In the current study, we used a multimodal approach to investigate the association between attachment representations and empathic stress in a situation of acute psychosocial stress in romantic couples. Deepening our understanding of these dynamics may be an important stepping stone in boosting resilience in romantic relationships, thus enabling couples to manage and recover from stress more effectively.
Our findings showed resonance in HF-HRV and cortisol stress responses, indicating the proportional de-activation of parasympathetic activity and release in cortisol in passive observers and directly stressed targets. Also, observer attachment was shown to significantly associate with stress-induced biobehavioral synchrony (BBS) in these two stress markers. Particularly, secure (vs. insecure) observer attachment was associated with higher dyadic HF-HRV and cortisol stress resonance. No resonance (or association with attachment) was found for subjective-psychological and sympathetic stress reactivity. Also, observers’ vicarious stress responses (a proxy of firsthand stress sensitivity, independent of target stress reactivity) were unrelated to their attachment representations.
Our finding of higher stress resonance in dyads with observers classified as securely attached is in line with our a priori hypothesis, which was based on attachment theory. Securely attached individuals typically hold positive inner working models (IWMs) of the self and others. They expect others to possess good intentions, trust them, and view themselves as competent in providing support (Mikulincer & Shaver, 2009). This positive belief system encourages care for others and prevents hostile behavior (Stern & Cassidy, 2018). In our study, the positive belief system may be reflected in relatively higher physiological attunement with the romantic partner’s acute stress. Conversely, insecure-dismissingly attached individuals tend to have negative perceptions of others (Bartholomew, 1990), exhibit reduced emotional understanding (Gallistl et al., 2024), and, in their aim to prioritize their own independence, may disregard others’ needs and emotions (Mikulincer et al., 2005). These tendencies are reflective of a deactivation of the attachment system to minimize emotional responsiveness and the emerging need to seek others’ support (Spangler & Zimmermann, 1999). Such deactivation may contribute to the lack of stress resonance seen in insecure-dismissingly attached partners.
A potential contributing factor to the attachment-related differences in stress resonance could be attachment-related differences in empathic abilities. We know from prior research that empathic abilities are crucial in modulating BBS in general (Atzil et al., 2014), and more specifically in the context of stress resonance (Blasberg et al., 2023; Buchanan et al., 2012; Engert et al., 2014). Moreover, in their meta-analysis, Xu et al.(2022) revealed a positive interaction between empathy and secure attachment, and a negative interaction between empathy and insecure-dismissing attachment. In a prior study within the same sample as tested here, we furthermore discovered that insecure-dismissing attachment, as measured with the Adult Attachment Interview (AAI), was linked to diminished cognitive empathy, specifically Theory of Mind abilities assessed in a behavioral video task, the EmpaToM (Gallistl et al., 2024; Kanske et al., 2015). However, Theory of Mind abilities did not interact with the link of attachment and stress resonance found in the current data. There are several possible explanations for this negative result. First, previous findings of an association between the occurrence of empathic stress and empathic abilities were questionnaire- rather than EmpaToM-based. Touching upon the same issue as found in the attachment field (i.e., the lack of a correlation between self-report questionnaires and narrative, arguably more objective measurement methods,Roisman et al., 2007), questionnaire-based and behavioral empathy data do not reliably correlate(Böckler et al., 2018). Further, attachment could influence both Theory of Mind and stress resonance independently, without a mediating role.
Emotional closeness is another potential contributing factor to the link between attachment and the tendency to exhibit stress resonance. Thus, when comparing romantic partners and strangers in our earlier study, it was a modulator in the occurrence of stress resonance (Engert et al., 2014b). Moreover, individuals with insecure-dismissing attachment have been found to perceive others as less emotionally close compared to securely attached individuals (Lee & Gillath, 2016). Unfortunately, only romantic partners were recruited in the current sample, and emotional closeness was not assessed.
The emotional context seems to be crucial for the occurrence of BBS, including physiological synchrony. In our study, the utilized stressor was relationship-unrelated, mimicking a situation of an external stressor affecting only one relationship partner. Therefore, it is not surprising that our results differ from studies using relationship-related stressors which found that higher cortisol synchrony was linked to lower relationship satisfaction (Laws et al., 2015), lower levels of support behavior (Ha et al., 2016), and was predictive of breakup (Schneiderman et al., 2014). When considering the resilience-promoting potential of attachment security, lower cortisol attunement in a relationship conflict may indicate that one partner is effectively managing the conflict by remaining calm and composed, thereby providing a de-stressing, stabilizing influence on the couple. Higher cortisol attunement when confronted with a relationship-unrelated, external stressor, on the other hand, may have multiple adaptive effects. Thus, it may allow for the mobilization of the necessary energetic resources to overcome adversity together (Dhabhar, 2014). Resonance in a situation of external stress influence may also foster a better understanding of the partner’s emotional state. By showing similar emotional and bodily experiences, individuals may better empathize and connect with the other, enhancing interpersonal relationships and promoting mutual support (Engert et al., 2019a). Further, stress resonance may serve as a warning signal in situations of danger. By perceiving and resonating with the stress experienced by others, individuals may detect potential risks and adjust their behavior accordingly, promoting safety and self-preservation (Savadori & Lauriola, 2021). Lastly, our models indicate that securely attached individuals exhibit cortisol resonance also in the absence of stress in both partners (i.e., low target stress correlates with low observer stress), indicating a heightened synchrony during non-stressful periods. These could suggest that securely attached individuals may regulate homeostatic processes in synchrony, encompassing both stressful and non-stressful phases. This propensity could serve as a protective mechanism, mitigating the long-term stress burden.
Despite its many protective properties, attachment security may bear potential risks when it comes to a context of increased suffering or stress exposure in the context of empathic stress. The permanent confrontation with the stress of others, conceivable in the partners or children of chronically stressed individuals, or in people working in care professions, may lead to an increased physiological stress load in the empathic observer. Given the critical role of heightened cortisol levels in the development of stress-associated disease (Chrousos, 2009; McEwen, 1998; Sapolsky et al., 2000), long-term stress resonance may put individuals at risk for adverse health outcomes themselves. Thus, while covariation in which couples influence each other toward homeostasis may promote health and well-being (Sbarra & Hazan, 2008), physiological linkage with chronically stressed loved ones may predispose affected individuals to HPA-axis dysregulation themselves.
Other than expected, we found no association of observer vicarious stress (i.e., observer firsthand stress sensitivity, independent of target stress reactivity) with observer attachment. Our a priori hypothesis of higher vicarious stress in insecurely attached individuals had been based on previous research demonstrating higher cortisol reactivity to direct stress exposure in individuals with an insecure-preoccupied attachment (Brooks et al., 2011; Diamond et al., 2008; Pietromonaco & Beck, 2019; Powers et al., 2006). Since insecure-preoccupied individuals were clearly underrepresented in our sample, there may have been insufficient power to detect the expected effect. In contrast to insecure-preoccupied individuals who are prone to the use of hyperactivating regulatory attachment strategies, insecure-dismissing individuals tend to distance themselves from others and rely more on themselves, employing deactivating attachment strategies such as suppressing their emotional needs and minimizing the importance of close relationships (Spangler & Zimmermann, 1999), which may render them less vulnerable to the experience of vicarious stress.
A limitation of our study is the lack of a large enough insecure-preoccupied participant group. Future studies should expand their scope to include individuals with insecure-preoccupied attachment, as these exhibit unique IWMs and behaviors. Further, measuring a marker for emotional closeness would have been beneficial to further understand the found association. Lastly, our study was limited by a lack of diversity in the sample, as we only included heterosexual couples.
In sum, we provide evidence for a link of attachment with the occurrence of empathic stress, such that dyads of romantic partners with securely (as opposed to insecure-dismissingly) attached observers exhibited higher resonance during acute psychosocial stress experience. This association between observer attachment and empathic stress was specific to HF-HRV and cortisol responding, and hence not observed in subjective-psychological or heart rate stress reactivity. Observer attachment was also unrelated to the strength of observer vicarious stress responding. Higher HF-HRV and cortisol resonance may have multiple adaptive functions, such as fostering mutual understanding and triggering partner support. Conversely, it may also have negative health implications in situations of chronic exposure to others’ stress. Further research is needed to explore the long-term consequences of attachment-related empathic stress on health and well-being, and to investigate the underlying mechanisms driving these effects. By elucidating the role of attachment in empathic stress, we may contribute to the development of targeted interventions aimed at promoting both stress resilience and healthier relationships.