The Relationship Between the Quality of Sleep and Emotion Regulation in Infants

The Relationship Between the Quality of Sleep and Emotion

Regulation in Infants

Name: Frederique Kunst Student number: 11697229 Daily supervisor: Martina Zaharieva

Supervisor: Ingmar Visser Submission date: 18-12-2020

Abstract

The current study examined the relationship between the quality of sleep and emotion regulation in infants because of the predictive effect that the development of emotion regulation has on later-life behavioural and emotional abilities. Infants (N = 19) participated with their caregivers in the Still-Face Paradigm and were afterwards coded on their facial expressions. Afterwards, their caregivers filled in the Brief Infant Sleep Questionnaire on their sleeping behaviour, from which three indicators were chosen: total sleep duration, the number and duration of nocturnal awakenings. Poor sleep quality was predicted to increase negative and decrease positive facial expressions during the entire paradigm and increase the carry-over effect, shown by an increased negative and decreased positive proportion of shown affect in the reunion compared to the baseline phase. The results of the Still-Face Paradigm showed an increase in negative affect after the baseline phase and a carry-over effect between the baseline and reunion phase. The results also showed a decrease in positive affect in the still-face phase, but subsequently, the positive affect increased again, leaving a more subtle, nonsignificant carry-over effect. However, these results on the negative and positive proportions of affect and the carry-over effects did not show a relation with the sleep indicators for sleep quality. Therefore, with these results, we can conclude that the relation between the quality of sleep and emotion regulation in infants does not exist. Due to the limitations of the experiment based on the sample and measurement tools, this relationship deserves further research attention.

The Relationship Between the Quality of Sleep and Emotion Regulation in Infants The development of emotion regulation is a widely investigated subject in developmental research. The development of this skill has its onset in early infancy and has been related to problems in later behavioural and emotional abilities, such as the development of social competence (Denham, Blair, DeMulder, Levitas, Sawyer, Auerbach-Major & Queenan, 2003; Thompson 1991). Furthermore, the development of emotion regulation could be a predictive factor for psychopathological problems in later life (Keenan, 2006). For instance, the development of inhibited behaviour, which implies socially withdrawn, fearful and clingy behaviour to the caregiver. This behaviour, which goes along with increasing cortisol levels, seems to be a risk for the development of anxiety disorders in childhood (Biederman, Rosenbaum, Hirshfeld, Faraone, Bolduc, Gersten, Meminger, Kagan, Snidman & Reznick, 1990).

There are multiple factors that influence the emotion regulation, among which the regulation of sleep. Poor sleep regulation is characterized by sleep deprivation and poor sleep quality. It seems to be linked to a decrease in flexibility of context-related emotional reactions and has influence on certain aspects of emotion regulation which results in an increase of negative affect and a decrease of positive affect (Feldman, 2009; Geva & Feldman, 2008; Mauss, Troy & LeBourgeois, 2013; Palmer & Alfano, 2017; Zohar, Tzischinsky, Epstein & Lavie, 2005). Because of the importance of good emotion regulation in infancy for later behavioural and emotional development, it is of interest to investigate the relationship between the sleep and emotion regulation in the first year of life.

Sleep regulation can influence emotion regulation in different ways, including the expression of negative and positive emotions. Emotion regulation infers to the extrinsic and intrinsic processes which provide the monitoring, evaluation and modification of emotional reactions (Thompson, 1991). These regulatory functions, along with the further integration of

aspects of emotional and attentional regulation, are controlled by a vertical-integrative hierarchical system consisting of the brainstem, the limbic system and the cortical systems (Weller & Feldman, 2003). One of the emotional regulatory functions is providing the ability to select emotional strategies and implement these strategies, which are developed by the intrinsic pathway of emotion regulation (Gross, 2015; Palmer & Alfano, 2017). This pathway develops the ability to experience and express negative and positive emotions with varying regulatory strategies. The pathway exists of the mid frontal area, including the anterior cingulate, that performs as a control system for negative emotions (Feldman, 2009; Posner & Rothbart, 2000).

The quality of sleep regulation is of influence on this control system of negative emotional expressions. Disturbed sleep seems to interrupt higher cognitive functions that control emotion regulation, among which the control of negative emotions, and seems to be associated with a higher chance on socio-emotional problems (Hysing, Siversten, Garthas-Niegel & Eberhard-Gran, 2016; Siversten, Harvey, Reichborn-Kiennerud, Torgersen, Ystrom & Hysing, 2015; Tucker, Whitney, Belenky, Hinson & van Dongen, 2010). With this interruption, a negative emotional reactivity arises, which increases goal-disruptive events and decreases goal-orientated events, and thereby also positive emotions (Geva & Feldman, 2008; Zohar et al., 2005). A study that looked into the underlying neural structure in adolescents showed that there was a functional decrease of the amygdala by the ventral anterior cingulate cortex, part of the negative control system, when there was sleep debt. This functional decrease elicited continuous emotional responses to negative emotional stimuli, which created an emotional instability in the control system of negative emotions (Motomura, Kitamura, Oba, Terasawa, Enomoto, Katayose, Hida, Moriguchi, Higuchi & Mishima, 2013). Thus, the control of negative emotions can be disturbed by difficulties in sleep regulation and thereby influence the emotional expression.

One widely used method to measure the regulation of emotion in infants, and thereby their emotional affect, is the Still Face Paradigm presented by Tronick, Als, Adamson, Wise, and Brazelton (1978). The paradigm is used because of the robustness of the expected patterns and behavioural reactions that the Still-Face Paradigm elicits compared to natural situations (Adamson & Frick, 2003; Mesman, Linting, Joosen, Bakermans-Kranenburg & Van IJzendoorn, 2013). It consists of three phases in which the face-to-face interactions between infant and caregiver are observed. The first phase is the baseline phase, in which the caregiver has an everyday, playful interaction period with the infant. In the second phase, called the still-face, the caregiver does not respond to the infant’s interaction and has a neutral facial expression for one minute. In the last phase the caregiver resumes their normal, playful interaction with the infant and is called the reunion phase. During the entire Still Face Paradigm, the facial expressions of the child are coded as neutral, positive or negative affect which will provide proportions of affect per face. Therefore, this paradigm is a useful tool to measure the emotional expressions within emotion regulation in infants. That is why we use it as an instrument in the current research.

Besides controlling negative emotions, emotion regulation skills help place the meaning and significance of emotional arousal in certain contexts (Thompson, 1991). According to a model of Gross (1998) and an elaboration of Palmer and Alfano (2017), one of the five processes of emotion regulation is the identification of emotion as something that should be regulated. This skill depends on emotional awareness, experience and context. Judging by an extended model of Gross (2015), a good sleep quality seems to contribute to the identification of emotion. It facilitates a mindfulness in adults that encourages better emotion regulation and is related to a better match between the needed emotion and the emotion that is felt in a certain situation (Palmer & Alfano, 2017). Sleep deprivation, a feature of poor sleep regulation, seems to disturb the ability to identify happy and angry facial expressions, which interrupts the

emotional awareness and thereby the emotion regulation (Gross, 2015; van der Helm, Gujar & Walker, 2010). Besides, sleep deprivation can decrease the implementation of the strategy and the flexibility of reaction around context (Palmer & Alfano, 2017). Thus, this ability to match the context with emotion that should be expressed is another feature of emotion regulation that can be disturbed by a poor sleep regulation.

In the Still Face Paradigm presented by Tronick et al. (1978), there are also three effects that can be measured. The first effect is called the still-face effect, which looks at the emotion regulation during the still-face phase. The changes in the caregiver’s behaviour seems to provoke the infant, which often elicit increased gaze aversion and negative affect and decreased positive affect compared to the normal face-to-face interaction. The duration of the change from the still-face phase to the reunion phase is known as the subsequent recovery effect. The final effect that can be obtained from this paradigm is the carry-over effect, which is measured by the comparison between the reunion phase and the baseline phase. These phases are similar in content, but the reunion phase follows the still-face in which the caregiver showed neutral emotion expressions instead of positive. Thus, this carry-over effect shows the difference in this emotion regulation between the two phases and is predicted to show a decrease in positive affect and an increase in negative affect.

The carry-over effect, with the changes in affect, could be a result of the recovery in the reunion period from this neutral emotion expression of the caregiver during the still-face phase or/and a mismatch between the context and the expressed emotions (Mesman, van IJzendoorn & Bakermans-Kranenburg, 2009; Tronick et al., 1978). This effect was found in research on the infant’s responses to maternal still-face. The 9-month-old infants showed a greater carry-over effect than the 4-month-old infants, which consisted of increased distress after the still-face phase in comparison to the baseline phase (Yato, Kawai, Negayama, Sogon, Tomiwa & Yamamoto, 2008). The carry-over effect also came forward in a meta-analysis research. The

results there showed a recovery of the positive affect during the reunion, but the overall positive affect was significantly higher in the baseline period. The negative affect however did not fully recover in the reunion period and stayed significantly higher in comparison to the baseline period (Mesman et al., 2009). These findings in the effects originated from the phases seem to be robust. Thus, in combination with the robustness in comparison to natural situations, the Still-Face Paradigm is a useful tool to measure this carry-over effect of emotion regulation in infants, which is of interest to look into in relation to the predicted increased context-related difficulties by poor sleep regulation.

To measure the sleep regulation in infants, Sadeh (2004) developed the Brief Infant Sleep Questionnaire, a measurement tool based on infant sleep literature. The questionnaire includes questions about the specific matters and changes of the sleep regulation in infants. It roughly consists of three main subjects: the duration of sleep, the nocturnal awakenings and the way the infant sleeps. The questionnaire showed significant and robust correlations between almost all the matters of sleep regulation in the BISW compared to actigraphy and daily sleep logs, on a clinical level. A study on the specificity of the BISQ sleep parameters in comparison to the caregivers’ reports on sleep-related problems in infancy confirms that the questionnaire is valid for epidemiological studies (Del-Ponte, Xavier, Bassani, Tovo-Rodrigues, Halal, Shionuma, Ulguim & Santos, 2020). Due to the good validity and robustness, the BISQ is a good and easy to administer measurement tool for the sleep regulation in infants and is therefore used in our research.

With respect to the previous studies about difficulties in sleep regulation in relation to emotion regulation, a poor sleep regulation seems to influence negative emotion regulation and context-related emotion regulation in infants. Since the development of emotion regulation in infancy is predictive for later on emotion regulation, it is of interest that the development goes well. If there are problems at an early age, it can provide information about the development of

psychopathology and potential insights on the pathways for early intervention (Berger, Miller, Seifer, Cares & LeBourgeois, 2012; Keenan, 2006; Sheppes, Suri & Gross, 2015). To gain a better understanding of the relationship between poor sleep regulation and emotion regulation in infancy, this report will answer the research question ‘What is the relationship between the quality of sleep regulation and the emotion regulation in early-infancy?’.

Given that previous researches showed the differences in emotion regulation due to difficulties in sleep regulation, we expect that poorer sleep regulation relates to an increased proportion of negative affect and a decreased proportion of positive affect during the whole Still-Face Paradigm. This expectation will be measured via the amount of positive and negative affect that the infant displays during the Still-Face Paradigm and the quality of sleep regulation reported by the parent in the BISQ. The infant’s data from the Still Face Paradigm interaction with the caregiver will be micro-coded using a coding scheme which focuses on multiple aspects of emotion regulation, including the emotional affect. This affect will be coded on positive, negative and neutral facial expressions expressed by the infant. The infant’s sleep regulation ability will be estimated on the basis of three outcome measures from the BISQ: total sleep duration (hours and minutes), number of night awakenings and nocturnal wakefulness (hours and minutes). These three independent variables of sleep regulation will then be checked on the correlation with the proportion of negative and positive affect during the entire Still Face Paradigm to test the hypotheses.

Given that difficulties in the sleep regulation disturb the contextual features and control of negative emotional reactions, we expect that the potential carry-over effect will increase in relation to poorer sleep regulation, which will be shown in a greater difference in negative and positive affect between the reunion and baseline phase. To test this hypothesis, we will calculate the difference in proportion of the negative and positive affect during the reunion and the

baseline period. After that, the correlation between this proportional difference and the three independent variables of sleep regulation will be checked to assess the hypothesis.

First, we expect to see a positive correlation between the proportion of negative affect during the entire Still-Face Paradigm and the number of night awakenings and nocturnal wakefulness. With the total sleep duration, we expect to see a negative correlation with the proportion of negative affect. Furthermore, we expect to see a negative correlation between the proportion of positive affect during the whole Still-Face Paradigm and the number of night awakenings and nocturnal wakefulness. With the total sleep duration, we expect to see a positive correlation with this proportion of positive affect. Besides, we expect that the difference between the proportion of negative affect in the reunion phase and the baseline phase will be greater when the night awakenings and nocturnal wakefulness are higher and the total sleep duration is lower. At last, we expect that the difference between the proportion of positive affect in the reunion phase and the baseline phase will be greater when the night awakenings and nocturnal wakefulness are higher and the total sleep duration is lower.

Methods Participants

For this research we recruited infants with their caregivers from the municipality of Amsterdam and surroundings by emailing day-care centres and midwife services, spreading flyers online and distributing flyers at physical locations. The infants were between 3 and 12 months old during the lab visit (M = 4.88 months, SD = 1.70). There are 21 (13 male, 8 female) infants in total that participated in the GEAR-project, from which 2 failed to complete the whole session or had data that could not be included in this research. The remaining infants (12 male, 7 female, N = 19) were included in our final sample.

Study Design and Procedure

This research is part of a prospective research that uses the GEAR (Growing Emotion and Attention Regulation) longitudinal design. The current research consists of two measurements, which were collected as part of a larger test battery: measuring the emotion regulation within the infant with the Still-Face Paradigm in a lab visit and filling out the BISQ about sleep regulation at home by the caregiver. Both the data from the Still-Face Paradigm and the BISQ were gathered for all infants.

Still-Face Paradigm

The procedure of the Still-Face Paradigm (see Figure 1) took place in the BabyLab of the University of Amsterdam, where the infant was placed in an infant seat and the caregiver in a chair, which were faced to each other. The interaction between the caregiver and infant was recorded during the three phases of the Still-Face Paradigm with a dual-lens mobile camera that provided a 360˚ view, which was placed on a tripod standing between the participants on an approximate distance of just a little further than the end of the infant’s seat. This tripod was

placed straight in the middle of the baby’s caregiver’s body and the horizontal part of the tripod horizontal. The camera then was attached to the tripod so that the lower part of the camera was aligned with the infant’s top of the head. In the baseline phase, the caregiver played with his/her infant for two minutes without any toys or other attributes. Subsequently, when a beeping noise was heard, the still-face phase started. Here the caregiver looked at the infant with a neutral facial expression and were not allowed to have any other reactions to the infant. After one minute of the still-phase phase, another beeping noise introduced the reunion phase, in which the caregiver resumed their behaviour of the first phase for two minutes (Tronick et al., 1978).

Figure 1

Order of events in the Still-Face Paradigm

The Brief Infant Sleep Questionnaire

A week after the lab visit, the caregivers received a questionnaire about the sleep regulation of the infant via Qualtrics. This questionnaire is the Brief Infant Sleep Questionnaire, a semi-structured questionnaire with seventeen questions from which ten are about the sleeping behaviour of the infant of the prior week (Sadeh, 2004). The questionnaire takes about 5-10 minutes to answer.

Instruments

The Brief Infant Sleep Questionnaire

To obtain insight in the quality of sleep problems of the infant, we looked in this research at three measurement outcomes of the questionnaire: total sleep duration (hours and minutes), the number of night awakenings and the nocturnal wakefulness (hours and minutes). These outcomes came from questions of the BISQ and these indicators were chosen for their high specificity (Del-Ponte et al. 2020; Sadeh, 2004). The total sleep duration is measured by the questions ‘How much time (in hours and minutes) does your child spend in sleep during the NIGHT (between 7 in the evening and 7 in the morning)?’ and ‘How much time (in hours and minutes) does your child spend in sleep during the DAY (between 7 in the morning and 7 in the evening)?’. The outcome of night awakenings is obtained with the question ‘What is the average number of night waking’s per night?’. The last outcome, nocturnal wakefulness, is obtained with the question ‘How much time (in hours and minutes) during the night does your child spend in wakefulness (from 10 in the evening to 6 in the morning)’.

Video-coding

The micro-coding of emotion regulation during the Still-Face Paradigm was performed with the program Observer XT – 15 at a 1s-interval. The video data of each infant and caregiver pair (recorded at 30 frames per second) were imported into the Observer. In the coding scheme, shown in Table 1, we distinguished between the three phases of the Still-Face Paradigm, the types of gaze, the types of facial expressions and if the infant/parent vocalized or not. These variables are mutually exclusive. Furthermore, extra behaviours were coded that are not mutually exclusive. All features were coded for the infant, but for the caregiver some extra behaviours and one gaze value were not coded.

Table 1

The Coding Scheme of the Still-Face Paradigm

Situation Gaze Affect Vocalisations/ verbalizations

Extra behaviour (not mutually exclusive) Child: Baseline Look at

face/torso

Neutral No voc Self-comforting Still-face Look at hands Positive Distancing

Reunion Look away Negative Voc Crying

Not visible Touching

Parent: Baseline Look at face/torso

Neutral No voc Distancing Still-face Look away Positive Voc Touching Reunion Not visible Negative

Note. Firstly you choose the participant you are observing. Then we distinguished between the three phases

of the Still-Face Paradigm: the baseline, still-face and reunion situations. After that, we coded the gaze of the child in one of four possible values: Look at face/torso, Look at hands, Look away and Not visible. For the parent we also code the gaze values, but without the Look at hands value. Affect was coded as either

Neutral, Positive or Negative facial expressions for both parties. The presence of vocalizations/verbalizations

were also coded. At last, there were extra behaviours included in the coding scheme, which were not mutually exclusive. These extra behaviours are self-comforting, distancing, crying and touching. All extra behaviour values were coded for the infant, but for the caregiver were the extra behaviours self-comforting and crying not coded.

Reliability

To obtain a sufficient reliability of the observation, 26% of the participants (n = 5) were coded by two observers with a sufficiently high agreement. To gain this reliability, a Cohen’s Kappa statistic reliability analysis was performed within the program Observer XT – 15. For the infant affect scores, there was an average percentage of agreement of 95%. The average Kappa of the infant’s affect was .90 (max. k =.96, min. k = .86).

Data analysis

With the obtained data from the Still-Face Paradigm we looked into the proportion of affect per phase. The negative and positive affect were analysed separately to test the hypotheses. The dependent variables, the proportions of negative and positive affect, were calculated by dividing the amount of seconds that the type of affect was shown with the total amount of seconds, which is obtained out of the video micro-coding. The independent variables are the three measured outcomes of the BISQ: total sleep duration (hours and minutes), number of night awakenings and nocturnal wakefulness (hours and minutes). The distribution of these variables were checked on normality (Shapiro-Wilk test) and afterwards on their multicollinearity (Pearson Correlation Coefficient if normality p < .05; Spearman’s Rank Correlation Coefficient if normality p > .05).

Proportion of Negative and Positive Affect Corresponding to the Infant’s Sleep Regulation For testing the first two hypotheses, in which we expect poorer sleep regulation resulting in an increased proportion of negative affect and a decreased proportion of positive affect, we needed the calculated proportion of negative and positive affect shown by the infant during the entire Still-Face Paradigm per participant. The normality of the distributions of these dependent variables was tested with a Shapiro-Wilk test (p < .05) to check the assumptions for a multiple regression analysis. Finally, this analysis was used because of the insight it provides in the relation between the shown negative and positive affect that is changing during the paradigm and the three fixed sleep regulation variables independently of each other; the total sleep duration, the number of night awakenings and the hours of nocturnal wakefulness.

Proportion of the Carry-Over Effect Corresponding to the Infant’s Sleep Regulation

The final hypothesis states that the carry-over effect will increase in relation to poorer sleep regulation, which will be shown in a greater difference in negative and positive affect between the reunion and the baseline phase. We first needed to calculate the difference in proportion of negative affect between the reunion and the baseline phase, which are obtained out of the video micro-coding. Afterwards we needed to check both the difference in negative and positive affect proportions on their normality by performing a Shapiro-Wilk test (p < .05). To test this hypothesis we performed two multiple regression analyses, one with the difference in proportion of negative affect between the baseline and reunion period as dependent variable and one with the difference in proportion of positive affect.

Results

Proportion of Negative and Positive Affect During the Phases of the Still-Face Paradigm First, we assessed whether the infants in our study showed the expected pattern of changes in the proportion of negative and positive affect during the phases of the Still-Face Paradigm. Based on previous research, we expected to find more negative and less positive affect during the reunion phase and the still-face phases relative to the baseline, with the proportion of negative affect in the still-face phase exceeding the observed proportion in the reunion phase.

The proportion of negative affect was compared across the three phases of the Still-Face Paradigm using a non-parametric Friedman test as the normality (Shapiro-Wilk test, W = .90, p < .001) requirement for a Repeated Measures ANOVA was not met. As expected, the proportion of negative affect shown by each infant differed significantly across the three phases (X2_{F(2) = 18.11, p < .001; for visual inspection, see Figure 2). Infants also showed the expected} pattern of change – the greatest proportion of negativity was observed in the still-face phase relative to both the baseline (Pairwise Wilcoxon Rank Sum test, p = .0022) and nonsignificant to the reunion phase (p = .80), with the negative affect in the reunion phase exceeding that in the baseline phase (p = .051).

A non-parametric Friedman test was also performed for the proportion of positive affect across the three phases because the normality (Shapiro-Wilk test, W = .93, p = .0034) requirement for a Repeated Measures ANOVA was not met. The proportion of positive affect shown by the infant differed, as expected, significantly across the three phases (X2_{F(2) = 22.38,} p < .001; for visual inspection, see Figure 2). Nevertheless, the expected pattern of change was not fully shown – the infants showed in the still-face phase, as expected, significantly less positive affect in comparison to the baseline phase (Pairwise Wilcoxon Rank Sum test, p <.001)

and the reunion phase (p = .0024), but there was no difference between the positive affect in the reunion phase and the baseline phase (p = 1.00; mean proportions showed in Table 2).

Figure 2

The Proportion of Affect During the Still-Face Paradigm

Note. The proportion of negative affect, shown in red, seems to have increased significantly in the still-face

phase and the reunion phase compared to the baseline phase. The proportion of positive affect, shown in blue, decreased significantly in the still-face phase compared to the other phases, but did not show a great difference between the baseline and reunion phase.

* p < .01

Table 2

Descriptive statistics of the infants that performed the Still-Face Paradigm. Total sample (N = 19)

M / p SD

Demographics

1. Female .37 -

2. Male .63 -

3. Mean age in months 4.88 1.70

Sleep quality indicators

4. Total sleep duration in hours 12.63 0.93

5. Number of awakenings at night 2.42 1.07

6. Nocturnal wakefulness in hours 0.64 0.58

Negative affect proportion p SD

Total .18 .19 Baseline phase .047 .080 Still-face phase .34 .32 Reunion phase .25 .30 Baseline to still-face 1. Decrease .05 - 2. Increase .95 - Still-face to reunion 1. Decrease .68 - 2. Increase .32 - Baseline to reunion 1. Decrease .05 - 2. No change .21 -

3. Increase (carry-over effect) .74 -

Positive affect proportion p SD

Total .34 .17 Baseline phase .41 .20 Still-face phase .063 .066 Reunion phase .40 .24 Baseline to still-face 1. Decrease 1.00 - 2. Increase .00 - Still-face to reunion 1. Decrease .05 - 2. No change .16 - 3. Increase .79 - Baseline to reunion

1. Decrease (carry-over effect) .42 -

2. Increase .58 -

Note. Continuous variables are summarized by sample means; categorical variables are summarized by

Sleep Quality Indicators

The sleep quality indicators are three outcome variables from questions on the BISQ from Sadeh (2004), under which the normally distributed total sleep duration in hours (M = 12.63, SD = 0.93; Shapiro Wilk test, W = .93, p = .16), the not normally distributed number of night awakenings (M = 2.42, SD = 1.07; W = .76, p < .001) and the also not normally distributed hours of nocturnal wakefulness (M = 0.64, SD = 0.58; W = .78, p < .001). These variables are displayed in Figure 3 on their frequency and correlations with the age of the infants in months. These correlations between the normally distributed age (W = .63, p <.001) and the total sleep duration (Pearson Correlation Coefficient, r(17) = .39, p = .098), the number of night awakenings (Spearman’s Rank Correlation Coefficient, r(17) = -.38, p = .11) and the hours of nocturnal wakefulness (Spearman’s Rank Correlation Coefficient, r(17) = -.22, p = .37) were not significant.

The three sleep indicators operate independently, which was confirmed with the not significant correlations between the total sleep duration and number of night awakenings (Spearman’s Rank Correlation Coefficient, r(17) = -.23, p = .35), the total sleep duration and hours of nocturnal wakefulness (r(17) = -.040, p = .87) and the number of night awakenings and hours of nocturnal wakefulness (r(17) = .37, p = .12).

Figure 3

Frequency of Sleep Indicators and Their Correlation With Age

(a) (b)

(c) (d)

(e) (f)

Note. (a) The frequency of hours of total sleep duration in 24 hours shown by the infants. (b) The hours of

total sleep duration do not seem to correlate with the age in months. (c) The frequency of night awakenings in 24 hours shown by the infants. (d) The number of night awakenings does not seem to correlate with the age in months. (e) The frequency of hours of nocturnal wakefulness in 24 hours shown by the infants. (f) The hours of nocturnal wakefulness do not seem to correlate with the age in months.

Relation Between Total Negative Affect and the Infant’s Sleep Regulation

A multiple regression analysis was performed to assess the relationship between the overall proportion of negative affect during the Still-Face Paradigm and the three indicators of the quality of sleep regulation; the total sleep duration in hours, the number of night awakenings and the hours of nocturnal wakefulness. The model did not provide a good overall fit to the data we observed in our sample (R2 _{= .0076, F(3,15) = .038, p = .99). The individual sleep indicators} were examined further and indicated that the total sleep duration (t = -.29, p = .78), the number of night awakenings (t = -.22, p = .83) and the duration of nocturnal wakefulness (t = -.073, p = .94) are all not significant predictors in the model. These effects are shown in Figure 4. Because the number of night awakenings had too little variability to be able to predict variance in the proportion of negative affect, we performed a follow-up regression analysis comparing the mean difference in the proportion of the negative affect between two groups (participants <=2 and participants >2 night awakenings). Nevertheless, the relation was still not significant (Mann-Withney U test, W = 41, p = .90).

Relation Between Total Positive Affect and the Infant’s Sleep Regulation

To assess the relationship between the three indicators of the quality of sleep regulation with the overall proportion of positive affect during the Still-Face Paradigm, another multiple regression analysis was performed. The model also did not provide a good overall fit to the data we observed in our sample (F(3,15) = .44, p = .73, R2 _{= .081). By examining the individual} predictors further, we saw that the total sleep duration (t = .60, p = .56), the number of night awakenings (t = -.76, p = .46) and the hours of nocturnal wakefulness (t = .99, p = .34) were all not significant predictors in the model (for visual inspection, see Figure 4). We also did a follow-up regression analysis comparing the mean difference in proportion of positive affect between two groups based on night awakenings (participants <=2 and participants >2 night

awakenings), which still did not show a significant relation (Mann-Withney U test, W = 39, p = 1.00).

Figure 4

Correlations Between the Negative- and Positive Affect and the Sleep Regulation Indicators

(a) (b)

(c) (d)

(e) (f)

Note: (a) A non-significant, subtle correlation between the total sleep duration in hours and the proportion

of negative affect. (b) A non-significant, subtle correlation between the total sleep duration in hours and the proportion of positive affect. c) A non-significant, subtle correlation between the number of night

awakenings and the proportion of negative affect. (d) A non-significant, subtle correlation between the number of night awakenings and the proportion of positive affect. (e) A non-significant, subtle correlation between the hours of nocturnal wakefulness and the proportion of negative affect. (f) A non-significant, subtle correlation between the hours of nocturnal wakefulness and the proportion of positive affect.

Relation Between the Carry-Over Effect and the Infant’s Sleep Regulation

We expected that infants who are better at regulating their sleep would show a less extreme increase in negative affect during the reunion relative to the baseline phase (and more attenuated decrease in positive affect, respectively). We assessed these hypotheses by applying two multiple regression models.

The first model looked at the difference in proportion of negative affect between the baseline and the reunion phase, which was expected to have an increased proportion in the reunion phase compared to the baseline phase, linked to the sleep regulation indicators. This model showed that there was no significant relationship between the difference in proportion on the infant’s negative affect between the baseline and the reunion phase, also known as the carry-over effect (F(3, 15) = .29, p = .83, R2 _{= -.13). The individual predictors were examined} further and indicated that the total sleep duration (t = -.65, p = .53), the number of night awakenings (t = .069, p = .95; with two groups of participants <=2 and 2<, Mann-Withney U test, W = 39.5, p = 1.00) and the hours of nocturnal wakefulness (t = -.60, p =.56) are all not significant predictors for the carry-over effect with negative affect.

The second model for this hypothesis showed the non-significant relation between the difference in proportion of the infant’s positive affect between the baseline and the reunion phase and the three sleep indicators (F(3, 15) = .047, p = .99, R2 _{= .0092), which was expected} to less in the reunion phase compared to the baseline phase. However, looking at the mean proportion of positive affect of the different phases, it was indicated there was not a significant carry-over effect to begin with. By examining the individual predictors further, the total sleep duration (t = -.36, p = .72), the number of night awakenings (t = .019, p = .99; with two groups of participants <=2 and 2<, Mann-Withney U test, W = 39, p = 1.00) and the hours of nocturnal wakefulness (t = -.046, p = .96) were all not significant predictors for the carry-over effect with positive affect.

Although the carry-over effect was not found in these data, there was a significant difference between the still-face phase and the other two phases, the baseline and reunion. For exploratory purposes, we looked into the models on these relations. Nevertheless, there was not a significant relation with the three sleep indicators and the difference between the still-face and baseline phases (F(3, 15) = 1.48, p = .26, R2 _{= .23) and the difference between the still-face} and reunion phases (F(3, 15) = 1.27, p = .32, R2 _{= .20).}

Figure 5

Correlations Between the Carry-Over Effects and the Sleep Regulation Indicators

(a) (b)

(e) (f)

Note: (a) A non-significant, subtle correlation between the total sleep duration in hours and the carry-over

effect with negative affect. (b) A non-significant, subtle correlation between the total sleep duration in hours and the carry-over effect with positive affect. c) A non-significant, subtle correlation between the number of night awakenings and the carry-over effect with negative affect (d) A non-significant, subtle correlation between the number of night awakenings and the carry-over effect with positive affect. (e) A non-significant, subtle correlation between the hours of nocturnal wakefulness and the carry-over effect with negative affect. (f) A non-significant, subtle correlation between the hours of nocturnal wakefulness and the carry-over effect with positive affect.

Discussion

In this study, we examined the relationship between infant quality of sleep regulation and the ability to regulate emotion in 3-12 months old infants. The ability to regulate emotion was measured with the proportion of shown affect by the infant during the Still-Face Paradigm.

From these data we expected to see the most proportion of negative and the least proportion of positive affect in the still-face phase, but also the carry-over effect by showing more negative and less positive affect during the reunion phase compared to the baseline phase. Besides these data, we gathered data about the sleep regulation in infants with the BISQ (Sadeh, 2004). The sleep indicators total sleep duration, number of night awakenings and duration of nocturnal wakefulness were tested on their relationship with the shown emotion regulation in the Still-Face Paradigm. We expected that when the sleep regulation was poorer, the total negative affect would be higher and the total positive affect lower. Besides, we expected that the carry-over effect would increase when the sleep regulation was poorer.

The Negative and Positive Affect Shown During the Still-Face Paradigm

The proportion of negative affect shown by the infants during the Still-Face Paradigm significantly differed between the baseline phase and the two other phases, the still-face and the reunion phase. These results tell us that the infant showed more negative facial expressions during the still-face and the reunion phase, which was expected due to the change in interaction with the caregiver. The proportion of positive affect differed between the still-face phase and the other two phases, the baseline and the reunion phase, which indicates that the infant showed less positive affect during the changed interaction with the caregiver in the still-face phase.

Looking at the difference in the proportion of shown affect between the phases, there is a carry-over effect with the negative affect. These proportions of the baseline and reunion phase differed significantly and the average proportion was higher in the reunion phase. However,

there was a more subtle carry-over effect shown in the proportion of positive affect that was not significant.

The Relationship Between the Quality of Sleep and Emotion Regulation

Despite that our sample showed the expected patterns of positive and negative affect during the Still-Face Paradigm, we did not find evidence to support a significant link between the sleep regulation indicators and emotion regulation in the Still Face Paradigm. The total sleep duration, number of night awakenings and hours of nocturnal wakefulness were not predictive of the overall positive and negative affect during the Still-Face Paradigm, as well as of the observed pattern of changes in positive and negative affect from the baseline to the reunion phase.

Although, when we look at previous research on these subjects, it is likely that the relationship between the difficulties in sleep regulation and emotion regulation exists. Therefore, it seems there were other factors in our research that manipulated the results.

The Development of Emotional Regulation

Looking into previous research on the development of emotional regulation, the development of emotional regulatory functions takes place primarily in the first year of life (Thompson, 1990). In these first months, the development of emotion regulation progresses relatively fast and therefore there are a lot of changes in emotion regulatory skills between these months. Considering that in our sample only 5% (n = 1) of the infants were older than 9 months, it is possible that some expected results of the Still-Face Paradigm were not seen in our research. As said before, research of Yato et al. (2008), who found the carry-over effect in their results, showed that 9-month-old infants showed a greater carry-over effect than the subtle carry-over effect shown by 4-month-old infants. Although this effect consisted of only the negative affect,

the 4-months-old also returned to the positive emotions in the reunion phase more easily than the 9-months-old. Therefore, in our sample, that includes only one infant older than 9 months, this aftermath of decrease in positive emotions during the reunion phase could be absent.

Hence, it would be of interest to look at the same procedure with more participants who also are more distributed in age. Besides, because of the fast progress of the emotion regulation development, it would be of interest to look into more aspects of the emotion regulation then only the affect, like the gaze direction or the self-soothing behaviours of the infant. With this expansion, other features of emotion regulation can be tested on their link with the quality of sleep.

The Measure Instruments for Sleep Regulation

Another aspect that has an impact on the results is our choice of sleep indicators: the total sleep duration, the number of night awakenings and the nocturnal wakefulness in hours. These indicators are the outcomes of the Brief Infant Sleep Questionnaire, which screens sleep problems in infancy on a clinical level. As described before in this paper, the research from Del Ponte (2020) on the robustness of the BISQ showed that there were three high specificity factors that supported the validity of parents’ reports on sleep-related problems in childhood. Based on their used cut-off scores that state that the quality of sleep regulation will be poor when one of the following criteria is satisfied: total sleep <9-24 h, >3 awakenings at night and nocturnal wakefulness> 1 h., there are a few infants in our sample that have a poor sleep regulation (n = 3) and few that are just under the limit (n = 6). Because of the high specificity results, we used the outcomes of these three indicators for measuring the sleep regulation in our research. However, the obtained robustness of these indicators was also based on a clinical level, just like the BISQ, while our sample consists of non-clinical participants. Therefore, it is possible that

these indicators are not the right factors for measuring the quality of sleep regulation within our participants.

It would be of interest to look at a different instrument to measure the sleep regulation, because of this clinical level the BISQ works on. Instead of a one-time questionnaire, the caregivers could fill in a sleep diary to track the infant’s sleeping behaviours daily over a longer time span (e.g., one or two weeks). This data is still easy to gather and can measure multiple sleep parameters in various contexts, just like questionnaires, but may be more accurate for the assessment of some of these parameters. It is also not based only on clinical cases yet, so the outcomes apply to our sample. For more precise outcomes, this measurement could be combined with sleep trackers on smart devices. The accuracy of these devices becomes better each day and therefore would give some extra information on the sleep regulation, like heart rates and breathing patterns. For further research it would be interesting to look into these devices.

Expectations Based on Adult Theories

When we discuss the hypotheses about the proportion of negative and positive affect during the entire Still-Face Paradigm, we expected to see more negative and less positive facial expressions when the sleep regulation was poorer. This expectation was based on the results of previous research on the emotion regulation in infants and the effect of difficulties in sleep regulation on them. As said before, difficulties in sleep regulation have an influence on the control of negative emotions. This control would decrease due to a poor sleep regulation and thereby increase more negative emotional reactivity and decrease positive emotions. The control system for negative emotions is included in the intrinsic pathway of emotion regulation, which eventually also provides the ability to select and implement emotional strategies. The development of this intrinsic control takes place in infancy and will develop into experiences

and expressing of negative and positive emotions with the strategies (Posner & Rothbart, 2000; Feldman, 2009). Although this development takes place in the early-infancy, the link with the sleep regulation was mostly based on the functions of emotion regulation stated by Gross (2015), that are constructed mostly on adults. Therefore, the ability to select and implement emotion strategies and thereby control the negative emotional expressions could be underdeveloped in our participants. Thus, the expectation of the link between poor sleep regulation and a decrease in the control of negative emotion that was based on this ability could be incorrect.

Study Limitations and Recommendations

Furthermore, a few factors decreased the stringency of the Still-Face Paradigm and thereby could have influenced the outcomes of the emotion regulation. First, there were infants who had a baby pacifier in their mouth or another self-soothing object in their hands (n = 4). This self-soothing behaviour could have manipulated the results on the infant’s emotion regulation, since it is a feature of emotion regulation. Besides, some caregivers were still touching their child during the still-face phase (n = 5), whilst there should not be any other interaction besides the neutral facial expression and eye-contact. This touching could have influenced the behaviour of the child, because the still-face phase did not change as much from the baseline and reunion phases as it should have been. At last, there were 2 infants who were not visible for many periods in the video’s. The first infant was not attached in the seat and constantly tried to get out of it and therefore was with his back to the camera. With regard to the other infant the camera was not placed correctly causing that most of the time the top of the head was filmed and the face was not visible. In the end, there was only 63% (n = 12) of the infants who did not have anything that went wrong according to the procedure plan. Due to the small amount of participants we had, we could not eliminate these participants from the sample.

To avoid these issues, it would be convenient to imply baby pacifiers and/or other objects that fulfilled the self-soothing behavioural aspect in the coding scheme in the future. Besides, the visualisation of the participants should be better checked before the paradigm starts. During the paradigm, it is also important that caregivers do not interact with their infant in any other way than prescribed during the Still-Face Paradigm. If they still touch the infant or use vocalisation as a way to communicate, they should be informed to do otherwise via for instance a speaker.

Further Perspectives

For further research it would be of interest to replicate this research with a larger sample where the ages differ more from each other. With these differences in age, the relation with the potential carry-over effect can be shown. When the carry-over effect then is shown in the negative and positive affect, the link with the sleep regulation can again be examined. If the group sample would be larger, the difficulties around some infants with respect to the stringency of the Still-Face Paradigm would also be straightened out. With respect to the developmental issue around the emotion regulation in infancy, it would also be of interest to include a longitudinal aspect to this research. If the infants were examined at the age of 3-6 and 9-12 months, the difference in emotion regulation between the infants who have had a poor and normal sleep regulation at a young age can be taken into the effects on a developmental perspective. Especially within this longitudinal design, it would be a valuable addition to imply more features of the emotion regulation that can be measured within the Still-Face Paradigm. With those features included, it would give information about the development of these skills and the potential relationship with sleep regulation on these developments, with more perspectives on the effects on later issues in emotion regulation.

It would also be of interest to look at a different instrument to measure the sleep regulation, because of the clinical level the BISQ works on. Instead of a questionnaire, the caregivers would fill in a sleep diary for each day a week long. This data is still easy to gather and can measure multiple sleep parameters in various contexts, just like questionnaires, but may be more accurate for the assessment of some of these parameters. It is also not based only on clinical cases yet, so the outcomes apply to our sample (Sadeh, 2015). It can also be combined with an actigraphy, which is an instrument that measures activity and gives accurate information about the total sleep duration (Camerota, Tully, Grimes, Gueron-Sela & Propper, 2018).

Conclusion

To conclude from the results of this paper, we did not find evidence supporting the relationship between the quality of sleep regulation and the emotion regulation. Nevertheless, there are some factors that could have limited our results like the distribution of age in our sample and thereby the limitations on the development of emotion regulation skills. The limitations of the BISQ and the choice of only three sleep indicators could have limited our results as well. Due to these limitations, this research did not give enough evidence to rule the relationship between the quality of sleep regulation and the emotion regulation out. Because of the importance of the development of emotion regulation on later life emotional and behavioural abilities, this relationship with sleep quality should still be examined to make it possible to intervene possible early-on issues with emotion regulation. Therefore, further research should be executed.

References

Adamson, L. B., & Frick, J. E. (2003). The Still Face: A History of a Shared Experimental Paradigm. Infancy, 4(4), 451–473. https://doi.org/10.1207/s15327078in0404_01

Biederman, J., Rosenbaum, J. E, Hirshfeld, D. R., Faraone, S. V., Bouldouc, E. A., Gersten, M., Meminger, S., Kagan, J., Snidman, N., & Reznick, J. S. (1990). Psychiatric Correlates of Behavioral Inhibition in Young Children of Parents With and Without Psychiatric Disorders. Archives of General Psychiatry, 47, 21–26.

https://doi.org/10.1001/archpsyc.1990.01810130023004

Berger, R. H., Miller, A. L., Seifer, R., Cares, S. R., & LeBourgeois, M. K. (2012). Acute sleep restriction effects on emotion responses in 30-to 36-month-old children. Journal of sleep research, 21(3), 235-246. https://doi.org/10.1111/j.1365-2869.2011.00962.x

Camerota, M., Tully, K. P., Grimes, M., Gueron-Sela, N., & Propper, C. B. (2018).

Assessment of infant Sleep: how well do multiple methods compare? Sleep, 41(10),

https://doi.org/10.1093/sleep/zsy146

Del-Ponte, B., Xavier, M. O., Bassani, D. G., Tovo-Rodrigues, L., Halal, C. S., Shionuma, A. H., Ulguim, K. F., & Santos, I. S. (2020). Validity of the Brief Infant Sleep

Questionnaire (BISQ) in Brazilian children. Sleep Medicine, 69, 65–70.

https://doi.org/10.1016/j.sleep.2019.12.018

Denham, S. A., Blair, K. A., DeMulder, E., Levitas, J., Sawyer, K., Auerbach–Major, S., & Queenan, P. (2003). Preschool Emotional Competence: Pathway to Social

Competence?. Child Development, 74(1), 238-256. https://doi.org/10.1111/1467-8624.00533

Feldman, R. (2009). The Development of Regulatory Functions From Birth to 5 Years: Insights From Premature Infants. Child Development, 80(2), 544-561.

https://doi.org/10.1111/j.1467-8624.2009.01278.x

Geva, R., & Feldman, R. (2008). A neurobiological model for the effects of early brainstem functioning on the development of behavior and emotion regulation in infants: implications for prenatal and perinatal risk. Journal of Child Psychology and Psychiatry, 49(10), 1031–1041. https://doi.org/10.1111/j.1469-7610.2008.01918.x

Gross, J. J. (1998). The Emerging Field of Emotion Regulation: An Integrative Review. Review of General Psychology, 2(3), 271–299. https://doi.org/10.1037/1089-2680.2.3.271

Gross, J. J. (2015). Emotion Regulation: Current Status and Future Prospects. Psychological Inquiry, 26(1), 1–26. https://doi.org/10.1080/1047840x.2014.940781

Hysing, M., Sivertsen, B., Garthus-Niegel, S., & Eberhard-Gran, M. (2016). Pediatric sleep problems and social-emotional problems. A population-based study. Infant Behavior and Development, 42, 111–118. https://doi.org/10.1016/j.infbeh.2015.12.005

Keenan, K. (2006). Emotion Dysregulation as a Risk Factor for Child Psychopathology. Clinical Psychology: Science and Practice, 7(4), 418–434.

https://doi.org/10.1093/clipsy.7.4.418

Mauss, I. B., Troy, A. S., & LeBourgeois, M. K. (2013). Poorer sleep quality is associated with lower emotion-regulation ability in a laboratory paradigm. Cognition & Emotion, 27(3), 567–576. https://doi.org/10.1080/02699931.2012.727783

Mesman, J., Linting, M., Joosen, K. J., Bakermans-Kranenburg, M. J., & van IJzendoorn, M. H. (2013). Robust patterns and individual variations: Stability and predictors of infant behavior in the still-face paradigm. Infant Behavior and Development, 36(4), 587–598.

https://doi.org/10.1016/j.infbeh.2013.06.004

Mesman, J., van IJzendoorn, M. H., & Bakermans-Kranenburg, M. J. (2009). The many faces of the Still-Face Paradigm: A review and meta-analysis. Developmental Review, 29(2), 120–162. https://doi.org/10.1016/j.dr.2009.02.001

Motomura, Y., Kitamura, S., Oba, K., Terasawa, Y., Enomoto, M., Katayose, Y., Hida, A., Moriguchi, Y., Higuchi, S., & Mishima, K. (2013). Sleep Debt Elicits Negative Emotional Reaction through Diminished Amygdala-Anterior Cingulate Functional Connectivity. PLoS ONE, 8(2), e56578. https://doi.org/10.1371/journal.pone.0056578

Palmer, C. A., & Alfano, C. A. (2017). Sleep and emotion regulation: An organizing, integrative review. Sleep Medicine Reviews, 31, 6–16.

https://doi.org/10.1016/j.smrv.2015.12.006

Posner, M. I., & Rothbart, M. K. (2000). Developing mechanisms of self-regulation. Development and Psychopathology, 12(3), 427-441.

https://doi.org/10.1017/s0954579400003096

Sadeh, A. (2004). A Brief Screening Questionnaire for Infant Sleep Problems: Validation and Findings for an Internet Sample. PEDIATRICS, 113(6), e570–e577.

https://doi.org/10.1542/peds.113.6.e570

Sadeh, A. (2015). III. SLEEP ASSESSMENT METHODS. Monographs of the Society for Research in Child Development, 80(1), 33–48. https://doi.org/10.1111/mono.12143

Sheppes, G., Suri, G., & Gross, J. J. (2015). Emotion Regulation and Psychopathology. Annual Review of Clinical Psychology, 11(1), 379–405.

https://doi.org/10.1146/annurev-clinpsy-032814-112739

Sivertsen, B., Harvey, A. G., Reichborn-Kjennerud, T., Torgersen, L., Ystrom, E., & Hysing, M. (2015). Later Emotional and Behavioral Problems Associated With Sleep

Problems in Toddlers. JAMA Pediatrics, 169(6), 575.

https://doi.org/10.1001/jamapediatrics.2015.0187

Thompson, R. A. (1991). Emotional regulation and emotional development. Educational Psychology Review, 3(4), 269–307. https://doi.org/10.1007/bf01319934

Tronick, E., Als, H., Adamson, L., Wise, S., & Brazelton, T. B. (1978). The Infant’s

Response to Entrapment between Contradictory Messages in Face-to-Face Interaction. Journal of the American Academy of Child Psychiatry, 17(1), 1–13.

https://doi.org/10.1016/s0002-7138(09)62273-1

Tucker, A. M., Whitney, P., Belenky, G., Hinson, J. M., & Van Dongen, H. P. A. (2010). Effects of Sleep Deprivation on Dissociated Components of Executive Functioning. Sleep, 33(1), 47–57. https://doi.org/10.1093/sleep/33.1.47

van der Helm, E., Gujar, N., & Walker, M. P. (2010). Sleep Deprivation Impairs the Accurate Recognition of Human Emotions. Sleep, 33(3), 335–342.

https://doi.org/10.1093/sleep/33.3.335

Weller, A., & Feldman, R. (2003). Emotion regulation and touch in infants: the role of cholecystokinin and opioids. Peptides, 24(5), 779–788. https://doi.org/10.1016/s0196-9781(03)00118-9

Yato, Y., Kawai, M., Negayama, K., Sogon, S., Tomiwa, K., & Yamamoto, H. (2008). Infant responses to maternal still-face at 4 and 9 months. Infant Behavior and Development, 31(4), 570–577. https://doi.org/10.1016/j.infbeh.2008.07.008

Zohar, D., Tzischinsky, O., Epstein, R., & Lavie, P. (2005). The Effects of Sleep Loss on Medical Residents’ Emotional Reactions to Work Events: a Cognitive-Energy Model. Sleep, 28(1), 47–54. https://doi.org/10.1093/sleep/28.1.47