Relations between Infant Sleep Quality, Physiological Reactivity, and Emotional Reactivity to Stress at 3 and 6 Months

Abstract

The study examines the association between infant sleep, physiological, and emotional reactivity at 3 and 6 months of age in 89 African American infants and their caregivers. Infant sleep was objectively measured at 3- and 6-months using actigraphy for 7 days and nights. At 6 months of age, dyads participated in the Still-Face Paradigm (SFP) (Tronick et al., 1978) to assess infants’ physiological reactivity (via respiratory sinus arrhythmia (RSA)) and emotional reactivity. Findings revealed that infant night wakings at 3 months was positively correlated with baseline RSA at 6 months (β = .35, p < .001). Night wakings at 3 months marginally predicted negative affect during the recovery episode of SFP (β = −.26, p = .057). Further, night wakings at 6-months-old predicted more positive affect during the recovery episode of SFP (β = .34, p = .007). We discuss potential explanations for these findings such as an exuberant temperament style, parenting behavior, and lack of sleep consolidation at this early age to be the focus of future studies in this area. The current study adds to the limited research examining the relationship between sleep and reactivity in the earliest months of development. Findings reveal that it is critical to take developmental timing into account as all results were unexpected in relation to the childhood literature. Moreover, this is the first study of its kind to focus on an African American sample.

A poll conducted by the National Sleep Foundation reported that many children in the United States are not getting adequate amounts of sleep (Hirshkowitz et al., 2014). Insufficient amounts of sleep have been linked to a variety of detrimental developmental outcomes, such as poor school adjustment, behavioral problems, and emotional issues (Alfano et al., 2009; Bates et al., 2002; Reid et al., 2009; Williams et al., 2016). Despite these associations, there is a lack of literature exploring how sleep, physiological reactivity, and emotional reactivity influence each other early in life to predict later development. The current study aims to fill this gap by exploring the association between infant physiological reactivity during calm and stressful situations and infant sleep quality, concurrently and over time, and the subsequent relationship between early sleep quality and emotional reactivity during a stressful situation at 6 months. Moreover, the current study focuses on African American mother-infant dyads, due to previous findings that African American children and adults get less sleep than their counterparts (e.g.,Crosby, 2005) and experience a greater prevalence of sleep disorders, such as insomnia and sleep apnea (Lichstein et al., 2004; Petrov & Lichstein, 2015). To date, limited research has examined the early origins of these disparities. Therefore, studying the ways in which early sleep quality is linked to physiological and emotional reactivity has critical implications for behavioral and emotional development in later childhood and into adulthood.

1.1. Sleep Development

In their first two years of life, infants spend more than half of their time sleeping (Nelson et al., 2006). During this developmental period, sleep patterns change substantially, with the most dramatic organization occurring in the first 6 months (Sadeh et al., 2015). Young infants sleep almost non-discriminately during the day and night (Iglowstein et al., 2003), spend nearly half of nightly sleep in rapid eye movement (REM) sleep (Burnham et al., 2002; El-sheikh & Sadeh, 2015), and wake intermittently every 3–4 hours to feed (DeLeon & Karraker, 2007). By 6 months, sleep continues to develop, but is substantially more diurnal; 6-month-olds can sleep through the night and tend to sleep primarily at night (Henderson et al., 2011; Sadeh et al., 2015). During the same developmental period, neural networks are rapidly forming, and an infant’s brain is developing to about 90% of its adult size (Nelson et al.,2006). The fact that evolution favors sleep during the same time that dramatic brain growth occurs suggests the importance of sleep for development (Dahl, 1996).

Moreover, sleep problems are thought to reflect a biobehavioral indicator of self-regulation (Williams et al., 2016). This suggests that sleep problems might result from variations in children’s ability to regulate, or, alternatively, poor sleep may lead to more difficulties in regulatory capacities. This may occur day-to-day (i.e., poor sleep on Day 1 may lead to poor regulation on Day 2; or an over-aroused child may have more trouble falling asleep on that day) or it could lead to changes over time (i.e., poor sleep for many months may alter brain and regulatory processes; or continued problems with regulation could lead to altered sleep patterns that may last into adulthood). Becker and Langberg (2014) found an association between children’s sleep problems and poor adjustment skills, behavioral competence, and social functioning and other researchers have found that disrupted sleep patterns predicted poor adjustment in preschool, even after controlling for family stress (Bates et al., 2002).

There are several ways to measure infant reactivity, including behavioral observations of emotional reactivity and subsequent regulation, as well as underlying psychophysiology that supports these capacities. The parasympathetic nervous system, as indexed by respiratory sinus arrhythmia, is a well-established indicator of underlying regulation at this young age. The current study will examine the links between sleep and each of these measures of reactivity.

1.2. Respiratory Sinus Arrhythmia, Self-Regulation, and Sleep

Self-regulation refers to the cognitive and behavioral process by which individuals control emotional, motivational, and cognitive arousal and is believed to have physiological underpinnings (Blair & Diamond, 2008). Much of the extant research on self-regulation focuses on the behaviors associated with regulation as well as the physiological underpinnings, primarily the activity of the autonomic nervous system (Calkins et al., 2013), which is composed of two branches: the sympathetic nervous system (SNS) and the parasympathetic nervous system (PNS). Whereas the SNS is responsible for redistributing metabolic energy throughout the body in response to perceived external threat, the PNS is responsible for maintaining homeostatic physiological function during times of rest and decreasing energy expenditures following the resolution of stressful experiences (Porges, 2007). Respiratory sinus arrhythmia (RSA) (measured as variability in heart rate over time) is a purported index the PNS control of cardiac functioning via the vagus nerve (Porges, 2007). Porges’ polyvagal theory of social engagement (Porges, 1995) asserts that during times of no external demand or challenge, the “vagal brake” is engaged and parasympathetic control over the heart allows the body to focus on internal demands, such as organ growth and restoration. Alternatively, during environmentally challenging scenarios this brake is removed (decreasing parasympathetic influence) so that the body can mobilize and focus on external demands (Propper et al., 2008). Thus, a decrease in RSA (or RSA withdrawal) is typical in situations that require coping. When normal environmental conditions reemerge, the typical response is for RSA to increase and return to baseline levels followed by decreased heart rate (Porges, 1995). As a physiological measure of a child’s reaction to a challenging situation, RSA is highly useful because it can be measured noninvasively and has meaningful variation that is observable even in early infancy (Dietrich et al., 2007).

Individual differences in RSA withdrawal during challenging situations have been implicated in long-term developmental outcomes. A meta-analysis (of 44 studies) on the effects of RSA on outcomes in children revealed that increased RSA withdrawal (i.e., lower RSA levels) in challenging situations was related to decreased externalizing and internalizing behaviors and fewer cognitive and academic problems (Graziano & Derefinko, 2013). However, it is important to note that there are mixed findings in the literature. For example, extremely low and high levels of RSA withdrawal have both been associated with negative outcomes. For instance, children at risk for externalizing behaviors have exhibited lower levels of RSA withdrawal compared to their peers (Calkins et al., 2007). In addition, excessively high RSA withdrawal has been linked to deficits in emotional functioning (Beauchaine, 2001). Lastly, a study found that 3.5-year-olds who showed moderate decreases in RSA during executive functioning tasks had increased performance compared to children who showed too little or too much RSA withdrawal (Marcovitch et al., 2010). Based on the current literature, it is evident that RSA baseline and RSA withdrawal during a challenge are important correlates of child reactivity, regulation, and overall functioning. As such, it is reasonable to conclude that because RSA baseline and RSA change are indicators of reactivity, this physiological system may be associated with sleep quality (a process that is dependent on being able to calm down, relax, and self-soothe).

Indeed, sleep problems have been associated with inadequate emotion regulation and therefore higher levels of reactivity (Dahl, 1996). Due to the association between RSA and child neurodevelopmental outcomes, as described above, the existing literature has examined RSA as a predictor of sleep in preschool children (Elmore-Staton et al., 2012). Findings revealed that higher levels of baseline RSA predicted less restlessness during sleep (as indicated by less activity and movement during sleep), more sleep efficiency, and more minutes spent asleep, as measured by actigraphy. Gueron-Sela et. al (2017) found that less infant RSA withdrawal during a stressful task predicted more infant sleep problems at 18 months of age. Similar relationships between physiological reactivity and regulation, emotional intensity, and future sleep quality have been found in elementary school children (El-Sheikh & Buckhalt, 2005). Similar to the preschool age group, in this age group, less RSA withdrawal in response to stressful scenarios is associated with increased sleep problems. Less RSA withdrawal, among other autonomic measures, has also been shown to increase the likelihood of sleep problems in children who experience adversity, measured by multiple major life events (Alkon, et. al, 2017).

However, while RSA has been associated with certain aspects of sleep, it is still not possible to establish directionality in the relationship between RSA and sleep quality. Sleep quality and RSA might impact each other bidirectionally such that atypical RSA levels may lead to poor sleep quality, which in turn may further alter parasympathetic (RSA) functioning (Elmore-Staton et al., 2012). It is also possible that early sleep patterns in the first months of life may shape RSA functioning by 6 months. Sadeh et al. (2015) posit a relationship between night waking and regulatory skills such that frequent sleep interruptions interrupt brain growth and development to disrupt later regulatory capabilities. Interestingly, adolescents who reported poorer emotion regulation also reported greater sleep problems, even when accounting for demographic characteristics and current stress (Palmer et al., 2018). Moreover, a high level of emotional reactivity has been associated with higher levels of sleep disturbances in preschool children (Kelmanson, 2013), supporting Dahl’s (1996) claim that problems in emotion regulation could affect sleep development. During infancy, mother-reported frequency of night wakings and total time awake at night have been correlated with more negative emotionality (DeLeon & Karraker, 2007; Spruty et al., 2008). Due to these known associations between RSA and reactivity in children (Gentzler et al., 2009; Fox, 1989), the current study examines these links at 3 months of age, which is earlier than has ever been examined before.

1.3. Sleep and RSA in African American Children

Researchers have recently started to look at race-related health disparities in sleep problems. Specifically, studies indicate that African American children experience shorter night-sleep durations, nap more frequently, and sleep more on the weekends compared to White children (Buckhalt et al., 2007; Crosby et al., 2005). Researchers have also found that racial/ethnic minority children sleep fewer hours during the first two years of life compared to White children (Nevarez et al., 2010). Furthermore, it is evident that racial differences in sleep problems persist into adulthood with African American adults having poorer sleep continuity and quality, excessively short or long sleep duration, greater sleep variability, and greater risk of sleep apnea compared to white adults ((Lichstein et al., 2004; Petrov & Lichstein, 2016). It should be noted that most of the existing literature on racial differences in RSA and sleep focuses on the effects of these differences on health, whereas fewer studies have examined the effects of these differences on behavior. Although there is clear evidence of racial differences in sleep problems in children and adults, to our knowledge there has yet to be a study to explore the origins of these differences starting in the first months of life. Whether these differences are due to cultural (e.g., attitudes surrounding sleep), contextual (e.g., work schedules), or biological (e.g., neurophysiology) factors, is still unknown. Nevertheless, it is critical to explore the early onset of these problems in infancy to further understand when and why these differences begin to emerge. Thus, the current study focuses on an African American sample to better understand these early associations with sleep development within this population.

1.4. The Current Study

This study takes a first step towards understanding the complex relationships between sleep, physiological, and emotional reactivity in the first months of life. More specifically, we hypothesize that (1) higher infant sleep quality at 3 months of age (i.e., longer sleep duration and fewer night wakings) will predict higher baseline RSA at 6 months of age; (2) higher infant sleep quality at 3 months of age would predict greater RSA withdrawal during a stressor task (the Still-Face Paradigm (SFP); Tronick et al., 1978) at 6 months of age; and (3) and higher infant sleep quality at 3 and 6 months of age would predict less emotional reactivity during the SFP at 6-months. The SFP may be particularly interesting for this question due to the physical and emotional separation from mother that occurs during the still-face episode that may mimic bedtime practices where mothers must physically separate themselves from infants to allow the infants to fall asleep. Moreover, this study is the first to look at these relationships in an African American sample, which could shed light on the earliest origins of individual differences in sleep patterns and reactivity that have been observed in older African American children and adolescents.

2. Method

2.1. Participants

Participants (N = 89) were drawn from the Neonatal and Pediatric Sleep (NAPS) Study, a longitudinal study of 103 African American caregivers and infants. Families who resided within a 50-mile radius of a large public university in North Carolina and had an infant less than 3 months of age were recruited via public birth records, social media, and flyers. Potential participants were excluded if mothers were younger than 18 years of age, did not identify as African American, did not speak English, or if infants had experienced serious medical complications at birth (e.g., NICU stay >7 days), or were part of a twin pair. Seven infants (7.4%) were born prematurely (e.g., gestational age <37 weeks). The home visits for premature infants were scheduled based on their corrected ages (i.e., their chronological age reduced by the number of days born before 37 weeks’ gestation). The average adjustment for prematurity was 13 days (range = 1–30). Infants were enrolled (52.81% male) and seen at 3-month (M = 3.56 months; SD = .46 months) and 6-month (M = 6.47 months; SD = .49 months) home visits. 68.5% of mothers were employed; 38.9% had a bachelor’s degree or higher; 42.7% of mothers were married.

2.2. Procedures

In-home data collection visits took place when infants were 3- and 6-months of age. During the visit, mother-child dyads participated in parent-child interaction tasks and mothers completed questionnaires. At the 6-month home visit, dyads participated in the Still-Face Paradigm (Tronick et al., 1978). Heart rate monitors were placed on participants by trained research assistants at the beginning of each home visit to record RSA throughout the visit. Beginning on the night after each home visit, families completed a 1-week sleep assessment, including 7 days and nights of actigraphy monitoring. At the end of each home visit, mothers were compensated up to $130 in the form of a gift card and infants received a small gift. All procedures were approved by an institutional review board, and written consent was obtained from participants prior to data collection.

2.3. Measures and Materials

2.3.1. Infant sleep quality

Actigraphy was used to measure infant daytime and nighttime sleep quality for one week following the 3- and 6-month home visits. Actigraphy is an objective measure of sleep that records the amount of activity in 15-second epochs. Actigraphy data was obtained via Actiwatch 2 motion watches, which infants wore on their left ankles for 7 continuous days except at bath-time. Data was edited and analyzed using Phillips Respironics software (version 6.0.7). To align with earlier studies (So, et. al, 2005), the automatic setting (.888 x average activity) was used as the minimum threshold of activity classifying an awake state. Further information about the procedures based on the current sample have previously been reported (Camerota et al., 2018). For our analyses, we examined the average sleep duration throughout the day and night as well as the average number of night wakings at 3- and 6-months of age.

2.3.2. Infant emotional reactivity

The Still Face Paradigm (SFP) was used to assess emotional reactivity during the 6-month home visit (Tronick et al., 1978). During this task, mothers and infants sat facing each other for three 2-minute episodes. In the first episode (normal episode), the dyads interacted normally. In the next episode (still-face episode), the mothers stared at infants with an expressionless face and were told to be unresponsive to infant cues. In the final episode (reunion episode), the mothers resumed normal interaction. Note that any episode of the SFP (normal, still-face, or reunion) could be cut short if the infant displayed extreme distress (hard crying for 15 seconds or more).

Interactions were recorded and later coded for infants and mothers across all 3 episodes of the SFP in 5- second intervals. For each interval, infant affect was coded as positive, negative, or neutral expressions. Infant self-regulation was coded only during the still-face and reunion episodes and included any rhythmic and purposeful action done by the infant on another object or on itself to soothe itself or direct attention away from a stressor. Examples of behaviors coded as self-regulatory include feet-grabbing, thumb-sucking, self-stroking, self-clasping, fidgeting with the car-seat or other item, rhythmic body rocking, or rhythmic arm motions. Additionally, composite ratios for infant affective expression and self-regulatory behavior codes were calculated for analyses. Composite ratios were created for each episode by dividing the total number of instances of each of the codes described above by 24 (total number of possible intervals in each of the 3 episodes). If the episode ended early, the ratio was adjusted accordingly.

2.3.3. Infant physiological reactivity

RSA was used to assess infants’ physiological reactivity during the 3- and 6-months home visit. At the beginning of each visit, researchers placed Actiwave Cardio heart-rate monitors developed by CamNtech on infants’ chest (from sternum to left rib) using neonatal electrodes to obtain RSA data. Once the monitor was securely in place and the infant was acclimated and in a calm state (approximately 5 minutes after electrode placement), the infant was seated on the mother’s lap and the mother was instructed not to interact with their infant 4 minutes so that stimulation was minimized, and baseline RSA could be measured. At the 6-month home visit, RSA was additionally collected during each episode of the SFP.

Once downloaded onto computers, the data from the 6-month visits were segmented into the baseline episode and the multiple episodes of the SFP. Prior to RSA calculation, R-waves were identified by a computer algorithm and any missing or incorrect R- waves were edited manually using the CardioEdit software (Porges, 1985). Incorrect R-waves were identified by outliers that were far above or far below the typical pattern of points that surrounded the outliers and that made up the wave. Incorrect R-waves were edited by summing, dividing, or averaging the outliers with each other or with adjacent points to make the outliers consistent with the typical pattern of the surrounding points of the R-wave. Outliers creating incorrect R-waves may have been from excess noise picked up by the heart rate monitors, which could be caused by a range of factors including infants or mothers tampering with the monitors and accidental removal of the monitors. RSA was calculated in 30-second epochs over the 4-minute baseline and in 30-second epochs over each 2-minute episode of the SFP using the Porges’ method (1985). Average RSA values across all epochs were then obtained per episode via CardioBatch Plus software for further analyses. CardioBatch applies a moving polynomial filter and quantifies the amplitude of RSA with age-specific parameters, sensitive to the maturational shifts in the frequency of spontaneous breathing (0.3–1.3 Hz for infants). Following previous research (Moore & Calkins, 2004; Propper et al., 2008), difference scores for the 6-month visit were computed by subtracting still-face episode of RSA from baseline RSA, such that sign indicated direction of change, with positive values indicating greater RSA withdrawal. The RSA variables used in the current study include RSA baseline at 3mo, RSA baseline at 6mo, and RSA change from baseline to the SF episode at 6mo.

2.3.4. Covariates

Demographic information (e.g., child sex, child age, maternal education, and feeding practices) was reported by mothers via questionnaires administered at the 3-month and 6-month home visits. Although study visits were conducted at 3 and 6 months of age, there was some variability around the timing of these visits so that they did not always fall directly on the 3^rd or 6^th month birthday. Due to the important developmental leaps that occur in the first months of life, we controlled for age in all analyses. A continuous measure of maternal years of education was used in the current analyses (e.g., high school degree = 12 years; four-year college degree = 16 years). A dichotomized variable of breastfeeding was also included (e.g., 0 = no breastfeeding; 1 = breastfeeding).

2.4. Analytic Plan

First, Pearson correlations were used to assess the associations among infant sleep quality measures and infant physiological reactivity variables. Pearson correlations were also used to assess the associations among infant sleep quality measures and infant emotional reactivity variables. Next, linear regressions were conducted to predict infant physiological and emotional reactivity from infant sleep quality measures after controlling for appropriate covariates. Descriptive statistics were conducted using SAS 9.4, and path analyses were conducted using Mplus 8.1.

From the larger sample, four dyads did not complete the 3-month visit, eight did not complete the 6-month visit, and two completed only the questionnaire portion of the 6-month visit. Current analyses include only the dyads for which data were available at the 3-month and 6-month time points (N = 89). Missing data were handled using full-information maximum likelihood (FIML).

3. Results

3.1. Descriptive Statistics

Descriptive statistics for all variables are presented in Table 1. Based on patterns of significant correlations, primary caregiver education, child sex, child breastfeeding status, and child age at 3- and 6-months were included as covariates.

Table 1.

Descriptive statistics for all study variables

	N	M	SD	Min	Max
Child age at 3mo	86	3.56	0.46	2.63	5.10
Child age at 6mo	84	6.47	0.49	5.52	5.52
Primary caregiver education (years)	84	14.75	2.22	10.00	18.00
Average sleep duration at 3mo	80	473.18	64.99	292.04	603.11
Night wakings at 3mo	80	2.23	0.85	0.71	5.50
Average sleep duration at 6mo	74	446.54	49.65	297.61	563.64
Night wakings at 6mo	74	1.64	0.74	0.00	3.86
RSA at baseline at 3mo	63	3.36	0.81	1.80	5.29
RSA at baseline at 6mo	72	3.64	0.99	1.16	6.12
RSA change from baseline at 6mo	52	−0.06	0.85	−1.70	2.62
Negative affect during still face episode	73	0.49	0.42	0.00	1.00
Negative affect during recovery episode	65	0.45	0.40	0.00	1.00
Positive affect during still face episode	73	0.06	0.12	0.00	.52
Positive affect during recovery episode	65	0.14	0.23	0.00	1.00
Self-regulation during still face episode	73	0.57	0.29	0.00	1.00
Self-regulation during recovery episode	62	0.30	0.32	0.00	1.00
	N			%
Child sex (Male)	47			52.81
Breast fed (Yes)	44			55.70

3.2. Early sleep and baseline RSA

Correlations between infant baseline RSA and sleep are presented in Table 2. Baseline RSA at 3-months was not associated with any sleep variable at 3- or 6-months (r = −.15 to .13; p > .05); further models were therefore not examined. In contrast, infant night waking at 3-months was positively correlated with baseline RSA at 6-months (r = .31, p = .013). Controlling for all covariates (Table 3), the relationship between night wakings at 3-months and baseline RSA at 6-months remained significant (β = .35, p < .001). The same pattern of findings held up controlling for concurrent (i.e., 6-month) night wakings (β = .40, p < .001; Table 3).

Table 2.

Associations between sleep variables, RSA variables, and SF variables

	1	2	3	4	5	6	7	8	9	10	11	12	13
1. ASD 3mo	−
2. NW 3mo	−.42***	−
3. ASD 6mo	.40***	−.21†	−
4. NW 6mo	−.26**	.44***	−.40***	−
5. RSA B 3mo	.01	.02	.09	−.15	−
6. RSA B 6mo	−.06	.31*	−.03	−.03	.43**	−
7. RSA C 6mo	.14	−.07	.16	−.14	−.31†	−.50***	−
8. Neg SF	.00	−.18	−.03	−.12	−.06	−.28	.10	−
9. Neg Re	.05	−.24†	−.01	−.10	−.15	−.36	.08	.79***	−
10. Pos SF	−.06	.13	−.24†	.07	.16	.40	−.06	−.39***	−.51***	−
11. Pos Re	−.08	.11	−.16	.25†	.12	.26	−.10	−.48***	−.54***	.63***	−
12. SR SF	−.20	.19	−.05	−.04	.12	.28	−.08	−.26*	−.34***	.09	.23†	−
13. SR Re	−.08	.06	.15	−.15	−.12	.31	.04	−.26*	−.27*	.12	.17	.50***	−

Note.

^†p < .10,

^*p < .05,

^**p < .01,

^***p < .001

Ns range from 35 to 80

ASD 3mo = Sleep duration at 3mo; NW 3mo = Night wakings at 3mo; SD 6mo = Sleep duration at 6mo; NW 6mo = Night wakings at 6mo; RSA B 3mo = RSA baseline at 3mo; RSA B 6mo = RSA baseline at 6mo; RSA C 6mo = RSA baseline change at 6mo; Neg SF = Negative affect during still face episode; Neg Re = Negative affect during recovery episode; Pos SF = Positive affect during still face episode; Pos Re = Positive affect during recovery episode; SR SF = Self-regulation during still face episode; SR Re = Self-regulation during recovery episode

Table 3.

Regression analyses for NW at 3mo predicting RSA baseline at 6mo and predicting negative affect during recovery episode

	RSA Baseline at 6mo		Negative during recovery
	β	SE	β	SE
Not controlling for concurrent sleep
Child age at 3mo	.06	.12	.13	.12
Child age at 6mo	.23	.10	−.16	.12
Primary caregiver education (years)	−.09	.12	.15	.14
Child Sex	.14	.11	−.23	.12
Breast fed	.12	.13	−.24	.16
NW 3mo	.35***	.10	−.26†	.14
Controlling for concurrent sleep
Child age at 3mo	.07	.12	.12	.12
Child age at 6mo	.20*	.10	−.15	.12
Primary caregiver education (years)	−.09	.12	.15	.14
Child sex	.12	.12	−.23*	.12
Breast fed	.12	.13	−.23	.16
NW 6mo	−.12	.13	.03	.17
NW 3mo	.40***	.09	−.25	.16

Note.

^†p < .10,

^*p < .05,

^**p < .01,

^***p < .001

3.3. Early sleep and physiological reactivity/regulation

Correlations between RSA withdrawal and sleep variables are presented in Table 2. RSA change from baseline to still face did not correlate with any sleep measures (i.e., sleep duration, night wakings) at either 3- or 6-months. Due to the lack of bivariate associations, further models were not estimated.

3.4. Early sleep and emotional reactivity

Correlations between sleep and emotional reactivity variables are presented in Table 2. Due to the lack of bivariate associations between sleep and self-regulatory behaviors, further models were not estimated for those behaviors. Night wakings at 3-months were marginally correlated with negative affect during the recovery episode of SFP (r = −.24, p = .072). In a regression model controlling for all covariates (Table 3), night wakings at 3-months marginally predicted negative affect during the recovery episode of SFP (β = −.26, p = .057). To test whether these findings held up when controlling for concurrent sleep the model was re-run to account for night wakings at 6-months. However, when controlling for all covariates as well as night wakings at 6-months, night wakings at 3-months did not predict negative affect during the recovery episode of SFP (β = −.25, p = .121).

Concerning concurrent relationships, average sleep duration at 6-months was marginally correlated with positive affect during the still face episode of SFP (r = −.24, p = .057). After adjusting for all covariates, average sleep duration at 6-months did not predict positive affect during the still face episode of SFP (β = −.273, p = .122). Night wakings at 6-months was marginally correlated with positive affect during the recovery episode of SPF (r =.25, p = .058). After adjusting for covariates (Table 4), night wakings at 6-months positively predicted positive affect during the recovery episode of SFP (β = .34, p = .007).

Table 4.

Regression analyses for NW at 6mo predicting positive affect during recovery episode

	Positive affect during recovery
	β	SE
Night Wakings 6mo
Child age at 3mo	−.08	.10
Child age at 6mo	.23*	.09
Primary caregiver education (years)	.06	.20
Child sex	.07	.10
Breast fed	−.10	.19
NW 6mo	.34**	.13

Note.

^†p < .10,

^*p < .05,

^**p < .01,

^***p < .001

4. Discussion

This study examined the relationships between infant sleep quality, physiological, and emotional reactivity during the first 6 months of life in an African American sample. Overall, although findings were inconsistent with hypotheses, several interesting associations were found. First, we hypothesized that high sleep quality at 3 months of age would predict higher baseline RSA at 6 months of age. Our findings did not support this hypothesis, but instead indicate that infant night wakings at 3 months were positively associated with baseline RSA at 6 months. Our second hypothesis was also not supported; higher sleep quality at 3 months of age was not associated with greater levels of RSA withdrawal during the still-face episode. Finally, we hypothesized that higher sleep quality at 3 and 6 months of age would predict less emotional reactivity during the SFP at 6 months. We found that infants who woke up more throughout the night at 3 months had lower levels of negative affect during the recovery episode of SFP at 6 months, however this effect went away after controlling for concurrent sleep, highlighting the importance of daily routines rather than earlier experience. Although our findings were unexpected, it is important to note that these questions have never been asked at this early age, thus our findings provide initial insight into how the relationship between sleep and reactivity may differ during infancy from findings in the extant literature.

Our first research question focused on the relationship between early sleep and developing physiological functioning during infancy. We hypothesized that better sleep at 3 months of age would be associated with higher baseline RSA at 6 months (i.e., a purported indicator of decreased reactivity and better regulation). Unexpectedly, we found the opposite to be the case; increased night wakings at 3 months predicted higher baseline RSA at 6 months. These findings are likely related to the developmental timing of this assessment. Although in childhood and into adulthood, higher baseline RSA has been related to appropriate emotional reactivity (Stifter & Fox, 1990; Beuchaine, 2001) it is indeed the case that in infancy, high baseline RSA has been linked to greater emotional expressivity (e.g., Stifter et al., 1989), greater behavioral reactivity (e.g., Beuchaine, 2001; DiPietro & Porges, 1991; Porges, 1991), and more active focused attention (e.g., Richards, 1987). In addition, mixed findings in the field include studies that reveal associations between high baseline RSA and emotional reactivity well into childhood as well (Cole et al., 1996; Eisenberg, et al., 1995). Moreover, researchers have found that more behaviorally reactive infants may also be at risk for poor sleep quality (De Marcas et al., 2015). Thus, it may be the case that, at this age, a behaviorally reactive infant may be less easy to soothe to sleep, more likely to wake through the night, and generally more engaged with the environment may also exhibit higher levels of baseline RSA.

Another important consideration here is the parent-child relationship. Advocates of co-sleeping have previously highlighted the beneficial physiological and psychological effects that nighttime maternal contact has on infants, linking it to increased maternal interactions and arousals and the formation of a positive attachment style – a prerequisite for healthy infant development (McKenna & McDade, 2005). Furthermore, increased nighttime mother-infant contact has been associated with mutual regulation and physiological interconnectedness as well as more frequent infant arousals (McKenna et al., 2007). Thus, infants who are waking up more throughout the night may have increased infant-caregiver interaction, leading to improved regulation of both behavior and physiology at 6-months whereas those who are waking less frequently may miss out on these skin-to-skin experiences including close proximity to caregivers. Future studies should examine parent-child interactions through the night to determine whether this may be the case, and whether this may differ for dyads with more positive or sensitive interactions.

Our second hypothesis was that better sleep quality at 3 months of age, including longer sleep duration and a fewer number of night wakings, would be associated with greater levels of RSA withdrawal during the still-face episode. Although this hypothesis was based on existing findings in the literature of comparable experiences in adolescents (e.g., El-Sheikh & Buckhalt, 2005), we did not find support for this hypothesis. Again, this difference may be due to this early developmental period. Nocturnal sleep becomes more efficient and less fragmented over time (Louis et al., 1997), and at 6 months of age poor sleep quality is still quite common. In other words, sleep-wake regulation is still coming online at this age, and fragmented sleep may be developmentally appropriate until consolidation becomes more normative around 12 months of age (Scher et al., 2004). Therefore, night wakings at 6 months of age may be common and related to many other factors beyond the stress response. For example, sleep fragmentation that occurs at 6 months of age may be related to changes in motor development (i.e., crawling) or changes in the caregiving routine (i.e, caregivers going back to work after maternity leave) (Scher et al., 2004). Future studies should examine these questions at 12 months of age and later.

Along these lines, there are likely many other unmeasured variables that are salient for RSA withdrawal during stress, rather than sleep alone, that were not measured here including the mother-infant relationship, infant temperament, or overall mood may influence infant response to the SFP. Indeed, maternal behavior has been related to infant responses in the SFP, with infants of sensitive mothers showing more regulatory behaviors, more positive affect and less avoidance and negative affect during the still-fact and reunion episodes (Mesman et al., 2008). Thus, infants who experience high levels of maternal sensitivity may develop strong regulatory behaviors that increase their resilience to stressors. Future studies should include a more comprehensive approach to understanding the various factors that may contribute to psychophysiological response to a stressor in infancy.

Our third, and final, hypothesis was that better sleep quality at 3 and 6 months (i.e., longer average sleep duration and fewer numbers of night wakings) would predict less emotional reactivity (i.e., more positive, less negative, and more self- regulatory behaviors) during the SFP at 6 months. Unexpectedly, we found that more night wakings at 3 months was associated with lower levels of negative affect during the recovery episode of SFP at 6 months. However, after controlling for concurrent sleep this effect went away, suggesting that concurrent sleep plays a critical role, perhaps more than then earlier sleep development. Moreover, closer examination of concurrent relationships revealed that average sleep duration at 6 months did not predict positive affect during the still face episode of the SFP. However, findings reveal that infants who are waking up more through the night exhibit higher levels of positive affect during the recovery episode of SFP.

Although we did not anticipate this finding, the results may be due to variations in infant temperament, or individual differences in reactivity and self-regulation surrounding affect, activity, and attention (Rothbart & Bates, 1998). Previous literature has suggested that temperament influences children’s sleep (Wolfson & Montgomery-Downs, 2013) in that reactivity to stimuli is a central aspect of temperament (De Marcas et al., 2015; Rothbart & Derryberry, 1981). Therefore, a temperamentally reactive child may have more difficulty disengaging from internal and external stimuli, making it more difficult for them to initiate and maintain sleep (De Marcas et al., 2015). Moreover, though the literature on the effects of infant temperament on responses to the SFP is scarce, two studies have reported temperamental differences in infants predicting still-face responses (Braugart-Reiker et al., 1998; Tarabulsy et al., 2003). Specifically, it was found that fussier infants were more likely to be associated with less self-comforting and object orientation during mother-infant still face (Braungart-Reiker et al., 1998). Thus, the association between more night wakings and positive affect during recovery may be related to the temperamental domain of exuberance, a profile reflecting the disposition toward positive affect and sociability (Fox et al., 2001). Compared to non-exuberant children, exuberant children are highly sociable, show limited signs of inhibition, and exhibit high approach behaviors. Exuberance is believed to be motivated by an underlying motivation to approach, where infants engage in high levels of positive reactivity to instigate social interactions (Fox, 1991; Gray, 1982). Linked to distinct neural profiles, found in patterns of the left frontal EEG asymmetry, exuberance is apparent across behavioral and biological levels and associated with adaptive, sociable behavior (Degnan et al., 2011; Fox, 1994). Thus, the association between more night wakings and positive affect during recovery suggests that these infants may be utilizing approach behaviors to enhance social interaction (typically with their mother) both during a stressor task and at night.

Alternatively, this behavior may be shaped by the emotional and behavioral responses of the caregiver. Caregivers influence the development of reactivity and regulation by providing infants with a supportive, responsive environment and by socializing culturally appropriate behavior (Thompson, 1994; Thompson, 1998). Infants’ actively shape the frequency of maternal interactions they receive by waking up more throughout the night, however, are passive agents regarding the quality of those maternal interactions. The additional opportunities to interact with caregivers may provide more experiences to learn about how to calm down following negative emotion and how to cope with stress. This assumes, however, that those interactions are sensitive, responsive, and positive in nature. Future studies should examine these nighttime interactions more closely to determine whether infant reactivity and response to stress differs based on dyadic interactions through the night. It may also be the case that those infants who more frequently wake at night are the ones who are seeking support from caregivers and therefore find comfort from those caregivers, and thus may be more likely to have positive interactions during the SFP.

4.1. Summary, Limitations, and Future Directions

These findings add to the existing literature on the relationship between sleep and emotional reactivity. The link between sleep problems and emotional problems in adulthood have been well established (American Psychiatric Association [APA], 1994), with most longitudinal sleep studies suggesting that sleep problems in childhood and/or adolescence predict later psychopathology (Gregory & Sadeh, 2012). However, studies on children have not found many causal relationships between sleep and reactivity or regulation. Of the correlative studies, less sleep and more restless sleep have been associated with inadequate emotion regulation, negative emotionality, and depressive or anxious tendencies (Smaldone et al., 2007; Wolfson & Carskadon, 1998). Early childhood may be a critical period for addressing sleep problems (Maski & Kothare, 2013) to maximize later developmental outcomes. However, few studies explore these associations in the first few months, and it is currently unclear when relations between sleep problems and emotion regulation emerge and the direction of influence (Williams et al., 2017). Our study provides some insight into these processes early in life and suggest that it may be developmentally and evolutionarily appropriate to wake at night at 6 months of age, and these additional wakings could lead to increased opportunities to be proximal to caregivers, and result in less negativity and higher baseline RSA in infancy. Age-appropriate sleep patterns may lead to better regulation skills (Dahl, 1996). Moreover, this study focuses on an African American sample of mother-infant dyads, which again is the first of its kind. Although it was beyond the scope of this paper to identify mechanisms that may lead to associations in this population, studies such as these are an integral first step in identifying processes that may differ across culture and race.

This study revealed unexpected and interesting findings; however, limitations must also be considered. Specifically, some of the missing SFP data may not be random, as the infants who are most distressed during the SFP may have had the most trouble completing tasks which could lead to problems with “messy” data. As such, future research should aim for a larger sample size and additional measures of stress. Although it is a strength that our sample was African American, it is also a limitation because we cannot generalize these findings to other racial groups. Future studies should look across race, as well as take into account mechanisms that may be at play here including cultural differences, variations in family structure, and attitudes about interactions, sleep, and stress. We also propose that parenting during the day and through the night may be an important predictor of improved reactivity and regulation in infancy, but the current study does not include parenting behavior (either through the night or during the SFP). Furthermore, understanding maternal sleep quality and its relation to infant sleep is an important next step to understanding how caregiving behaviors may influence infants’ reactivity and regulation. For co-sleeping dyads, poor maternal sleep (i.e., increased movement through the night) may lead to additional night wakings for infants. It may also be the case that mothers who are not getting an adequate amount of sleep may be more negative towards their infant during interactions, as the result of exhaustion and decreased patience. Indeed, previous work from our group (Lerner et al., 2020) found that mother-infant co-sleeping influenced behavioral regulation in infants. Specifically, it was found that mother-infant co-sleeping at 3 months was associated with more self-regulatory behaviors at 6 months. Therefore, future studies should consider the influence of mother-infant sleep arrangements (e.g., co-sleeping) on infant social-emotional development, as this additional contact and interaction, for better or worse, may lead to changes in the physiological underpinnings of self-regulation across infancy. Additionally, we focused on two objective measures of sleep: sleep duration and night wakings. While these are important indicators of sleep at this age, there are several other measures that may have been useful to include, including sleep consolidation, efficiency, and sleep across a 24-hour period (including naps). Finally, although the current examination of RSA during a stressful situation provides new findings regarding the link between sleep and physiological regulation in infants, future studies should include a more precise investigation of RSA change across stress. A time series analysis of RSA and behavior would allow for a better understanding of how RSA changes moment-to-moment in relation to stress, leading to a clearer picture of whether infants are doing an effective job regulating or soothing themselves across the stressful situation, or whether they are becoming increasingly aroused and reactive to the stressor. Understanding this intrinsic process, as well as the way in which mothers may influence this response (via synchronous interactions with infants or coregulation), will offer valuable information about individual differences in infant social-emotional development as the result of sleep, stress response, and relationships with caregivers.

Despite this, the results of the current study are important for understanding how infant sleep and reactivity may develop across the first year. Overall, results indicate that infants who wake more through the night may not be at risk for increased reactivity. Instead, these night waking behaviors may be developmentally appropriate and perhaps adaptive. It is crucial to examine these associations across time, as the links between sleep and behavior will change cross development and sleep becomes increasingly consolidated. However, at this age, there are likely many other factors (including temperament and parent-child interactions) that may influence physiological and behavioral response to stress beyond sleep patterns alone. It is likely that a third variable, such as an exuberant temperament, may lead infants to wake more at night and also to respond with more positivity to the SFP. Although this is the first study to look at these links within an African American sample, future studies of this population should focus on mechanisms and whether these associations would look different in dyads within families of other races.

Highlights.

• Infants who are waking up more throughout the night and who have the shortest longest sleep period (measured via actigraphy) are not struggling in terms of self-regulatory skills.

• These behaviors may be both adaptive and resilient by aiding infants in obtaining increased maternal interactions.

• More night wakings may lead to more scaffolding opportunities that aid in the development of self-regulation.

• Findings imply that infants are both active and passive agents in the development of their own regulatory skills.