The relationship of attachment insecurity to subjective stress and autonomic function during standardized acute stress in healthy adults

Abstract

Objective

The purpose of this study was to test predicted relationships between adult attachment and stress using subjective and physiological measures.

Methods

Sixty-seven healthy adults completed measures of adult attachment and perceived chronic stress. Subjective stress and the high-frequency (HF) and low-frequency (LF) spectral bandwidths of heart rate variability (HRV) were measured during a standardized stress protocol.

Results

Attachment anxiety is associated with between-subject differences in chronic perceived stress (P=.001) and subjective acute stress (P=.01). There is a main effect of attachment avoidance on between-subject differences in HF HRV (P=.004). Attachment avoidance is inversely associated with HF HRV, independent of age and variability in respiration.

Conclusion

Attachment anxiety is associated with self-reported distress. Attachment avoidance is inversely associated with HF HRV, a marker of vagal influence on cardiac activity, but is not associated with subjective stress.

Introduction

In a synthesis of the literature regarding the link between attachment security, stress, and physical health, we found that evidence supports a model in which attachment insecurity contributes to increased stress by several mechanisms, including increased perception of stress and perturbations of stress physiology among insecure individuals. However, the evidence for these associations is suggestive rather than conclusive [1]. We now report on tests of the predictions of that model regarding the relationships between adult attachment style, perceived stress, and a marker of autonomic function during acute stress.

Attachment theory and the rationale for a connection between adult attachment style, stress, and health are described in detail in the earlier review [1]. In brief, attachment theory proposes that a biologically based system of signals and behaviors regulates proximity between an infant and its primary caretakers with the goal of increasing the infant's chance of survival in the face of external threats [2]. Adult attachment theory expands on this basic tenet by proposing that expectations and responses to interpersonal situations learned in early relationships provide a template for relatively stable and enduring patterns of intimate interpersonal relationships, which are referred to as adult attachment style. While the infant attachment system regulates goals such as physical proximity and soothing in response to distress, the adult system regulates mentalized constructs related to proximity and intimacy in close relationships such as trust, dependency, and comfort expressing affect within the relationship. It is postulated that adult attachment style remains linked to the psychological and biological systems that regulate threat, stress response, and recovery from stress.

The most common tools to assess adult attachment style are structured interviews, particularly the Adult Attachment Interview [3], and self-report questionnaires [4]. The Adult Attachment Interview and some self-report measures of attachment insecurity assign individuals to a category of attachment style (e.g., secure/autonomous, dismissing, preoccupied, unresolved/disorganized [3], or secure, preoccupied, dismissing, fearful [5]). Alternatively, many self-report adult attachment instruments measure two primary dimensions of attachment style, attachment anxiety and attachment avoidance [4]. Attachment anxiety is characterized by intense feelings about the partner, fear of loss, care-seeking, and perceived dependency. Attachment avoidance is characterized by a relative deactivation of attachment signals, distrust of intimacy, and a preference for autonomy. High scores on either attachment dimension are described as indicating insecurity. Among self-report measures of adult attachment, recent analyses [4], [5], [6] and expert opinion [7] suggest that the dimensional approach has greater validity than the categorical approach.

Although evidence supports a relationship between attachment style and stress response [1], this evidence is compromised because the relationship between attachment and stress has not been studied in adults with both a validated measure of attachment and a physiological measure of stress. Typically, studies use self-report measures of both stress and attachment [8], [9], leaving unresolved the possibility that both measures refer to a singular underlying construct, such as a propensity to express distress (e.g., neuroticism) that causes self-reports of both stress and attachment anxiety to be elevated. Self-reports of avoidant attachment, on the other hand, may be compromised because avoidant attachment is conceptually related to under-reporting of distress. In attachment theory, attachment avoidance is the result of a relative deactivation of distress signals in the service of maintaining an attachment relationship with a parent who does not tolerate a child's distress easily [10].

Physiological measures of stress have been studied in animal models [11], [12], in which inferences about attachment are made based on developmental observations of maternal–infant interaction or experimental manipulation of these interactions [12]. In humans, a study of cardiovascular responses to stress was reported in which inferences about attachment were made based on the presence of parental loss during childhood and a measure of family disturbance, but this study did not measure attachment directly [13]. In children in the Strange Situation, an anxiety-provoking procedure that reliably categorizes attachment type, children with avoidant attachment appear behaviorally to be less distressed than secure and anxious-ambivalent children when they are separated from mother and exposed to a stranger [14]. In an early study, heart-rate data collected by telemetry suggested that, in spite of the lack of overt distress, avoidant children experience tachycardia during separation, which is similar to other attachment types, and fail to show the deceleration after reunion with mother that is typical of secure children. In fact, tachycardia persisted long into the reunion phase in avoidant children [15]. A recent study replicated the finding that tachycardia occurs during separation in insecure-avoidant children in spite of a lack of overt signs of distress, but did not find any difference between attachment styles with respect to heart deceleration after reunion between child and parent [16]. Physiological tests of stress response in adults may be useful both to see beyond the distress-reporting biases associated with attachment style and to explore possible psychophysiological correlates of attachment style.

If there is a causal relationship between attachment insecurity and physiological stress in adults, it could occur in a number of ways. The possibilities include a relationship of attachment insecurity to (i) baseline or tonic levels of hypothalamic-pituitary-adrenal activity (HPA) or autonomic nervous system (ANS) activity (sympathetic or parasympathetic-vagal), (ii) the threshold to activation of the HPA or autonomic response when threat is perceived (“stress response”), (iii) the intensity of stress response, and (iv) recovery from stress response. Although it is not known whether a relationship between insecurity and stress would be similar for attachment anxiety and attachment avoidance, the previous review suggested that physical pathology may be more strongly associated to attachment avoidance than attachment anxiety [1].

This study is designed to address the methodological limitations of previous research by testing the relationships between each dimension of adult attachment insecurity (avoidance and anxiety) and (i) measures of subjective stress, chronically and during a standardized procedure, and (ii) a marker of autonomic influences on cardiac reactivity during stress. Thus, the study design allows testing of congruence or noncongruence between subjectively experienced (self-report) stress and the autonomic response to a stressful situation.

The physiological measure used is frequency-domain analysis of heart rate variability (HRV). Measurement of HRV provides a noninvasive probe of the function of systems that influence heart rate, especially the sympathetic and parasympathetic (vagal) divisions of the ANS, by analyzing temporal patterns in the small variations that occur in the interval from one heart beat to the next (R-R interval). A review of HRV, its physiological interpretation, and standards for HRV measurement has been provided by a joint task force of the European Society of Cardiology and the North American Society of Pacing Electrophysiology [17]. Spectral power analysis of fluctuations in R-R interval allows measurement of the power of high-frequency, fast-acting influences on heart rate (HF HRV, 0.15–0.40 Hz), which are primarily due to vagal activity, and low-frequency, slightly slower acting influences on heart rate (LF HRV, 0.04–0.15 Hz), which are due to a more complex set of regulatory factors including sympathetic and parasympathetic contributions. The very low frequency spectral band of HRV (<.04 Hz, i.e., factors whose influence on heart rate have a periodicity of 25 s or more) has a less clear physiological interpretation and cannot be assessed adequately over short periods of measurement (≤5 min) [17], but has recently received attention in psychosomatic research [18].