Childhood adversity is one of the main risk factors for affective disorders and related emotion processing biases. These biases usually reflect atypical and maladaptive processing of emotional information that can result in e.g., heightened threat vigilance, altered attention to emotional stimuli, impaired emotion recognition, and emotion dysregulation (e.g., Bodenschatz et al.,
2019; Briggs-Gowan et al.,
2015; Harms et al.,
2019; Lakshman et al.,
2020; McCoy et al.,
2016; Nelson et al.,
2020; Pollak,
2015a). Such atypicalities in emotion processing might arise from early experiences of threatening and unstable environments and contribute to the development of poor cognition and psychopathology (e.g., Pollak,
2015a). Caregiver trauma and mental health problems are also risk factors for poor child psychosocial functioning (e.g., Mirzaaghasi et al.,
2014). As a result of the growing number of conflicts and security threats in the world, there are currently 110 million forcibly displaced people worldwide, over 43 million of whom are children (UNHCR,
2023). Displaced adults and children suffer the consequences of war trauma exposure and face post-displacement difficulties, which often result in increased risk of mental illness (Cratsley et al.,
2021; McEwen et al.,
2023). Notwithstanding the vast number of children affected by armed conflict, the consequences of these experiences on children’s wellbeing and emotion processing are still poorly understood.
Refugee children face a plethora of displacement-related stressors which place them at risk for emotional and behavioural problems, disruptions in emotion regulation and recognition, enhanced attention to threat, and overall poorer emotion processing (Burkhouse & Kujawa,
2023; Durà-Vilà et al.,
2012; Gredebäck et al.,
2021a; Hodes & Vostanis,
2018; Khamis,
2019; Michalek et al.,
2022; Scherer et al.,
2020; Yayan et al.,
2020, but see Michalek et al.,
2023). Whilst these socio-emotional impairments can occur as a direct result of war exposure, displacement, and continuing adversity, parental trauma and mental health problems are also likely to influence refugee children’s development (Miller & Rasmussen,
2017). Indeed, it has been shown that caregiver trauma and psychopathology affect children’s emotional and behavioural outcomes in non-refugee populations (Clavarino et al.,
2010; Goodman et al.,
2011; Lambert et al.,
2014; Mirzaaghasi et al.,
2014; Morris et al.,
2012), and emerging research suggests similar effects in refugee families. Studies in post-conflict areas, war zones, and refugee settings highlight strong associations between child and caregivers’ distress, PTSD symptoms, and internalising and externalising problems (Betancourt et al.,
2012,
2015; Field et al.,
2013; McEwen et al.,
2023; Meyer et al.,
2017; Thabet et al.,
2009). For example, Syrian refugee parents’ poorer mental health predicted their children’s emotional and behavioural problems (Eruyar et al.,
2018), whilst parental war trauma exposure has been linked to child conduct problems and hyperactivity in refugee families (Bryant et al.,
2018; Eruyar et al.,
2018). Interestingly, parenting styles and parental displays of anger (often resulting from their own experiences of trauma and PTSD) also play a role in this context (Hinton et al.,
2009; Sim et al.,
2018; Thabet et al.,
2009). For instance, maternal PTSD predicted children’s poorer identification of emotional expressions amongst Syrian refugees, with children’s impaired emotion recognition linked to harsher parenting (Gredebäck et al.,
2021a; Peltonen et al.,
2022). A recent meta-analysis reported that parental war-related trauma was linked to harsher parenting styles, which in turn mediated the association between parental trauma and child adjustment, including emotional symptoms, social problems, and quality of life (Eltanamly et al.,
2021). On the other hand, higher levels of family acceptance and lower levels of community stigma were linked to fewer internalising problems among adolescents living in post-conflict Sierra Leone (Betancourt et al.,
2015). Similarly, Syrian children and adolescents, living in Jordan and Lebanon, who reported higher family cohesion and lower levels of family conflict also reported fewer internalising and externalising problems, indicating that positive family environment is linked to higher mental wellbeing in refugee youth (Khamis,
2021). Few studies so far have focused on emotional processing biases in refugee children (e.g., Gredebäck et al.,
2021b; Michalek et al.,
2022,
2023) but research with non-refugee children suggests that such biases might be instrumental in linking parental and child mental health. Taken together, parental trauma, mental illness, poor parenting strategies, and unstable family environments may jointly contribute to impairments in emotional, cognitive, and behavioural outcomes in refugee children, although little is known about the mechanisms underlying these effects.
According to integrative models of familial risk for psychopathology, cognitive patterns exhibited by depressed or anxious parents (e.g., enhanced threat vigilance, negative affect, interpreting ambiguous situations as negative) might be a mechanism of intergenerational transmission of anxiety and depression (Bögels & Brechman-Toussaint,
2006; Goodman et al.,
2011; Goodman & Gotlib,
1999; Hadwin & Field,
2010). Caregivers’ displays of atypical emotion processing could result in their children developing similar emotion processing biases, and in turn, symptoms of affective disorders. For instance, maternal anxiety levels have been linked to heightened attention to threat in infants and children (Aktar et al.,
2019; Morales et al.,
2017), and children of depressed mothers were found to misattribute sadness to other emotions (Kluczniok et al.,
2016), display an attention bias to sad stimuli (Owens et al.,
2016), perceive more negative affect in maternal emotional states (Luebbe et al.,
2013), and display impaired emotion recognition overall (Priel et al.,
2020). The link between caregiver-child emotion processing, however, is less clear, with some studies suggesting that maternal biases are predictive of their children’s biases (de Lijster et al.,
2020; Waters et al.,
2015) and others finding no relationship (Aktar et al.,
2019; Platt et al.,
2021), highlighting the complexity of the cognitive mechanisms involved in familial psychopathology transmission. Taken together, it is possible that the cognitive biases displayed by caregivers, combined with changes in family dynamics and negative parenting strategies which show associations with maladaptive emotion regulation, might exacerbate mental health symptoms, behavioural problems, and emotion processing disturbances in their children.
Discussion
In our study, we examined refugee mother-child mental health and whether attention biases towards emotional facial expressions in mother-child dyads were associated with each other or with psychopathology symptoms. As expected, we found that mothers’ trauma and mental health were related to their child’s mental health. We also found that refugee mothers and their children displayed threat hypervigilance, although we found no correlation between mother and child attention biases, and attention biases were also unrelated to maternal trauma exposure, or maternal or child mental health.
Our findings add to the growing body of research highlighting the importance of maternal trauma exposure and mental health on their child’s development in a refugee context (e.g., Bryant et al.,
2018; Eruyar et al.,
2018; McEwen et al.,
2023). Overall, mothers reported high levels of trauma exposure, PTSD, and depression symptoms, indicative of potential heightened psychopathology risk (Radloff,
1977; Weatheres et al.,
2013), however their children’s internalising, externalising, and attention problems were suggestive of lower psychopathology risk in this sample (based on cut-off scores from previous studies in non-refugee samples, e.g., Jellinek et al.,
1988). Mental health difficulties and problematic behaviours amongst refugee children are reportedly linked to increased traumatic experiences (Karadag & Ogutlu,
2021; Yayan et al.,
2020), and to poorer parental mental health (Bryant et al.,
2018; Meyer et al.,
2017). We found here that maternal depression and anxiety symptoms were overall better predictors of their child’s mental health than maternal trauma and PTSD, with the latter only linked to child internalising and attention, but not externalising problems, contrary to previous reports (e.g., Bryant et al.,
2018). It is possible that externalising or conduct problems in refugee children could depend more on other factors previously reported for non-refugee populations, such as negative parenting styles (Schulz-Heik et al.,
2010; Viding et al.,
2009). It is worth noting that the relatively low scores on psychopathology symptoms for the refugee children in the current study might be unique to the studied population, as previous findings suggest high prevalence of symptoms of PTSD, depression, and anxiety among refugee children (Henley & Robinson,
2011; McEwen et al.,
2023; Panter-Brick et al.,
2018a). Our results might be linked to the young age of the participating children, and the fact that the majority of them were born in displacement and would not have had any direct exposure to war. However, these findings might also hint at potential protective factors related to the family environment and contextual resilience (Panter-Brick, Hadfield, Panter-Brick et al.,
2018a,
b).
Both mothers and their children displayed an attention bias towards angry faces only, consistent with previous reports linking adversity with attention biases towards threat (e.g., Briggs-Gowan et al.,
2015; Powers et al.,
2019; Roy et al.,
2008). The attention bias to anger reflected a heightened vigilance to threatening faces (albeit with small effect sizes), with no difficulties in disengaging from the stimulus. This is somewhat at odds with an earlier experiment we conducted with a separate group of Syrian refugee children, where we found that higher trauma in children was not linked to vigilance to threat (children showed initial avoidance of angry and happy expressions) but was related to an increased sustained attention to anger, suggesting difficulties disengaging from threat (Michalek et al.,
2022). However, with substantial methodological differences between the two experiments it is difficult to directly compare these results. Indeed, in our previous experiment we measured attention using eye-tracking only and children were shown 4 different emotional faces at a time with no task other than to simply view the faces (‘free viewing’). Our dot probe task here is also confined to the initial stages of attention allocation since stimulus presentation was very short (500ms), whereas in the previous experiment we measured children’s eye movements (scan paths) over 4000ms. It is likely that being required to perform a task (specifically to give a speeded response) and a limited viewing time tap into slightly different cognitive mechanisms than simply free viewing images for an extended period.
Our results are consistent with a growing body of work reporting hypervigilance to threat in children who are at risk for poorer mental health following early adversity (e.g., Briggs-Gowan et al.,
2015; Pollak et al.,
2001; Roy et al.,
2008), suggesting that heightened initial detection of threat might be an important mechanism in the formation of detrimental cognitive patterns (Harms et al.,
2019). Threat hypervigilance has been previously linked to adversity and psychopathology in non-refugee populations (Abend et al.,
2018; Bar-Haim et al.,
2007; Felmingham et al.,
2011, although see Lisk et al.,
2020) and could possibly reflect more general cognitive control difficulties (Cisler & Koster,
2010; Eysenck et al.,
2007) and potential emotion regulation problems. Attention biases in refugee children could point to maladaptive cognitive strategies, which might lead to development of internalising and externalising symptoms in later life (Harms et al.,
2019). Interventions or strategies which alter threat attention biases could represent a potential target to increase socio-emotional processing abilities and improve mental wellbeing in refugee children.
Surprisingly, we found that the attention biases to emotional expressions were largely unaffected by maternal trauma and mother/child psychopathology symptoms. These results differ from the previously well-established links between adversity, mental health, and attention biases, both in children and adults (e.g., Bodenschatz et al.,
2019; Cisler & Koster,
2010; Hadwin et al.,
2003; Pollak,
2015b; Reid et al.,
2006), although recent meta-analyses and reviews suggest some inconsistencies (e.g., see Kruijt et al.,
2019; Lisk et al.,
2020). Threat hypervigilance might be reflective of the chronic stress of displacement and general life difficulties related to the refugee experience. Although it is surprising that this hypervigilance bias was not linked to poorer mental health here, the overall higher level of wellbeing and mental health of children in the current study might explain this lack of association.
We also did not find the expected associations between child attention biases and their mothers’ mental health. Previous studies of non-refugee populations show that children of depressed mothers usually display a biased attention to dysphoric stimuli (e.g., sad expressions), and children of anxious mothers display an enhanced attention to threat (Burkhouse et al.,
2015; Kujawa et al.,
2011; Morales et al.,
2017; Owens et al.,
2016). Interestingly, Gibb et al. (
2022) found that maternal depression moderated children’s attention to sad faces, where younger children of depressed mothers displayed attentional avoidance while older children in this group displayed enhanced attention to sad stimuli. This hints at important developmental differences in the direction of attention biases in children at risk for depression and highlights potential developmental trajectory of emotion regulation atypicalities. In refugee populations, maternal PTSD symptoms have been previously linked to lower recognition accuracy of sad and happy emotions in their children (Gredebäck et al.,
2021a). Yet, we found no evidence that affective attention biases were associated with the transmission of psychopathology risk, although the non-clinical nature of the mental health measures in our study, and a lack of control (non-refugee) group make comparisons difficult.
It is likely that trauma experienced (and reported) by the child affects the child’s emotional development more than trauma experienced by the mother. Children in the current study may be unaware of their mother’s traumatic experiences, and they are unlikely to have directly experienced war-related trauma themselves, as most children were born in displacement (82%). Considering the young age of the children, the length of displacement, and the current living conditions (mostly urban neighbourhoods) of the refugee families in our study, it is likely that the participants experienced other types of adversity not captured by our trauma measure, such as food insecurity, debt, unemployment, limited service provision, and lack of social support (e.g., Hall,
2022), and these post-displacement difficulties might have greater influence on children’s emotional processing than mothers’ trauma. Furthermore, maternal trauma and mental health itself might be less predictive of child facial emotion processing, with other family factors - such as parenting strategies and social support - playing a more important role, especially in humanitarian context (Peltonen et al.,
2022). Caregiver mental health might influence refugee children’s development through parenting styles and parent-child relationship (e.g., Bryant et al.,
2018; Eltanamly et al.,
2021), rather than through cognitive biases. Programmes targeting positive parenting strategies, increasing family cohesion, and improving parent-child relationship through parenting training might be particularly beneficial for refugee children’s mental health and emotional development (Bosqui et al.,
2022; Bryant et al.,
2018; Eruyar et al.,
2018; Khamis,
2021).
We found no relationship between mother and child attention biases for either angry or sad facial expressions. With few studies investigating these associations, and the variability of findings in the literature, our results are contrary to some reports (de Lijster et al.,
2020; Waters et al.,
2015) and in line with others (Aktar et al.,
2019; Platt et al.,
2021). It is possible that the higher variability in the distribution of children’s biases (average
SD = 0.14) as compared to the biases of their mothers (average
SD = 0.04) might blur the association. Whilst research in non-refugee populations largely supports the importance of emotion processing biases in the transmission of psychopathology risk within families (e.g., Kluczniok et al.,
2016), our results suggest that other aspects of cognition and behaviour could play a more important role in this transmission in humanitarian settings. Including measures of mother’s
expressions of anger and sadness in everyday life, rather than only their
perception of or
attention to these emotions might help clarify potential impact of their affective processing on emotion biases in their children, as mothers’ expressive styles and displays of facial affect have been shown to influence children’s emotion recognition and regulation in non-refugee populations (Camras et al.,
1990; Nelson et al.,
2012). Overall, our results suggest that maternal attention biases are not related to their child’s attention biases, and that transgenerational effects of familial psychopathology may be influenced by other factors, such as parenting or attachment styles (e.g., Thabet et al.,
2009).
Our study has some limitations. Firstly, despite the dot probe task being widely used to study attention biases, it generally has poor test-retest reliability and poor internal consistency (Brown et al.,
2014; Macleod et al.,
2019; Schmukle,
2005; Staugaard,
2009; Xu et al.,
n.d.), although some have reported good internal consistency (e.g., Bar-Haim et al.,
2007). Longitudinal attention bias data would be helpful in determining the consistency of emotion processing in our sample. Furthermore, as attention can shift within the first 200ms of stimuli presentation (Kappenman et al.,
2014; Müller & Rabbitt,
1989), it is possible that children might disengage their attention between the presentation of the stimulus and the presentation of probe, although it is unclear if this very rapid shift of attention could occur in children. Secondly, since mothers reported on their child’s behaviours, it is possible that those mothers with worse mental health might have perceived their child’s mental health more negatively, and thereby reported higher child psychopathology. Parental emotional distress and anxiety levels have been shown to influence their reporting of their child’s anxiety (Krain & Kendall,
2000; Niditch & Varela,
2011), although parents are often thought to be reliable reporters of child mental health, with parental reports used across many studies (e.g., Abate et al.,
2018; Murphy et al.,
2012). It is also important to note that although our study is cross-sectional in design, the data from the mothers were collected approximately 3 months after the collection of the children’s measures. Given the stability of the families’ living conditions, we expected biases to remain stable, but this time difference might have contributed to the lack of associations between children’s and mothers’ attentional biases. Future studies should investigate mental health outcomes and emotion processing biases measured at multiple timepoints in both children and caregivers to examine developmental trajectories in emotional processing in refugee youth.
Taken together, our findings highlight the important effects of mother’s war trauma exposure and mental health on their children’s wellbeing. The attention bias displayed by mothers and children to angry faces reveals a hypervigilance to threatening stimuli. Surprisingly, this hypervigilance is unaffected by mother or child mental health, suggesting that other potential cognitive mechanisms of intergenerational psychopathology transmission should be explored in the refugee context with complex trauma exposure.