Stress, sleep and autonomic function in healthy adolescent girls and boys: Findings from the NCANDA Study

Abstract

Objectives:

Starting in adolescence, female sex is a strong risk factor for the development of insomnia. Reasons for this are unclear but could involve altered stress reactivity and/or autonomic nervous system (ANS) dysregulation, which are strongly associated with the pathophysiology of insomnia. We investigated sex differences in the effect of stress on sleep and ANS activity in adolescents, using the first night in the laboratory as an experimental sleep-related stressor.

Design:

Repeated measures (first night vs. a subsequent night) with age (older/younger) and sex (males/females) as between factors.

Setting:

Recordings were performed at the human sleep laboratory at SRI International.

Participants:

106 healthy adolescents (Age, mean±SD: 15.2±2.0 years; 57 boys).

Measures:

Polysomnographic sleep, nocturnal heart rate (HR) and frequency-domain spectral ANS HR variability (HRV) indices.

Results:

Boys and girls showed a first night effect, characterized by lower sleep efficiency, lower %N1 and %N2 sleep, more wake after sleep onset and %N3 sleep, altered sleep microstructure (increased high-frequency sigma and Beta1 electroencephalographic activity), and reduced vagal activity (p<0.05) on the first laboratory night compared to a subsequent night. The first night ANS stress effect (increases in HR and suppression in vagal HRV during rapid eye movement sleep) was greater in girls than boys (p<0.05).

Conclusions:

Sleep and ANS activity were altered during the first laboratory night in adolescents, with girls exhibiting greater ANS alterations than boys. Findings suggest that girls may be more vulnerable than boys to sleep-specific stressors, which could contribute to their increased risk for developing stress-related sleep disturbances.

Introduction

Poor sleep quality affects physical and mental health (1). About one-third of the population complains about having trouble sleeping, with 10% meeting the diagnostic criteria for an insomnia diagnosis (2) and with more women than men suffering from insomnia across the lifespan (3). While epidemiological studies show that women frequently report poorer sleep quality and more sleep disturbance than men throughout their lives (4, 5), polysomnographic (PSG) indices of sleep suggest better objective sleep quality in women (6, 7), characterized by longer total sleep duration, more slow wave sleep (N3 sleep), and fewer arousals during the night compared to men (7). Differences between objective and subjective sleep constructs, as well as greater manifestation of anxiety, low mood, and negative emotion in women, which are known to affect sleep perception, may drive these discrepancies and contribute to a higher prevalence of insomnia in women (8). As shown with animal models, differences in the sex hormone milieu can also contribute to sex differences in sleep as well as changes in sleep according to reproductive stage (6). Indeed, women are susceptible to the development of insomnia at particular life-stages, including adolescence and menopause (9, 10).

Insomnia disorder frequently emerges in adolescence (10), putting adolescents at risk for developing mental health and medical conditions later in life (11). Adolescent insomnia needs to be contextualized within the normal developmental changes in sleep and circadian bioregulatory processes occurring across adolescence. There are dramatic reductions in slow wave activity (reflecting changes in brain reorganization), along with a shift in chronotype toward preferences for evening activities in late compared to early adolescence (12). In addition, external factors such as electronic media use at bedtime (i.e. watching TV, using mobile phones) (13), consumption of stimulants (e.g., caffeine, energy drinks) (14), and early school start times, can all impact sleep in adolescents.

The life-long preponderance of insomnia in the female sex becomes evident in the post pubertal period with a male:female prevalence of 1:2. The biological basis for this sex difference is still unknown. It is hypothesized that some individuals have a trait-like vulnerability to insomnia driven by an augmented response to stress (15). Stress exposure predicts insomnia, particularly in individuals with high sleep reactivity who tend to have difficulty sleeping in stressful situations (16).

The autonomic nervous system (ANS) is a primary system involved in the stress response (17) and is considered to play a key role in the pathophysiology of insomnia (18). Compared to wakefulness, sleep is characterized by a shift in ANS sympathovagal balance towards vagal dominance. There are also sleep-stage dependent changes in ANS activity (greater vagal dominance in NREM compared to REM sleep) and transient ANS shifting (increased sympathetic activity and vagal withdrawal) in response to arousals from sleep (19). Poor sleep affects ANS activity and vice versa (19). In insomnia, altered ANS activity, such as elevated sympathetic ANS activity, has been reported during the falling asleep process and throughout the sleep period, and is thought to interfere with sleep processes (20, 21). Also, an altered relationship between sleep electroencephalographic (EEG) activity and cardiac function (decreases in coherence, i.e. the amount of linear coupling between measures of cardiac vagal activity and EEG Delta activity) has been shown in insomnia sufferers (22), suggesting an altered bidirectional communication between central and autonomic function during sleep in this population (19).

ANS flexibility is particularly important for the adaptive stress response. In healthy adolescents (12–22y), our team recently found significant age- and sex- dependent changes in sleep ANS activity, with adolescent girls exhibiting an upregulation of the ANS with increasing age and a greater HR acceleration in response to sleep stage shifting at night when compared to adolescent boys (23). It is possible that different trajectories of ANS maturation across adolescence in boys and girls may put girls at high risk for stress-related sleep disruption. Adolescent girls compared with boys tend to be exposed to more stressors (24), have higher stress levels, and also have a greater stress reactivity (25–27). Other studies revealed significant sex differences towards a greater experience of depressive symptoms and stronger reactions to similar stress events in adolescent girls compared to boys (27). To our knowledge, it is unknown whether there is also a sex difference in sleep autonomic stress reactivity in adolescents.

This study aimed to investigate sex- and age- specific differences in stress-related sleep disturbances (vulnerability to sleep disruption) in a sample of healthy adolescents by using a well-known sleep-disrupting challenge, i.e. the first night effect (28). Due to the unfamiliar surroundings, the first night in a sleep laboratory is considered to be more stressful than subsequent nights, being characterized by greater sleep disruption as documented mostly in healthy young adults (29–31). Even though an earlier study investigated a mixed adolescent and young adult population and found effects similar to those described above (29), studies investigating the first night effect in adolescent samples are rare.

In the current study, PSG sleep macrostructure, microstructure, and nocturnal cardiac ANS profiles were compared between the first night in the laboratory and a subsequent recording night. We hypothesized that adolescents would exhibit alterations in sleep macrostructure, microstructure, and ANS activity during the first night compared to the subsequent recording night, with girls showing magnified sleep and ANS stress-responses (greater sleep disruption and ANS upregulation) as compared to boys.

Methods

Participants

106 healthy adolescents (49 girls; Age, mean ± SD: 15.2 ± 1.9 y), from the National Consortium on Alcohol and Neurodevelopment in Adolescence (NCANDA) sleep sub-study at SRI International, Menlo Park, California participated. NCANDA is a longitudinal multi-site study involving yearly assessments of brain structure and function, sleep behavior, psychosocial functioning, and clinical measures in adolescents. Details about NCANDA methodology are described elsewhere (32). The study was approved by the Institutional Review Board at SRI International. All participants were recruited from the local community and compensated for their participation.

Participants were screened by a trained clinical expert using the Semi-Structured Assessment of the Genetics of Alcoholism (SSAGA) (33) and were determined to be free of major mental (e.g. depression) and physical (e.g. diabetes, cancer) conditions. None of the participants had insomnia (self-reported) or other sleep disorders (e.g. apnea), as confirmed by in-lab clinical PSG sleep evaluations. Sample characteristics were obtained from data release version: NCANDA_DATA_00010_V2, and are provided in Table A.1. All participants completed the Pubertal Development Scale (PDS) (34), a well-validated self-administered instrument to noninvasively assess pubertal status (from prepubertal to postpubertal); scores correlate strongly with physical exams of pubertal development (35). Body mass index was calculated from measurements of weight and height while depressive symptoms were assessed with the Center for Epidemiologic Studies Depression Scale (CES-D) (36). CES-D scores can range from 0 to 60, with values above 15 corresponding to clinical symptoms of depression. Sleep quality was assessed using the Pittsburgh Sleep Quality Index (PSQI) (37), which is a valid self-administered instrument to assess sleep quality. Scores can range from 0 to 21, with higher scores indicating worse sleep quality.

Table A.1.

Characteristics of the sample. Data are reported as numbers or mean ± standard deviation (SD), separately for boys and girls.

	Boys	Girls
	Mean ± SD	Mean ± SD
Sample, No.	57	49
Age, years	15.0 ± 1.9	15.3 ± 2.0
Caucasian, No.	45	37
BMI, kg × m−2	21.4 ± 4.1	21.4 ± 4.7
PDS, total scores	2.7 ± 0.7	3.3 ± 0.7*
PSQI, total scores	3.8 ± 2.2	4.1 ± 1.8
CES-D, total scores	5.1 ± 3.5	7.2 ± 7.4

^*, significant (p < 0.05) group differences. Body mass index (BMI), Pubertal Development Scale (PDS), Pittsburgh Sleep Quality Index (PSQI), Center for Epidemiologic Studies Depression Scale (CES-D).

Procedure

All participants underwent a laboratory PSG adaptation/clinical first night and a subsequent PSG recording night at the Human Sleep Lab at SRI International. Nights were scheduled at the participants’ convenience, meaning that the interval between the two nights was variable between subjects (mean ± SD, 25.0 ± 19.3 days). All in-lab overnights were scheduled on weekdays, and participants were asked to maintain their usual sleep schedule for at least 5 days before the overnight. Bedrooms were sound-attenuated and temperature-controlled. Participants chose their bedtimes and wake-up times in the lab. At each lab visit, all participants had to have negative breath alcohol tests (S75 Pro, BACtrack Breathalyzers, San Francisco, CA, USA) and urine drug tests (10 Panel iCup drug test kit, Instant Technologies, Inc., Chicago, IL, USA) for the recording to continue. Girls were scheduled irrespective of menstrual cycle phase. The majority of girls (n=38) were postmenarchal but only 5 girls had detectable progesterone levels in saliva, consistent with the luteal phase of their menstrual cycle.

Sleep and cardiac autonomic assessment

Standard PSG was performed using the Compumedics Grael HD-PSG system (Compumedics, Abbotsford, Victoria, Australia). Sleep stages, awakenings and brief arousals were manually scored by experienced scorers in 30-s epochs according to the American Academy of Sleep Medicine (AASM) guidelines (38). Time in bed (TIB, min), was calculated as the time between lights-out and lights-on. Total sleep time (TST, min) was calculated as the time spent asleep after sleep onset. Sleep efficiency (SE; %) was calculated as TST / TIB * 100, and sleep onset latency (SOL, min) was defined as the time from lights-out until the first epoch of N2 sleep. Total amount of wake after sleep onset (WASO) was also calculated. Time spent in each sleep stage (N1, N2, N3 and rapid-eye-movement [REM]) was computed and reported as percentage of TST. The number of awakenings per hour of sleep (awakening index) was calculated. Before statistical analysis, SOL and WASO were log transformed to improve normality.

Sleep EEG power spectra were analyzed within non-REM (NREM) sleep (N2+N3), matching ANS analysis. Spectral power was calculated from central electrodes (C3, C4) using the EEGLAB toolbox (39) for MATLAB (MathWorks, Natick, MA, USA), as previously described (40). EEG was re-referenced to the average mastoid and filtered (0.3–36 Hz). Fast Fourier Transform analysis was conducted on each arousal-free 30s epoch using a 4s sliding Hanning window to calculate power density values with 0.125 Hz resolution. Power density (μV²/Hz) values were calculated across delta (0.3– <= 4 Hz), theta (>4– <= 8 Hz), alpha (>8– <= 12Hz), sigma (>12– <= 15 Hz), low beta (>15– <= 23 Hz) and high beta (>23– <=30 Hz) frequency bands, and EEG relative power (ratios) calculated for each of the frequency bands as a function of the EEG total power (1Hz to <=30Hz), and then averaged across C3 and C4.

Cardiac ANS activity (HR and HRV outcomes) was calculated using artifact- and arousal-free, 2-minute bins of undisturbed REM or NREM (N2 and N3 combined) sleep sampled across the night according to de Zambotti et al. (23). R-waves were automatically detected from the ECG signal (512 Hz sampled) and manually checked, and normal-to normal inter-beat intervals (ms) were calculated. HR (bpm), and frequency-domain HRV measures, including narrow absolute HF power (HF_na, ms²/Hz (41)) indicating vagal activity and normalized HF power (HF_nu: HF absolute power divided by the total power [sum of high and low frequency absolute power]) indicating vagal dominance, were calculated separately for REM and NREM sleep and then averaged across the night for each sleep stage. In this analysis of stable sleep epochs, potentially direct effects of sleep disturbance on HR and HRV were limited. Before statistical analysis, HF_na was log transformed to improve normality. In a second analysis, HR (bpm) was calculated across the first 7 hours of the night (minimum time available for analysis guaranteeing 100% data availability across the sample, which is in line with previous analyses of sleep ANS activity in a larger NCANDA sample) (23), irrespective of the presence of arousals, awakenings, and sleep stage shifting. In this analysis, HR reflects combined effects of sleep stages and first-night stress on ANS activity.

Statistical analyses

To analyze the effect of first-night laboratory stress on sleep macro- and micro-structure and ANS function, repeated measurement GLM’s were employed. In all models, night (first night vs. subsequent recording night) was used as the within-participant-factor; sex (male vs. female) and age (old vs. young) were used as between-participant-factors. Age was included as a dichotomous variable to compare older vs. younger participants. The variable was generated by a median split resulting in two equally distributed groups: older participants (n=51, M=16.8 years, ± 1.0 years, age range: 15.1 – 19.9 years) and younger participants (n=52, M=13.4 years, ± 0.8 years, age range: 12.1 – 14.8 years).

We explored the relationship between BMI and HRV and PSG responses to the first night stress effect. None of the relationships was significant (p >.01) and therefore BMI was not included in the models.

For the analysis of arousal free, all-night sleep HR and HRV frequency domain metrics, sleep stage (REM, NREM) was included in the models as a within-participant-factor. To investigate the first night stress effect on the night-time HR profile irrespective of sleep stage composition and presence of awakenings and arousals, a separate GLM model, night × sex × age × time (h1, h2, h3, h4, h5, h6, h7) was used. All analyses were performed using IBM SPSS 25 Statistics Software (https://www.ibm.com/products/spss-statistics). A p-value of 0.05 was considered significant. When appropriate, Fisher’s Least Significant Difference (LSD) procedure was used to conduct post-hoc comparisons.

Three participants were identified as statistical outliers (exceeding mean ±3SD) and excluded from HRV analyses.

Results

PSG sleep macro- and micro-structure

Sleep macro- and micro-structure variables are provided separately for boys and girls, for each night, in Table A.2. Detailed analyses of age and sex differences in PSG data in a larger sample of NCANDA participants are described elsewhere (42). Here, we focus on the first night stress effect and interactions between stress, sex, and age. In both boys and girls, SE (F_1,105 = 10,6, p = 0.002) was lower and WASO (F_1,105 = 10.0, p = 0.002) was higher during the first night compared to the subsequent recording night. The first night was also characterized by lower percentages of N1 (F_1,105 = 6,1, p = 0.015) and N2 sleep (F_1,105 = 6,8, p = 0.010), and a higher percentage of N3 sleep (F_1,105 = 13,7, p = 0.001), compared to the subsequent recording night. Analysis of EEG power spectral data indicated that NREM low beta (F_1,105 = 4,9, p = 0.029) and sigma (F_1,105 = 4.7, p = 0.032) power were elevated during the first night compared to the subsequent recording night.

Table A.2.

Polysomnographic sleep macro- and microstructure variables on the first night in the laboratory and subsequent recording night in the laboratory. Data are reported as mean ± standard deviation (SD), separately for boys (N=57) and girls (N=49). Data are provided as mean and standard deviation (SD) of the mean.

	Boys		Girls
	First night	Subsequent night	First night	Subsequent night
	Mean ± SD	Mean ± SD	Mean ± SD	Mean ± SD
TIB, min	485.3 ± 36.7	482.5 ± 41.7	481.0 ± 44.7	474.1 ± 53.2
TST, min	433.5 ± 40.1	443.5 ± 47.3	435.4 ± 44.2	437.5 ± 51.3
WASO, min	38.2 ± 29.4	27.5 ± 21.2	29.7 ± 23.5	20.9 ± 16.4*
SE, %	89.4 ± 6.3	91.9 ± 5.9	90.6 ± 5.4	92.3 ± 4.7*
SOL, min	17.2 ± 12.2	14.5 ± 14.4	19.9 ± 13.1	20.8 ± 17.3
N1, %TST	5.3 ± 3.0	5.7 ± 3.4	4.6 ± 2.7	5.4 ± 2.4*
N2, %TST	43.8 ± 8.6	46.5 ± 8.2	48.0 ± 8.2	49.4 ± 6.1*
N3, %TST	30.7 ± 9.6	27.5 ± 8.4	26.4 ± 8.0	24.3 ± 6.5*
REM, %TST	20.1 ± 4.5	20.2 ± 4.7	21.1 ± 4.8	20.8 ± 4.6
Awakening Index	0.3 ± 0.1	0.3 ± 0.1	0.2 ± 0.1	0.2 ± 0.1
EEG delta, relative power	0.935 ± 0.017	0.935 ± 0.016	0.931 ± 0.020	0.933 ± 0.019
EEG theta, relative power	0.041 ± 0.009	0.041 ± 0.009	0.043 ± 0.011	0.042 ± 0.011
EEG alpha, relative power	0.013 ± 0.006	0.013 ± 0.006	0.012 ± 0.004	0.114 ± 0.004
EEG sigma, relative power	0.010 ± 0.001	0.009 ± 0.005	0.123 ± 0.007	0.012 ± 0.007*
EEG low beta, relative power	0.001 ± 0.001	0.001 ± 0.001	0.012 ± 0.001	0.001 ± 0.001*
EEG high beta, relative power	0.001 ± 0.001	0.001 ± 0.001	0.001 ± 0.001	0.001 ± 0.001

^*, significant night main effects (p < 0.05). Time in bed (TIB), total sleep time (TST), wake after sleep onset (WASO), sleep efficiency (SE), sleep onset latency (SOL), rapid-eye-movement- (REM), elecroencephalography (EEG). Awakening index is calculated as number of awakenings per hour of sleep.

No significant interaction effects between the first night stress and sex or age were found.

Sleep autonomic nervous system activity in artifact-free NREM and REM sleep

There were significant main first night effects for HF_nu (F_1,102 = 4,2, p = 0.043) and HR (F_1,102 = 4,8, p = 0.031), in addition to a trending main effect for HF_na (F_1,102 = 3,9; p = .051), indicating lower HF_nu and HF_na, and higher HR during undisturbed NREM and REM sleep on the first night compared to the subsequent recording night.

A significant night × sex interaction for HR (F_1,102 = 4,4, p = 0.039) indicated that only in girls, HR was higher during the first night compared to the subsequent recording night (p = 0.005; Figure A.1). Further, a significant interaction effect was found between night, sex, and sleep stage for HF_na (vagal activity) (F_1,102 = 5,6, p = 0.020), indicating that only in girls, HF_na was lower during REM sleep on the first night compared to the subsequent recording night (p = 0.040; Figure A.1).

Figure A.1.

Mean heart rate (HR) and high frequency heart rate variability in narrow absolute power (HF_na) during REM (left) and NREM (right) sleep in boys and girls for the first night (black) and subsequent recording night (grey). Error bars depict standard errors. Only girls showed a higher HR during the first night as compared to the subsequent recording night (significant night*sex interaction), and a significant reduction in vagal tone during REM sleep on the first night compared to the subsequent recording night (night*sex*sleep stage interaction).

There were main effects of sleep stage for HR (F_1,102 = 79,9, p = 0.001), which was higher in REM sleep, and for HF_na (F_1,102 = 49,2, p = 0.001) and HF_nu (F_1,102 = 285,3, p = 0.001), which were lower in REM compared to NREM sleep in both male and female adolescents.

Night-time HR pattern irrespective of sleep stage composition, arousals and awakening

Averaged hourly HR patterns for the first 7 hours of the night irrespective of sleep stage composition, arousals, and awakenings are shown for the first and subsequent recording nights in Figure A.2. Analysis of hourly averages of HR showed a significant main first night effect (F_1,101 = 5.5, p = 0.021) and an interaction effect of night × hour (F_1,101 = 3.4, p = 0.002). Post hoc tests indicated that HR was particularly accelerated on the first night compared to the subsequent recording night in the first (p = 0.004) and second (p = 0.009) hours of the night. In addition, on both nights, HR declined significantly from h1 to h7, in both boys and girls (p < 0.01).

Figure A.2.

Heart rate (HR) pattern across the first 7 hours of the night, irrespective of awakenings, arousals, and sleep stage in boys (left) and girls (right) for the first night (black) and subsequent recording night (grey). Error bars depict standard errors.*, significantly higher during the first night compared to the subsequent recording night (p < 0.05); ^δ, significantly lower than the first hour of the night, for both first and subsequent recording night (p < 0.01)

Discussion

Here, we examined sleep and ANS activity during the first night in the sleep lab, a well-known laboratory stressor, compared to a subsequent overnight recording in a group of healthy adolescent boys and girls. Our results indicate that participants had a more disturbed sleep (more wakefulness and lower sleep efficiency), altered sleep composition (reduced %N1 and %N2 sleep and more %N3 sleep), increased NREM EEG sigma and low beta relative powers, and increased HR and altered ANS activity (reduced vagal activity) during sleep on the first night relative to a subsequent night. While sleep responses to the first night were similar between boys and girls, sex differences in the ANS responses were evident. Only girls showed a significant increase in nocturnal HR, as measured during undisturbed REM and NREM sleep, in association with greater cardiac vagal suppression particularly during REM sleep, on the first night compared to the subsequent night. Results suggest a greater ANS sensitivity to the first time sleeping in a new environment (considered to be a stressful situation) in girls compared to boys.

The stress effect of the first night in the laboratory on sleep macrostructure is a broadly reported phenomenon in healthy and clinical adult groups (29, 43–45). As shown in an early study by Agnew et al. (29), adolescents (16 years and older) and young adults had more WASO and a poorer SE during the first night in the lab as compared to a subsequent overnight. The first night effect on sleep in children and younger adolescents is less studied. Scholle et al. (46) examined 131 children and adolescents with suspected sleep disordered breathing and found more awakenings and a lower SE in older adolescents (14–17 years old) during the first night than a subsequent night in the laboratory, reflecting a similar first night effect in adolescents as observed in adults, and similar findings to our results in healthy adolescents. Our finding of a greater percentage of slow wave sleep (%SWS) on the first night despite the increased wakefulness is surprising and may reflect differences in distribution of SWS. For example, in our study the between-night differences in the hourly wake/N3 ratio follow different patterns for the first (h1: 0.56, h2: 0.12, h3: 0.22, h4: 0.36, h5: 0.35, h6: 0.53, h7: 0.52) and subsequent (h1: 0.49, h2: 0.15, h3: 0.14, h4: 0.26, h5: 0.29, h6: 0.29, h7: 0.46) nights, toward greater between-night divergencies in the second half of the night. A recent investigation, focusing on the first night effect in children and adolescents with suspected sleep disordered breathing, also reported higher amounts of SWS in adolescents aged 13 to 17 during the first night (47). However, the reasons for this are still unclear and further studies are needed to investigate potential differences in homeostatic pressure dissipation as a function of stress.

We also examined the first night stress effect on sleep microstructure, which has not received as much attention in the literature. Specifically, the participants in our study had more NREM sigma and low beta (15–23 Hz) activity but no other differences in spectral activity during the first compared to the subsequent night. Sigma activity encompasses the fast spindle frequency band and partly reflects characteristics of sleep spindles, hypothesized to promote the stabilization of sleep (see 48). Possibly, the increased EEG activity in the sleep spindle band reflects an attempt to stabilize sleep in the novel environment. Our finding of increased low beta EEG activity on the first night is similar to that of a small study in adults, which analyzed similar low beta ranges (15–24Hz) to our study, although the effect they found was isolated to the posterior EEG sites (45) whereas we examined EEG activity at central sites. Activity in the EEG beta range is thought to reflect cortical hyperarousal and has been shown by others to be increased in adults with insomnia (49) in association with high levels of perceived stress (50) and following a pre-sleep anticipatory stressor (51). However, the beta range expression of hyperarousal (low- and high-beta cut-offs) varies widely across these studies, making the results challenging to compare. Further research is needed to explore the manifestation of hyperarousal within the beta range of the sleep EEG in response to stress.

Few studies have investigated the first night effect on sleep HR and ANS activity in adults or adolescents. In a recent investigation of adult women, Virtanen et al. (52) found a small increase in vagal activity, as measured by non-linear HRV analysis, from the first night to a second, consecutive night. In contrast, despite showing a first night stress effect on sleep macrostructure, Scholle et al. (46) reported no significant differences in HR between the first night (stress) and the second night in older adolescents (14–17 years old). However, authors detected a significant decrease in HR from the stress to the no-stress night in children and younger adolescents (7–13 years old). Our results show altered HR and vagal ANS measures on the first night: while HR showed the expected nocturnal decline across the night (41, 53), it was higher on the stress night, particularly in the first couple of hours, as compared to the no-stress night, and particularly in girls. The discrepancies between our study and Scholle et al. (46) could be due to differences in methodology and populations studied. We studied healthy adolescents on a first night in the laboratory compared with a subsequent nonconsecutive night. On the other hand, Scholle et al. (46) studied children and adolescents with suspected sleep-related respiratory disorders on two consecutive nights. Differences between studies could also relate to the sex × stress interaction effect that we found, which indicated that girls showed a greater ANS response to the first night than did boys. Scholle et al. (46) did not investigate possible sex differences.

Our hypothesis that girls would show greater sleep disruption and ANS upregulation than boys was partially supported. Sleep responses to the first night were similar between boys and girls; however, girls showed a more prominent cardiac response to the first night than did boys, with a faster HR during undisturbed REM and NREM sleep and diminished vagal activity (characterized by lower HF_na in REM sleep) during the first night as compared to the subsequent recording night. There was not a sex-interaction effect for HR as measured across the night regardless of arousals, awakenings, and sleep-stages, however, this HR measure went in the same direction as for HR in undisturbed sleep, with girls tending to show a bigger HR stress response than boys (Figure A.2). These findings are in line with former studies indicating higher stress levels and a greater stress reactivity in girls (25–27), further suggesting that there may be a female-specific vulnerability to stress.

Our findings should be considered within the context of study limitations. We requested that participants maintain a regular sleep-wake schedule during the week before their overnight recordings, however, this was not confirmed with sleep logs or actigraphy. We took advantage of the first night in the laboratory as a type of stress. We focused on differences in PSG and autonomic measures on the first night compared to a subsequent night and did not include measures of subjective stress responses. Future work is needed to examine effects of other types of stressors on sleep and ANS activity, as well as subjective stress levels, in male and female adolescents. For example, Hall and colleagues used an acute experimental anticipatory stress task before bedtime to investigate effects of stress on ANS activity, and found that young male and female adults had lower vagal activity during sleep, compared to an age- and sex-matched control group (54); possible sex × stress interaction effects were not examined. Our sample consisted mainly of adolescents in later pubertal maturation stages. Further studies are needed to assess whether sex differences in sleep ANS responses to stress differ across pubertal developmental stages.

Our finding of a greater female-specific ANS sensitivity to stress, coupled with our prior evidence of sex × age differences in ANS maturation (toward ANS upregulation in older girls) (23), may be relevant in the context of insomnia disorder. Altered cardiac ANS activity has been found in insomnia (18, 20, 55) and has been interpreted as a sign of autonomic hyperarousal, a key factor in the pathophysiology of the disorder. Longitudinal data would be required to determine whether repeated acute exaggerated stress responses lead to trait levels of ANS upregulation in girls, and whether this trait vulnerability could predict different trajectories in the development of insomnia disorder.

It is important to note that stress in teenagers is a serious problem and is on the rise in today’s 24-hour society. Chronic stress during adolescence may lead to changes in neurobiology and behavioral constructs, further increasing the risk of developing severe psychiatric disorders such as depression, anxiety or substance use disorders (56). The investigation of sex-specific developmental changes and specific mechanisms underlying stress in teens may help to better understand, diagnose, and treat stress-related disorders during adolescence such as insomnia, which is still under-diagnosed and under-treated in today’s society.

Conclusions.

The present study suggests a female-specific autonomic vulnerability to stress-related sleep disruption. The finding of an exacerbated autonomic reactivity to stress in healthy girls compared to boys, could contribute to their increased risk for developing insomnia.

Abbreviations

ANS

Autonomic nervous system

BMI

Body Mass Index

CESD

Center for Epidemiologic Studies Depression Scale

EEG

Electroencephalography

HF_na

Narrow absolute high frequency power

HF_nu

Normalized high frequency power

HR

Heart rate

HRV

Heart rate variability

NCANDA

National Consortium on Alcohol and Neurodevelopment in Adolescence

NREM

Non-Rapid-Eye-Movement

PDS

Pubertal Development Scale

PSG

Polysomnography

PSQI

Pittsburgh Sleep Quality Index

REM

Rapid-Eye-Movement

SE

Sleep efficiency

SOL

Sleep onset latency

SWS

Slow wave sleep

TIB

Time in bed

TST

Total sleep time

WASO

Wake after sleep onset