The default mode network (DMN) is a cortical network that becomes more active and synchronized during rest and deactivates when one is engaged in a goal-oriented task (1). The DMN comprises seven brain regions, including the left and right angular gyrus, ventro- and dorso-medial pre-frontal cortex, posterior cingulate cortex/pre-cuneus, right medial pre-frontal cortex, and left inferior temporal gyrus (2), and its organization and functioning during rest is considered to define the personal “fingerprint” of the brain (3). In adults, the DMN is thought to play a key role in sustaining the sense of self and in organizing the individual's experience of the world; it supports self-referential thinking, introspection (4), internal narratives (4), mentalizing activity (5), the autobiographical self and self-projection (6). Highly synchronized and smooth activity of the DMN is critical for a sense of well-being, personal identity, and resilience (7–11).

The importance of the DMN for proper brain functioning is widely-accepted. The DMN is involved in overall brain activity as well as in attentive, cognitive, and social processes, and associations have been repeatedly shown between DMN dysfunction and a host of psychiatric disorders in adults (12–16). In contrast, very little is currently known about the development and functional connectivity of the DMN in children and particularly lacking are data related to DMN maturation under high-risk conditions, with even more limited knowledge on how structures of the DMN cohere in children reared under various high-risk conditions. A study utilizing resting state functional connectivity MRI and graph theory methods to characterize the maturation of the DMN found that in contrast to the strong interhemispheric connectivity between homolog brain regions in adult, DMN connectivity in children is sparse and becomes more integrated with development (17). The authors interpret their findings in relation to the ever-maturing functions sustained by the DMN, such as consolidation of the sense of self that is still immature in children. For example, the ability to encode and retrieve stored memories is fully functioning at an early age but episodic memory continues to improve with age. This improvement does not stem from the mere development of encoding and retrieval but from the ability to incorporate more complex strategies to encode self-relevant information, which is supported by the progressive integration of the DMN into a unified network. These notions are supported by a study indicating that DMN connectivity in children and adolescents is related to the quality of reconstruction of past experiences and the ability to imagine the future (18).

Very few studies tested the maturation of DMN connectivity across childhood and adolescence. A longitudinal fMRI study found significant changes in DMN connectivity between the ages 10 and 13, suggesting that it continues to consolidate throughout development (19). This is consistent with cross-sectional studies that compared DMN connectivity in children, adolescents, and adults (17, 20–23). Notably, the findings that age-related differences are observed even within the narrow age-range of 10–13 lends further support to the hypothesis that early adolescence may be a particularly important time-window for the integration of the DMN into a coherent network.

While very little research has tested DMN connectivity using magnetoencephalography (MEG), MEG has unique benefits for the study of DMN functionality due to its high temporal resolution which enables the decomposing of neural signals into frequency bands (24). Using MEG/EEG studies it was found that whereas in adults, the strongest activity in the DMN during rest is in the alpha band (8–12 Hz), during childhood the dominant frequency is the theta band (4–7 Hz), which gradually shifts to alpha as children grow (25–28). In a previous longitudinal study we utilized MEG to investigate the effect of early life stress (ELS) on DMN connectivity in preadolescents and their mothers and found that ELS decreased theta DMN connectivity in adolescents and alpha DMN connectivity in mothers (29); hence, we focused here on theta connectivity in youth by assessing Phase Locking Value [PLV; Lachaux et al. (30)], the most commonly used index to measure phase coupling in EEG/MEG experiments (31, 32). PLV quantifies the degree to which signals rise and fall together over time in absolute values; high PLV (values closer to 1) implies more synchronized signals while low PLV (values closer to 0) denotes lower synchrony. We found that the experience of anxious, intrusive parenting across the first decade and higher cortisol levels predicted DMN connectivity impairment, indicating that neurobiological, clinical, and relational factors that index stress longitudinally shape DMN connectivity in children.

The neuropeptide oxytocin (OT) has been repeatedly implicated in attachment, bonding, and sociality across mammalian species (33, 34). In addition to its involvement in bond formation, OT plays a critical role in stress regulation. In response to emotional, physical, or pharmacological stress, OT is released into the bloodstream as well as within hypothalamic and extrahypothalamic limbic regions where it regulates HPA activity and modulates response to stress (35–37). Since the DMN is highly sensitive to stress, its functionality is probably impacted by the OT system and evidence points to the association between OT and DMN activity in adults (38–40). One study describes associations between allelic variability on the oxytocin receptor gene (OXTR) and DMN connectivity (40). Following research indicating that 40% of the variance in DMN connectivity is heritable, Wang et al. (40) found that compared with individuals with OXTRrs2254298 GG genotype, individuals carrying the AA/AG genotypes displayed lower DMN connectivity.

To further understand the association between genetic variability on the OXTR and stress-related conditions, studies have suggested that a cumulative genetic risk index—a measure combining multiple single nucleotide polymorphisms (SNPs) on oxytocin-pathway genes linked with social dysfunction—provides a better risk estimate than the study of individual SNPs (41–44). Three OXTR SNPs have been typically included in this cumulative index, the rs2254298, rs53576, and rs1042778 alleles. In addition to these three alleles, CD38, a multifactorial molecule implicated in cell proliferation, differentiation, and migration that plays a critical role in regulating OT release (45) and vasopressin AVPR1a, a closely linked neuropeptide to OT that is considered to evolve from the same ancient vasotocin molecule (46), have been incorporated into the oxytocin-pathway genes cumulative index. This cumulative risk index, combining the three OXTR SNPs, CD38, and AVPR1a, has been shown to predict PTSD chronicity in children exposed to chronic trauma (47) and to correlate with empathic failure among new romantic partners (48). Furthermore, the cumulative risk approach is more statistically powerful and enables the evaluation of a continuum of genetic risk in relation to stress-related conditions.

In light of the above, the current study utilized a unique cohort of stress-exposed children living in Sderot, Israel, a small town located on the Gaza boarder which has been subjected to continuous, life-threatening trauma for nearly two decades. Sderot has been the target of continuous rocket attacks since 2001, with six periods of exacerbations when the town was under attack by dozens of daily missiles for several days to several weeks. During attacks, citizens hear a siren warning, leaving them only 15 s to reach shelters before an explosion. Children were followed at several time-points from early childhood and at the transition to adolescence (11–13 years) underwent MEG scanning to assess DMN connectivity. Our goal was to examine the cumulative contributions of genetic susceptibility on the OT system and anxiety-related symptoms, which are known to jointly impact neural responses (49, 50), to the development of DMN connectivity in children growing in the context of chronic trauma. We thus examined stress-related factors across childhood and hypothesized that cumulatively these factors may place children at risk for maturation of the DMN. Specially, we considered genetic variability on the OT system, avoidant manifestations in early childhood, and the consolidation of a full-blown anxiety disorder in late childhood as factors which, when occurring together across development, may tip children toward a risk trajectory. Our conceptual model is presented in Figure 1.

In early childhood, we focused on children's avoidant responses to trauma reminders. Avoidance is a core symptom cluster in the diagnosis of PTSD and has shown to predict PTSD chronicity in children and adults (47, 51). The diagnosis of PTSD in early childhood, before children can verbally recount the trauma, requires clinical observation of the child's behavioral response to trauma reminders [DC: 0–3R; Zero to Three (52)]. Young children's behavioral response to trauma reminders is characterized by four types of responses; regressive-clinging, aggressive, hyper-anxious, and avoidant (53), of which the avoidant type is the most detrimental for later development (47). In terms of anxiety disorders, longitudinal studies has shown that temperamental avoidance and reticence in early childhood increase the risk for the consolidation of anxiety disorders by late childhood and adolescence, and only those children who showed both reticence in early childhood and anxiety disorders later on exhibited alterations in neural functioning (54–57). We thus examined the joint contribution of early behavioral avoidance and later anxiety disorders to DMN connectivity at the transition to adolescence.

Several hypotheses were formulated. First, we expected that greater cumulative risk index on oxytocin-pathway genes would be associated with reduced connectivity of the DMN (hypothesis 1). We also expected that avoidant response to trauma reminders in early childhood would be linked with impaired DMN connectivity (hypothesis 2). Similarly, the existence of a full-blown anxiety disorder, which increases the likelihood of a chronic course of PTSD (51), was expected to predict lower DMN connectivity (hypothesis 3). Finally, these factors in combination—genetic risk, avoidant response, and anxiety disorder, were expected to jointly explain variance in DMN connectivity in preadolescence (hypothesis 4).

While the consolidation of the DMN into a unified network marks a developmental achievement critical for proper brain functioning [for a systematic review and meta-analysis see Mak et al. (85)], very little research examined DMN connectivity across childhood and adolescence, with nearly none testing its maturation in relation to stress-related symptoms and markers of risk and resilience. In the current study, we examined the association between genetic variability on the OT system and DMN connectivity at the transition to adolescence. Our goal was to expand knowledge on how ELS and psychological manifestations of stress undermine the DMN's ability to function as a coherent network whereas better innate functionality of the OT system may facilitates DMN maturation. Utilizing a prospective longitudinal study and following children across the first decade of life, we found that exposure to ELS, cumulative genetic risk on oxytocin-pathway genes, avoidant behavioral response to trauma reminders in early childhood, and the diagnosis of an anxiety disorder in late childhood cumulatively explained 25% of the variance in DMN connectivity in preadolescence. Overall, our findings point in four important directions. First, greater inborn functionality of the OT system sustains maturation of the DMN and may serve to buffer the effects of chronic early stress on DMN connectivity. Second, young children's avoidant response to trauma reminders in the context of chronic trauma serves as an important early indicator of difficulties in neural maturation a decade later. Third, the existence of a well-defined anxiety disorder bears negative consequences for the development of the DMN, a core system sustaining the emerging self. Finally, independent of ELS exposure, these factors cumulatively predict DMN connectivity in preadolescence and when appearing together across childhood may chart a risk trajectory for the maturation of the DMN.

Our findings are the first to demonstrate how the interplay of genetic factors and stress-related symptoms as they mutually influence each other over time shapes DMN connectivity at the transition to adolescence. The DMN is known to be involved in processes implicated in stress regulation, such as hypervigilance and alertness, alterations of intrinsic and extrinsic attention, self-referential mental activity, and recollection of personal experiences (1, 86). Studies in adults have shown that following stress, DMN activation is both reduced (87) and less integrated (88), and various ELS conditions, such as a history of poverty reduces DMN connectivity (88). Much further longitudinal research is thus required to tease various forms of ELS and to chart specific developmental trajectories for each condition that may impair the functioning of core neural networks.

DMN connectivity has been repeatedly associated with PTSD symptoms and anxiety disorders in adults. PTSD has been linked with reduced connectivity of the DMN (13, 14, 89) and a history of chronic early trauma has been associated with reduced DMN connectivity in adulthood, with the degree of impairment increasing as a function of the severity of dissociative symptoms (89). Furthermore, mindfulness-based psychotherapy helped increase DMN connectivity in veterans suffering from war-related PTSD and the improvement correlated specifically with reduction in avoidant symptoms (13), consistent with our findings on the specific associations between DMN connectivity and avoidant symptoms. Similarly, PTSD patients exhibited shorter activations of the DMN after exposure to trauma reminders that correlated with the severity of symptoms (90). Trauma-focused cognitive behavior therapy rescued DMN activity duration of activation and brought it back to normality and other stress-related psychopathologies linked with alterations in DMN activation patterns in adults (91, 92). Our findings highlight avoidant response to trauma reminders in early childhood, which is associated with PTSD chronicity in childhood (47), as an early marker of disrupted neural maturation and underscores the need to identify young children with avoidant symptoms and avoidant behavior in contexts of chronic stress and trauma as those who require the most rapid and targeted interventions.

In addition to PTSD, anxiety and stress-related disorders in adults have been linked with disruptions to DMN functionality (91, 92) and our findings are the first to show similar associations in children. Comorbidity between depression and anxiety was associated with reduced connectivity in the anterior regions of the DMN and increased connectivity in the posterior regions. During task, when DMN activity is typically suppressed, psychosomatic patients with anxiety disorders showed reduced activation in selected DMN nodes (92). In adults, disruptions to DMN alpha connectivity were related to anxious and depressive symptoms (29). The current findings indicate that when an anxiety disorder is consolidated by late childhood, this poses a risk for the development of DMN connectivity during the 10–13-years-old period described by Sherman et al. (19) as a sensitive window for the maturation of the DMN.

The current study expands our previous knowledge by linking genetic variability on the extended OT system with functionality of the DMN in children in the context of chronic early life stress. Based on the literature linking OT to stress-related psychopathologies, such as PTSD and anxiety (70), the OT system has been considered as a potential route for prevention and intervention in stress-related disorders (93). Our findings show that cumulative genetic risk on the extended OT system and avoidant symptoms in early childhood were each uniquely predictive of disruptions to DMN connectivity in early adolescence, above and beyond stress exposure, adding to the discussion on the role of the OT system in brain maturation in the context of chronic early stress. These findings suggest that there may be non-circumstantial factors that pose a risk to functionality of the DMN. Even among children reared in low-risk contexts, those born with more risky dispositions and develop avoidant responses that mature into a full-blown anxiety disorder can suffer a protracted maturation of the DMN. We have previously shown that in these children greater salivary oxytocin was associated with better immune biomarkers and both oxytocin and immune-system functionality buffered the effect of trauma on the development of anxiety disorders (94). This suggests that multiple indices of OT-system functionality may serve as indices of resilience and buffer the effects of stress on outcome in children exposed to chronic stress, trauma, and adversity.

Our findings have several clinical implications for children growing up amidst chronically stressed conditions. First, boosting the child's OT system, whether through intranasal administration or via bolstering parent-child synchrony, parental touch, and physical contact which have shown to increase OT production (95, 96), may provide a buffer against stress-related psychopathologies. Thus, early interventions aimed at increasing child-parent synchronization and physical touch or nasally administrating OT may serve as protective factors for brain development. Second, interventions focusing on reducing avoidant behavior in response to ELS may protect children from the long-lasting effects of ELS on the DMN. Indeed, studies show that therapeutic strategies, whether cognitive or emotional focusing on reducing avoidant behavior are highly successful in reducing the severity of GAD in adults (97) and our findings suggest that the same mechanisms may operate in children.

Several study limitations should be taken into consideration in the interpretation of the findings. First, consistent with much prior research we treated the DMN as a unitary system; however, several recent studies have demonstrated that the DMN comprises several interrelated subsystems (1, 98–100). Similarly, we treated the cumulative genetic risk system as a unitary factor in which each gene received equal weight and it is possible that one or more genes are more important in the current context. Future research should address this possibility by statistically determining the importance of each gene across subjects toward the computation of a more nuanced cumulative variable. In addition, we did not target other neural systems associated with the DMN or OT. Further, there are external factors that predict DMN connectivity in children that were not included in current study. For example, maternal depression, anxiety, and post-traumatic symptoms, which we previously found to be linked with DMN maturation (29). Thus, in order to draw a more comprehensive picture, future research should take into account, in addition to child's characteristics, also the child's surrounding envelope. Finally, some of the findings are only marginally significant, such as the difference in DMN connectivity between children high and low in avoidant behavior and children high and low in anxiety disorder, and these findings should be treated with caution. Nevertheless, it is important to note that while these variables were not significant when tested independently, taken together, they had a significant cumulative effect and when appearing on top of each other across development were found to impair the development of the DMN.

Despite the above limitations, the current study expends previous knowledge by demonstrating the long-lasting effect of ELS on the developing brain. This is the first study to longitudinally test the cumulative impact of genetic, behavioral, and clinical factors as they unfold across development on maturation of a core neural system. Our findings highlight the role of the oxytocin system for brain maturation during a sensitive period of the transition to adolescence and may have important implication for both research on the developing brain and the construction of targeted interventions for children exposed to chronic stress, war, and trauma.

The datasets analyzed in this article are not publicly available. Request to access the datasets should be directed to feldman.ruth@gmail.com.

The studies involving human participants were reviewed and approved by Ethics Committee of Bar Ilan University. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

RF designed the longitudinal study and wrote the paper. JL designed and conducted the MEG experiment. MZ-W analyzed the MEG data and wrote the paper. RE conducted the genetic analysis.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.