According to Bowlby's original proposal, significant disruptions in attachment bonds occurring in early childhood would be associated with an increased risk of depression, while repeated experiences of temporary separation, or threats of separation, could predispose to anxiety (Bowlby, 1970, 1998; Holmes, 2013). Subsequent research suggests has allowed a better delineation of the effects of disrupted attachment on the clinical profile of these disorders. In depression, insecure attachment has been associated with greater symptom severity, a higher relapse rate, suicidal behavior, reluctance to seek help and poor adherence to medication (Ozer et al., 2015; Hooper et al., 2016; Conradi et al., 2018; Huang Y. C. et al., 2019; Cacciola and Psouni, 2020). In anxiety disorders, insecure attachment has also been associated with a poorer response to both pharmacological and psychological therapies (Milrod et al., 2016; Nielsen et al., 2019; Petrowski et al., 2019). Lower levels of secure attachment also appear to predict comorbid anxiety disorder in depressed patients (Marganska et al., 2013; Huang Y. L. et al., 2019). This association appears to be most specific for social anxiety disorder, in which insecure attachment is associated with both the presence and the severity of comorbid depression (Adams et al., 2018, 2019; Elling et al., 2022). These findings are consistent with a role for disturbances in childhood attachment in the onset, persistence and co-occurrence of depressive and anxiety disorders. The above studies, and other relevant observational research on insecure attachment, are summarized in Table 2. From these findings, it is clear that both preoccupied and avoidant patterns of adult attachment are associated with comorbid anxiety and depression, though there may also be an association with disorganized attachment when this pattern is associated with significant levels of fear. Having established this result with some consistency, it is now possible to examine the underlying cellular and neurobiological processes at work in patients with these disorders.

Insecure attachment also influences fundamental cognitive and affective processes implicated in the pathogenesis of anxiety and depression. These include impairments in emotion recognition (Monti and Rudolph, 2014), emotional regulation (Malik et al., 2015; Conrad et al., 2021), and theory of mind skills (Ozer et al., 2015; Koelkebeck et al., 2017). These deficits have been documented in patients with either disorder, as well as in patients with comorbid anxiety and depression (Marques et al., 2018; Radetzki et al., 2021). In addition, experiences of disrupted attachment appear to mediate the relationship between other environmental risk factors, such as childhood poverty or sexual abuse, and the subsequent risk of depression or anxiety (Briere and Jordan, 2009; Fearon et al., 2017).

The neural mechanisms underlying these higher-order processes are better understood now than they were in Bowlby's time. Attachment security appears to be encoded in both amygdalae (Lemche et al., 2006) and severe degrees of insecure attachment are associated with increased amygdalar volume (Moutsiana et al., 2015; Lyons-Ruth et al., 2016) and hippocampal volume (Hidalgo et al., 2019). Insecure attachment is also associated with altered functioning in the prefrontal, anterior cingulate, insular and parietal cortices, basal ganglia and amygdala during the performance of tasks related to social cognition or reward. The changes observed in these regions are suggestive of inefficient information processing and heightened reactivity to external situations (Warren et al., 2010; Moutsiana et al., 2014; Schneider-Hassloff et al., 2015; Quevedo et al., 2017; Miller et al., 2020). Avoidant and anxious attachment have both been associated with specific alterations in the functioning of a “social aversion circuit” involving the dorsal anterior cingulate cortex and middle temporal gyrus; such changes may be relevant to the fear of criticism or rejection, negative cognitions and avoidance behavior that are seen in patients with comorbid anxiety and depression (Krause et al., 2016). Insecure attachment representations have also been associated with increased asymmetrical activation of the parietal cortices when processing emotional memories, potentially leading to increased arousal, impaired emotion regulation, and an increased vulnerability to both anxiety and depression (Kungl et al., 2016). Unresolved attachment has also been associated with reduced white matter integrity in the corpus callosum, leading to cognitive impairments that could serve as a “general” vulnerability factor for both anxiety disorders and depression (Riem et al., 2019). The results of these imaging studies support the hypothesis of a close link between insecure attachment, altered or impaired cognition, and an increased vulnerability to comorbid anxiety and depression.

At a neurochemical level, these changes appear to correlate with alterations in dopaminergic circuitry (Strathearn, 2011) and peptidergic transmission, particularly involving endogenous opioid peptides and oxytocin (Muller et al., 2019). In particular, mu-opioid (μ) receptor availability has been found to correlate with adult attachment (Turtonen et al., 2020); lowered availability of these receptors has been associated with symptoms of both depression and anxiety (Nummenmaa et al., 2020). Likewise, oxytocin has been found to modulate neural circuits involved in fear as well as those involved in the processing of social information. As oxytocin is a key mediator of attachment behavior, insecure attachment could alter the influence of this transmitter on the activity and connectivity of brain regions such as the amygdala. This could be particularly relevant to the comorbidity between specific anxiety disorders, such as social anxiety disorder, and depression (Kirsch et al., 2005).

Insecure attachment has also been associated with alterations in autonomic nervous functioning, increased activity of the hypothalamic-pituitary-adrenal axis, and immune dysregulation (Jaremka et al., 2013; Abtahi and Kerns, 2017), all of which have been linked to the pathogenesis of comorbid depression and anxiety (McQuaid, 2021). Among these processes, alterations in cardiac autonomic functioning may be particularly important, as they represent a potential mechanism of “embodiment” through which experiences of attachment and separation can influence fear conditioning and sensitivity to both anxiety and depression (Battaglia and Thayer, 2022; Battaglia et al., 2022a,b; Gander et al., 2022). Cross-talk between central and peripheral structures involved in fear responses, such as the ventromedial prefrontal cortex, are also crucially influenced by attachment security (Eisenberger et al., 2011; Battaglia et al., 2022a). Likewise, the effects of attachment on the functioning of the HPA axis may explain the relationship between chronic stress and the onset of depression and anxiety in insecurely attached individuals (Smyth et al., 2015). The effect of attachment security on adult personality and mental health may itself be modified by genetic factors, such as functional polymorphisms of the oxytocin receptor (Schneider-Hassloff et al., 2016). There is also preliminary evidence that attachment security can influence the balance between pro- and anti-inflammatory cytokines; secure attachment stimulates the expression of anti-inflammatory genes (Stanton et al., 2017), while anxious attachment is associated with increased levels of inflammatory markers (Ehrlich et al., 2019). These changes could result in central nervous system inflammation and increased oxidative stress, providing another potential mechanism linking comorbid anxiety and depression with attachment (Tanaka and Vecsei, 2021; Tanaka et al., 2022b,c).

An integration of the above findings is presented in Figure 2. Early childhood attachment is determined by non-linear interactions involving several factors, including childhood temperament (itself the result of gene x environment interactions), parental attachment style and mental health, and broader socio-cultural factors (Barry et al., 2008; Toepfer et al., 2019; Takegata et al., 2021). Through processes that are likely to involve a combination of embodiment and higher-order representation, attachment influences the development of higher-order cognitive and affective processes related to the pathogenesis of anxiety and depression. If attachment is secure, the outcome of these processes is a secure adult attachment style and psychological resilience (Long et al., 2020). However, various patterns of insecure attachment can lead to inefficiency or biases in these higher-order processes, leading to insecure adult attachment, reduced relationship stability and satisfaction, and a sensitivity to both anxiety and depression, which are themselves components of the triphasic “separation response” (Bowlby, 1988, 1998). Also important in this model are the interactions between insecure attachment and other forms of adversity, which include social disadvantage, childhood abuse, discord or violence within adult intimate relationships, and other forms of chronic stress (Briere and Jordan, 2009; Fearon et al., 2017).

Though tentative, this proposal is consistent with our current understanding of the pathophysiology of anxiety (Nolte et al., 2011) and depression (Holmes, 2013). With the accumulation of further clinical and translational evidence, it should be possible to refine the above model in several ways. For example, the specific sub-type(s) of insecure attachment that predispose to comorbid anxiety and depression could be identified more consistently. Likewise, it is likely that distinct attachment-related neural and cognitive processes could be involved in the comorbidity between depression and specific anxiety disorders, such as social anxiety or panic disorder (Torpey et al., 2007; Elling et al., 2022). Finally, biomarkers related to insecure attachment, reflecting altered information processing in the brain circuitry involved in affect regulation, problem-solving and social cognition, could be identified in patients and assessed as predictors of outcome (Gander and Buchheim, 2015; van Hoof et al., 2019).

The integration of perspectives and findings from attachment theory and cognitive neuroscience is of more than theoretical interest in patients with comorbid anxiety and depressive disorders. Such an approach could optimize treatment outcomes through the selection of attachment-based therapies that address either maladaptive attachment styles or the cognitive errors associated with them (Newman et al., 2015; Gunlicks-Stoessel et al., 2019; Cortes-Garcia et al., 2020). Such therapies could integrate existing psychodynamic or cognitive behavioral approaches under an attachment framework (Fonagy and Target, 2007; Holmes, 2013). For example, it has been shown that cognitive training focused on the modification of biases in information processing can modulate attachment anxiety (Doolan and Bryant, 2021). Couples or family therapies based on attachment theory may also be effective in treating adults with comorbid depression and anxiety, while minimizing the burden experienced by a spouse or partner (Johnson and Greenman, 2006). A better understanding of the neural and molecular mechanisms linking insecure attachment and comorbid anxiety and depression could also inform innovative approaches to biological treatment, such as pharmacological modulation of attachment-related peptidergic mechanisms (Kormos and Gaszner, 2013) or attachment-related alterations in immune or endocrine function (Hennessy et al., 2019), and even stimulation therapies that target attachment-related autonomic dysfunction (Fanselow, 2013). It should also not be assumed that the mechanisms discussed in this paper are the only ones linking attachment with anxiety and depression. For example, recent research suggests that experiences of attachment security may be “biologically embedded” in infants through alterations in DNA methylation (Merrill et al., 2021); such alterations involve changes in the expression of pro-inflammatory genes, which may be related to subsequent anxiety and depression.

Finally, pursuing this line of research could lead to a deeper understanding of the developmental roots of this comorbidity. It is well known that parental depression can affect infant attachment, and parental depression may itself reflect attachment patterns and experiences in the mother or father's own childhood (Sliwerski et al., 2020). More generally, patterns of attachment are known to be stable across generations, and this may reflect a bidirectional link with anxiety and depression: the presence of these disorders in a parent may both reflect their own childhood experiences and insecure attachment, and predict insecure attachment and subsequent anxiety or depression in their children (Galbally et al., 2022). Such an “intergenerational” transmission of attachment security has also been demonstrated in animal models, and been linked to altered expression of specific genes in the brain (Alyamani et al., 2021). A circular process of this sort offers opportunities for early intervention (e.g., identification and management of internalizing symptoms in children of depressed mothers) or even prevention (e.g., attachment-based therapies for parents whose depression and anxiety is related to attachment insecurity, thereby preventing the “transmission” of an insecure attachment style). Such an approach may be especially fruitful in “high-risk” families with an increased genetic risk for depression, or in families exposed to social disadvantage or chronic stress (Lecompte et al., 2018).

Attachment theory offers a promising explanatory framework for our understanding of comorbid anxiety and depression. The coherence between the constructs of attachment theory and the findings of cognitive neuroscience, which was envisioned by early workers in this field, can now be placed on a more secure footing. The available evidence suggests that insecure attachment in childhood can crucially alter the subsequent functioning of key neurocognitive, neuroendocrine, and neuroimmune processes, leading to changes at both the physical level (such as increased inflammation and an altered stress response) and in higher-order cognitive functioning (cognitive errors and biases). These changes can predispose to the development and persistence of subsequent anxiety and depression. This framework may be heuristically useful both in terms of deepening our understanding of specific genetic, epigenetic, neural, immune and endocrine mechanisms, and in guiding the development of more effective treatment approaches, as well as opening up avenues for prevention and early intervention in childhood.

The author confirms being the sole contributor of this work and has approved it for publication.