Trauma exposure is extremely common in countries affected by conflicts or natural disasters. In recent years, several studies have pointed to intergenerational trauma transmission (ITT) in countries affected by massive traumatic events (i.e., Burundi, Rwanda, Cambodia, Sierra Leone) and underlined the need to understand the mechanisms of transmission, in order to limit the potential negative impact on an entire community or even region, through generations (2–5).

Several studies in recent decades have tried to explain the transmission of parental trauma to the child. Many of these studies referred to collective historical trauma: the Holocaust, September 9/11, Vietnam Veterans, the Armenian genocide, etc., in which the impact of parents' post-traumatic stress disorder (PTSD) was measured in older children or in the second generation (6–10). Various other backgrounds have been explored in the process of ITT across generations, in different contexts of populations at risk for PTSD: low-income (11), abused mothers (12), adverse Childhood experiences (13) etc. Research has highlighted several possible models that may help explain the pathways of intergenerational trauma transmission.

Kellermann (14) proposed to distinguish the process of transmission, and what is transmitted. He described four models: psychodynamic, sociocultural, family system, and biological. Since this paper, two directions have been especially explored. Biological models have tried to understand the intergenerational transmission of stress (15) and the epigenetic mechanisms of stress and trauma transmission (16). Another set of studies concerns how the infants' sociocultural and family environment, the parents' representations, behavior and patterns of communication with their infant, can be considered as important factors involved in the process of ITT, a mechanism that appears to be a complex interaction between several co-processes [Perinatal Interactional Model of ITT, (17)]. Parental PTSD is characterized by physiological and emotional dysregulation, trauma related beliefs, and avoidance/withdrawal. This impacts fetal programming, and has repercussions on parent-child interaction through social learning, and maladaptive parenting. These phenomena decrease the child's regulatory capacity, increase child distress, and can lead to poorer child mental health, which in turn participates in a vicious circle.

The attachment theory (18) offers a theoretical framework to explain how caregiving behavior is impacted by parents' traumatic experience and can directly affect the child. Parents with unresolved trauma experience dissociation phenomena. They show frightened emotional expressions to the child who is unable to make the causal link between the terror of his/her parents and the trauma. The child responds with emotional and behavioral disorganization (19) and disorganized attachment behavior (20).

Maternal cognitions and representations also impact caregiving behavior, through reflective functioning. Schechter et al. (21) found that greater maternal PTSD severity was associated with unrealistic expectations of the child and distorted attributions of child intentions. These distorted maternal mental representations of the child (22, 23) will increase the frightening behaviors in response to the child, who consequently will react with disorganized behaviors. Child distress and reactions can reactivate traumatic memories in the mother, leading to a vicious circle. This can lead to specific offspring attachment orientations, in particular to higher attachment insecurities, either anxiety or avoidance, in association to self-amplifying cycle of PTSD and attachment insecurities of parents following trauma exposure (24).

Emotional availability (25) focuses more specifically on the emotional signaling between mother and child as a major mediator of trauma transmission. High scores indicate adequate maternal emotion regulation, and the open expression of negative emotions results in lower scoring. It has been proposed as a process involved in ITT, in spite of the fact that it was not found as a mediator between maternal trauma and infant psychosocial functioning in a sample of refugee mothers (26).

The PTSD mechanism was confirmed in a recent longitudinal study of ITT, which suggested a strong effect of maternal PTSD on the attachment relationship and consequently on child development; however, the need to identify the mechanisms by which maternal PTSD has an impact on the mother-infant relationship was underlined (27). The interpersonal neurobiological model of attachment and relational trauma developed by Schore (28, 29) can help to explain maternal PTSD transmission to the infant through a mechanism involving the rhythmic pattern behind mother-infant interactions (visual, tactile, vocal), which are negatively affected when mothers are traumatized. Mothers cannot modulate their stimulation and their response to their infant's needs, leading to overstimulation or on the contrary, to neglect of the child. Both behaviors can be traumatic for the infant who will have difficulty self-regulating his affective state to protect himself from becoming overwhelmed by hyperstimulation or the absence of stimulation from the mother. In a refugee sample, mothers' intrusion and avoidance explained individual differences in extremely insensitive parenting, which had direct negative effects on children's attachment organization (26).

In line with this mechanism, through the concept of “Affective Attunement,” Stern (30) proposed a pathway of transmission based on the sharing of emotional states between mother and child by means of inter-modal exchanges. D. Stern uses the term “intermodal” when in the dyadic exchanges an affect is expressed according to one modality (vocal, visual, bodily) then is “translated” and associated with another modality.

Affective attunement is a process centered on the mother's intermodal transformation of the baby's affective state. The mother tends to complement the baby's actions with gestures, gaze, and vocalizations, while the infant matches the mother's gesture behavior and perceives this as an affirmation of the continuity of his own affective state. The mother's transformation of the infant's affective state through behavior emphasizes the infant's recognition of his internal state. Traumatized mothers can transmit their negative affective states through their behavior, and these behaviors can modify the process of infant internalization of mothers' intentions.

More recently, a paradigm shift has been proposed for ITT in complex settings (war, migration). Intermediary dimensions have been introduced between parental PTSD and their effects on parent-child relationships. When parents were less able to draw on secure attachment representations, symptoms of PTSD increased the risk of insensitive parenting in a sample of Dutch asylum seekers and refugee parents (31). It is difficult to isolate the contribution of mothers' Post-Traumatic Stress Disorder (PTSD) in ITT, as in most cases, similarities with depressed or anxious mothers are frequent.

Studies of the direct mechanisms behind the transmission of parental traumatic experience to young children through caregiving have been undertaken recently, but are still rare (6, 20, 23, 26, 27, 32–35).

To investigate trauma transmission in a population at high risk of trauma exposure, we chose population from Central African Republic (CAR), which has been affected by collective violence since 2013, forcing the population to live in internal displaced camps or to migrate to refugee camps in border countries such as Chad or Cameroon. The current study addresses the issue of psychological trauma of the mother, which occurred in her life before the child was born. The objective was to identify the direct specific processes of mother-to-infant trauma transmission in traumatized mothers in humanitarian contexts, through the comparison of inter-modal mother-child interactions, occurring during different paradigmatic moments: moments of trauma recall (possibly leading to dissociation), compared to neutral moments, using a microanalysis method. We hypothesized that mothers affected by maternal post-traumatic stress symptoms would show lower levels of availability to respond to their infants' needs and show poor or inappropriate mother-infant interactions. When faced with the traumatic experiences or their recall (36), mothers may show specific behaviors (absent gaze, disconnection from reality, etc. as manifestations of dissociative states), which will result in a lack of attunement in interactions: non-synchrony (temporal coordination) or non-qualitative contingency (lack of inter-modality) between mother and infant behaviors. We hypothesized that during the narration of their traumatic experience in presence of the child, mothers may experience episodes of dissociation, that impact the infant's reactions and the interactions between mother and infant. Observing this specific moment in comparison with a neutral moment could highlight the specific mechanisms involved in trauma transmission.

The purpose of this study was to shed light on the process of mother to child trauma transmission in mothers who had experienced severe traumas in humanitarian contexts. As far as we know, this is the only study to investigate the quality of the dyadic interaction, by observing and conducting a microanalysis of the mother-infant interactions in a trauma reactivation situation. The mother-child interaction during a trauma recall situation was compared with mother-child interaction during a more neutral interview.

We found three main results, confirming our hypothesis.

The first result was that the mothers' behavior toward the child was specifically affected during trauma recall narration compared to the two more neutral moments of interaction. During the “traumatic narration,” mothers appeared to be emotionally affected by memories: they touched the infant less than in the session before, they looked at the infant less than in the session after, they looked more absent and smiled less than “before” and “after” moments. Moreover, they looked at the interviewer less during the traumatic narration than before and after. This is in favor of the relevance of our setting, which tried to catch these dissociative moments. When they were faced with the trauma recall, the mothers' absent look suggested that they were re-experiencing the event and were unable to focus on the present situation and on the infant's requests. They showed a behavioral pattern linked to the traumatic experience: they withdrew from interaction, smiled, and looked less at others, and seemed to be absent. These signs may reflect the maternal traumatic dissociation the infants face when their mother is lost in the traumatic experience. The mothers seem cut off from the present experience without being able to perceive external reality and the needs expressed by their child (48). This suggests that the mothers had difficulty in properly assessing the verbal and non-verbal expressions of infant arousal because they were overwhelmed by their own emotional state. This “absence of an appropriate response” emphasizes the transmission of the mothers' negative emotional state to the infants, who can potentially internalize this affective state as their own (30).

The majority of children in this study had no access to the direct meaning of their mother's words as they still were in a pre-verbal developmental stage, and even for those aged between 12 and 30 months, their mothers' words could not attack their symbolic world since they were not yet able to organize Internal Working Models, due to their developmental stage (49). Consequently, what is traumatic for them is more related to the absence, the lack of response to their requests and queries, which can potentially create external and internal sensory chaos (50).

The second finding is that the infants' behavior was impacted by their mothers' traumatic experience. During the traumatic narration, infants looked at the interviewer less than in the session before, and they sucked the breast more than before. Moreover, the mother-child interaction was also affected: “Infant self-stimulation—Mother looks absent” and “Infant sucks the breast—Mother looks absent” occurred more during the traumatic narration than in the other two sessions. These interactive behaviors reflect the children's coping strategies of self-regulation: they perceive these moments, and try to compensate with self-stimulation or by clinging to the satisfaction of primary needs. To suck the breast is a behavior which maintains close contact with the mother, and offers a primary satisfaction that the child does not find in the interaction. The lack of correlation between the age of the child and suckling the breast suggests that in the case of older infants (more than 12 months), this request was not simply linked to the need to be fed, but to a possible coping strategy for self-regulation in a stressful situation. These two coping strategies are efficient in providing self-comfort and limiting the negative effect of distress (51). This is in line with studies by Tronick (52), who stated that infant self-stimulation is a coping behavior that reduces stressful situations and offers self-calming, in particular when mothers are “still faced” (53). The mothers' “absent look” is a kind of ecological “still face” situation. These interactive behaviors reflect the children's coping strategies of self-regulation: they perceive these moments, and try to compensate with self-touch or by clinging to the satisfaction of primary needs. A previous research has shown that during a still-face procedure, when mothers were unavailable, infants spent more time touching themselves, supporting the regulatory and exploratory roles of infant touch, especially during periods of maternal unavailability (54).

It should be noted that, according to Main and Solomon "Indices of Disorganization and Disorientation (Main and Solomon, unpublished manuscript), the children did not show disorganized behavior during their mothers' “absence,” but rather showed adaptive behaviors such as self-stimulation. Several explanations can be put forward for this. First, mothers had received psychological care from NGO psychosocial workers, which had possibly helped them to reduce their traumatic experience; our paradigm of trauma recall may perhaps not have reflected the fact that their trauma was reduced. Second, infant self-stimulation may be the first stage of coping; if the mother's “absence” is repeated, their adaptive strategies may prove insufficient to handle the mother's negative experience and may lead to internalized working models of disorganization. Third, the fact that the mothers talked very little to their child could limit the trauma transmission. Last, cultural differences might explain the way they cope with traumatic experiences. In traditional societies, children are not only raised by their mothers (55). They grow up surrounded by several caregivers, who complete the construction of the world and the child's identity. This aspect of shared mothering could be a protective factor in the transmission of trauma from mother to child. The presence of a co-mothering system could mitigate or compensate for the effects of the mother's lack of response to the child's requests when she is absorbed by traumatic intrusions.

The third finding was that there were no significant associations between interaction and infant gender and age, the severity of traumatic experience assessed by IES, and mothers' depression and anxiety symptoms assessed by HAD. This is maybe due to the small sample. Concerning age, it has however been shown (34) that the period when child locomotion develops might represent a critical time window for mother-child interaction in dyads with mothers having a history of abuse.

Not even the country of residence had an influence on the interactions between the mother and infant. The only difference was linked to the impact of the traumatic event and the dyads' country of residence. The highest level of PTSD symptoms was associated with those living in the Central African Republic (CAR). This can be explained by the fact that people living in CAR are in an ongoing extremely traumatogenic context, whereas people who migrated and are living in refugee camps are more protected against new traumatizing situations.

Despite the fact that we found a significant positive association between the severity of the impact of the traumatic event (IES-R) and the time that had elapsed since the traumatic events, neither the severity of the traumatic experience nor the severity of anxiety or depression were associated with interactive behaviors. This may indicate that among the war traumatized and refugees, the theoretical frame to understand ITT has to be enlarged. Mothers' general psychological distress, but not PTSD, was directly associated with negative parenting and child psychosocial difficulties in a sample of 291 Syrian refugee mothers in Lebanon who had been exposed to war trauma in the past (56). These results argue in favor of taking into account a broader framework for complex trauma in conflict and post-conflict areas, including the psychosocial framework, with the trauma focused approach. We did not evaluate the impact of other psychosocial conditions.

The main limitation of the study is the size of the sample. The recruitment in humanitarian contexts involves several challenges linked to security constraints and difficulties of access to subjects. A larger sample would be more representative of the observed mechanisms of transmission and make our results more significant. However, even with the small sample, our results point to particular lines of enquiry to pursue in future studies. The cultural particularities of mother and child relationships require investigation to design an appropriate framework for analysis as a key to understanding the affection linked to behaviors and interactions. The second main limitation is the absence of a control group. But our aim was to explore how trauma recall can affect mothers' presence to the child and the way they interact with the child, and not to globally qualify the interactive mother-child process. That is why we compared each dyad in three different settings, each dyad being its own control. A control group could have allowed us to compare mother-child interaction in play sessions. But by definition, it is not relevant to explore trauma recall with mothers who have not experienced traumatization. Another limitation is the context in which interaction took place, which was set up to catch the direct effects of trauma reactivation. The mothers were probably more involved in the interview than in playing with the child; however, this setting was close to an ecologic situation of daily life. Future studies should include longitudinal observation of mother and child dyads, with control groups, to identify the long-term impact of mothers' inappropriate response impacted by dissociation, when trauma is recalled, to their infants' needs and child development.

This research is unique in its approach to mother-infant interactions: in the field explored (ITT in humanitarian contexts), by the setting (trauma recall), and by the microanalysis of interaction. Our results shed light on the process of trauma transmission, focusing on the caregiver reactions during trauma recall, and suggest that trauma affects a mother's availability to interact with her infant and to regulate his/her state of emotional arousal. The infant experiences this sudden lack of proper responses from the mother when she faces events or thoughts that trigger trauma memories, which can have an impact on the infant's perception of his/her own emotional status. A new and surprising finding was that children did not react by disorganization as has been previously described, but with self-stimulation, representing possible coping strategies in reaction to the mother's enigmatic reaction. This adaptive behavior may compensate for the mothers' psychic absence and for the lack of “attunement” between mother and child.

Our results strongly reflect the increasing need to understand trauma transmission mechanisms in highrisk populations, particularly when they live in unremitting traumatogenic conditions.

In particular, the question about self-stimulation during mother's “absence” should be considered in further longitudinal studies as this initial adaptive mechanism could lead to internalized working model of disorganization, if repeated.

To prevent or to limit the impact of maternal trauma transmission, we recommend psychological interventions with mothers, starting from their pregnancy, and continuing after childbirth during the important period for the development of the child, until 2 years. Psychological support for at-risk mothers should aim to allow them to resolve the effects of the trauma in order to limit the transmission of trauma in interactions with the child. Interventions focusing on parental skills, their sensitivity, their ability to decipher, and respond to the specific needs of their child like video feed back interventions (57) are also recommended. But psychological support, which has been proposed for mothers with children born of sexual violence (58), has to go beyond trauma focused interventions, and provide a holistic and community embedded approach that can be applied across settings.

The datasets generated for this study are available on request to the corresponding author.

The studies involving human participants were reviewed and approved by the institutional ethical review boards in CAR and in Cameroon, and by the Ministère de la Femme, de l'Action Sociale et de la Solidarité Nationale in Chad. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

EDo was the principal investigator of this research and she was supervised by MF, MMo, CB, and LO during data collection and analysis. EDo, MMa, MMo, EDr, ML, and LO all contributed to the definition of the research methodology and contributed to the data analysis and interpretation. All authors participated in the writing and revision of the article.

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.