Maternal Childhood Trauma and Perinatal Distress Predict Infants’ Attention From 6 to 18 Months

Maternal distress is repeatedly reported to have negative impacts on the cognitive 2 development in children. Studies examining the association between maternal distress and the 3 development of attention in infancy are few. This study investigated the longitudinal 4 relationships between maternal distress (depressive symptoms, anxiety symptoms, and 5 exposure to childhood trauma) and the development of attention in infancy in 118 mother-infant 6 dyads. We found that maternal exposure to non-interpersonal traumatic events in childhood and 7 a large degree of anxiety during the 2 nd trimester was associated with less attention of the infants 8 to audio-visual stimuli at 6, 10, and 18 months. In addition, exposure to interpersonal traumatic 9 events in childhood was identified as a moderator of the negative effect of maternal anxiety 10 during the 2 nd trimester on the development of attention in infants. We discuss the possible 11 mechanisms accounting for these cross-generational effects. Our findings underscore the 12 importance of maternal mental health to the development of attention in infancy and address 13 the need for early screening of maternal mental health during pregnancy.


Introduction 1
Attention in infancy is an important cognitive operation involving alerting, orienting, 2 and attending to information in the environment 1 . Attention develops rapidly in the first 3 postnatal year 2 and continues to develop into adulthood 3 , playing a fundamental role in 4 learning 4,5 . Attention is also linked to the development of self-regulation and executive function 5 in childhood 6,7 . Attention is seen as a predictor of social development 8 , cognitive functioning 9 , 6 language development 10 , and academic skills 11 . Poor attention skills are related to several 7 neurodevelopmental disorders, such as attention-deficit/hyperactivity disorder (ADHD) 12 and 8 autism spectrum disorder (ASD) 13 . Taken together, identifying the risk factors of the 9 development of attention in early years is crucial for prevention and intervention. 10 Maternal distress is seen to influence the trajectories of attentional development in 11 childhood. Extensive evidence from large cohort groups has shown that maternal depressive 12 and anxiety symptoms are associated with attention problems in offspring at the ages of 2 13 years 14 , 3 and 4 years 15 , as well as 5, 6.5, and 14 years 16,17 . Maternal distress is also linked to 14 ADHD symptoms at the age of 4 and 8-9 years 18,19 . Moreover, recent studies also reported that 15 maternal childhood adverse experiences contribute to ADHD and ASD in children [20][21][22] . 16 In nonhuman primates, exposure to mild stress during pregnancy is related to less visual 17 exploration and higher distractibility of offspring 23 . In humans, maternal stress during 18 pregnancy has a negative impact on infants' attention shifting at the age of 18 months 24 . It is 19 been shown that infants whose mothers perceived higher stress during pregnancy needed more 20 time than others to process visual information at the age of 7.5 months and looked away from 21 the tasks significantly more than infants whose mothers had low perceived stress during 22 pregnancy 25 . Preliminary evidence also suggests that this cross-generational association 23 between maternal distress and infant development might be linked to trauma exposure prior to 24 pregnancy 26,27 . Moreover, infants of depressed mothers have less synchronous mutual gaze with 25 their mothers than infants of non-depressed mothers 28 . In turn, mutual gaze has been associated 26 with visual attention in the first postnatal year of life 29 . The impact of maternal distress on 1 mother-infant interactions 30 and maternal sensitivity 31 have been related to infants' selective 2 attention 32 and gaze-following ability 33 . 3 Maternal distress is often symptomized by an unbalanced and/or strained emotional 4 state from pregnancy to postpartum commonly including depression and/or anxiety 34,35 . The 5 severity of psychological distress is strongly linked to exposure to traumatic experiences earlier 6 in life 36,37 . Evidence also suggests different types of traumatic events, such as interpersonal and 7 non-interpersonal trauma, have different impacts on mental distress and psychiatric 8 symptoms 38,39 . Overall, the literature suggests that maternal distress is multifaceted and it 9 contributes to the development of attention in children. However, contrary to compelling 10 evidence indicating the impact of specific variables related to maternal psychological distress 11 on the development of attention in childhood, there is only very little evidence investigating the 12 impact of multiple dimensions of maternal distress on infants and following the full path from 13 mother's childhood, pregnancy to 1 st postnatal year. The analysis of multiple risk factors 14 together is essential due to the high likelihood of co-morbidity and high correlations between 15 risk factors. To bridge this gap, in the current study, we took depression, anxiety, and trauma 16 exposure into consideration. The main aim of this longitudinal study was to investigate how 17 maternal childhood trauma and maternal distress are related to infants' attention from 6 to 18 18 months of age. We hypothesize that maternal childhood trauma exposure and maternal distress 19 are negatively associated with the infant's attention. In addition, besides identifying the risk 20 factors that might hinder infants' attention, we also preliminarily explored the supportive 21 factors that might buffer against the negative impact of maternal distress. Methods and results 22 are reported in the supplementary information. 23 24

Results 25
Multivariate regression analysis 26 Insert Figure 1. 5 Moderation analysis 8 Following the results described above, exposure to interpersonal traumatic events in 9 childhood was examined as a moderator of the relationship between the anxiety level during 10 the 2 nd trimester and the infants' look percentage after adjusting for infant sex and mother's 11 education. Figure 2 displays the slopes for the anxiety level during antenatal 17 weeks and the 12 levels of the exposure to interpersonal traumatic events predicting infants' attention. As 13 indicated by the change in the direction, the effect is moderated by interpersonal traumatic 14 events (F(5, 103) = 2.916, R 2 = 0.124, p = 0.018). In other words, with a higher exposure of 15 interpersonal traumatic events, infants' attention decreases as the mother's anxiety level 16 increases. 17

Discussion 22
The primary goal of the current study was to investigate whether maternal distress 23 affects the development of attention in infancy. We found that exposure to non-interpersonal 24 and interpersonal traumatic experiences childhood has cross-generational effects on infants' 25 attention. Moreover, childhood interpersonal trauma experience moderates the maternal anxiety 26 level during the 2 nd trimester and hinders the development of attention in infancy. In other 1 words, our results indicate that both traumatic experiences in mothers' childhood and perinatal 2 anxiety increase the vulnerability of mothers and elevate the risk for poor attention of infants. 3 From the standpoint of prevention, our findings underscore the importance of early screening 4 and intervention for mental health issues to support mothers and infants and prevent long-term 5 consequences, even before the pregnancy starts. 6 In our longitudinal data across pregnancy to early infancy, we found a particular 7 vulnerability in the 2 nd trimester. There are two main possible explanations for this finding. The 8 first explanation is that the fetal brain is vulnerable to the in-utero environment due to the 9 critical period of neurogenesis. Exposure to risk factors during this stage of development leads 10 to altered neuron connectivity. Compared to the 1 st and 3 rd trimesters, exposure to ethanol in 11 the 2 nd trimester has been reported to cause great neuronal loss in rodents 40 , attenuated cerebral 12 blood flow 41 , and long-lasting alternations in synaptic plasticity 42 in the human fetus. In 13 children, a previous study also reported that maternal anxiety during the 2 nd trimester, but not 14 later during pregnancy, is associated with gray matter reduction in several brain areas in 15 children (6-9 years old) 43 , including the prefrontal lobe, which is a crucial area in cognitive 16 development 44 and controls attention 45 . The second explanation is associated with the elevated 17 cortisol levels in mothers during pregnancy. Compared to mothers who do not experience 18 maternal anxiety, childhood maltreatment, or an adverse environment, mothers who experience 19 anxiety or these traumatic experiences have a higher level of cortisol 46-48 . Scheinost et al. 20 reported that increased cortisol levels during the 2 nd trimester and increased subjective maternal 21 distress in the 3 rd trimester are associated with weaker connectivity of the anterior cingulate 22 cortex of neonates 49 . The anterior cingulate cortex has been linked to infant's attention 50,51 and 23 ADHD in children 52,53 and adults 54 . In addition, one previous study investigating infants' 24 cognitive development at 12 months of age reported that infants with higher cognitive 25 performance were born to mothers with lower maternal cortisol levels in the 2 nd trimester but 1 higher cortisol levels in the 3 rd trimester 55 . 2 Interestingly, maternal depression showed no association with infants' attention. 3 However, previous literature has shown that infants of depressed mothers have a less 4 synchronous gaze in the mother-infant interaction 28,56 that may affect the development of 5 attention 29 . Similar to two well-controlled studies investigating cognitive development, 6 maternal depression during pregnancy and infancy did not affect cognitive development at the 7 age of 3 years 57 and 18 months 58 , respectively. In the context of the current study, there are 8 several plausible reasons for this finding. First, the association between maternal depression 9 and infants' attention may not exist. However, using the same dataset investigating gaze  Keeping these alternatives in mind, we cautiously propose another reason. Given the 23 high comorbidity of depression and anxiety in our data (Table 4) and the literature 62,63 , we 24 propose that anxiety may be the driving force behind peripartum depression. For example, when 25 we examined depressive and anxiety symptoms separately (Table 2), they showed a unique 26 effect during the 2 nd trimester. When we further combined all dimensions and examined the 1 effect while simultaneously controlling others, anxiety dominated the effect. To the best of our 2 knowledge, maternal depression and anxiety are rarely combined and related to child 3 development, meaning that the importance of maternal anxiety may have been interpreted as 4 an effect of depression in prior work. However, the complexity and dynamics between 5 traumatic experiences, depression, and anxiety and how the dynamics change over time are 6 beyond the scope of the current study. Future studies are needed to help us understand how 7 maternal mental health affects infants' attention. Most importantly, it will provide us more 8 knowledge on promoting maternal mental health and infant development. 9 Finally, and especially due to our limited sample size, our results must be interpreted in 10 light of some limitations. Firstly, we could not compare clinically severe cases due to the 11 relatively small number of severely depressed mothers. To deal with the relatively small sample 12 size and the significant collinearity between depression and anxiety, we calculated the factor 13 scores for depression and anxiety separately at four different time points. This may prevent the 14 plausible interactions at different stages and different levels from being observable in our results. 15 Moreover, our sample is limited to a homogenous population in Uppsala (Sweden), with more 16 than half of participating mothers having education levels of university or higher. In addition, 17 we did not control for the possible influence of partners' mental health on mothers' well-being 18 and infants' attention. As our results indicate the important influence of interpersonal traumatic 19 experiences, future studies should consider this interpersonal aspect and its dynamics with 20 regard to mothers' well-being. 21 Our findings add to the growing body of research, suggesting that prevention and 22 intervention should start before pregnancy for both mothers and infants. Lastly, the findings 23 describe a previously undocumented connection between maternal early trauma, anxiety, and 24 the development of attention in infants. Treating pregnant women's anxiety, especially if she 25 has experienced traumatic events in the past, may not only improve the lives of mothers, but 1 also support positive development of their children from infancy onwards. 2 3 Methods 4 Participants 5 The final data included 118 mother-infant dyads from the BASICchild cohort as part of 6 a longitudinal study (the BASIC Child Project) 64 of a subsample of the population-based 7 BASIC study "Biology, Affect, Stress, Imaging, and Cognition (BASIC)" 65 collected from 2014 8 to 2018. Characteristics of the mother-infant dyads are shown in Table 3. Only healthy pregnant 9 women > 18 years old who received a routine examination at Uppsala University Hospital were 10 invited to participate in the projects. Mothers who consented to participate were invited to fill 11 out a series of questionnaires online at 17 and 32 gestational weeks, and postpartum at 6 weeks, 12 6 months, and 12 months. Mothers and infants who took part in the BASIC Child Project visited 13 the Uppsala Child and Baby Lab when the infants were aged 6 (n = 118; mean = 185 days, SD 14 = 7.5 days, 59 boys), 10 (n = 110; mean = 302 days, SD = 9.2 days, 53 boys), and 18 months 15 (n = 104; mean = 544 days, SD = 12.1 days, 53 boys). All infants were reported healthy. Sixty-16 five percent of the mothers held a university degree. All procedures in the study were conducted 17 in accordance with the 1964 Declaration of Helsinki ethical standards and approved by the  Table 3. 25

Measures of maternal distress 2
Symptoms of depression were measured using the Swedish version of the Edinburgh 3 Postnatal Depression Scale (EPDS) 66,67 . The EPDS includes 10 questions scored from 0 to 3. 4 Thus, the total score ranges from 0 to 30, with higher scores indicating more severe symptoms. 5 The reliability and validity of the EPDS has been shown to be adequate 68,69 . Symptoms of 6 anxiety was measured using the Beck Anxiety Inventory (BAI) 70 . The scale consists of 21 items, 7 with participants indicating the extent to which they were bothered by each item. The total score 8 for each item ranges from 0 to 63, with higher scores indicating more server symptoms 71 . A 9 high level of internal consistency and a good test-retest correlation have been reported 70 .

Measure of infants' attention 26
Infants' attention was measured by the look percentage (defined as the total fixation 1 duration of the stimuli divided by the total duration of all tasks at each age) across a variety of 2 free-looking tasks at the age of 6, 10, and 18 months. Attention is assumed to be closely linked 3 to oculomotor movement and oculomotor control 75,76 and, in this study, look percentage is used 4 as a proxy for sustained attention 77,78 . The mean look percentage at 6, 10, and 18 months was 5 73.63% (SD = 9.84), 73.47% (SD = 9.36), and 79.24% (SD = 6.86), respectively. The Pearson 6 correlation coefficients (Table 4) of attention, look percentage, between different ages were 7 0.33 (6-10 months, n = 110, p <0.001), 0.21 (6-18 months, n = 103, p = 0.04), and 0.31 (10-8 18 months, n = 100, p < 0.01), suggesting the stability and internal consistency of attention 9 during the course of development. In the current study, the composite score of look percentage 10 was calculated by averaging each participant's look percentage measured at three time points 11 and used as the dependent variable. All tasks were recorded using an eye-tracker with a 12 sampling rate of 60 Hz following a 5-point calibration (Tobii TX300, Tobii Technology AB, 13 Sweden). 14 Insert Table 4. 16

Statistical analysis 19
We used multivariate linear regression models and a moderator analysis to examine the 20 association between multiple predictors across different time points and the outcome measure. 21 To assess the reliability of the maternal scale instruments, we calculated the internal consistency 22 coefficient, Cronbach's alpha for each tool: EPDS, 0.87, good; BAI, 0.81, good; and LITE, 0.9, 23 excellent. Before adjusting their scores, the zero-order Pearson correlations (with Benjamini-24 Hochberg correction), skewness, and kurtosis of all variables were calculated (Table 4). The 25 variance inflation factor (VIF) was calculated based on the assumption that infants' look percentage is predicted by 10 variables from the EPDS (4 time points), BAI (4 time points), 1 and LITE (1 time point). As seen in Table 4, raw scores for anxiety symptoms during antenatal 2 17 weeks and postpartum 6 weeks are not in the acceptable range of the kurtosis index. The 3 raw scores of the EPDS, BAI, and LITE did not reach the range of approximate symmetric 4 distribution (kurtosis index acceptable range, -2 to +2; skewness index accetable range -0.5 to 5 + 0.5) 79 . In addition, the literature has shown that comorbidity of depression and anxiety is 6 common 62,63 , so we expected to detect a potential multicollinearity from the raw data. As seen 7 in Table 4, the raw scores of the EPDS and the BAI during antenatal 17 weeks fit the strict analysis. In step 1, considering that maternal trauma exposure prior to pregnancy (both IP and 24 nIP) may interact with depression or anxiety, we separated variables into four groups as listed 25 in Table 2 and analyzed four linear regression models independently. Applying a backward 26 stepwise method, the number of variables in each model was reduced (3 rd column, Table 2). In 1 step 2, based on the statistical selection shown in Table 2, we combined the significant variables 2 and 2-degree interaction from all models to assess how they jointly predict infants' attention 3 (measured by look percentage; see Model A, Table 1). Based on Model A, we selected 4 significant variables for Model B (see Table 1). In the third step, we added the sex of infants 84,85 , 5 mothers' smoking habits 86 , education 87 , and the maternal age at birth 88 to the analysis (Model 6 C, Table 1). All tests were two-sided tests with p < 0.05 considered significant. All statistical 7 analyses were performed using R 4.0.3 89 . 8    8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47 Table 4. Pe arson's zero order correlations between all variables using raw scores    Figure 1 Illustration of the multivariate linear regression model after adjusting for the sex of the infant, mother's education level, smoking history, and the maternal age at birth. Non-interpersonal traumatic experiences in mother's childhood and maternal anxiety in early pregnancy had a direct impact on the infants' look percentage Co = 0.014 and 0.006, respectively). When anxiety at week 17 of pregnancy interacts with interpersonal traumatic exposure in childhood, the negative impact on the infants' look percentage is highly signi cant (p <0.001). LITE, Lifetime Incidence of Traumatic Events; IP, interpersonal events; nIP, non-interpersonal events.

Figure 2
The relationship between maternal anxiety (Beck Anxiety Inventory at antenatal 17 weeks, BAI week 17) and infants' attention, look percentage (LP), is moderated by the level of interpersonal traumatic events (IP) in childhood measured by Life Incidence of Traumatic Events (LITE). Level 1 (solid line) represents mothers who exposed to less trauma in childhood compared to those at level 2 (dotted line).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementaryInformationSP.pdf