Regular Research Paper
J Sleep Res. (2016)
Mapping changes in cortical activity during sleep in the first 4 years of life LUANA NOVELLI1, AURORA D’ATRI1, CRISTINA MARZANO1, E L E N A F I N O T T I 2 , M I C H E L E F E R R A R A 3 , O L I V I E R O B R U N I 4 and LUIGI DE GENNARO1 Department of Psychology, University of Rome ‘La Sapienza’, Rome, Italy; 2Department of Biomedical and Neuromotor Sciences, University of Bologna, Bologna, Italy; 3Department of Life, Health and Environmental Sciences, University of L’Aquila, L’Aquila, Italy; 4Department of Developmental and Social Psychology, University of Rome ‘La Sapienza’, Rome, Italy 1
Keywords brain maturation, children, local sleep, power spectra, regional electroencephalogram Correspondence Luigi De Gennaro, Department of Psychology, Section of Neuroscience, University of Rome ‘La Sapienza’, Via dei Marsi, 78, 00185 Rome, Italy. Tel.: (+39)-06-49917647; fax: (+39)-06-49917711; e-mail:
[email protected] Accepted in revised form 7 December 2015; received 13 July 2015 DOI: 10.1111/jsr.12390
SUMMARY
A coherent body of evidence supports the notion that sleep is a local and use-dependent process. Significant changes in brain morphology and function occur in the first years of life, revealing a postero–anterior trajectory of cortical maturation. On this basis, a recent study demonstrated that regional cortical maturation between early childhood and late adolescence is reflected in regional changes of sleep slow wave activity (SWA) during non-rapid eye movement (NREM) sleep. Our hypothesis is that changes of electroencephalogram (EEG) rhythms during sleep from birth to childhood are also mirrored by parallel regional changes in the EEG rhythms of sleep according to the assumption of a postero–anterior gradient in cortical maturation. We studied all-night EEG of 39 healthy, full-term, infants and children aged between 0 and 48 months, evaluating regional differences in NREM sleep. We confirmed the strictly local nature of sleep with frequency-specific regional differences. Specifically, we found a general shift of maxima of the upper alpha activity from occipital to prefrontal regions, expressed mainly by the ~11 Hz frequency. This shift corresponds to a postero–anterior trajectory of the socalled ‘slow spindles’. The theta and alpha EEG activity of the frontal cortex exhibits a clear, positive, correlation with age. We conclude that specific local differences during NREM sleep, parallel cortical maturation also in the first 4 years of life.
INTRODUCTION A growing body of evidence supports the notion that sleep, from its onset to the stage of awakening, is a local and usedependent process (Ferrara and De Gennaro, 2011). In adults, the onset of sleep occurs initially in frontal areas; a frontal predominance of low-frequency electroencephalogram (EEG) power (0.5–8.0 Hz) persists during the early part of the night, when homeostatic processes are dominant (Marzano et al., 2010). Frontal areas are crucially involved in sleep homeostasis (Ferrara and De Gennaro, 2011). According to the local use-dependent theory (Krueger et al., 2008), this should be interpreted in terms of a higher sleep need of the frontal cortex due to its greater diurnal activity. When do these local mechanisms develop between birth and early childhood? Sleep exhibits substantial changes from ª 2016 European Sleep Research Society
birth to the first 3–4 months of life (Jenni and Carskadon, 2004) and from early childhood throughout adolescence, when important changes in brain morphology and function occur (Campbell and Feinberg, 2009). A salient feature of quiet sleep (i.e. the equivalent of NREM in infants) is the emergence of a peak in the sigma band at the age of 2– 3 months that corresponds to the appearance of sleep spindles (Jenni et al., 2004). There is a stable high level of low-frequency delta activity between 2 and 9 months, and high sigma activity is detectable at the age of 6 months (Jenni et al., 2004). Although several studies have analysed quantitative EEG in neonates, infants and children (Jenni et al., 2004; Korotchikova et al., 2009; Paul et al., 2003; Scher et al., 1994), evidence of local changes during sleep remains scarce. As a notable exception, age-related sleep EEG
1
2
L. Novelli et al.
changes have been mapped from early childhood to late adolescence (2.4–19.4 years) during non-rapid eye movement (NREM) sleep (Kurth et al., 2010). Frequency-specific regional distributions of EEG power were observed in all subjects, indicating how regional cortical maturation from early childhood to late adolescence is reflected in regional changes of sleep slow wave activity (SWA). More specifically, Kurth et al. (2010) reported a shift of SWA along the postero–anterior axis, which parallels findings indicating that cortical maturation starts earlier in posterior areas and spreads rostrally over the frontal cortex. Also, Feinberg et al. (2011), studying the maturational trajectories (spanning ages 9–18 years) of EEG power, showed that delta activity and, to a lesser extent, theta activity become frontally dominant during early adolescence. If cortical maturation from early childhood is mirrored by specific sleep EEG changes, it is likely that the relevant changes in cortical maturation from birth to early childhood are reflected by parallel regional changes of sleep EEG (Huber and Born, 2014). Indeed, pronounced developments in brain structure and behaviour occur during the first 2 years of life (Johnson, 2001). The volume of grey matter increases rapidly up to approximately 4 years of age (Pfefferbaum et al., 1994), while the most rapid changes of white matter occur during the first 2 years [Dubois et al., 2008; (1–4 months); Gilmore et al., 2012; (0–2 years); Kulikova et al., 2015 (1–5 months)]. Maturational trends from posterior to anterior cortical areas have also been already described in preterm neonates (Melbourne et al., 2014). Here we investigate changes of EEG rhythms during sleep from birth to childhood under the assumption that postero– anterior gradients of cortical maturation will be mirrored by parallel regional changes in sleep EEG rhythms. METHODS Subjects Thirty-nine full-term participants (14 females and 25 males) aged between 0 and 48 months (mean age 15.0 months) were enrolled for the purpose of this study. Participants were recruited from paediatric clinics at the Universities of Rome (n = 4) and Padua (n = 17) and through private sources (n = 18). All infants and children had a regular birth course with uncomplicated postnatal adaptation, Apgar score at 5 min between 7 and 10 and a normal physical examination at birth. The inclusion criteria for this study were: (i) no history of severe health problems; (ii) absence of epileptic seizures and mental retardation; and (iii) no ongoing medication. Additionally, all infants and children had no history of major sleep problems as confirmed by sleep polysomnography [apnea–hypopnea index (AHI) < 1; periodic limb movement index (PLMI) < 5]. For the purposes of this study, subjects were divided into four groups, according to their age: group 1 (n = 14): 0–3 months [mean age = 2.05 months, standard
deviation (SD) = 1.69]; group 2 (n = 7): 4–12 months (mean age = 5.70 months, SD = 1.30); group 3 (n = 11): 13– 24 months (mean age = 20.74 months, SD = 6.05); and group 4 (n = 7): 25–48 months (mean age = 41.32 months, SD = 7.50). All parents signed a consent form before performing the study. The study was approved by the local Institutional Ethics Committee and conducted in accordance with the Declaration of Helsinki. Procedure Each polysomnographic study was performed at home at regular nightly bedtime, in the interval between 20:00 and 08:00 hours. This time window was chosen because there are substantial differences in the amount and distribution of sleep within the 24-h period, depending on the age of the children. For example, newborns sleep during most of the day and night, whereas this phenomenon tends to disappear during the first year of life with the consolidation of night sleep and the circadian organization of naps (Kleitman, 1987). All children were following regular napping schedules as appropriate for their age, and no attempt to modify such schedule was made. Parents were asked to: note any periods in which infants and children were being fed; keep the environment in which the child was asleep in a state of low/absent brightness whenever possible; maintain a stable and comfortable room temperature; maintain a quiet sleeping environment for the child; and take the child in their arms only if absolutely necessary. All recordings were performed at home in the routine sleep environment. Due to the different settings in which recordings were acquired and the intrinsic characteristics of polyphasic sleep after birth, the values of total sleep time (TST) span a wide range. Table 1 reports the main polysomnographic variables of the four age groups. Polysomnographic (PSG) recording The unipolar PSG montage included: EEG (at least 12 channels with the electrodes positioned according the 10–20 International System: Fp1, Fp2, Fz, Cz, Pz, Oz, F3, F4, C3, C4, O1, O2); left and right electro-oculogram (EOG), chin electromyogram (EMG), left and right tibialis EMG, electrocardiogram (ECG), thorax and abdominal effort, nasal cannula, peripheral oxygen saturation, pulse and position sensors. Reusable sintered Ag/AgCl electrodes were used for the EEG signals. EEG abrasive skin preparation gel (Nuprep; D. O. Weaver & Co., Aurora, CO, USA) was applied using a cotton bud before positioning the electrodes. In order to obtain a reliable signal for sleep monitoring, the adhesive paste Ten20 (D. O. Weaver and Co.) was also applied. Disposable Ag/AgCl electrodes Neuroline 710 (Ambu A/S, ª 2016 European Sleep Research Society
Sleep EEG differences in the first 4 years of life
3
Table 1 Means and SE of polysomnographic variables in the four age groups
TSP (min) TST (min) QS/NREM (%) AS/REM (%) WASO (% SPT) N1 N2 N3
0–3 months Mean (SE)
4–12 months Mean (SE)
13–24 months Mean (SE)
268.4 216.2 64.5 35.5 10.4 – – –
349.6 307.4 76.2 23.8 7.7 1.6 37.2 37.4
349.2 327.2 81.7 18.3 3.6 1.2 49.3 31.2
(72.0) (52.1) (3.6) (3.6) (3.3)
(93.8) (79.9) (5.2) (5.2) (3.1) (0.8) (12.8) (9.9)
(109.7) (100.6) (3.8) (3.8) (1.8) (0.7) (10.0) (10.4)
25–48 months Mean (SE) 359.7 325.4 85.2 14.8 8.0 4.7 60.8 19.7
(74.7) (69.1) (3.4) (3.4) (2.8) (1.4) (5.6) (5.3)
TSP, total sleep period; TST, total sleep time; QS, quiet sleep; AS, active sleep; SE, standard error; WASO, wake after sleep onset; REM, rapid eye movement; NREM, non-REM; SPT, sleep period time.
Ballerup, Denmark) were employed for the EMG and EOG signals. A linked-mastoid reference (A1–A2) was used for the EEG and EOG signals. Electrode impedance was below 10 kOhm at the start of each recording. An Embla titanium portable polygraph was used for recordings. Sleep EEG signals were band-pass filtered between 0.15 Hz and 220 Hz using a 40 dB per decade anti-aliasing hardware input filter. Data were collected by a 12-bit resolution A/D with a sampling frequency of 8192 Hz per channel. A further 40 dB per decade anti-aliasing digital filter was applied to the low-pass filtered data below 120 Hz. The digitized and filtered EEG was down-sampled to 256 Hz and archived in European Data Format (EDF). Sleep stage scoring In accordance with standard guidelines (Grigg-Damberger et al., 2007), the scoring of sleep stages was performed following the procedures outlined by Anders et al. (1971) at age 0–8 Hz frequency bins, while correlation between EEG power and age was always significant for the adjacent lower frequencies (7.00 Hz: r = 0.68; 7.25 Hz: r = 0.63; 7.50 Hz: r = 0.60; 7.75 Hz: r = 0.57). On the other bound of this range, correlation was significant at 11.25 Hz (r = 0.55; P = 0.0003), 11.50 Hz (r = 0.58; P = 0.0002) and 11.75 Hz (r = 0.60; P < 0.0001). Correlation coefficients were not significant in the upper bins (>12.00 Hz), excluding the 12 Hz frequency (r = 0.58). Hence, the positive relation between the alpha band and age is due mainly to the EEG activity at ~11 Hz frequency. DISCUSSION The regional EEG differences of NREM sleep show specific changes of activity
