! waking states during the first six months of life. Exp Neurol 1981 ;72:294-307. 5. Thoman EB. Early development of sleeping behaviors in infants. In: Ellis N, ed. Aberrant development in infancy: human and animal studies. Miltsdale, NY: Lawrence Erlbaum Associates, 1975. 6. Kellaway P, Crawley JW. A primer ofelectroencephalography of infants, sections I and II; methodology and criteria for normal. Houston, TX: Baylor University College of Medicine, 1964. 7. Lombroso CT. Neonatal electroencephalography. In: Nieder·· meyer E, Lopez-Desilva F, eds. Electroencephalography brain. Basic principles and clinical applications in related fields. Bal· timore and Munich: Erban and Schwarzenberg, 1989:559-637. 8. Precthl HFR. Behavioral states of the newborn infant. Brain Res 1974;76:185-212. 9. Tharp BR. Neonatal electroencephalography. In: Korobkin R, Guilleminault C, ed. Progress in perinatal neurology. Williams and Wilkins, 1981:31-64. 10. Werner S, Stockard JB, Bickford RJ. Atlas of neonatal electro· encephalography, 2nd ed. New York: Raven Press, 1991. II. Stem E, Parmelee A, Harris MA. Sleep state periodicity in premature and young infants. Dev PsychoI1973;6:357-65.
12. Dubowitz LMS, Dubowitz V, Goldberg C. Clinical assessment of gestational age in the newborn infant. J Pediatr 1970;77:110. 13. SPSS, Inc. Statistical Package for the Social Sciences, Inc. 14. Joseph JP, Lesevre N, Dreyfus-Brisac C. Spatio-temporal organization ofEEG in premature infants and full-term newborns. Electroencephalogr Clin Neurophysiol 1976;40: 153-68. IS. Booth C, Leonard H, Thoman E. Sleep states and behavioral patterns of preterm and full-term infants. Neuropediatrics 1980; 11:354-64. 16. Booth C, Morin W, Waite S, Thoman E. Periodic and nonperiodic sleep apnea in premature and full-term infants. Dev Med Child NeuroI1983;25:283-96. 17. Stem E, Parmelee AN, Akiyama Y, Schultz MZ, Wener WHo Sleep cycle characteristics in infants. Pediatrics 1969;43:65-70. 18. Dreyfus-Brisac C. Ontogenesis of sleep in human prematures after 32 weeks of conceptional age. Dev PsychobioI1970;3:91121. 19. Parmelee A. Neurophysiological and behavioral organization of premature infants in the first months of life. Bioi Psychiatry 1975; 10:501-12. 20. Watt JE, Strongman KT. The organization and stability of sleep states in full-term, preterm, and small for gestational age infants: a comparative study. Dev Psychobiol1985; 18: 151-62. 21. Tomioka K, Tomioka F. Development of circadian sleep wakefulness rhythmicity of three infants. J Interdiscipl Cycle Res 1991 ;22:71-80. 22. Hakamada AS, Watanabe K, Miyazaki S. Development of the motor behavior during sleep in newborn infants. Brain and Dev 1981;3:345-50. 23. Curzi-Daskalova L, Peirano P, Vincent G. Body motility according to sleep states in normal newborn infants: a preliminary study. In: Jones CT, Nathaniel SZPW. The physiological development of the fetus and newborn. London: Academic Press, 1985:817-20. 24. Ellingson RJ. Electroencephalograms of normal full-term newborns immediately after birth with observations on arousal and visual evoked responses. Electroencephalogr Clin Neurophysiol )958;10:31-50. 25. Curzi-Daskalova L, Peirano P, Morel-Kahl F. Development of sleep states in normal premature and full-term newborns. Dev PsychobioI1988;21:431-44. 26. Petre-Quadens 0, de Lee C. Eye movements during sleep: a common criteria oflearning capabilities and endocrine activity. Dev Med Child Neuro11970; 12:730-40. 27. Scher MS. Developmental marker of central nervous system maturation: part I. Pediatr NeuroI1988;4:265-73. 28. Scher MS. Developmental marker of central nervous system maturation: part II. Pediatr NeuroI1988;4:???-??? 29. Anders TF, Keener M. Developmental course of night time sleep/wake patterns in full-term and preterm premature infants during the first year of life. I. Sleep 1985;8: 173-92.
Sleep, Vol. 15, No.5, 1992 Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
© 1992 American Sleep Disorders Association and Sleep Research Society
Comparison of EEG Sleep Measures in Healthy Full-Term and Preterm Infants at Matched Conceptional Ages *tMark S. Scher, *Doris A. Steppe, tRonald E. Dahl, *tShobha Asthana and *Robert D. Guthrie *Departments of Pediatrics and tNeurology, Magee- Womens Hospital and Children's Hospital of Pittsburgh, Pittsburgh, Pennsylvania 15213, U.S.A.; and t.Department of Psychiatry, Western Psychiatric Institute and Clinic, University of Pittsburgh, School of Medicine, Pittsburgh, Pennsylvania 15213, U. S.A.
Summary: Continuous electroencephalogram (EEG) sleep studies were obtained on healthy full-term and preterm infants at matched conceptional ages. Studies were recorded under environmentally controlled conditions. Eighteen healthy preterm infants were matched to 18 full-term infants based on conceptional age, sex, race and socioeconomic class. The initial 3 hours of a 12-hour recording were simultaneously recorded on paper and computer. The visually scored data based on the paper recordings for sleep architecture and continuity measures were studied. Differences in each sleep organization for the preterm infants included the following: a longer ultradian sleep cycle (70 minutes vs. 53 minutes, p = 0.02) was noted. More abundant trace altemant (34% vs. 28%, p = 0.02) and less abundant low-voltage irregular active sleep (13% vs. 17%, p = 0.05) were noted. Although no differences were observed for sleep latency and efficiency, the preterm infants had fewer numbers and shorter durations of arousals, fewer body movements and rapid eye movement (REM) (p < 0.0 I), particularly during quiet sleep. The extrauterine experience or the earlier birth of the preterm infant may influence specific sleep architecture and continuity measures when compared with the sleep of full-term infants who experienced a complete intrauterine gestation. Key Words: EEGsleep- N eonates- Preterm - Fullterm.
For more than three decades, investigators have explored the use of sleep measures as a route to understanding early neonatal development (1-9,17). Studi,es have included longitudinal and cross-sectional designs delineating the maturational patterns of sleep in the preterm infant maturing to term. It has been generally accepted that the sleep cycle in the infant born prematurely is highly variable until at least 36 weeks gestation (11), after which time it will resemble the sleep cycle ofthe appropriate-for-age full-term infant. It has also traditionally been assumed that the ultradian sleep pattern of the full-term infant is the "gold standard" to which the sleep rhythm of the preterm infant at term should be compared (1,17). Even when matched for conceptional age, however, the premature infant has achieved the final aspects of development in an extrauterine environment, with a variety of experiences that may directly impact the development of the brain as expressed by the neonatal sleep cycle.
The present study reports on electroencephalogram (EEG) sleep in two cohorts of infants: a normal group of full-term infants and a group of healthy premature infants at corrected term ages. Groups were matched for sex, race, conceptional age, and socioeconomic class, and EEG sleep measures of architecture and continuity were compared. The primary hypothesis of the study was that despite comparable conceptional ages, preterm infants would show differences in specific sleep architecture and continuity parameters, reflecting the influences of extrauterine development or their premature birth.
METHODS Patient population
Healthy preterm infants of < 32 weeks estimated gestational age (EGA) were recruited from a neonatal population admitted to the neonatal intensive care unit of a large obstetrical center. Selection was based on a Accepted for publication May 1992. Address correspondence and reprint requests to Mark S. Scher, review of maternal and neonatal medical records and M.D., Magee-Womens Hospital, Developmental Neurophysiology Laboratory, Halket and Forbes Avenue, Pittsburgh, Pennsylvania on consultation with the attending neonatologist. Infants with any significant medical illnesses were ex15213, U.S.A.
442
Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
EEG SLEEP MEASURES IN INFANTS cluded, including those with sepsis, respiratory disease, intracranial hemorrhage, or other major organ system illnesses. Infants were enrolled only after 7 days oflife, permitting verification of their healthy medical condition. At least one cranial ultrasound was available to document the absence of significant intracranial hemorrhage or peri ventricular echodensities. When each preterm subject reached a conceptional full-term age (i.e. estimated gestational age at birth plus weeks of life), an appropriate-for-gestational-age fullterm infant was selected from the well-child nurseries. Careful review of the medical records and physical examinations were carried out to verify the healthy status of these full-term infants. Each control subject was matched to a preterm infant at term by conceptional age (within 1 week), sex, race and socioeconomic class (Hollingshead, unpublished). All infants received a neurodevelopmental assessment on the day of each recording (12). Information regarding feeding schedules, sleep behavior and medical care were documented. After completion of the sleep recording portion of the study, all infants were followed clinically for at least 18 months and judged to be healthy, including ageappropriate neurodevelopment milestones. Psychometric testing included Bayley Motor and Mental Performance Scales, Carey Temperament Scales, and Vineland Social Maturity Scales. EEG Sleep Recording Sessions
Twelve-hour EEG sleep Twelve-hour EEG sleep studies were carried out in an environmentally controlled setting in which sound, light, humidity and tactile stimulation were carefully monitored. All infants were studied while they slept in open beds. Preterm infants had received monthly studies from birth until their term conceptional age. The full-term subjects received a single sleep study between 2 and 4 days of life. For analyses comparing preterm subjects with fUll-term control infants as presented in this study, only the term conceptional age data were used.
Recording parameters Continuous recordings began after a feeding and diaper change between 2000 to 2100 hours and ended between 0800 and 0900 hours of the following day. The subject was observed in a prone position, which was the usual sleeping position in the nursery. While the entire 12-hour study was digitized on an Apollo computer workstation (Hewlett-Packard Corp.), the initial 3 hours were simultaneously recorded on paper using a 21-channel EEG machine (Nihon-Kohden,
443
model 4100) for visual scoring. Results in this paper encompass only the visually scored sleep architecture and continuity measures from this 3-hour recording obtained at a term conceptional age. For the 3-hour paper recording, 14 EEG channels were utilized, with 11 cerebral electrodes placed in the standard 10-20 system. N oncerebral channels included one tonic chin electromyogram (EMG) channel, two electrooculographic channels, two respiratory channels (thoracic and abdominal positions), electrocardiogram (EKG), skin and rectal temperature and pulse oximetry. All movements were recorded directly on the paper and into a computer-driven event-marker file operated by the technologist. Sensitivity was selected at 7 m V/ mm, with a time constant of 0.3 and a paper speed of 15 mm/second. The EEG technologist documented all relevant events regarding subject behavior, manipulations (i.e. feedings, diaper changes, etc.) and sources of artifact. A neonatal research nurse was responsible for the clinical care of the infant during each recording session. Sleep and feeding behavior, diaper changes, medication administration and technical comments (i.e. equipment malfunctions, environmental measurements, etc.) supplemented the technologist's comments. All infants were medication-free at the time of the sleep studies.
Data entry and verification procedures Demographic, clinical, technical and EEG sleep data were recorded on hand-written data sheets and then entered into a relational data base (Interbase, Inc.). An electroencephalographer (M.S.S.) assigned a score for EEG sleep state, and arousal number and duration for each minute of the 180 minutes of paper recording. Accurate data entry into the computer was verified. In order to assess scoring reliability, 10% of the records were rescored by one individual (M.S.S.) for state, arousal, motility and REM. There was a 96% agreement with the original scoring of all minutes. Sleep architecture and continuity measures were tabulated for each of the initial 180 minutes of the allnight recording (Table 1). These measures included I-minute numerical values for the following: EEG sleep state; arousal number and duration; total body, facial, head, small and large body movement counts; and rapid eye movement counts (REM). Statistical Analysis
EEG sleep variable definitions Nine measures were defined based on the minuteby-minute tabulated data (Table 1): Sleep, Vol. 15, No.5, 1992
Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
444
M. S. SCHER ET AL. TABLE 1. Sleep architecture and continuity measurements Architecture
Continuity
Phasic events
EEG sleep state (ESS) Mixed-frequency (M) High-voltage slow (HVS) Trace alterant (T A) Low-voltage irregular (LVI) Cycle length (minutes)
Arousals Sleep latency Slleep efficiency
Body movements (BM) Face/suck (F) Head (H) Small (S) Large (L) REMs
EEG sleep state (ESS). Six state categories defined traditional EEG state (I, 10): two active sleep segments (mixed-frequency and low-voltage irregular), two quiet sleep segments (high-voltage and trace alterant), indeterminate sleep and wakefulness. An awake state was defined as 3 minutes of continuous eye-opening associated with an appropriate EEG pattern. Less than 1% of the minutes could not be scored because excessive artifact prevented an electrographic interpretation. Arousals (A). Transient epochs of EEG activity during which a desynchronization of the EEG background activity was noted with more than a 50% reduction illl amplitude and greater than 90% reduction in slow EEG frequency activity. Arousal durations were recorded to the nearest second. Coincidental movements or sources of artifact were also noted. Spontaneous arousals were distinguished from evoked arousals, based on coincident artifact-producing events. Body movements (BM). Total body, facial (including suck), head, small and large body movements were summed for each minute. Rapid eye movements (REMs). The determination of REM was based on synchronized out-of-phase pen deflections in two electrooculographic channels (left and right outer canthi) of at least 20 mv in amplitudl~. Sleep onset. This measure was defined as the first 3 minutes of continuous eye closure after "lights out". Sleep latency (SL). This measure was defined as the time from "lights out" to sleep onset. "Lights out" was determined when the lighting was turned out after the infant was placed in a prone position following a feeding and diaper change. Sleep efficiency (SE). This measure was the ratio of total sleep time (minutes) over the total recording time (minutes). REM latency (RL). This measure was the time from sleep onset to the first REM period. TABLE 2.
Cyclicity (C). This measure was the time (in minutes) between 2 consecutive periods of the same sleep stage (a minimum of 3 consecutive minutes of asleep stage was designated to denote a "period"). For example, cyclicity is the time between quiet to quiet sleep states, trace to trace alterant, etc. Calculations were based on the first cycle after the infant initiated sleep. Statistical analyses
All computer analyses were performed using a statistical package (13) that was configured to our relational data base. Exploratory calculations included tabulation of average numbers and durations of all measures for each subject. Scatterplots and histograms were also used to define outliers and patterns of distribution. Two-tailed, matched-pair t tests were performed on each preterm-full-term pair, matched for conceptional age, sex, race and socioeconomic class. Nonparametric calculations (i.e. Mann-Whitney calculation) were used when distributions were non-normal. Chi-squared calculations were used for categorical data in demographic comparisons.
RESULTS Demographic and clinical data
Eighteen preterm at term (PTT) and full-term (FT) pairs were selected. All neonates were appropriate for gestational age at birth and demonstrated normal growth parameters. Table 2 lists the range of gestational age, conceptional age at the time of the EEG sleep study, birth weight, length and occipital-frontal circumferences. Maternal histories for the preterm group included six women with preeclampsia and 12 with premature
Group comparisons Preterm group
Estimated gestational age (weeks) Conceptional age (weeks) Birth weight (g) Length (cm) Occipital frontal circumference (cm)
Full-term group
Mean
Range
Mean
Range
29.3 40.9 3,698 50.9 36.5
27-32 38-42 2,705-3,950 47-52 34.3-37.5
40.5 40.5 3,591 51.8 34.9
38-43 38-43 2,810-4,020 47.5-53 34.5-37.8
Sleep. Vol. 15. No.5. 1992 Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
EEG SLEEP MEASURES IN INFANTS TABLE 3. Comparisons ofsleep architecture measurements in preterm and full-term infants (using paired t tests) % Preterm group (± SD)
EEG sleep state Mixed (M) High-voltage slow (HVS) Trace alterant (TA) Low-voltage irregular (LVI) Indeterminate sleep (IS) Waking (W) M + LVI HVW + TA Cyclicity (C) (minutes) MtoM TA to TA LVI to LVI
a NS
=
% Full-term Signifgroup (± SD) icance
34.9 (±9.6)
34.2 (± 11.6)
NSa
5.6 (±3.6) 33.9 (±S.7)
4.7 (±2.3) 27.S (±7.1)
NS 0.02
12.7 (±4.7)
17.4 (±S.6)
0.05
13.2 (±6.0) 2.1 (±4.3) 46.S (±9.4) 39.5 (±S.7)
15.5 (±6.S) 3.5 (±7.2) 51.6 (±9.1) 32.5 (±S.O)
NS NS 0.07 0.02
62.1 (± 12.4) 70.2 (± 15.9) 66.9 (±25.2)
52.5 (± IS.9) 52.9 (±19.6) 50.7 (± IS.6)
0.07 0.02 0.05
not significant.
445
Sleep architecture measures EEG sleep state
Significant differences between PTT and FT groups were noted for percentages of trace alterant (TA) and low-voltage irregular sleep segments (LVI), as shown in Table 3. The PTT group showed a significantly greater percentage of T A (34% vs. 28%, p = 0.02) and less LVI than the FT group (13% vs. 17%, p < 0.05). When the percentages of both quiet sleep and active sleep segments were added together, the quiet sleep percentage remained significantly higher in the PTT than in the FT group. It is notable that the state percentages of mixedfrequency active sleep, indeterminate sleep, high-voltage slow quiet sleep and waking segments were not significantly different between the two groups. Cyclicity
rupture of membranes. The average gestational age at birth was 29.5 weeks. Average Apgar scores at 1 and 5 minutes were 5 and 8. Twelve infants were delivered vaginally and six were delivered by Cesarean section, four for placental abruption or prolapsed cord, one for chorioamnionitis, and one for breech presentation. Six infants were black and 12 were white. Nine males and nine females were included in this group. Sixteen infants had an initial oxygen requirement for an average of 3 days. Two infants had grade I or II intraventricular hemorrhage. Pediatric and neurological examinations for this group were all normal. Maternal histories for the full-term group were unremarkable. Six infants were black, 12 were white. Nine males and nine females were included in this group. Five were delivered vaginally and 13 by Cesarean section, eight for previous Cesarean section or failure to progress, three for signs of fetal distress, and two for cephalopelvic disproportion. Average Apgar scores at 1 and 5 minutes were 6 and 9. All control infants were studied between the second and fourth day of life and demonstrated normal examination at the time of the EEG sleep study.
The length of the ultradian EEG sleep cycle was significantly longer in the PTT group compared to the FT group. The PTT group's sleep cycle length between two consecutive T A segments was 70.2 minutes compared to the FT group's length of 52.9 minutes (p < 0.02). Similarly, the sleep cycle length between two consecutive LVI segments was 66.9 minutes for the PTT group, compared to only 50.7 minutes in the FT group ' (p < 0.05). Sleep continuity measures Arousals
The number and durations of spontaneous episodes of transient electrographic reactivity (i.e. arousals) were significantly lower in the PTT group (Table 4). The average number of arousals was decreased throughout the entire ultradian sleep cycle (i.e. 0.13 vs. 0.2l/minute, p = 0.0 I). However, duration of arousals was also significantly decreased in the quiet sleep segments (i.e. 0.8 seconds vs. 3.0 seconds, p = 0.000). No significant differences were noted between the two groups for sleep latency or efficiency.
TABLE 4. Comparisons of sleep continuity measurements in preterm and full-term infants (using paired t tests) Arousals
Preterm group (± SD)
Full-term group (± SD)
Significance
Average no. arousals/study Average no. arousals/quiet sleep Average duration (seconds)/study Average duration /quiet sleep
0.13 (±0.04) 0.06 (±0.05) 3.2 (±I.3) 0.77 (±0.7l)
0.21 (±O.ll) 0.30 (±0.15) 3.5 (±2.1) 3.0 (±2.0)
0.009 0.000 NSa 0.000
12.3 (± 1.23) 0.S5 (±0.09)
16.1 (±IS.I) O.SI (±0.1O)
NS NS
Sleep latency (minutes) Sleep efficiency a
NS
=
not significant. Sleep, Vol. 15, No.5, 1992
Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
M. S. SCHER ET AL.
446
TABLE 5. Comparisons of body movements and REM during sleep in preterm and full-term neonates (using paired t tests)
Body movements (BM) Total BM Total head (H) Active sleep Quiet sleep Total face (F) Active sleep Quiet sleep Total small (S) Active sleep Quiet sleep Total large (L) Active sleep Quiet sleep Rapid eye movements (REMs) Total REMs/study REMs/minute a NS = not significant.
Preterm group
Full-term group
Minutes (± SO}
Minutes (± SD)
Significance
0.64 (±0.33) 0.13 (±0.06) O.IS (±0.09) 0.04 (±0.05) 0.29 (±0.33) 0.34 (±0.19) 0.22 (±0.51) 0.02 (±0.02) 0.03 (±0.03) 0.007 (±0.02) 0.19 (±O.IO) 0.24 (±0.14) 0.03 (±0.02)
0.S4 (±0.34) 0.25 (±0.16) 0.30 (±O.IS) 0.16 (±O.IS) 0.30 (±0.24) 0.33 (±0.22) 0.26 (±0.36) 0.05 (±0.05) 0.05 (±0.07) 0.04 (±0.06) 0.24 (±0.13) 0.23 (±0.12) 0.16 (±0.14)
0.06 0.007 0.06 0.01
S55.S (±352.3) 5.2 (±2.1)
0.003 0.003
447.7 2.7
(±299.9) (± I.S)
Body movements and REM Significantly fewer body movements and REMs were noted during sleep in the PTT group. As shown in Table 5, the PTT group showed significantly fewer head movements during both active and quiet sleep (0.13 vs. 0.25/minute, p = 0.007). The PTT group had fewer small and large body movements during quiet sleep (i.e. 0.007 vs. 0.04/minute, p = 0.02, and 0.03 vs. 0.16/minute, p = 0.001, respectively). The total score for body movements as well as facial movements, however, was not significantly different from Ff infants (0.64 vs. 0.84, p = 0.06). Rapid eye movements (REMs) were significantly lower in the PTT infants. Reductions were seen both in the total REM counts for the entire 3-hour study period (i.e. 0.447 vs. 0.855, p = 0.003) as well as the REM counts per minute for each active sleep segment (i.e. 2.7 vs. 5.2, p = 0.003). DISCUSSION Our study has delineated both differences and similarities in the sleep architecture and continuity measures of PTT and Ff neonates. Although the basic ultradian sleep rhythm is similar for the PTT infant, the length of the first quiet to quiet sleep interval is increased by almost 113 for the PTT infant. There are also changes noted in the duration of two segments within the sleep cycle: a longer trace alterant quiet sle,~p and a shorter low voltage irregular active sleep segment. The PTT infant also demonstrates fewer and shorter arousals, fewer body movements and fewer REMs. These findings have important implications for an understanding of normal sleep ontogeny for an inSleep, Vol. 15, No.5, 1992 Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
NSa NS
NS 0.06 NS
0.02 NS NS
0.001
fant deprived of the intrauterine environment during the last months of fetal development. Differences in EEG features between PTT and Ff neonates have been previously demonstrated. Longer bursts during trace alterant, more immature patterns for the stated conceptional age, better phase stability and earlier sleep spindle development at older ages have been described for PTT infants (14). Differences in polysomnographic features of sleep for PTT and Ff infants include longer quiet sleep segments and lower correlation among sleep state measures for the PTT group (15-19). The basic order of the ultradian sleep cycle and most of the sleep state percentages are comparable between PTT and Ff groups. Our findings reaffirm earlier reports that quiet sleep is longer in the PTT infant. However, we have now identified that the trace alterant segment, rather than the high-voltage slow segment, contributes to this increase in quiet sleep percentage. In fact, the high-voltage slow segment is brieffor both the Ff and PTT infant. No significant changes were noted for the combined active sleep percentage (i.e. mixed-frequency plus low-voltage irregular), but a shorter low-voltage irregular active sleep segment was noted for the PTT infant. We also reported that the overall EEG sleep cycle length is longer for the PTT and Ff infants as measured by specific EEG sleep patterns. Unlike previous reports, which defined this cycle length on behavioral or polygraphic criteria (20), we defined this biologic rhythm using both EEG and sleep criteria. A number of confounding variables may contribute to the specific differences noted in EEG sleep measures between PTT and Ff infants. Comparatively higher
EEG SLEEP MEASURES IN INFANTS
numbers of arousals and body movements in the FT group may have occurred in part because of the stress of a more recent delivery. These children were born within days of the recording. The experience of delivery may have contributed to greater instability in the child's sleep behavior, resulting in more arousals and body movements. However, this explanation is not supported by our findings of similar state percentages for active sleep, transitional sleep and wakefulness between PTT and FT groups. We should have expected increases in these state percentages in the FT neonate, whose sleep cycle was disrupted by the recent experience of parturition. Later recordings during the second week of life might have helped answer this question, but such studies were not performed. Adaptation effect may have contributed to group differences. That is, the PTT infants had undergone previous recordings with EEG and, in general, were more accustomed to a wide variety of stimuli and manipulations because of their extrauterine experiences in the preceding weeks. Thus, the PTT group may have shown smaller adaptation effects to the process of recording EEG sleep in our laboratory. The entire issue of adaptation is complex and difficult to address methodologically. One way to address the possible confound of adaptation effects would be to use noninvasive monitoring techniques, such as infra-red video recordings. However, these techniques do not permit the more quantitative and state-scoring measurements described in this paper. A third variable concerns the circadian effect on the neonatal EEG sleep cycle. We have not yet analyzed the other 9-hour segments throughout each recording period. Sleep architecture and continuity measures may be affected by a circadian effect (21). Despite a significant decrease in phasic activities (arousal, body movements and REMs) in PTT infants, these differences do not affect the overall ability of the PTT infant to efficiently achieve and maintain sleep similar to the FT infant. There is a predictable ontogeny of body motility during development (22). Myoclonic and large body movements predominate in preterm infants, but small, localized movements occur more frequently in term infants. It has not been previously reported, however, that specific body movements, principally small and large body movements, are fewer in PTT infants during quiet sleep than in FT infants. Only head movements are decreased in all sleep states. These differences in motility between the two groups exist by a term age despite the early occurrence of body movements within active rather than quiet sleep segments in preterm infants as young as 31 weeks conceptional age (23). Although neuronal systems responsible for movements determine when during the sleep cycle such activity will occur (i.e. as early
447
as 12 weeks premature), other neurophysiological mechanisms may determine the quantity expressed in each specific movement category. Few reports have addressed transient electrographic arousal periods on neonatal EEG sleep recordings. Although there is an increase in reactivity to external stimuli during maturation for the preterm infant (24), no comparative studies of PTT and FT infants exist with respect to this particular physiological measure. Our findings of decreased arousal numbers and durations, particularly during quiet sleep in PTT infants, have not been previously reported. Transient electrographic arousals may have an important function by which infants can shift between segments of the sleep cycle. The longer T A and shorter LVI segments may, in part, result from an underdeveloped arousal system in the PTT infant. The PTT infant cannot shift as promptly from T A to LVI segments as the FT infant. Consequently, the T A segment is longer and the LVI segment is shorter for the PTT infant. By contrast, behavioral arousal to wakefulness abruptly ends all sleep activity. However, this physiological response may involve different alerting mechanisms from transient electrographic micro arousals, which only shift between sleep stage segments. Rapid eye movements represent the principal physiological feature identified with the active sleep state, even in preterm neonates as early as 31 weeks conceptional age (25). Although reports document increasing numbers ofREMs during ontogenesis (26), no comparative studies at term between PTT and FT infants have yet been reported. Our finding ofdecreased REMs during active sleep in PTT infants reflects the same decrease in phasic activities, as with body movements and arousals. Neuronal aggregates responsible for these phasic activities may develop differently for the PTT infants. Sleep is the most frequent state of consciousness of the neonate, interrupted by brief arousal and waking periods. Consequently, EEG sleep patterns are the dominant neurophysiological expressions of cerebral function at this early stage of brain development (27,28). Neurophysiological maturation in the brain will occur either in an intrauterine or extrauterine environment. However, differences in sleep development between the PTT and FT infants may reflect different congenital or environmental influences on developmental neurobiological processes. Behavioral differences in sleep architecture are largely resolved by one year corrected age in healthy PTT infants (29). However, the differences noted between PTT and FT infants in sleep architecture and continuity measures suggest the potential for continued divergent adaptation during the first year of postnatal brain development. If specific medical illnesses, toxin exposures or enviSleep. Vol. 15. No.5. 1992
Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
448
M. S. SCHER ET AL.
ronmental conditions continue beyond birth into infancy, deviations from expected patterns of sleep maturation may persist. Acknowledgements: This study was supported in part by grants NS 01110 and NS 26793 to Mark S. Scher from the Scaife Family Foundation and The Twenty Five Club of Magee-Womens Hospital. This paper was presented in part at the June 1991 meeting of the Association of Professional Sleep Societies in Toronto, Canada. Elaine Artinger and Lisa Bonnaure prepared the manuscript.
REFERENCES I. Anders T, Parmelee RA. A manual ofstandardized terminology, technique and criteria for scoring states ofsleep and wakefulness in newborn infants. Los Angeles: UCLA Grant Information Service, NINDS, Neurological Information Network, 1971. 2. Dreyfus-Brisac C. Neonatal electroencephalography. In: Scarpelli EM, Cosmi EV, eds. Review in perinatal medicine, Vol. 3. New York: Raven Press, 1979:397-430. 3. Ellingson RJ. Studies of electrical activity of the developing human brain. In: Himwich WA, EH, eds. The developing brainprogress in brain research. Amsterdam: Elsevier Press, 1964:26-· 53. 4. Harper RN, Leake B, Miyahara L, Mason J, Hoppenbrouwers T, Sterman MB, Hodgman J. Temporal sequencing in sleep and! waking states during the first six months of life. Exp Neurol 1981 ;72:294-307. 5. Thoman EB. Early development of sleeping behaviors in infants. In: Ellis N, ed. Aberrant development in infancy: human and animal studies. Miltsdale, NY: Lawrence Erlbaum Associates, 1975. 6. Kellaway P, Crawley JW. A primer ofelectroencephalography of infants, sections I and II; methodology and criteria for normal. Houston, TX: Baylor University College of Medicine, 1964. 7. Lombroso CT. Neonatal electroencephalography. In: Nieder·· meyer E, Lopez-Desilva F, eds. Electroencephalography brain. Basic principles and clinical applications in related fields. Bal· timore and Munich: Erban and Schwarzenberg, 1989:559-637. 8. Precthl HFR. Behavioral states of the newborn infant. Brain Res 1974;76:185-212. 9. Tharp BR. Neonatal electroencephalography. In: Korobkin R, Guilleminault C, ed. Progress in perinatal neurology. Williams and Wilkins, 1981:31-64. 10. Werner S, Stockard JB, Bickford RJ. Atlas of neonatal electro· encephalography, 2nd ed. New York: Raven Press, 1991. II. Stem E, Parmelee A, Harris MA. Sleep state periodicity in premature and young infants. Dev PsychoI1973;6:357-65.
12. Dubowitz LMS, Dubowitz V, Goldberg C. Clinical assessment of gestational age in the newborn infant. J Pediatr 1970;77:110. 13. SPSS, Inc. Statistical Package for the Social Sciences, Inc. 14. Joseph JP, Lesevre N, Dreyfus-Brisac C. Spatio-temporal organization ofEEG in premature infants and full-term newborns. Electroencephalogr Clin Neurophysiol 1976;40: 153-68. IS. Booth C, Leonard H, Thoman E. Sleep states and behavioral patterns of preterm and full-term infants. Neuropediatrics 1980; 11:354-64. 16. Booth C, Morin W, Waite S, Thoman E. Periodic and nonperiodic sleep apnea in premature and full-term infants. Dev Med Child NeuroI1983;25:283-96. 17. Stem E, Parmelee AN, Akiyama Y, Schultz MZ, Wener WHo Sleep cycle characteristics in infants. Pediatrics 1969;43:65-70. 18. Dreyfus-Brisac C. Ontogenesis of sleep in human prematures after 32 weeks of conceptional age. Dev PsychobioI1970;3:91121. 19. Parmelee A. Neurophysiological and behavioral organization of premature infants in the first months of life. Bioi Psychiatry 1975; 10:501-12. 20. Watt JE, Strongman KT. The organization and stability of sleep states in full-term, preterm, and small for gestational age infants: a comparative study. Dev Psychobiol1985; 18: 151-62. 21. Tomioka K, Tomioka F. Development of circadian sleep wakefulness rhythmicity of three infants. J Interdiscipl Cycle Res 1991 ;22:71-80. 22. Hakamada AS, Watanabe K, Miyazaki S. Development of the motor behavior during sleep in newborn infants. Brain and Dev 1981;3:345-50. 23. Curzi-Daskalova L, Peirano P, Vincent G. Body motility according to sleep states in normal newborn infants: a preliminary study. In: Jones CT, Nathaniel SZPW. The physiological development of the fetus and newborn. London: Academic Press, 1985:817-20. 24. Ellingson RJ. Electroencephalograms of normal full-term newborns immediately after birth with observations on arousal and visual evoked responses. Electroencephalogr Clin Neurophysiol )958;10:31-50. 25. Curzi-Daskalova L, Peirano P, Morel-Kahl F. Development of sleep states in normal premature and full-term newborns. Dev PsychobioI1988;21:431-44. 26. Petre-Quadens 0, de Lee C. Eye movements during sleep: a common criteria oflearning capabilities and endocrine activity. Dev Med Child Neuro11970; 12:730-40. 27. Scher MS. Developmental marker of central nervous system maturation: part I. Pediatr NeuroI1988;4:265-73. 28. Scher MS. Developmental marker of central nervous system maturation: part II. Pediatr NeuroI1988;4:???-??? 29. Anders TF, Keener M. Developmental course of night time sleep/wake patterns in full-term and preterm premature infants during the first year of life. I. Sleep 1985;8: 173-92.
Sleep, Vol. 15, No.5, 1992 Downloaded from https://academic.oup.com/sleep/article-abstract/15/5/442/2749312 by guest on 30 May 2018
