- Email: [email protected]
Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants
G Model
ARTICLE IN PRESS
BBR-10496; No. of Pages 9
Behavioural Brain Research xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Behavioural Brain Research journal homepage: www.elsevier.com/locate/bbr
Research report
Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants Roseriet Beijers ∗ , Marianne Riksen-Walraven, Katharina Sebesta, Carolina de Weerth Department of Developmental Psychology, Behavioral Science Institute, Radboud University, The Netherlands
h i g h l i g h t s • • • • •
Do secure and insecure infants use different regulatory behaviors when stressed? Insecure infants fussed less when physiologically stressed. Secure infants showed more self-soothing when physiologically stressed. 29% explained variance in cortisol by regulatory behaviors in mother’s absence. Behavior and cortisol are differentially associated in secure and insecure infants.
a r t i c l e
i n f o
Article history: Received 28 July 2016 Received in revised form 15 September 2016 Accepted 4 October 2016 Available online xxx Keywords: Attachment Regulation Behavior Cortisol HPA-axis Stress
a b s t r a c t In this study we examined whether securely versus insecurely attached infants use different regulatory behaviors in absence of their mother and whether these regulatory behaviors are differentially associated with physiological stress responses in secure versus insecure infants. Participants were 193 one-year-olds and their mothers. During three 3-min episodes of separation from the mother in the Strange Situation Procedure (SSP) [1], the following infant regulatory behaviors were observed: crying, fussing, self-soothing, manipulation of toys, and manipulation of the door. Salivary cortisol was measured at baseline and 25, 40, and 60 min after the SSP to measure reactivity and recovery. Additionally, infants were classified as securely or insecurely attached to their mothers. During the mother’s absence, secure infants engaged more in manipulation of the door than insecure infants. Furthermore, in insecure (but not secure) infants less fussing was associated with higher cortisol reactivity, while in secure (but not insecure) infants more self-soothing was associated with higher cortisol reactivity. In total, 29% of the variance in cortisol reactivity was explained by infant regulatory behaviors in the mother’s absence. Cortisol recovery was not predicted by infant regulatory behaviors in the mother’s absence. To conclude, the results show differential associations between behavioral and cortisol responses to a stressor in secure versus insecure infants. This might indicate that secure and insecure infants apply different behavioral regulatory strategies when physiologically stressed. © 2016 Elsevier B.V. All rights reserved.
1. Introduction The notion that infant attachment quality is related to later socio-emotional development and mental health is one of attachment theory’s most well-known contributions to the field of developmental psychology. For example, securely attached children are considered better able to regulate themselves, i.e. more capable of applying behaviors that modulate their physiological
∗ Corresponding author at: Developmental Psychology, Behavioral Science Institute, Radboud University, P.O. Box 9140, 6500 HE Nijmegen, The Netherlands. E-mail address: [email protected] (R. Beijers).
stress [2,3]. However, the specific types of regulatory behaviors that secure versus insecure infants use to regulate their physiological stress are as yet unclear. The purpose of the present study was to first investigate whether securely versus insecurely attached infants use different regulatory behaviors in absence of their mother, and second whether these regulatory behaviors are differentially associated with physiological stress responses in secure versus insecure infants. By the end of the first year, infants can be classified as securely versus insecurely attached to a caregiver by observing them in the Strange Situation Procedure (SSP) [1]. In the SSP, children are exposed to a series of mild stressors (i.e., confrontation with a stranger and two brief separations from the caregiver), to observe
http://dx.doi.org/10.1016/j.bbr.2016.10.008 0166-4328/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9 2
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
whether they can use the caregiver as a ‘secure base’ in times of uncertainty or stress. When distressed, securely attached infants seek proximity and contact with the attachment figure. The attachment figure can then comfort them and the infants can return to exploration and play. By their attachment behaviors, secure infants apparently elicit external regulation of stress by the attachment figure. Such ‘co-regulation’ is assumed to protect the infant from excessive levels of distress and to promote the infant’s own emergent capacity to regulate arousal [3]. Regulatory capacities gradually evolve into more elaborated and self-initiated processes during the first years of life as a result of increasing control capacities, and are important to cope with developmental challenges. Through a history of co-regulation, secure infants are therefore suggested to develop better self-regulatory capacities, when compared to insecure infants [2]. As yet, it is however unclear what types of specific regulatory behaviors secure and insecure infants use when the attachment figure is not present and whether these regulatory behaviors are differentially associated with physiological stress responses in secure versus insecure infants. Spangler and Schieche took a first step in examining this question [4]. They suggested two models to explain how behavior can be associated with physiological reactions in secure and insecure attached infants. An arousal model implies that stressful events cause a general stress response, resulting both in physiological and behavioral stress responses [5]. Contrarily, in a coping model, stressful situations elicit physiological stress responses only when appropriate behavioral coping strategies are not available or effective. In their research, Spangler and Schieche studied the association between negative emotional expressions (sum of crying, fussing, and sad faces) and cortisol reactivity to the SSP in 12-month-old infants [4]. Cortisol is the end product of the hypothalamic-pituitary-adrenocortical (HPA)-axis, which is activated when confronted with uncertain, unpredictable and stressful situations Ref. [6]. Spangler and Schieche found a positive association between negative emotional expressions and cortisol reactions in insecure infants, but not in secure infants, who showed no association between negative emotional expression and cortisol reactivity [4]. The authors explained this lack of association in secure infants using a coping model, whereas they interpreted the positive behavior-physiology association of insecure infants using an arousal model. According to this reasoning, insecure infants would lack appropriate behavioral coping strategies in the face of maternal separation distress. Negative emotional expressions in this group would be a sign of general stress responses. In contrast, secure infants would show negative emotional expression as a coping strategy, for example to reestablish proximity with the attachment figure, with the consequence that a physiological response does not occur. However, in situations where coping strategies result ineffective, secure infants would also show negative emotional expressions as a sign of a general stress response. Spangler and Schieche did not examine other types of regulatory behaviors [4]. Also, cortisol recovery was not included, while one may speculate that regulatory behaviors are not only important for regulating the stress response, but also for recovering from stress. Moreover, the groundwork laid by Spangler and Schieche was unfortunately not elaborated by other attachment researchers [4]. More recently, however, the association between affective states and cortisol reactivity in response to another type of stress-inducing procedure, namely the still-face procedure, was studied in very low birth weight (VLBW) and full-term infants [7]. They found a significant negative association between positive as well as negative affect and cortisol reactivity for full-term infants. Thus, the expression of both more positive and more negative affect was related to lower cortisol levels in these infants. By contrast, the associations in the VLBW group were limited. The authors
explained their findings by suggesting that less synchrony (i.e., less correlation, either positive or negative) between a behavioral and a physiological response would correspond to a lack of coordination among stress systems, and that the lack of synchrony would represent greater dysregulation [7]. This explanation may also be applicable to a study that showed a greater dissociation between behavioral distress and cortisol reactions after an inoculation in 2-month-olds who were later classified as insecurely attached, as compared to securely attached [8]. The first goal of the present study was to investigate whether secure and insecure infants engage in different regulatory behaviors when separated from the mother. During the episodes of maternal separation within the Strange Situation Procedure (SSP) [1], we observed behaviors that have been suggested to have a regulatory function. Next to crying and fussing, we extended previous research by also including self-soothing, manipulation of toys and manipulation of the door as regulatory behaviors [9]. The second goal of this study was to investigate whether the associations between cortisol responses (reactivity and recovery) and infant regulatory behaviors during the mother’s absence are moderated by infant attachment status (secure or insecure). When examining the differential associations, no hypotheses were made as three alternative models existed: (1) the arousal model expects a positive association between physiological and behavioral stress responses in both securely and insecurely attached infants [5], (2) the coping model expects a positive association between physiological and behavioral stress responses only in insecurely attached infants [4], (3) the model of synchrony expects a positive or negative association between physiological and behavioral stress responses only in securely attached infants [7]. Although securely attached children are considered better able to regulate themselves [2,3], the specific types of regulatory behaviors that secure versus insecure infants use to regulate their physiological distress are as yet unclear. The results will provide more knowledge on the normative development of children’s regulatory strategies within the secure attachment relationship and the nature of regulation difficulties identified in insecure infants. 2. Methods 2.1. Participants This study is part of a prospective longitudinal project starting in pregnancy and examining early caregiving factors and their influences on children’s development and health across the first years of life. The subjects were healthy infants, whose mothers were recruited during pregnancy through flyers dispersed in midwife practices in and around the Dutch cities of Nijmegen and Arnhem. Inclusion criteria were an uncomplicated and singleton pregnancy without drug use, a clear understanding of the Dutch language, and no current physical and mental health problems. The study was approved by the Faculty Ethical committee and all mothers gave written informed consent before starting. A total of 220 mothers enrolled voluntarily in the study. The final sample was limited to 193 women and their infants, as 8 women were excluded due to medical reasons and 19 discontinued the study during the first three postpartum months because of personal circumstances. No differences in demographics were found between mothers who took part in the study and the 19 women who dropped out. 2.2. Procedure At 12 months of age (M = 53 weeks and 6 days, SD = 19 days), mother-infant pairs participated in the Strange Situation Procedure (SSP) [1] in the laboratory of the Radboud University, Nijmegen.
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
This 21-min procedure consists of eight sequenced episodes. The episodes contain mildly stressful situations of different kinds, which are assumed to be increasingly disturbing. Respectively, they contain exposure to an unfamiliar environment, exposure to a stranger, and two brief separations from the mother (each followed by a reunion). The unfamiliar environment corresponded to a large child-proof room with a one-way screen, two cameras, toys for the infant to play with, and two chairs for the mother and the stranger. The stranger was always a trained female psychology student. During the first separation, the infant was left alone with the stranger for one episode. The second separation entailed the infant being left alone for one episode and then being together with the stranger for a subsequent episode. The two cameras were operated by an assistant in the control room. The assistant switched from one camera to the other to obtain the best picture of the child. The SSP was recorded on videotape for later classification of infantmother attachment and coding infant regulatory behaviors. Saliva samples were taken before the SSP (T1) and at 25 (T2), 40 (T3), and 60 min (T4) after the SSP to measure infants’ cortisol reactivity and recovery.
Table 1 Definitions of the observed categories of infant regulatory behavior in the absence of the mother. Code
Definition
A Negative emotional expression
All negative vocal emotional expressions of the infant. Infant negative vocal expression with high intensity and for longer than 3 of the 5 s. Infant negative vocal expression with low intensity or cries not longer than 3 of the 5 s.
A1 Crying
A2 Fussing
B Visual orientation B1 to toys B2 to doora C Manipulation
C1 of toys C2 of door
2.3. Measures
D Self-soothing
2.3.1. Infant cortisol responses Infant saliva samples were collected with Sorbette eye sponges [10]. The samples were taken pre-stressor (T1) and 25, 40 and 60 min post-stressor (T2–T4). After collection, the samples were stored in a −25 ◦ C freezer. Cortisol analysis was conducted at the Laboratory of Endocrinology of the University Medical Center of Utrecht University, making use of radioimmunoassay. The lower limit of detection was 1.0 nmol/l, and inter-assay and intra-assay variation were below 10% [10]. Acceptable time ranges for the poststressor samples were 20–30 min for T2, 35–50 min for T3, and 55–75 min for T4. Valid cortisol samples were available for 176 infants at T1, 171 infants at T2, 173 infants at T3, and 148 infants at T4. Cortisol reactivity to the SSP was represented by the highest post-stressor concentration (T2 or T3) and corrected for prestressor concentration (T1) to take baseline values into account [11]. An independent-samples t-test indicated that the magnitude of the infants’ peak at T2 (M = 9.23, SD = 4.45) was not different from the magnitude at T3 (M = 8.43, SD = 4.09, t(174) = 1.24, p = 0.219), and half of the infants (51%) showed peak values at T2, and the other half at T3. In line with Beijers et al. [11], cortisol recovery was measured by obtaining the difference between the last post-stressor sample and the pre-stressor sample (T4–T1), and correcting for the peak (T2 or T3) in the analyses on recovery. This measure reflects infants’ recovery from the cortisol reaction, with higher scores reflecting slower recovery. Peaking at T2 versus T3 was not related to recovery (Pearson’s r = 0.12, p > 0.10). 2.3.2. Infant regulatory behaviors The video recordings of the SSP were observed by two independent assistants who were trained and reliable, and blind to the infant cortisol responses and other data. The inter-rater reliability was good (Spearman’s rho’s ranging between 95.2% and 97.5%). The occurrence of seven specific behaviors reflecting the infant’s negative emotional expression, visual orientation, and manipulation behavior (see Table 1) was coded in 5-s intervals. With the use of a software program (The Observer, Noldus Information Technology, Wageningen, The Netherlands), the coders rated every 5 s whether the infant had shown the specific behavior (1) or not (0). For each behavioral category the number of infant behaviors was summed up per episode, divided by the total number of intervals for that episode and multiplied by 100, resulting in a percentage between 0 and 100. To assess infant regulatory capacities in
3
a
Every looking behavior that is directed at something. Infant looks at the toys or in their direction. Infant looks at the door or in its direction. Visual orientation and action (touching) that is directed at something. Infant directs attention and action to the toys. Infant directs attention and action to the door. Infant directs attention and action to self or clothing. Examples are thumb sucking and plucking at clothes.
The door is the room’s only door and thus the door through which the mother
left.
absence of the mother, the regulatory behaviors shown during the three episodes in which the infant was separated from the mother were summed up (episodes 3, 5, and 6). To be able to control for infant regulatory capacities in presence of the mother, the behaviors during the episodes in which the mother was present were summed up (episodes 1, 2, 4 and 7). Visual orientation to the toys and manipulation of the toys were positively correlated (r = 0.84, p < 0.001) as well as visual orientation to the door and manipulation of the door (r = 0.42, p < 0.001), and these positively correlated and related behaviors were therefore averaged. This resulted in the investigation of five behavioral responses: crying, fussing, selfsoothing, visual orientation to and manipulation of the toys, and, visual orientation to and manipulation of the door.
2.3.3. Infant attachment The video recordings of all SSP’s were coded by trained and supervised examiners at the Institute of Child Development, University of Minnesota. Every infant was assigned a secure (B) or insecure (avoidant (A), resistant (C), or disorganized (D)) attachment pattern [1,12]. Inter-observer reliability, based on a subset of 46 infants, was good (Cohen’s K of 0.82, intraclass correlation of 0.86; see also Beijers et al. [13]).
2.3.4. Confounders The time of day when the SSP took place and the time since the last feeding and sleep before the start of the SSP were included as confounders. As this study specifically focused on the regulatory behaviors infants use when physiologically distressed and the mother is absent, we also controlled for these regulatory behaviors (crying, fussing, self-soothing, visual orientation to and manipulation of the toys, and, visual orientation to and manipulation of the door) when the mother was present during the SSP (episodes 1, 2, 4 and 7).
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
4
Table 2 Descriptive statistics of the study variables, separately for securely and insecurely attached infants. Secure attachment N = 127
Insecure attachment N = 56
Mean
SD (range)
Mean
SD (range)
Demographics Maternal age
32.17
3.82 (21.10–42.90)
33.15
3.57 (25.10–40.30)
Maternal educational level (%) Primary education Secondary education College or university
2.4 20.8 76.8
1.9 22.7 75.4
Maternal marital status (%)a Living with partner
98.4
96.4
Infant sex (%) Girl
44.9
48.2
Birth order (%) Firstborn
46.4
30.9
Infant cortisol regulation Pre-stressor (T1) Reactivity (highest value T2/T3, nmol/l)* Recovery (T4-T1, nmol/l)b
7.19 8.38 −0.20
3.22 (3.30–22.00) 4.10 (3.10–24.00) 4.07 (−15.10–10.50)
7.90 9.90 −0.35
3.64 (3.70–21.00) 4.53 (4.40–22.00) 4.51 (−14.90–9.20)
Regulatory behaviors in mother’s absence Crying† Fussing Self-soothing Orientation to and manipulation of toys Orientation to and manipulation of door*
28.22 11.07 5.35 65.92 18.36
30.10 (0.00–100.00) 12.58 (0.00–58.33) 10.55 (0.00–52.22) 23.08 (11.67–100.00) 13.48 (0.00–71.67)
37.26 9.98 7.51 61.26 14.12
34.62 (0.00–100.00) 12.05 (0.00–50.51) 15.07 (0.00–66.67) 27.40 (0.00–100.00) 12.24 (0.00–50.00)
Regulatory behaviors in mother’s presence Crying** Fussing** Self-soothing Orientation to and manipulation of toys** Orientation to and manipulation of door
3.19 3.27 1.32 87.14 3.24
5.10 (0.00–30.88) 4.09 (0.00–23.96) 3.34 (0.00–20.96) 10.96 (51.09–100.00) 2.69 (0.00–10.42)
6.77 6.01 2.47 81.04 2.63
8.82 (0.00–41.67) 5.65 (0.00–20.49) 5.18 (0.00–24.07) 13.09 (51.77–100.00) 2.70 (0.00–10.13)
Remaining confounders Time of the day SSP (h) Hours since last feeding before SSP (h) Hours since last sleeping before SSP (h)
11:18 1:14 1:34
2:01 (8:52–17.31) 43 min (0:00–3:07) 1:01 (0:00–3:36)
11:15 1:23 1:37
1:57 (8:52–15:07) 45 min (0:00–3:21) 1:10 (0:00–4:48)
†
a b † * **
In wedlock or unmarried. Lower recovery scores indicate quicker recovery. p < 0.10. p < 0.05. p < 0.01.
2.4. Statistical analyses Of the 193 infants who participated in the project, 186 took part in the SSP [1]. The demographic characteristics of this group are presented in Table 2. Three infants were not classified as securely or insecurely attached because of doubt. In the event of skewed data, square root (self-soothing) or a logarithmic transformation (cortisol peak reactivity) was applied when appropriate, and the transformed variables were used for the analyses. No outliers were detected in the data. Subsequently, a dummy variable was created for attachment security, with secure infants serving as the reference group. Independent-samples t-tests were used to test whether secure and insecure infants engaged in different regulatory behaviors when exposed to a stressor when their mother was not present. To test whether specific behavioral responses are differently related to cortisol reactivity and cortisol recovery among secure and insecure infants, two standard multiple hierarchical regression analyses were conducted for cortisol reactivity and cortisol recovery. Infant behavioral variables were centered for their use in the creation of the interaction variables, which were calculated by multiplying the attachment security dummy by the centered behavioral variables [14]. The first regression analysis contained all predictors, interaction variables and confounders. Those variables that explained at least 1% of the variance, calculated as (part
correlation)2 × 100, were further used for the second regression analysis, in which all maintained confounders were entered in hierarchical step 1 and all maintained predictors were entered in hierarchical step 2. This approach was chosen to eliminate irrelevant confounders as well as predictors and to thereby increase power [11,13]. 3. Results 3.1. Preliminary analyses See Table 2 for the descriptive statistics of the study variables. Of the 183 infants, 127 (69%) were classified as securely attached and 56 (31%) as insecurely attached (9 insecure-avoidant, 21 insecureresistant, and 26 insecure-disorganized). Independent-samples t-tests revealed that insecure infants showed higher cortisol reactivity to the SSP (t(175) = −2.53, p = 0.01). During the mother’s presence, insecure infants cried and fussed more (t(79.47) = −3.00, p = 0.004 and t(81.43) = −3.27, p = 0.002, respectively), and also engaged less in orientation to and manipulation of the door (t(90.49) = 3.05, p = 0.003). Table 3 shows the Pearson correlations for cortisol reactivity, cortisol recovery, and the infant regulatory behaviors, presented separately for the secure and insecure infants. In secure infants, more crying and self-soothing was associated with higher corti-
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
5
Table 3 Pearson correlations between infant regulatory behaviors in mother’s absence and cortisol variables. 1.
2.
3.
4.
5.
6.
7.
Total sample 1. Reactivity 2. Recovery 3. Crying 4. Fussing 5. Self-soothing 6. Orientation to and manipulation of toys 7. Orientation to and manipulation of door
– 0.39** 0.45** 0.01 0.19* −0.39** 0.13†
– 0.18* −0.15† 0.20* −0.20* 0.05
– −0.03 0.20** −0.78** 0.39**
– 0.06 −0.12† 0.29**
– −0.18* 0.00
– −0.49**
–
Securely attached infants 1. Reactivity 2. Recovery 3. Crying 4. Fussing 5. Self-soothing 6. Orientation to and manipulation of toys 7. Orientation to and manipulation of door
– 0.41** 0.48** 0.08 0.36** −0.40** 0.17†
– 0.28** −0.06 0.26** −0.24* 0.12
– −0.07 0.15† −0.81** 0.40**
– −0.02 −0.18* 0.31**
– 0.17† −0.01
– −0.54**
–
Insecurely attached infants 1. Reactivity 2. Recovery 3. Crying 4. Fussing 5. Self-soothing 6. Orientation to and manipulation of toys 7. Orientation to and manipulation of door
0.38* 0.34* −0.14 −0.16 −0.36** 0.10
−0.05 −0.33* 0.08 −0.14 −0.13
0.06 0.27* −0.72** 0.44**
0.22 −0.01 0.18
−0.19 0.02
−0.43**
† * **
p < 0.10. p < 0.05. p < 0.01.
sol reactivity, whereas more orientation to and manipulation of toys was associated with less cortisol reactivity. In insecure infants, more crying was related to higher cortisol reactivity, whereas more orientation to and manipulation of the toys was related to less cortisol reactivity. With regard to cortisol recovery, in secure infants more crying and self-soothing were related to slower cortisol recovery, whereas more orientation to and manipulation of toys was related to quicker cortisol recovery. For insecure infants, in contrast, more fussing was related to quicker cortisol recovery.
3.2. Main analyses Independent-samples t-tests revealed that insecure infants engaged less in orientation to and manipulation of the door during the mother’s absence (t(181) = 2.02, p = 0.045), when compared to secure infants (see Table 2). A marginally significant difference was also found, indicating that insecure infants cried more during the mother’s absence (t(181) = −1.79, p = 0.079), when compared to secure infants. The regression models, which only contain the variables that individually explained at least 1% of the variance in the first analyses, are summarized in Table 4. This table shows that none of the regulatory behaviors in the mother’s presence were predictive of infant cortisol reactivity, while the regulatory behaviors in the mother’s absence explained 29% of the variance in cortisol reactivity to the SSP. Furthermore, the regression results showed that higher reactivity was significantly predicted by more crying and more self-soothing during the mother’s absence. Moreover, higher pre-stressor cortisol values significantly predicted infant cortisol reactivity to the SSP, and insecure infants tended to show higher cortisol reactivity than secure infants. Furthermore, significant interaction effects were obtained for attachment × fussing and attachment × self-soothing. Figs. 1 and 2 depicts these interaction effects. The simple slopes of the regression lines reflecting the relation between regulatory behaviors and cortisol reactivity for infants with an insecure or secure attachment
were tested for significance, using the method of Hayes and Matthes [15]. For the interaction attachment × fussing, only the slope for the insecure infants was significant (b = − 0.00, t = −2.21, p < 0.05), but not the slope for the secure infants (b = 0.00, t = −1.13, p > 0.05). This indicates that infants with an insecure attachment fuss less when their cortisol reactivity concentrations are higher. For the interaction attachment × self-soothing, only the slope for the secure infants was significant (b = 0.02, t = 3.76, p < 0.001), but not the slope for the insecure infants (b = −0.01, t = −1.937, p > 0.05). This indicates that infants with a secure attachment manipulated themselves more (e.g. thumb-sucking, plucking at their clothes) when their cortisol reactivity concentrations were higher. For cortisol recovery, only the cortisol reactivity peak concentrations were able to significantly predict recovery, indicating that cortisol recovery concentrations were higher when the cortisol peak was higher.
4. Discussion We examined whether securely versus insecurely attached infants use different regulatory behaviors in absence of their mother and whether these regulatory behaviors are differentially associated with physiological stress responses in secure versus insecure infants. When compared to secure infants, insecure infants engaged less in manipulation of the door, and tended to cry more when the mother is absent. When controlling for infant regulatory behaviors in the mother’s presence, cortisol reactivity was predicted by the interaction effects attachment × fussing and attachment × self-soothing. These interaction effects indicated that insecure (but not secure) infants fussed less when their cortisol reactivity was higher, while secure (but not insecure) infants showed more self-soothing when their cortisol reactivity was higher. In total, 29% of the variance in cortisol reactivity was explained by infant regulatory behaviors in the mother’s absence. Cortisol recovery was not predicted by these regulatory behaviors.
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9 6
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
Table 4 Final multiple hierarchical regression models for the prediction of infant cortisol responses to the SSP. B

p-value
Reactivity Step 1 0.013 0.234 0.000 Pre-stressor cortisol concentrations 0.048 0.117 0.067 Attachment (1 = secure, 2 = insecure) Step 2 Crying in mother’s absence 0.001 0.516 0.000 Fussing in mother’s absence 0.001 0.103 0.167 Self-soothing in mother’s absence 0.015 0.293 0.000 Step 3 Attachment × crying in mother’s absence 0.000 −0.136 0.102 Attachment × fussing in mother’s absence −0.002 −0.168 0.028 Attachment × self-soothing in mother’s absence −0.021 −0.250 0.004 Excluded variables: The time of day when SSP took place – hours since last feeding before SSP – hours since last sleeping before SSP – crying, fussing, self-soothing, orientation to and manipulation of the toys, and, orientation to and manipulation of the door in mother’s presence – orientation to and manipulation of the toys, and, orientation to and manipulation of the door in mother’s absence and their interactions with attachment. Recovery Step 1 8.132 0.372 0.000 Peak cortisol reactivity −0.710 −0.079 0.392 Attachment (1 = secure, 2 = insecure) −8.064 −0.061 0.496 Hours since last feeding before SSP −0.020 −0.152 0.117 Orientation to and manipulation of door in mother’s presence Step 2 0.005 0.115 0.393 Crying in mother’s absence −0.012 −0.109 0.329 Fussing in mother’s absence 0.004 0.078 0.488 Orientation to and manipulation of door in mother’s absence Step 3 −0.013 −0.177 0.124 Attachment × crying in mother’s absence −0.024 −0.114 0.284 Attachment × fussing in mother’s absence Excluded variables: The time of day when SSP took place – hours since last sleeping before SSP – crying, fussing, self-soothing, and orientation to and manipulation of the toys in mother’s presence – self-soothing and orientation to and manipulation of the toys in mother’s absence and their interactions with attachment – interaction between attachment and orientation to and manipulation of door in mother’s absence.
R2
F change
p-value
0.077
7.112
0.001
0.204
15.810
0.000
0.086
7.401
0.000
0.163
5.213
0.001
0.029
1.254
0.294
0.029
1.902
0.155
change
Notes: B = regression coefficient,  = standardized regression coefficient, R2 change = explained variance by the added step. Excluded variables explained individually less than 1% variance in cortisol reactivity and recovery in preliminary regression analyses.
Fig. 1. Graphical representation of the interaction effect between attachment and fussing on cortisol reactivity (log transformed).
Our results resemble earlier findings that insecure infants tend to cry more during the mother’s absence as compared to secure infants [4,9]. The marginally significant differences seen in crying are probably driven by the insecure-resistant infants that show high levels of negative vocalizations when separated from the mother [4,9]. These two studies did not observe infants’ behaviors directed to the door. However, our finding that secure infants engaged more in visual orientation to and manipulation of the door during the mother’s absence reflects Ainworth’s description of ‘search behaviors’ [1]. When distressed and alone, secure infants often use active attachment behaviors to establish proximity with the attachment
figure in order to elicit external regulation. Part of these active attachment behaviors are search behaviors including orienting to the door, following the mother to the door, and banging on the door [1]. While insecure-avoidant infants refrain from establishing proximity with the attachment figure, resistant and disorganized infants might be too distressed to mobilize these active search behaviors. The second aim of this paper was to examine whether infant regulatory behaviors and cortisol responses are differentially associated in secure versus insecure infants. When examining these associations, three alternative models existed to explain our findings: (1) the arousal model implies that stressful events cause a
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
7
Fig. 2. Graphical representation of the interaction effect between attachment and self-soothing on cortisol reactivity (log transformed).
general stress response resulting in a positive association between physiological and behavioral stress responses [5]; thus a positive association would be expected in both securely and insecurely attached infants, (2) the coping model implies that stressful situations elicit physiological stress responses only when appropriate behavioral regulatory strategies are not available or effective [4]; hence a positive association is expected only in insecurely attached infants, (3) the model of synchrony implies that the behavioral and physiological stress systems should be coordinated [7], thus a positive or negative association is expected only in securely attached infants. Our results do not seem to fit one-to-one with any one of these models. Rather, support for a specific model depended on the specific regulatory behavior under investigation. While higher cortisol reactivity was related to more infant crying during mother’s absence in both secure and insecure infants, fussing showed a differential association with cortisol reactivity in secure versus insecure infants. Insecure (but not secure) infants fussed less when their cortisol reactivity was higher. Previous studies combined fussing and crying in one measure of behavioral distress [4,7,8]. However, our results suggest that fussing is distinct from crying. Not only was there no correlation between infant crying and fussing in the SSP (r = −0.03, p > 0.05), the effects found for crying and fussing reveal more differences. Apparently, crying can be explained with the arousal model [5] and be seen as a signal of a general stress response in both secure and insecure infants, with higher cortisol reactivity being related to more crying. With respect to fussing, it is less clear which of the models apply. Insecure infants, in particular resistantly attached infants, are known to maximize their distress [16,17]. The lower cortisol concentrations might indicate that the insecure infants were not that distressed. However, through a history of insensitive and inconsistent care during the first year of life, these insecure infants might have learned to act distressed in order to elicit caregiving. Previous studies have found that fussing occurs more frequently in insecure infants during the first year of life [18] and that more hours of fussing, but not more hours of crying, was predictive of more inattention, emotional reactivity, and externalizing behavior in 3–8-year old children [19]. It might be that insecure infants more effectively adopted fussing as a functional strategy in order to elicit caregiving during the first year of life. Consequently, this behavior of fussing might underlie the link between insecure attachment and later socio-emotional and behavioral problems [20,21], but this hypothesis clearly needs further testing. More self-soothing, including thumb sucking and plucking at clothes, was related to higher cortisol reactivity, but only for secure infants. As secure infants are supposed to be better capable of reg-
ulating stress, these findings seem to correspond to the model of synchrony [7], which suggests that a correlation between behavior and physiology reflects a well-functioning stress system. But how could physiological stress responses that are synchronized with regulatory behaviors be beneficial? First, synchrony might enable a better coping with the stressor. One can speculate that secure infants that are physiologically stressed will experience a reduction in unpleasantness and negativity when certain specific behaviors are applied, such as self-soothing [22]. A better coordination among the stress systems would thus benefit the emotional state of the infant. This hypothesis is difficult to test in infants that cannot indicate their own experienced emotions, but it can be tested in older children and adults. A second benefit of synchrony could be to elicit external regulation from the attachment figure, often the mother. Physiological distress can prompt the use of regulatory behaviors, including self-soothing. When the physiological distress becomes too high, the infants’ own ability to engage in regulation might be undermined and the infants need to signal their distress to their mothers, for example by crying. Additionally, through the mechanism of synchrony between mothers and their infants, mothers react physiologically in accordance with their children [23]. Infant behavioral signals of stress, accompanied with increased maternal physiological stress levels, might activate mothers to optimally comfort and care for their children. Almost 30% of the variance in cortisol reactivity to the SSP was explained by infant regulatory behaviors when the mother was absent, while the regulatory behaviors during the mother’s presence were not predictive of cortisol reactivity. This result confirms the survival function of the attachment system. Infants are helpless and completely dependent on adults for care and protection during the first years of life [24]. Bowlby [25] therefore proposed that infants are endowed with this strong motivation to remain close to their attachment figures as their primary source of safety. When separated from the attachment figure, infants feel threatened and probably the HPA-axis is activated as a consequence. Moreover, as infants have no one else to depend on for regulation but themselves, infant regulatory behaviors during separations from the attachment figure are mostly correlated to their cortisol reactivity. Infants who are highly physiologically distressed are probably prompted to regulate more. These regulatory behaviors might also modulate the physiological stress response by reducing the unpleasantness and negativity of the situation or by signaling their stress to elicit external regulation from the attachment figure. In contrast, the regulatory behaviors in the mother’s presence were not predictive of infant cortisol reactivity. From an evolutionary point of view, these episodes are supposed to be less threatening to the infant as
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
8
their attachment figure is present. Moreover, the infant is no longer only dependent on his own regulatory capacities as the mother is present to provide co-regulation. This together might explain why infant regulatory behaviors in the mother’s presence were not predictive of infant cortisol reactivity to the SSP. Cortisol recovery from the SSP was not predicted by infant regulatory behaviors in mother’s absence nor in mother’s presence. An explanation could be that infant regulatory behaviors during stress do not aid cortisol recovery, at least not in this short time period and at this age. It could also be that the maternal co-regulatory behaviors during infant stress are more important to facilitate infant recovery from a stressor. Maternal co-regulation is part of maternal caregiving behavior. We have shown in our lab that higher quality of maternal caregiving behavior (sensitivity and cooperation) was related to a quicker recovery from a mildly stressful bathing session in 3-months-olds [26], and that more weeks of breastfeeding in the first six months of life were related to quicker recovery from the SSP in the same group of infants as currently used in this study [11]. Whether maternal co-regulation, and the specific maternal coregulatory behaviors, aid infant cortisol recovery is topic for future research. This study is among the first to investigate differential associations between a range of regulatory behaviors and cortisol reactions in secure versus insecure infants. Strong points are the expansion of regulatory behaviors, the focus on the separation episodes, and the inclusion of both cortisol reactivity and recovery. Some limitations also need to be addressed. First, the sample of participating mothers and infants was quite homogeneous: the majority was Dutch, highly educated and married or living together with their partner. This constrains the generalizability of the study. Second, the sample size was limited. Studies with larger samples could investigate the associations between behavioral responses and cortisol reactivity to the SSP separately in the three insecure attachment groups. This would offer a deeper understanding of infant behaviors under stress, as for example avoidantly attached infants, known to minimize attachment behaviors towards their mothers [27], may also display other behaviors when left alone when compared to the other groups of insecure infants. Lastly, the observational study design does not allow any conclusions on causality. It is unclear whether the regulatory behaviors proceed or are a result of heightened cortisol concentrations. In summary, the present study showed differential associations between behavioral and cortisol responses to a stressor in secure versus insecure infants. This might indicate that secure and insecure infants apply different behavioral regulatory strategies when physiologically stressed. Before definite conclusions can be drawn, the study findings need to be replicated and enriched by future research. Research differentiating between specific types of insecurity is needed as different profiles of physiology and behavior likely exist for avoidant, resistant, and disorganized infants. In addition, research including other physiological stress measurements that pertain to the parasympathetic and sympathetic nervous system, such as heart rate variability, respiratory sinus arrhythmia and salivary alpha-amylase, could complement the present study findings. Also, follow-up studies are needed to examine the longitudinal manifestation of these stress response associations and the potential consequences for later child development. Finally, the nature of these relations in response to other stress paradigms than the SSP could be examined, such as during the transition to childcare where separation from the mother is prolonged [28].
Conflict of interest There were no conflicts of interest, financial or otherwise to declare.
Role of the funding source This research was supported by a personal grant of the Netherlands Organization for Scientific Research (NWO) awarded to Carolina de Weerth (grant number 452-04-320). Acknowledgements We thank the families who kindly participated in the BIBO study, and all research assistants and students for their assistance with data collection. References [1] M.D.S. Ainsworth, M.D. Blehar, E. Waters, S. Wall, Patterns of Attachment. A Psychological Study of the Strange Situation, Erlbaum, Hillsdale, NJ, 1978. [2] S.D. Calkins, E.M, Leerkes, Early attachment processes and the development of emotional self-regulation, in: K.D. Vohs, R.F. Baumeister (Eds.), Handbook of Self-regulation: Research, Theory, and Applications, 2nd edn., The Guilford Press, New York, NY, 2011, pp. 355–373. [3] A.N. Schore, Effects of a secure attachment relationship on right brain development, affect regulation, and infant mental health, Inf. Mental Health J. 22 (2001) 7–66, http://dx.doi.org/10.1002/1097-0355(200101/ 04)22:1<7:AID-IMHJ2>3.0.CO;2-N. [4] G. Spangler, M. Schieche, Emotional and adrenocortical responses of infants to the strange situation: the differential function of emotional expression, Int. J. Behav. Dev. 22 (1998) 681–706, http://dx.doi.org/10.1080/016502598384126. [5] H. Selye, The Story of the Adaptation Syndrome, Acta, Montreal, 1952. [6] R.M. Sapolsky, L.M. Romero, A.U. Munck, How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions, Endocrinol. Rev. 21 (2000) 55–89. [7] S.J. Erickson, P. MacLean, C. Qualls, J.R. Lowe, Differential associations between infant affective and cortisol responses during the still face paradigm among infants born very low birth weight versus full-term, Inf. Behav. Dev. 36 (2013) 359–368, http://dx.doi.org/10.1016/j.infbeh.2013.03.001. [8] M.R. Gunnar, L. Brodersen, M. Nachmias, K. Buss, J. Rigatuso, Stress reactivity and attachment security, Dev. Psychobiol. 29 (1996) 191–204, http://dx.doi. org/10.1002/(SICI)1098-2302(199604)29:3<191:AID-DEV1>3.0.CO;2-M. [9] J.M. Braungart, C.A. Stifter, Regulation of negative reactivity during the strange situation: temperament and attachment in 12-month-old infants, Infant Behav. Dev. 14 (1991) 349–364, http://dx.doi.org/10.1016/01636383(91)90027-P. [10] C. De Weerth, J. Jansen, M.H. Vos, I. Maitimu, E.G. Lentjes, A new device for collecting saliva for cortisol determination, Psychoneuroendocrinology 32 (2007) 1144–1148, http://dx.doi.org/10.1016/j.psyneuen.2007.07.005. [11] R. Beijers, J.M.A. Riksen-Walraven, C. De Weerth, Cortisol regulation in 12-month-old human infants: associations with the infants’ early history of breastfeeding and co-sleeping, Stress 16 (2013) 267–277, http://dx.doi.org/ 10.3109/10253890.2012.742057. [12] M. Main, J. Solomon, Discovery of a disorganized/disoriented attachment pattern, in: T.B. Brazelton, M.W. Yogman (Eds.), Affective Dev. Inf., Ablex, Norwood, NJ, 1986, pp. 95–124. [13] R. Beijers, J. Jansen, J.M.A. Riksen-Walraven, C. De Weerth, Attachment and infant night waking: a longitudinal study from birth through the first year of life, J. Dev. Behav. Pediatr. 32 (2011) 635–643, http://dx.doi.org/10.1097/DBP. 0b013e318228888d. [14] L.S. Aiken, S.G. West, Multiple Regression: Testing and Interpreting Interactions, Sage, Thousand Oaks, CA, 1991. [15] A.F. Hayes, J. Matthes, Computational procedures for probing interactions in linear and logistic regression SPSS and SAS implementations, Beh. Res. Methods 41 (2009) 924–936, http://dx.doi.org/10.3758/BRM.41.3.924. [16] E.A. Carlson, M.C. Sampson, L.A. Sroufe, Implications of attachment theory and research for developmental-behavioral pediatrics, J. Dev. Beh. Pediatr. 24 (2003) 364–379. [17] J. Cassidy, J.D. Jones, P.R. Shaver, Contributions of attachment theory and research: a framework for future research, translation, and policy, Dev. Psychopath 25 (2013) 1415–1434. [18] J. Belsky, M. Rovine, D.G. Taylor, The Pennsylvania infant and family development project III: The origins of individual differences in infant-mother attachment: maternal and infant contributions, Child Dev. 55 (1984) 718–728, http://dx.doi.org/10.1111/j.1467-8624.1984.tb03810.x. [19] A. Desantis, W. Coster, R. Bigsby, B. Lester, Colic and fussing in infancy, and sensory processing at 3–8 years of age, Inf. Mental Health J. 25 (2004) 522–539, http://dx.doi.org/10.1002/1mhj.20025. [20] R.P. Fearon, M.J. Bakermans-Kranenburg, M.H. van IJzendoorn, A.M. Lapsley, G.I. Roisman, The significance of insecure attachment and disorganization in the development of children’s externalizing behavior: a meta-analytic study, Child Dev. 81 (2010) 435–456. [21] A.M. Groh, G.I. Roisman, M.H. van IJzendoorn, M.J. Bakermans-Kranenburg, R.P. Fearon, The significance of insecure and disorganized attachment for children’s internalizing symptoms: a meta-analytic study, Child Dev. 83 (2012) 591–610, http://dx.doi.org/10.1111/j.1467-8624.2011.01711.x.
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
[22] A.J. Oldehinkel, J. Ormel, N.M. Bosch, E.M.C. Bouma, A.M. Van Roon, J.G.M. Rosmalen, H. Riese, Stressed out? Associations between perceived and physiological stress responses in adolescents: the TRAILS study, Psychophysiol 48 (2011) 441–452, http://dx.doi.org/10.1111/j.1469-8986. 2010.01118.x. [23] R. Feldman, Bio-behavioral synchrony: a model for integrating biological and microsocial behavioral processes in the study of parenting, Parenting: Sci. Pract. 12 (2012) 154–164, http://dx.doi.org/10.1080/15295192.2012.683342. [24] J.S. Chisholm, The evolutionary ecology of attachment organization, Human Nat. 7 (1996) 1–37. [25] J. Bowlby, Attachment and loss Attachment, vol. 1, Basis Books, New York, 1969.
9
[26] E.M. Albers, J.M.A. Riksen-Walraven, F.C.G.J. Sweep, C. de Weerth, Maternal behavior predicts infant cortisol recovery from a mild everyday stressor, J. Child Psychol. Psychiatry 49 (2008) 97–103, http://dx.doi.org/10.1111/j.14697610.2007.01818.x. [27] J. Cassidy, Emotion regulation: influences of attachment relationships, Monogr. Soc. Res. Child Dev. 59 (1994) 228–249, http://dx.doi.org/10.1111/j. 1540-5834.1994.tb01287.x. [28] L. Ahnert, M.R. Gunnar, M.E. Lamb, M. Barthel, Transition to child care: associations with infant-mother attachment, infant negative emotion, and cortisol elevations, Child Dev. 75 (2004) 639–650, http://dx.doi.org/10.1111/j. 1467-8624.2004.00698.x.
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
ARTICLE IN PRESS
BBR-10496; No. of Pages 9
Behavioural Brain Research xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Behavioural Brain Research journal homepage: www.elsevier.com/locate/bbr
Research report
Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants Roseriet Beijers ∗ , Marianne Riksen-Walraven, Katharina Sebesta, Carolina de Weerth Department of Developmental Psychology, Behavioral Science Institute, Radboud University, The Netherlands
h i g h l i g h t s • • • • •
Do secure and insecure infants use different regulatory behaviors when stressed? Insecure infants fussed less when physiologically stressed. Secure infants showed more self-soothing when physiologically stressed. 29% explained variance in cortisol by regulatory behaviors in mother’s absence. Behavior and cortisol are differentially associated in secure and insecure infants.
a r t i c l e
i n f o
Article history: Received 28 July 2016 Received in revised form 15 September 2016 Accepted 4 October 2016 Available online xxx Keywords: Attachment Regulation Behavior Cortisol HPA-axis Stress
a b s t r a c t In this study we examined whether securely versus insecurely attached infants use different regulatory behaviors in absence of their mother and whether these regulatory behaviors are differentially associated with physiological stress responses in secure versus insecure infants. Participants were 193 one-year-olds and their mothers. During three 3-min episodes of separation from the mother in the Strange Situation Procedure (SSP) [1], the following infant regulatory behaviors were observed: crying, fussing, self-soothing, manipulation of toys, and manipulation of the door. Salivary cortisol was measured at baseline and 25, 40, and 60 min after the SSP to measure reactivity and recovery. Additionally, infants were classified as securely or insecurely attached to their mothers. During the mother’s absence, secure infants engaged more in manipulation of the door than insecure infants. Furthermore, in insecure (but not secure) infants less fussing was associated with higher cortisol reactivity, while in secure (but not insecure) infants more self-soothing was associated with higher cortisol reactivity. In total, 29% of the variance in cortisol reactivity was explained by infant regulatory behaviors in the mother’s absence. Cortisol recovery was not predicted by infant regulatory behaviors in the mother’s absence. To conclude, the results show differential associations between behavioral and cortisol responses to a stressor in secure versus insecure infants. This might indicate that secure and insecure infants apply different behavioral regulatory strategies when physiologically stressed. © 2016 Elsevier B.V. All rights reserved.
1. Introduction The notion that infant attachment quality is related to later socio-emotional development and mental health is one of attachment theory’s most well-known contributions to the field of developmental psychology. For example, securely attached children are considered better able to regulate themselves, i.e. more capable of applying behaviors that modulate their physiological
∗ Corresponding author at: Developmental Psychology, Behavioral Science Institute, Radboud University, P.O. Box 9140, 6500 HE Nijmegen, The Netherlands. E-mail address: [email protected] (R. Beijers).
stress [2,3]. However, the specific types of regulatory behaviors that secure versus insecure infants use to regulate their physiological stress are as yet unclear. The purpose of the present study was to first investigate whether securely versus insecurely attached infants use different regulatory behaviors in absence of their mother, and second whether these regulatory behaviors are differentially associated with physiological stress responses in secure versus insecure infants. By the end of the first year, infants can be classified as securely versus insecurely attached to a caregiver by observing them in the Strange Situation Procedure (SSP) [1]. In the SSP, children are exposed to a series of mild stressors (i.e., confrontation with a stranger and two brief separations from the caregiver), to observe
http://dx.doi.org/10.1016/j.bbr.2016.10.008 0166-4328/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9 2
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
whether they can use the caregiver as a ‘secure base’ in times of uncertainty or stress. When distressed, securely attached infants seek proximity and contact with the attachment figure. The attachment figure can then comfort them and the infants can return to exploration and play. By their attachment behaviors, secure infants apparently elicit external regulation of stress by the attachment figure. Such ‘co-regulation’ is assumed to protect the infant from excessive levels of distress and to promote the infant’s own emergent capacity to regulate arousal [3]. Regulatory capacities gradually evolve into more elaborated and self-initiated processes during the first years of life as a result of increasing control capacities, and are important to cope with developmental challenges. Through a history of co-regulation, secure infants are therefore suggested to develop better self-regulatory capacities, when compared to insecure infants [2]. As yet, it is however unclear what types of specific regulatory behaviors secure and insecure infants use when the attachment figure is not present and whether these regulatory behaviors are differentially associated with physiological stress responses in secure versus insecure infants. Spangler and Schieche took a first step in examining this question [4]. They suggested two models to explain how behavior can be associated with physiological reactions in secure and insecure attached infants. An arousal model implies that stressful events cause a general stress response, resulting both in physiological and behavioral stress responses [5]. Contrarily, in a coping model, stressful situations elicit physiological stress responses only when appropriate behavioral coping strategies are not available or effective. In their research, Spangler and Schieche studied the association between negative emotional expressions (sum of crying, fussing, and sad faces) and cortisol reactivity to the SSP in 12-month-old infants [4]. Cortisol is the end product of the hypothalamic-pituitary-adrenocortical (HPA)-axis, which is activated when confronted with uncertain, unpredictable and stressful situations Ref. [6]. Spangler and Schieche found a positive association between negative emotional expressions and cortisol reactions in insecure infants, but not in secure infants, who showed no association between negative emotional expression and cortisol reactivity [4]. The authors explained this lack of association in secure infants using a coping model, whereas they interpreted the positive behavior-physiology association of insecure infants using an arousal model. According to this reasoning, insecure infants would lack appropriate behavioral coping strategies in the face of maternal separation distress. Negative emotional expressions in this group would be a sign of general stress responses. In contrast, secure infants would show negative emotional expression as a coping strategy, for example to reestablish proximity with the attachment figure, with the consequence that a physiological response does not occur. However, in situations where coping strategies result ineffective, secure infants would also show negative emotional expressions as a sign of a general stress response. Spangler and Schieche did not examine other types of regulatory behaviors [4]. Also, cortisol recovery was not included, while one may speculate that regulatory behaviors are not only important for regulating the stress response, but also for recovering from stress. Moreover, the groundwork laid by Spangler and Schieche was unfortunately not elaborated by other attachment researchers [4]. More recently, however, the association between affective states and cortisol reactivity in response to another type of stress-inducing procedure, namely the still-face procedure, was studied in very low birth weight (VLBW) and full-term infants [7]. They found a significant negative association between positive as well as negative affect and cortisol reactivity for full-term infants. Thus, the expression of both more positive and more negative affect was related to lower cortisol levels in these infants. By contrast, the associations in the VLBW group were limited. The authors
explained their findings by suggesting that less synchrony (i.e., less correlation, either positive or negative) between a behavioral and a physiological response would correspond to a lack of coordination among stress systems, and that the lack of synchrony would represent greater dysregulation [7]. This explanation may also be applicable to a study that showed a greater dissociation between behavioral distress and cortisol reactions after an inoculation in 2-month-olds who were later classified as insecurely attached, as compared to securely attached [8]. The first goal of the present study was to investigate whether secure and insecure infants engage in different regulatory behaviors when separated from the mother. During the episodes of maternal separation within the Strange Situation Procedure (SSP) [1], we observed behaviors that have been suggested to have a regulatory function. Next to crying and fussing, we extended previous research by also including self-soothing, manipulation of toys and manipulation of the door as regulatory behaviors [9]. The second goal of this study was to investigate whether the associations between cortisol responses (reactivity and recovery) and infant regulatory behaviors during the mother’s absence are moderated by infant attachment status (secure or insecure). When examining the differential associations, no hypotheses were made as three alternative models existed: (1) the arousal model expects a positive association between physiological and behavioral stress responses in both securely and insecurely attached infants [5], (2) the coping model expects a positive association between physiological and behavioral stress responses only in insecurely attached infants [4], (3) the model of synchrony expects a positive or negative association between physiological and behavioral stress responses only in securely attached infants [7]. Although securely attached children are considered better able to regulate themselves [2,3], the specific types of regulatory behaviors that secure versus insecure infants use to regulate their physiological distress are as yet unclear. The results will provide more knowledge on the normative development of children’s regulatory strategies within the secure attachment relationship and the nature of regulation difficulties identified in insecure infants. 2. Methods 2.1. Participants This study is part of a prospective longitudinal project starting in pregnancy and examining early caregiving factors and their influences on children’s development and health across the first years of life. The subjects were healthy infants, whose mothers were recruited during pregnancy through flyers dispersed in midwife practices in and around the Dutch cities of Nijmegen and Arnhem. Inclusion criteria were an uncomplicated and singleton pregnancy without drug use, a clear understanding of the Dutch language, and no current physical and mental health problems. The study was approved by the Faculty Ethical committee and all mothers gave written informed consent before starting. A total of 220 mothers enrolled voluntarily in the study. The final sample was limited to 193 women and their infants, as 8 women were excluded due to medical reasons and 19 discontinued the study during the first three postpartum months because of personal circumstances. No differences in demographics were found between mothers who took part in the study and the 19 women who dropped out. 2.2. Procedure At 12 months of age (M = 53 weeks and 6 days, SD = 19 days), mother-infant pairs participated in the Strange Situation Procedure (SSP) [1] in the laboratory of the Radboud University, Nijmegen.
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
This 21-min procedure consists of eight sequenced episodes. The episodes contain mildly stressful situations of different kinds, which are assumed to be increasingly disturbing. Respectively, they contain exposure to an unfamiliar environment, exposure to a stranger, and two brief separations from the mother (each followed by a reunion). The unfamiliar environment corresponded to a large child-proof room with a one-way screen, two cameras, toys for the infant to play with, and two chairs for the mother and the stranger. The stranger was always a trained female psychology student. During the first separation, the infant was left alone with the stranger for one episode. The second separation entailed the infant being left alone for one episode and then being together with the stranger for a subsequent episode. The two cameras were operated by an assistant in the control room. The assistant switched from one camera to the other to obtain the best picture of the child. The SSP was recorded on videotape for later classification of infantmother attachment and coding infant regulatory behaviors. Saliva samples were taken before the SSP (T1) and at 25 (T2), 40 (T3), and 60 min (T4) after the SSP to measure infants’ cortisol reactivity and recovery.
Table 1 Definitions of the observed categories of infant regulatory behavior in the absence of the mother. Code
Definition
A Negative emotional expression
All negative vocal emotional expressions of the infant. Infant negative vocal expression with high intensity and for longer than 3 of the 5 s. Infant negative vocal expression with low intensity or cries not longer than 3 of the 5 s.
A1 Crying
A2 Fussing
B Visual orientation B1 to toys B2 to doora C Manipulation
C1 of toys C2 of door
2.3. Measures
D Self-soothing
2.3.1. Infant cortisol responses Infant saliva samples were collected with Sorbette eye sponges [10]. The samples were taken pre-stressor (T1) and 25, 40 and 60 min post-stressor (T2–T4). After collection, the samples were stored in a −25 ◦ C freezer. Cortisol analysis was conducted at the Laboratory of Endocrinology of the University Medical Center of Utrecht University, making use of radioimmunoassay. The lower limit of detection was 1.0 nmol/l, and inter-assay and intra-assay variation were below 10% [10]. Acceptable time ranges for the poststressor samples were 20–30 min for T2, 35–50 min for T3, and 55–75 min for T4. Valid cortisol samples were available for 176 infants at T1, 171 infants at T2, 173 infants at T3, and 148 infants at T4. Cortisol reactivity to the SSP was represented by the highest post-stressor concentration (T2 or T3) and corrected for prestressor concentration (T1) to take baseline values into account [11]. An independent-samples t-test indicated that the magnitude of the infants’ peak at T2 (M = 9.23, SD = 4.45) was not different from the magnitude at T3 (M = 8.43, SD = 4.09, t(174) = 1.24, p = 0.219), and half of the infants (51%) showed peak values at T2, and the other half at T3. In line with Beijers et al. [11], cortisol recovery was measured by obtaining the difference between the last post-stressor sample and the pre-stressor sample (T4–T1), and correcting for the peak (T2 or T3) in the analyses on recovery. This measure reflects infants’ recovery from the cortisol reaction, with higher scores reflecting slower recovery. Peaking at T2 versus T3 was not related to recovery (Pearson’s r = 0.12, p > 0.10). 2.3.2. Infant regulatory behaviors The video recordings of the SSP were observed by two independent assistants who were trained and reliable, and blind to the infant cortisol responses and other data. The inter-rater reliability was good (Spearman’s rho’s ranging between 95.2% and 97.5%). The occurrence of seven specific behaviors reflecting the infant’s negative emotional expression, visual orientation, and manipulation behavior (see Table 1) was coded in 5-s intervals. With the use of a software program (The Observer, Noldus Information Technology, Wageningen, The Netherlands), the coders rated every 5 s whether the infant had shown the specific behavior (1) or not (0). For each behavioral category the number of infant behaviors was summed up per episode, divided by the total number of intervals for that episode and multiplied by 100, resulting in a percentage between 0 and 100. To assess infant regulatory capacities in
3
a
Every looking behavior that is directed at something. Infant looks at the toys or in their direction. Infant looks at the door or in its direction. Visual orientation and action (touching) that is directed at something. Infant directs attention and action to the toys. Infant directs attention and action to the door. Infant directs attention and action to self or clothing. Examples are thumb sucking and plucking at clothes.
The door is the room’s only door and thus the door through which the mother
left.
absence of the mother, the regulatory behaviors shown during the three episodes in which the infant was separated from the mother were summed up (episodes 3, 5, and 6). To be able to control for infant regulatory capacities in presence of the mother, the behaviors during the episodes in which the mother was present were summed up (episodes 1, 2, 4 and 7). Visual orientation to the toys and manipulation of the toys were positively correlated (r = 0.84, p < 0.001) as well as visual orientation to the door and manipulation of the door (r = 0.42, p < 0.001), and these positively correlated and related behaviors were therefore averaged. This resulted in the investigation of five behavioral responses: crying, fussing, selfsoothing, visual orientation to and manipulation of the toys, and, visual orientation to and manipulation of the door.
2.3.3. Infant attachment The video recordings of all SSP’s were coded by trained and supervised examiners at the Institute of Child Development, University of Minnesota. Every infant was assigned a secure (B) or insecure (avoidant (A), resistant (C), or disorganized (D)) attachment pattern [1,12]. Inter-observer reliability, based on a subset of 46 infants, was good (Cohen’s K of 0.82, intraclass correlation of 0.86; see also Beijers et al. [13]).
2.3.4. Confounders The time of day when the SSP took place and the time since the last feeding and sleep before the start of the SSP were included as confounders. As this study specifically focused on the regulatory behaviors infants use when physiologically distressed and the mother is absent, we also controlled for these regulatory behaviors (crying, fussing, self-soothing, visual orientation to and manipulation of the toys, and, visual orientation to and manipulation of the door) when the mother was present during the SSP (episodes 1, 2, 4 and 7).
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
4
Table 2 Descriptive statistics of the study variables, separately for securely and insecurely attached infants. Secure attachment N = 127
Insecure attachment N = 56
Mean
SD (range)
Mean
SD (range)
Demographics Maternal age
32.17
3.82 (21.10–42.90)
33.15
3.57 (25.10–40.30)
Maternal educational level (%) Primary education Secondary education College or university
2.4 20.8 76.8
1.9 22.7 75.4
Maternal marital status (%)a Living with partner
98.4
96.4
Infant sex (%) Girl
44.9
48.2
Birth order (%) Firstborn
46.4
30.9
Infant cortisol regulation Pre-stressor (T1) Reactivity (highest value T2/T3, nmol/l)* Recovery (T4-T1, nmol/l)b
7.19 8.38 −0.20
3.22 (3.30–22.00) 4.10 (3.10–24.00) 4.07 (−15.10–10.50)
7.90 9.90 −0.35
3.64 (3.70–21.00) 4.53 (4.40–22.00) 4.51 (−14.90–9.20)
Regulatory behaviors in mother’s absence Crying† Fussing Self-soothing Orientation to and manipulation of toys Orientation to and manipulation of door*
28.22 11.07 5.35 65.92 18.36
30.10 (0.00–100.00) 12.58 (0.00–58.33) 10.55 (0.00–52.22) 23.08 (11.67–100.00) 13.48 (0.00–71.67)
37.26 9.98 7.51 61.26 14.12
34.62 (0.00–100.00) 12.05 (0.00–50.51) 15.07 (0.00–66.67) 27.40 (0.00–100.00) 12.24 (0.00–50.00)
Regulatory behaviors in mother’s presence Crying** Fussing** Self-soothing Orientation to and manipulation of toys** Orientation to and manipulation of door
3.19 3.27 1.32 87.14 3.24
5.10 (0.00–30.88) 4.09 (0.00–23.96) 3.34 (0.00–20.96) 10.96 (51.09–100.00) 2.69 (0.00–10.42)
6.77 6.01 2.47 81.04 2.63
8.82 (0.00–41.67) 5.65 (0.00–20.49) 5.18 (0.00–24.07) 13.09 (51.77–100.00) 2.70 (0.00–10.13)
Remaining confounders Time of the day SSP (h) Hours since last feeding before SSP (h) Hours since last sleeping before SSP (h)
11:18 1:14 1:34
2:01 (8:52–17.31) 43 min (0:00–3:07) 1:01 (0:00–3:36)
11:15 1:23 1:37
1:57 (8:52–15:07) 45 min (0:00–3:21) 1:10 (0:00–4:48)
†
a b † * **
In wedlock or unmarried. Lower recovery scores indicate quicker recovery. p < 0.10. p < 0.05. p < 0.01.
2.4. Statistical analyses Of the 193 infants who participated in the project, 186 took part in the SSP [1]. The demographic characteristics of this group are presented in Table 2. Three infants were not classified as securely or insecurely attached because of doubt. In the event of skewed data, square root (self-soothing) or a logarithmic transformation (cortisol peak reactivity) was applied when appropriate, and the transformed variables were used for the analyses. No outliers were detected in the data. Subsequently, a dummy variable was created for attachment security, with secure infants serving as the reference group. Independent-samples t-tests were used to test whether secure and insecure infants engaged in different regulatory behaviors when exposed to a stressor when their mother was not present. To test whether specific behavioral responses are differently related to cortisol reactivity and cortisol recovery among secure and insecure infants, two standard multiple hierarchical regression analyses were conducted for cortisol reactivity and cortisol recovery. Infant behavioral variables were centered for their use in the creation of the interaction variables, which were calculated by multiplying the attachment security dummy by the centered behavioral variables [14]. The first regression analysis contained all predictors, interaction variables and confounders. Those variables that explained at least 1% of the variance, calculated as (part
correlation)2 × 100, were further used for the second regression analysis, in which all maintained confounders were entered in hierarchical step 1 and all maintained predictors were entered in hierarchical step 2. This approach was chosen to eliminate irrelevant confounders as well as predictors and to thereby increase power [11,13]. 3. Results 3.1. Preliminary analyses See Table 2 for the descriptive statistics of the study variables. Of the 183 infants, 127 (69%) were classified as securely attached and 56 (31%) as insecurely attached (9 insecure-avoidant, 21 insecureresistant, and 26 insecure-disorganized). Independent-samples t-tests revealed that insecure infants showed higher cortisol reactivity to the SSP (t(175) = −2.53, p = 0.01). During the mother’s presence, insecure infants cried and fussed more (t(79.47) = −3.00, p = 0.004 and t(81.43) = −3.27, p = 0.002, respectively), and also engaged less in orientation to and manipulation of the door (t(90.49) = 3.05, p = 0.003). Table 3 shows the Pearson correlations for cortisol reactivity, cortisol recovery, and the infant regulatory behaviors, presented separately for the secure and insecure infants. In secure infants, more crying and self-soothing was associated with higher corti-
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
5
Table 3 Pearson correlations between infant regulatory behaviors in mother’s absence and cortisol variables. 1.
2.
3.
4.
5.
6.
7.
Total sample 1. Reactivity 2. Recovery 3. Crying 4. Fussing 5. Self-soothing 6. Orientation to and manipulation of toys 7. Orientation to and manipulation of door
– 0.39** 0.45** 0.01 0.19* −0.39** 0.13†
– 0.18* −0.15† 0.20* −0.20* 0.05
– −0.03 0.20** −0.78** 0.39**
– 0.06 −0.12† 0.29**
– −0.18* 0.00
– −0.49**
–
Securely attached infants 1. Reactivity 2. Recovery 3. Crying 4. Fussing 5. Self-soothing 6. Orientation to and manipulation of toys 7. Orientation to and manipulation of door
– 0.41** 0.48** 0.08 0.36** −0.40** 0.17†
– 0.28** −0.06 0.26** −0.24* 0.12
– −0.07 0.15† −0.81** 0.40**
– −0.02 −0.18* 0.31**
– 0.17† −0.01
– −0.54**
–
Insecurely attached infants 1. Reactivity 2. Recovery 3. Crying 4. Fussing 5. Self-soothing 6. Orientation to and manipulation of toys 7. Orientation to and manipulation of door
0.38* 0.34* −0.14 −0.16 −0.36** 0.10
−0.05 −0.33* 0.08 −0.14 −0.13
0.06 0.27* −0.72** 0.44**
0.22 −0.01 0.18
−0.19 0.02
−0.43**
† * **
p < 0.10. p < 0.05. p < 0.01.
sol reactivity, whereas more orientation to and manipulation of toys was associated with less cortisol reactivity. In insecure infants, more crying was related to higher cortisol reactivity, whereas more orientation to and manipulation of the toys was related to less cortisol reactivity. With regard to cortisol recovery, in secure infants more crying and self-soothing were related to slower cortisol recovery, whereas more orientation to and manipulation of toys was related to quicker cortisol recovery. For insecure infants, in contrast, more fussing was related to quicker cortisol recovery.
3.2. Main analyses Independent-samples t-tests revealed that insecure infants engaged less in orientation to and manipulation of the door during the mother’s absence (t(181) = 2.02, p = 0.045), when compared to secure infants (see Table 2). A marginally significant difference was also found, indicating that insecure infants cried more during the mother’s absence (t(181) = −1.79, p = 0.079), when compared to secure infants. The regression models, which only contain the variables that individually explained at least 1% of the variance in the first analyses, are summarized in Table 4. This table shows that none of the regulatory behaviors in the mother’s presence were predictive of infant cortisol reactivity, while the regulatory behaviors in the mother’s absence explained 29% of the variance in cortisol reactivity to the SSP. Furthermore, the regression results showed that higher reactivity was significantly predicted by more crying and more self-soothing during the mother’s absence. Moreover, higher pre-stressor cortisol values significantly predicted infant cortisol reactivity to the SSP, and insecure infants tended to show higher cortisol reactivity than secure infants. Furthermore, significant interaction effects were obtained for attachment × fussing and attachment × self-soothing. Figs. 1 and 2 depicts these interaction effects. The simple slopes of the regression lines reflecting the relation between regulatory behaviors and cortisol reactivity for infants with an insecure or secure attachment
were tested for significance, using the method of Hayes and Matthes [15]. For the interaction attachment × fussing, only the slope for the insecure infants was significant (b = − 0.00, t = −2.21, p < 0.05), but not the slope for the secure infants (b = 0.00, t = −1.13, p > 0.05). This indicates that infants with an insecure attachment fuss less when their cortisol reactivity concentrations are higher. For the interaction attachment × self-soothing, only the slope for the secure infants was significant (b = 0.02, t = 3.76, p < 0.001), but not the slope for the insecure infants (b = −0.01, t = −1.937, p > 0.05). This indicates that infants with a secure attachment manipulated themselves more (e.g. thumb-sucking, plucking at their clothes) when their cortisol reactivity concentrations were higher. For cortisol recovery, only the cortisol reactivity peak concentrations were able to significantly predict recovery, indicating that cortisol recovery concentrations were higher when the cortisol peak was higher.
4. Discussion We examined whether securely versus insecurely attached infants use different regulatory behaviors in absence of their mother and whether these regulatory behaviors are differentially associated with physiological stress responses in secure versus insecure infants. When compared to secure infants, insecure infants engaged less in manipulation of the door, and tended to cry more when the mother is absent. When controlling for infant regulatory behaviors in the mother’s presence, cortisol reactivity was predicted by the interaction effects attachment × fussing and attachment × self-soothing. These interaction effects indicated that insecure (but not secure) infants fussed less when their cortisol reactivity was higher, while secure (but not insecure) infants showed more self-soothing when their cortisol reactivity was higher. In total, 29% of the variance in cortisol reactivity was explained by infant regulatory behaviors in the mother’s absence. Cortisol recovery was not predicted by these regulatory behaviors.
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9 6
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
Table 4 Final multiple hierarchical regression models for the prediction of infant cortisol responses to the SSP. B

p-value
Reactivity Step 1 0.013 0.234 0.000 Pre-stressor cortisol concentrations 0.048 0.117 0.067 Attachment (1 = secure, 2 = insecure) Step 2 Crying in mother’s absence 0.001 0.516 0.000 Fussing in mother’s absence 0.001 0.103 0.167 Self-soothing in mother’s absence 0.015 0.293 0.000 Step 3 Attachment × crying in mother’s absence 0.000 −0.136 0.102 Attachment × fussing in mother’s absence −0.002 −0.168 0.028 Attachment × self-soothing in mother’s absence −0.021 −0.250 0.004 Excluded variables: The time of day when SSP took place – hours since last feeding before SSP – hours since last sleeping before SSP – crying, fussing, self-soothing, orientation to and manipulation of the toys, and, orientation to and manipulation of the door in mother’s presence – orientation to and manipulation of the toys, and, orientation to and manipulation of the door in mother’s absence and their interactions with attachment. Recovery Step 1 8.132 0.372 0.000 Peak cortisol reactivity −0.710 −0.079 0.392 Attachment (1 = secure, 2 = insecure) −8.064 −0.061 0.496 Hours since last feeding before SSP −0.020 −0.152 0.117 Orientation to and manipulation of door in mother’s presence Step 2 0.005 0.115 0.393 Crying in mother’s absence −0.012 −0.109 0.329 Fussing in mother’s absence 0.004 0.078 0.488 Orientation to and manipulation of door in mother’s absence Step 3 −0.013 −0.177 0.124 Attachment × crying in mother’s absence −0.024 −0.114 0.284 Attachment × fussing in mother’s absence Excluded variables: The time of day when SSP took place – hours since last sleeping before SSP – crying, fussing, self-soothing, and orientation to and manipulation of the toys in mother’s presence – self-soothing and orientation to and manipulation of the toys in mother’s absence and their interactions with attachment – interaction between attachment and orientation to and manipulation of door in mother’s absence.
R2
F change
p-value
0.077
7.112
0.001
0.204
15.810
0.000
0.086
7.401
0.000
0.163
5.213
0.001
0.029
1.254
0.294
0.029
1.902
0.155
change
Notes: B = regression coefficient,  = standardized regression coefficient, R2 change = explained variance by the added step. Excluded variables explained individually less than 1% variance in cortisol reactivity and recovery in preliminary regression analyses.
Fig. 1. Graphical representation of the interaction effect between attachment and fussing on cortisol reactivity (log transformed).
Our results resemble earlier findings that insecure infants tend to cry more during the mother’s absence as compared to secure infants [4,9]. The marginally significant differences seen in crying are probably driven by the insecure-resistant infants that show high levels of negative vocalizations when separated from the mother [4,9]. These two studies did not observe infants’ behaviors directed to the door. However, our finding that secure infants engaged more in visual orientation to and manipulation of the door during the mother’s absence reflects Ainworth’s description of ‘search behaviors’ [1]. When distressed and alone, secure infants often use active attachment behaviors to establish proximity with the attachment
figure in order to elicit external regulation. Part of these active attachment behaviors are search behaviors including orienting to the door, following the mother to the door, and banging on the door [1]. While insecure-avoidant infants refrain from establishing proximity with the attachment figure, resistant and disorganized infants might be too distressed to mobilize these active search behaviors. The second aim of this paper was to examine whether infant regulatory behaviors and cortisol responses are differentially associated in secure versus insecure infants. When examining these associations, three alternative models existed to explain our findings: (1) the arousal model implies that stressful events cause a
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
7
Fig. 2. Graphical representation of the interaction effect between attachment and self-soothing on cortisol reactivity (log transformed).
general stress response resulting in a positive association between physiological and behavioral stress responses [5]; thus a positive association would be expected in both securely and insecurely attached infants, (2) the coping model implies that stressful situations elicit physiological stress responses only when appropriate behavioral regulatory strategies are not available or effective [4]; hence a positive association is expected only in insecurely attached infants, (3) the model of synchrony implies that the behavioral and physiological stress systems should be coordinated [7], thus a positive or negative association is expected only in securely attached infants. Our results do not seem to fit one-to-one with any one of these models. Rather, support for a specific model depended on the specific regulatory behavior under investigation. While higher cortisol reactivity was related to more infant crying during mother’s absence in both secure and insecure infants, fussing showed a differential association with cortisol reactivity in secure versus insecure infants. Insecure (but not secure) infants fussed less when their cortisol reactivity was higher. Previous studies combined fussing and crying in one measure of behavioral distress [4,7,8]. However, our results suggest that fussing is distinct from crying. Not only was there no correlation between infant crying and fussing in the SSP (r = −0.03, p > 0.05), the effects found for crying and fussing reveal more differences. Apparently, crying can be explained with the arousal model [5] and be seen as a signal of a general stress response in both secure and insecure infants, with higher cortisol reactivity being related to more crying. With respect to fussing, it is less clear which of the models apply. Insecure infants, in particular resistantly attached infants, are known to maximize their distress [16,17]. The lower cortisol concentrations might indicate that the insecure infants were not that distressed. However, through a history of insensitive and inconsistent care during the first year of life, these insecure infants might have learned to act distressed in order to elicit caregiving. Previous studies have found that fussing occurs more frequently in insecure infants during the first year of life [18] and that more hours of fussing, but not more hours of crying, was predictive of more inattention, emotional reactivity, and externalizing behavior in 3–8-year old children [19]. It might be that insecure infants more effectively adopted fussing as a functional strategy in order to elicit caregiving during the first year of life. Consequently, this behavior of fussing might underlie the link between insecure attachment and later socio-emotional and behavioral problems [20,21], but this hypothesis clearly needs further testing. More self-soothing, including thumb sucking and plucking at clothes, was related to higher cortisol reactivity, but only for secure infants. As secure infants are supposed to be better capable of reg-
ulating stress, these findings seem to correspond to the model of synchrony [7], which suggests that a correlation between behavior and physiology reflects a well-functioning stress system. But how could physiological stress responses that are synchronized with regulatory behaviors be beneficial? First, synchrony might enable a better coping with the stressor. One can speculate that secure infants that are physiologically stressed will experience a reduction in unpleasantness and negativity when certain specific behaviors are applied, such as self-soothing [22]. A better coordination among the stress systems would thus benefit the emotional state of the infant. This hypothesis is difficult to test in infants that cannot indicate their own experienced emotions, but it can be tested in older children and adults. A second benefit of synchrony could be to elicit external regulation from the attachment figure, often the mother. Physiological distress can prompt the use of regulatory behaviors, including self-soothing. When the physiological distress becomes too high, the infants’ own ability to engage in regulation might be undermined and the infants need to signal their distress to their mothers, for example by crying. Additionally, through the mechanism of synchrony between mothers and their infants, mothers react physiologically in accordance with their children [23]. Infant behavioral signals of stress, accompanied with increased maternal physiological stress levels, might activate mothers to optimally comfort and care for their children. Almost 30% of the variance in cortisol reactivity to the SSP was explained by infant regulatory behaviors when the mother was absent, while the regulatory behaviors during the mother’s presence were not predictive of cortisol reactivity. This result confirms the survival function of the attachment system. Infants are helpless and completely dependent on adults for care and protection during the first years of life [24]. Bowlby [25] therefore proposed that infants are endowed with this strong motivation to remain close to their attachment figures as their primary source of safety. When separated from the attachment figure, infants feel threatened and probably the HPA-axis is activated as a consequence. Moreover, as infants have no one else to depend on for regulation but themselves, infant regulatory behaviors during separations from the attachment figure are mostly correlated to their cortisol reactivity. Infants who are highly physiologically distressed are probably prompted to regulate more. These regulatory behaviors might also modulate the physiological stress response by reducing the unpleasantness and negativity of the situation or by signaling their stress to elicit external regulation from the attachment figure. In contrast, the regulatory behaviors in the mother’s presence were not predictive of infant cortisol reactivity. From an evolutionary point of view, these episodes are supposed to be less threatening to the infant as
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
8
their attachment figure is present. Moreover, the infant is no longer only dependent on his own regulatory capacities as the mother is present to provide co-regulation. This together might explain why infant regulatory behaviors in the mother’s presence were not predictive of infant cortisol reactivity to the SSP. Cortisol recovery from the SSP was not predicted by infant regulatory behaviors in mother’s absence nor in mother’s presence. An explanation could be that infant regulatory behaviors during stress do not aid cortisol recovery, at least not in this short time period and at this age. It could also be that the maternal co-regulatory behaviors during infant stress are more important to facilitate infant recovery from a stressor. Maternal co-regulation is part of maternal caregiving behavior. We have shown in our lab that higher quality of maternal caregiving behavior (sensitivity and cooperation) was related to a quicker recovery from a mildly stressful bathing session in 3-months-olds [26], and that more weeks of breastfeeding in the first six months of life were related to quicker recovery from the SSP in the same group of infants as currently used in this study [11]. Whether maternal co-regulation, and the specific maternal coregulatory behaviors, aid infant cortisol recovery is topic for future research. This study is among the first to investigate differential associations between a range of regulatory behaviors and cortisol reactions in secure versus insecure infants. Strong points are the expansion of regulatory behaviors, the focus on the separation episodes, and the inclusion of both cortisol reactivity and recovery. Some limitations also need to be addressed. First, the sample of participating mothers and infants was quite homogeneous: the majority was Dutch, highly educated and married or living together with their partner. This constrains the generalizability of the study. Second, the sample size was limited. Studies with larger samples could investigate the associations between behavioral responses and cortisol reactivity to the SSP separately in the three insecure attachment groups. This would offer a deeper understanding of infant behaviors under stress, as for example avoidantly attached infants, known to minimize attachment behaviors towards their mothers [27], may also display other behaviors when left alone when compared to the other groups of insecure infants. Lastly, the observational study design does not allow any conclusions on causality. It is unclear whether the regulatory behaviors proceed or are a result of heightened cortisol concentrations. In summary, the present study showed differential associations between behavioral and cortisol responses to a stressor in secure versus insecure infants. This might indicate that secure and insecure infants apply different behavioral regulatory strategies when physiologically stressed. Before definite conclusions can be drawn, the study findings need to be replicated and enriched by future research. Research differentiating between specific types of insecurity is needed as different profiles of physiology and behavior likely exist for avoidant, resistant, and disorganized infants. In addition, research including other physiological stress measurements that pertain to the parasympathetic and sympathetic nervous system, such as heart rate variability, respiratory sinus arrhythmia and salivary alpha-amylase, could complement the present study findings. Also, follow-up studies are needed to examine the longitudinal manifestation of these stress response associations and the potential consequences for later child development. Finally, the nature of these relations in response to other stress paradigms than the SSP could be examined, such as during the transition to childcare where separation from the mother is prolonged [28].
Conflict of interest There were no conflicts of interest, financial or otherwise to declare.
Role of the funding source This research was supported by a personal grant of the Netherlands Organization for Scientific Research (NWO) awarded to Carolina de Weerth (grant number 452-04-320). Acknowledgements We thank the families who kindly participated in the BIBO study, and all research assistants and students for their assistance with data collection. References [1] M.D.S. Ainsworth, M.D. Blehar, E. Waters, S. Wall, Patterns of Attachment. A Psychological Study of the Strange Situation, Erlbaum, Hillsdale, NJ, 1978. [2] S.D. Calkins, E.M, Leerkes, Early attachment processes and the development of emotional self-regulation, in: K.D. Vohs, R.F. Baumeister (Eds.), Handbook of Self-regulation: Research, Theory, and Applications, 2nd edn., The Guilford Press, New York, NY, 2011, pp. 355–373. [3] A.N. Schore, Effects of a secure attachment relationship on right brain development, affect regulation, and infant mental health, Inf. Mental Health J. 22 (2001) 7–66, http://dx.doi.org/10.1002/1097-0355(200101/ 04)22:1<7:AID-IMHJ2>3.0.CO;2-N. [4] G. Spangler, M. Schieche, Emotional and adrenocortical responses of infants to the strange situation: the differential function of emotional expression, Int. J. Behav. Dev. 22 (1998) 681–706, http://dx.doi.org/10.1080/016502598384126. [5] H. Selye, The Story of the Adaptation Syndrome, Acta, Montreal, 1952. [6] R.M. Sapolsky, L.M. Romero, A.U. Munck, How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions, Endocrinol. Rev. 21 (2000) 55–89. [7] S.J. Erickson, P. MacLean, C. Qualls, J.R. Lowe, Differential associations between infant affective and cortisol responses during the still face paradigm among infants born very low birth weight versus full-term, Inf. Behav. Dev. 36 (2013) 359–368, http://dx.doi.org/10.1016/j.infbeh.2013.03.001. [8] M.R. Gunnar, L. Brodersen, M. Nachmias, K. Buss, J. Rigatuso, Stress reactivity and attachment security, Dev. Psychobiol. 29 (1996) 191–204, http://dx.doi. org/10.1002/(SICI)1098-2302(199604)29:3<191:AID-DEV1>3.0.CO;2-M. [9] J.M. Braungart, C.A. Stifter, Regulation of negative reactivity during the strange situation: temperament and attachment in 12-month-old infants, Infant Behav. Dev. 14 (1991) 349–364, http://dx.doi.org/10.1016/01636383(91)90027-P. [10] C. De Weerth, J. Jansen, M.H. Vos, I. Maitimu, E.G. Lentjes, A new device for collecting saliva for cortisol determination, Psychoneuroendocrinology 32 (2007) 1144–1148, http://dx.doi.org/10.1016/j.psyneuen.2007.07.005. [11] R. Beijers, J.M.A. Riksen-Walraven, C. De Weerth, Cortisol regulation in 12-month-old human infants: associations with the infants’ early history of breastfeeding and co-sleeping, Stress 16 (2013) 267–277, http://dx.doi.org/ 10.3109/10253890.2012.742057. [12] M. Main, J. Solomon, Discovery of a disorganized/disoriented attachment pattern, in: T.B. Brazelton, M.W. Yogman (Eds.), Affective Dev. Inf., Ablex, Norwood, NJ, 1986, pp. 95–124. [13] R. Beijers, J. Jansen, J.M.A. Riksen-Walraven, C. De Weerth, Attachment and infant night waking: a longitudinal study from birth through the first year of life, J. Dev. Behav. Pediatr. 32 (2011) 635–643, http://dx.doi.org/10.1097/DBP. 0b013e318228888d. [14] L.S. Aiken, S.G. West, Multiple Regression: Testing and Interpreting Interactions, Sage, Thousand Oaks, CA, 1991. [15] A.F. Hayes, J. Matthes, Computational procedures for probing interactions in linear and logistic regression SPSS and SAS implementations, Beh. Res. Methods 41 (2009) 924–936, http://dx.doi.org/10.3758/BRM.41.3.924. [16] E.A. Carlson, M.C. Sampson, L.A. Sroufe, Implications of attachment theory and research for developmental-behavioral pediatrics, J. Dev. Beh. Pediatr. 24 (2003) 364–379. [17] J. Cassidy, J.D. Jones, P.R. Shaver, Contributions of attachment theory and research: a framework for future research, translation, and policy, Dev. Psychopath 25 (2013) 1415–1434. [18] J. Belsky, M. Rovine, D.G. Taylor, The Pennsylvania infant and family development project III: The origins of individual differences in infant-mother attachment: maternal and infant contributions, Child Dev. 55 (1984) 718–728, http://dx.doi.org/10.1111/j.1467-8624.1984.tb03810.x. [19] A. Desantis, W. Coster, R. Bigsby, B. Lester, Colic and fussing in infancy, and sensory processing at 3–8 years of age, Inf. Mental Health J. 25 (2004) 522–539, http://dx.doi.org/10.1002/1mhj.20025. [20] R.P. Fearon, M.J. Bakermans-Kranenburg, M.H. van IJzendoorn, A.M. Lapsley, G.I. Roisman, The significance of insecure attachment and disorganization in the development of children’s externalizing behavior: a meta-analytic study, Child Dev. 81 (2010) 435–456. [21] A.M. Groh, G.I. Roisman, M.H. van IJzendoorn, M.J. Bakermans-Kranenburg, R.P. Fearon, The significance of insecure and disorganized attachment for children’s internalizing symptoms: a meta-analytic study, Child Dev. 83 (2012) 591–610, http://dx.doi.org/10.1111/j.1467-8624.2011.01711.x.
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008
G Model BBR-10496; No. of Pages 9
ARTICLE IN PRESS R. Beijers et al. / Behavioural Brain Research xxx (2016) xxx–xxx
[22] A.J. Oldehinkel, J. Ormel, N.M. Bosch, E.M.C. Bouma, A.M. Van Roon, J.G.M. Rosmalen, H. Riese, Stressed out? Associations between perceived and physiological stress responses in adolescents: the TRAILS study, Psychophysiol 48 (2011) 441–452, http://dx.doi.org/10.1111/j.1469-8986. 2010.01118.x. [23] R. Feldman, Bio-behavioral synchrony: a model for integrating biological and microsocial behavioral processes in the study of parenting, Parenting: Sci. Pract. 12 (2012) 154–164, http://dx.doi.org/10.1080/15295192.2012.683342. [24] J.S. Chisholm, The evolutionary ecology of attachment organization, Human Nat. 7 (1996) 1–37. [25] J. Bowlby, Attachment and loss Attachment, vol. 1, Basis Books, New York, 1969.
9
[26] E.M. Albers, J.M.A. Riksen-Walraven, F.C.G.J. Sweep, C. de Weerth, Maternal behavior predicts infant cortisol recovery from a mild everyday stressor, J. Child Psychol. Psychiatry 49 (2008) 97–103, http://dx.doi.org/10.1111/j.14697610.2007.01818.x. [27] J. Cassidy, Emotion regulation: influences of attachment relationships, Monogr. Soc. Res. Child Dev. 59 (1994) 228–249, http://dx.doi.org/10.1111/j. 1540-5834.1994.tb01287.x. [28] L. Ahnert, M.R. Gunnar, M.E. Lamb, M. Barthel, Transition to child care: associations with infant-mother attachment, infant negative emotion, and cortisol elevations, Child Dev. 75 (2004) 639–650, http://dx.doi.org/10.1111/j. 1467-8624.2004.00698.x.
Please cite this article in press as: R. Beijers, et al., Differential associations between behavioral and cortisol responses to a stressor in securely versus insecurely attached infants, Behav Brain Res (2016), http://dx.doi.org/10.1016/j.bbr.2016.10.008