Characterizing Social Communication Difficulties in Young Children within a Longitudinal Ecological Systems Framework

Table 1 presents descriptive statistics for study variables. Table S5 presents bivariate correlations between study variables.

We leveraged multi-method assessments across pregnancy and the first two years of life to characterize risk factors for child social communication difficulties. Drawing on Bronfenbrenner’s ecological systems model of development (Bronfenbrenner, 1992), we investigated child (i.e., gestational age), family (i.e., household income, maternal postpartum depression or anxiety, maternal bonding), and neighborhood (i.e., neighborhood resources) factors. Consistent with prior research (van Noort-van der Spek et al., 2012), earlier gestational age at birth was associated with more social communication difficulties at age 2. Although the effect of gestational age was not qualified by interaction with neighborhood resources, it is worth noting that few of the children included in the present study were born preterm (i.e., gestational age < 37 weeks; n = 16, 6.4%). Thus, effects of gestational age may have been underestimated. Future work is needed to examine interactive effects of preterm birth and low neighborhood resources on social communication difficulties among larger samples of preterm children.

Consistent with prior research (Le Bas et al., 2020), impaired maternal postpartum bonding related to more child social communication difficulties at age 2. Prior research suggests that mothers who report stronger bonds with their infant experience less parenting stress (Escribano et al., 2022) and implement more positive parenting strategies (Thomson et al., 2014), which have, in turn, been linked to improved child socioemotional skills (Batool & Lewis, 2020), cognition and language use (Prime et al., 2023), and effortful control (Klein et al., 2018) across childhood. Our results expand this literature by focusing on very early maternal bonding (i.e., 10–15 weeks postpartum), pointing to an important window for targeting preventative intervention efforts to improve the parent–child relationship and promote positive parenting.

Although maternal bonding was not directly related to social engagement difficulties at age 1, there was a significant interaction between neighborhood resources in pregnancy and impaired maternal bonding in the postpartum period, such that the effects of low neighborhood resources on age 1 social engagement difficulties were only significant for dyads with greater impairments in bonding. This finding is consistent with prior research documenting critical interactions between parenting and context in relation to child outcomes (Hyde et al., 2020). For example, links between neighborhood disadvantage and school violence and child behavior problems were buffered by positive parenting (Roche et al., 2007). Interestingly, fewer neighborhood resources only related to social engagement difficulties in the context of low maternal bonding. That is, better maternal bonding, characterized by mothers’ positive feelings about their child and beliefs about their ability to provide nurturing parenting (Freeman, 2017), served to buffer the influences of fewer neighborhood resources on social engagement difficulties at age 1.

Of note, after controlling for observed sensitive parenting at age 2, impaired postpartum bonding was unrelated to social communication difficulties at age 2, suggesting that concurrent parenting experiences may buffer negative early bonding impacts (or indexing some shared method variance, since observed social communication included observer ratings from the same tasks used to rate sensitive parenting). However, results suggest that early impairments in bonding may reflect more short-lived adjustment difficulties in the context of this community sample of dyads. Importantly, prior research shows that sensitive caregiving in the first three years of life is critical to establish healthy attachment relationships and positive child development (DePasquale & Gunnar, 2020). Extant research also indicates worse cognitive and social outcomes for toddlers whose mothers report chronic depression and persistently low parenting sensitivity (NICHD Early Child Care Research Network, 1999). Taken together, our findings provide support for prevention efforts aimed at increasing sensitive parenting and maternal-child bonding throughout infancy and toddlerhood.

Finally, lower household income during the postpartum period was associated with more child social communication difficulties at age 2. Again, there was an interaction with the broader context, such that low household income in concert with fewer neighborhood resources were related to more social communication difficulties at age 2. Household income continued to be related to child social communication difficulties at age 2 even controlling for sensitive parenting at age 2. Together, results are consistent with evidence that families in lower income households situated within already economically disadvantaged neighborhoods suffer the cumulative effects of financial and social stress (e.g., working multiple shifts, experiencing health problems), but without the benefits of protective neighborhood resources (e.g., access to healthy food, safe play spaces), which might otherwise promote positive child development (Acevedo-Garcia et al., 2014). Findings expand our knowledge of how living in a neighborhood with few economic and social resources can increase chronic stress within the family (Masarik & Conger, 2017). These stressors may be compounded when accompanied by early maternal bonding difficulties, leading to parents’ decreased beliefs about their ability to effectively soothe their child or meet their needs (Franco et al., 2010). Our findings are among the first to document how the interplay of early environmental influences, such as parenting and neighborhood context, shape risk for psychopathology before overt symptoms can be reliably detected (Wakschlag et al., 2019).

Together, our findings suggest that preventative interventions deployed within the peripartum period could be especially impactful if they target positive parenting and family level factors (e.g., supporting parents to cope with parental stress and improve infant bonding), particularly among families living in lower resourced neighborhoods (Masarik & Conger, 2017; Roche et al., 2007). Such efforts could adopt a tiered approach, composed of a universal primary prevention program followed by a targeted secondary prevention for families with greater needs (Canfield et al., 2023). Other efforts include cash-based interventions during infancy, such as the Baby’s First Years study, which is testing how poverty reduction can improve child development outcomes through the provision of unconditional cash transfers to families (Noble et al., 2021). Preliminary outcome data suggest that infants of families randomized to receive high-cash gifts showed differential brain activation in regions linked to cognitive functioning (Troller-Renfree et al., 2022). Alongside our findings, similar poverty reduction efforts aimed at individual families could help benefit children’s early social communication skills, including by reducing parental stress or helping parents to invest in resources that enrich a child’s early environment (Noble et al., 2021). Importantly, although we did not find an interaction between perceived discrimination and neighborhood resources, 90% of mothers who identified as Black and/or Latina/e/o/x lived in neighborhoods below the mean for the neighborhood resources factor. Future work is needed to further evaluate how living in poverty impacts the emergence of child social communication difficulties, including among minoritized families who, because of the legacy of structural racism, may be more likely to live in neighborhoods with fewer resources (Njoroge et al., 2023).

Despite many strengths (i.e., strong theoretical framework, repeated prospective assessments from pregnancy through the second year of life), findings from our study must be considered within the context of several limitations. First, findings did not fully replicate across ages 1 and 2, which could have been due to unavoidable differences in measurement approach, including shared method variance at age 2, but not age 1 (e.g., report measures at age 2 that were not collected at age 1). Second, we relied solely on an observational design, prohibiting causal claims about relationships between variables. Moreover, models did not capture inherent bidirectionality over time between maternal bonding and child social communication abilities. Thus, we did not model whether children with lower affiliative behavior evoke less social input from parents over time (Klahr et al., 2013). Randomized controlled trials targeting parent–child bonding and parenting (e.g., Family Check-Up; Shaw et al., 2019) or more sophisticated longitudinal study designs (e.g., ecological momentary assessment) (Russell & Gajos, 2020) could help disentangle these pathways. Third, we assessed a community sample with relatively low levels of clinically severe social communication problems. Results should be replicated in clinic-referred samples of children. Fourth, maternal depression/anxiety was unrelated to child social communication difficulties. Future work needs to replicate or explore similar models in treatment-seeking parents or samples with a high proportion of individuals meeting screening thresholds for postpartum depression or anxiety. Fifth, given our sample size and the number of parameters investigated, we estimated a latent neighborhood resources factor rather than examine each COI factor independently. Future research using larger samples is needed to examine finer-grained neighborhood influences. Sixth, our cohort was originally recruited during the pandemic, which had widespread consequences for child development and mental health (Binagwaho & Senga, 2021). For example, many families contended with financial instability and the full-time care and education of their children during their pregnancy and the early postpartum phase. Of note, in supplemental analyses, we replicated our findings after controlling for mother-reported pandemic worries. Future research, however, should attempt to replicate our findings among samples of children who were not born at the height of a once-in-a-generation stressful event. Finally, although we sampled for equal representation of individuals identifying their race as Black and non-Latino/a/e/x White, our sample is not representative of the US population and results may not generalize to other racial groups.

In sum, we provide new insights into the interplay between the child, family, and neighborhood risk factors in relation to child social communication difficulties, which can signal risk for various forms of psychopathology, including depression (Meagher et al., 2009), disruptive behavior (Donno et al., 2010), and autism spectrum disorder (Costescu et al., 2022). Our results emphasize the importance of implementing early preventative interventions to help support early maternal bonding within the first few months of life, along with interventions that target families living within less advantaged neighborhoods (Leventhal et al., 2006). Established preventative interventions are effective in addressing many of these pathways, including those that target parent–child bonding and parental stress (Shieh et al., 2023), financial insecurity (Noble et al., 2021), and families living within lower resourced neighborhoods (Shaw et al., 2019). Research is needed to establish whether such interventions can be adapted at scale and implemented on a widespread population level, perhaps through a tiered approach (Canfield et al., 2023) or beginning during pregnancy (Shieh et al., 2023) to best support the parent–child dyad across development.