Setting the Stage: Biopsychosocial Predictors of Early Childhood Externalizing Behaviors

Abstract

Research has indicated that biological, (self-regulation), psychological (temperament), and social factors (maternal parenting behaviors) predict childhood externalizing behaviors. Few studies, however, have evaluated psychological, biological, and social factors in conjunction as predictors of childhood externalizing behaviors. Further, limited research has examined whether these biopsychosocial predictors during infancy and toddlerhood predict the onset of behaviors in early childhood. The present study aimed to examine the longitudinal relations between biopsychosocial predictors of child externalizing behaviors. Children and their mothers (n = 410) participated when children were 5-, 24- and 36-months old. Child self-regulation was assessed via baseline respiratory sinus arrhythmia (RSA) at age 5-months and child psychology was measured via maternal report of effortful control at age 24-months. Additionally, maternal intrusiveness was assessed during a mother-child interaction at age 5-months. At 36-months, mothers reported on child externalizing behaviors. Longitudinal path modeling was used to examine the direct and indirect effects of maternal intrusiveness and child effortful control on child externalizing behavior, as well as whether these effects were conditional upon child baseline RSA. Results showed a significant indirect effect of maternal intrusiveness on externalizing behavior through effortful control, and this pathway was moderated by baseline RSA after controlling for orienting regulation at age 5-months. These results suggest that early childhood externalizing behaviors are jointly affected by biological, psychological, and social factors during toddlerhood.

Externalizing behaviors typically arise in early childhood as children refine their emerging self-regulatory and social skills (Barnes et al., 2013; Bornstein et al., 2010) and begin to decline as children get older due to marked improvements in self-regulation (e.g., Hill et al., 2006; Olson et al., 2017). For some children, however, externalizing behaviors will persist and manifest as maladaptive problems (e.g., peer rejection or aggression; Campbell, 2006) that may culminate in child psychopathology (Kuja-Halkola et al., 2015; Loeber & Burke, 2011). Much of the existing literature on externalizing behaviors has focused on older children (Eisenberg et al., 2005; Hinnant & El-Sheikh, 2009; Yan et al., 2019). Research with younger samples, however, is necessary to better understand at-risk child trajectories and to identify potential avenues for early intervention. Additionally, given the long-term impact externalizing behaviors can have on child development, understanding the early underpinnings that contribute to the development of later externalizing behaviors is critical to both clinical and developmental research.

To date, empirical research has illustrated a relation between child externalizing problems and social factors, such as maternal parenting (e.g., McKee et al., 2008). The mechanism(s) that may account for this association are not well understood, however, and investigations into whether certain groups of children (e.g., children with differing levels of physiological regulation) are more susceptible to the detrimental effects of disruptive parenting are limited in the early childhood literature. Evidence linking aspects of early self-regulation, such as respiratory sinus arrhythmia (RSA) and effortful control, with a host of critical socio-emotional outcomes later childhood may inform our current understanding of the relation between parenting and externalizing behaviors (Eisenberg et al., 2015; Hinnant & El-Sheikh, 2009; Yan et al., 2019). That is, maternal behavior may influence children’s early attentional/behavioral inhibition skills (effortful control), which in turn may predict child externalizing problems later in development (e.g., Chang et al., 2011). Furthermore, this proposed, indirect pathway may not emerge across all groups of children. In other words, individual differences in physiological regulation (RSA) and early temperamental regulation (e.g., orienting sensitivity) may result in an increased vulnerability to the effects of adverse parenting for some children in comparison to others (e.g., Dyer et al., 2016). Therefore, the goal of our study was to examine how biobehavioral child self-regulation and maternal parenting relate early in development (i.e., in infancy and toddlerhood), and whether these factors collectively predict externalizing behavior in early childhood.

Social Contributors to Early Externalizing Behaviors

A number of social factors have been shown to contribute to the development of early externalizing behaviors during childhood, including parenting behaviors, peer relations, and teacher support (Eisenberg et al., 2005; Hendrickx et al., 2022; Sturaro et al., 2011). Across the first few postnatal years, many children spend the majority of their time around their mothers (e.g., Fagan et al., 2014). Given our study’s emphasis on infancy, toddlerhood, and the early childhood years, we focused on parenting behaviors as a measure of children’s early social context as it relates to childhood externalizing behaviors. Though previous research has demonstrated that a variety of maternal behaviors can contribute to or protect against the development of externalizing behaviors and child psychopathology more generally (e.g., Penela et al., 2012; Yan et al., 2019), we were guided by the research literature to focus on one maternal behavior in particular: maternal intrusiveness.

Maternal Intrusiveness

Maternal behaviors that are sensitive to children’s cues foster supportive and engaging reciprocal interactions that promote early self-regulatory skills in infants and young children (Calkins & Johnson, 1998; Fox & Calkins, 2003). Consequently, maternal intrusiveness refers to parenting behavior that is over-controlling and disruptive, and it is thought to limit children’s developing autonomy and regulatory capabilities (Eisenberg et al., 2015; Graziano et al., 2010). Although developmental research illustrates that children who are exposed to adverse parenting behaviors early in life are at an increased risk of exhibiting behavioral problems later in development (Akcinar & Baydar, 2014; McCullough & Shaffer, 2014; Yan et al., 2019), there are significant gaps in the existing literature.

First, much of the current work examining social predictors of child externalizing behavior has focused on other aspects of “negative” maternal behaviors that are related but distinct from maternal intrusiveness, including harsh parenting (Akcinar & Baydar, 2014) and negative affect (McCullough & Shaffer, 2014). Given previously reported links between various aspects of child socio-emotional functioning and intrusive parenting (Eisenberg et al., 2015; Yan et al., 2019), further research is necessary examining the association between maternal intrusiveness and child externalizing behavior to develop a more robust understanding of the ways in which adverse parenting behaviors influence child development. Second, many studies examining predictors of child externalizing behavior have been conducted concurrently and/or with older child samples (Price et al., 2013; Perry & Price, 2018; Yan et al., 2019). Because externalizing behaviors begin to emerge in early childhood, it is important to develop a richer understanding of how social factors that are present early in development contribute to the onset of these behavioral problems. Finally, the mechanisms that may account for the association between early maternal intrusiveness and later child externalizing problems have received relatively little attention within the developmental literature. Indeed, children who experience high levels of intrusive parenting tend to struggle with adaptive self-regulatory skills and have lower levels of effortful control (Graziano et al., 2010; Smith & Day, 2018). For instance, using data from a sample of preschool- and kindergarten-aged children, Eisenberg and colleagues (2015) reported direct effects between maternal intrusiveness and child effortful control, and between child effortful control and externalizing problems. Child effortful control may therefore play an important role in linking early parenting behavior with children’s behavioral problems observed later in development, and evidence supporting effortful control as a potential mechanism behind this association is discussed in greater detail below. Taken together, we aimed to address these limitations by examining both the direct and indirect links between maternal intrusiveness and child externalizing behavior using a younger sample of children assessed across two developmental timepoints.

Psychological Self-Regulatory Contributors to Early Externalizing Behaviors

Self-regulation abilities develop across infancy and early childhood (Rothbart et al., 1994), and they play a crucial role in preventing the development of externalizing behaviors (Perry et al., 2018). Self-regulation depends on a number of factors, including biological processes (Bell & Deater-Deckard, 2007) and temperament (Rothbart et al., 2006), as well as family contexts (Kiss et al., 2014). We focused on one psychologically based aspect of self-regulation (effortful control; e.g., Murray & Kochanska, 2002) and one biologically-based aspect of self-regulation, (RSA; Calkins & Dedmon, 2000), as it relates to parenting and externalizing problems in early childhood. Additionally, we included orienting regulation during infancy as a covariate in an effort to capture early individual differences in self-regulation, as orienting regulation is related to other measures of self-regulation as children get older (e.g., Gartstein et al., 2013).

Effortful Control

Effortful control is a temperamental factor developed by Rothbart and colleagues (Rothbart, 1989; Rothbart & Bates, 2006) that has been defined as a child’s ability to inhibit a dominant response in favor of activating a subdominant one. Effortful control begins to develop late during the first year of life in concordance with developing brain areas associated with attention (Rothbart et al., 1994), and it is sensitive to the influence of intrusive parenting behaviors in toddlerhood (Taylor et al., 2013). Once effortful control begins to develop, it continues to do so very rapidly (Posner & Rothbart, 2000) and remains relatively stable across childhood (Kochanska et al., 2000). Children who exhibit high levels of effortful control tend to be better at self-regulation overall and tend to have fewer behavioral problems, compared to children who exhibit lower levels of effortful control (e.g., Gartstein et al., 2012). Effortful control during early childhood may be considered a psychological measure of self-regulation, given its ability to be assessed via maternal reports (e.g., Kochanska et al., 2000).

Research has demonstrated direct associations between effortful control and externalizing behaviors in early childhood. For instance, Olson and colleagues (2005) reported that both maternal and behavioral reports of child effortful control were directly and negatively associated with externalizing behaviors, such that less effortful control in 3-year-old children predicted more externalizing behaviors at the same age. Similarly, and with a longitudinal focus, it has also been shown that less effortful control during toddlerhood (around age 24-months) directly predicts more externalizing behaviors both during toddlerhood and early childhood (around age 36-months; Gartstein et al., 2012). Given evidence supporting relations between intrusive parenting behavior and effortful control, and between effortful control and externalizing behavior, it follows that maternal intrusiveness may indirectly predict externalizing behavior problems at a later point in childhood via child effortful control. In support of this interpretation, Eisenberg et al. (2015) showed that effortful control in 11-year-old children mediated the association between warm parenting observed when children were 9-years-old and externalizing behaviors measured when children were 13-years-old. In other words, greater parental warmth predicted higher levels of child effortful control, which then predicted fewer externalizing behaviors. Of course, the absence of maternal warmth does not inherently imply that mothers are engaging in adverse caregiving behaviors, and therefore an examination of maternal intrusiveness in particular is warranted. Furthermore, supplemental research seeking to replicate this indirect pathway with a sample of much younger children and their mothers is necessary given the efficacy of early intervention.

RSA

RSA is a biological measure of self-regulation that assesses cardiac vagal tone (Porges, 1991). RSA is a benign heart arrhythmia easily detectable in a child’s electrocardiogram (ECG) that occurs in tandem with respiration; specifically, RSA is a measure of the increase in heart rate during inhalation and the decrease in heart rate during exhalation (Grossman & Taylor, 2007). Polyvagal Theory, which postulates that RSA is a measure of parasympathetic nervous system (PNS) activity, suggests that RSA operates via control of the vagus nerve and works to inhibit the fight-or-flight responses for which the PNS is responsible and instead facilitates more adaptive responses given the environmental context (Porges, 2007). For instance, the vagus nerve can act as a break to slow heart rate when an individual’s vagal tone is high, allowing them to engage in more adaptive behaviors, and not inhibit heart rate when vagal tone is low, as it is not necessary to do so (Porges, 2001). As a measure of self-regulation, RSA remains relatively stable across childhood (Calkins & Keane, 2004).

In the event that maternal intrusiveness predicts externalizing behavior through child effortful control, then an important question becomes whether the strength of this indirect effect differs across groups of children with differing levels of RSA. In line with the differential susceptibility model (Belsky & Pluess, 2009), some children may be predisposed to be more sensitive than others to the effects of environmental factors, such as intrusive parenting behavior. In this sense, individual differences in children’s parasympathetic nervous system functioning, which are detectable and relatively stable early in life, may moderate the association between maternal behavior and child outcomes. For the most part, higher baseline levels of RSA are associated with better self-regulation and psychological functioning overall (Porges, 2001), which has implications for childhood behavioral difficulties. Indeed, Calkins & Dedmon (2000) demonstrated that 2-year-old children who were at a high risk of developing externalizing behavior problems (assessed via scores on the Child Behavior Checklist) were more likely to have a lower baseline RSA than children of the same age who were identified to have a lower risk of developing externalizing behavior problems. Additionally, 6-year-old children with a low baseline RSA had higher levels of externalizing problems at age 8 (Hinnant & El-Sheikh, 2009). Other work has also supported this link between lower baseline RSA and more externalizing problems, but has suggested that it is only significant in adolescents (Beauchaine et al., 2007).

In contrast, some researchers have failed to find a direct effect of baseline RSA on externalizing behavior problems, but instead have demonstrated interactive effects when various environmental factors are also taken into consideration. For example, 2-year-old children who were identified as being at risk for externalizing behavior problems who participated in family-based interventions had lower levels of externalizing behavior problems by age 5, however, this association was only significant when moderated by baseline RSA; specifically, the results were only significant when children exhibited lower levels of baseline RSA (Bagner et al., 2012). More recent work has suggested that both high and low levels of RSA across childhood can predict more externalizing problems depending on the family context (e.g., maternal engagement in harsh and/or emotionally supportive parenting; Ugarte et al., 2021).

Taken together, the current literature on the relation between baseline RSA and externalizing behaviors during childhood appears to yield contradictory findings based in part on study methodology. When taking into consideration children’s greater social context early in development, however, an interaction between baseline RSA and parenting may best account for variability in children’s temperamental self-regulation and/or externalizing behavior. Investigating whether certain groups of children with differing levels of baseline RSA are more vulnerable to the effects of intrusive maternal behavior is critical to identifying at-risk child trajectories. Thus, we aimed to better understand these early relations by examining the moderating role of baseline RSA on the association between social (maternal intrusiveness) and psychologically-based regulatory predictors (effortful control) of early child externalizing behaviors.

Orienting Regulation

Self-regulation has its roots in infancy, with the temperamental trait of orienting regulation (orienting; Gartstein et al., 2013). Infants who are high in orienting have a greater capacity for sustained attention during play, demonstrate positive affect while being held by a caregiver and engaging in soothing activities with a caregiver, and are easily soothed when distressed (Gartstein & Rothbart, 2003). Research has demonstrated that higher levels of orienting during infancy are associated with higher levels of effortful control during early childhood (Gartstein et al., 2013; Rothbart et al., 2011). Recently, it has been reported that infant orienting might predict different parenting behaviors, such as permissiveness (Wittig & Rodriguez, 2019) and could moderate the association between infant negative emotions and maternal intrusiveness (Wu & Feng, 2020). It has also suggested that low levels of infant orienting are associated with higher levels of childhood behavioral problems (Toffol et al., 2019), but others have failed to find any association between the two (Gartstein et al., 2012).

Given the conflicting findings regarding infant orienting and childhood behavioral problems (Gartstein et al., 2012; Toffol et al., 2019), we did not expect to find an association between infant orienting and early childhood externalizing behaviors at 36-months. However, because research has strongly argued that infant orienting plays an essential role in later self-regulation capacities (Gartstein et al., 2013), such as effortful control, we decided to control for infant orienting in our analyses in the current study. In doing so, our findings will address whether early caregiving behavior is related to toddler effortful control after accounting for rank-order stability in children’s emerging self-regulatory skills.

Current Study

Early biosocial factors play an important role in augmenting and/or minimizing children’s risk for exhibiting behavioral difficulties later in development. As noted, the available literature regarding the associations among self-regulation, maternal intrusiveness, and externalizing behaviors in childhood is limited in many ways. First, although adverse parenting behaviors have been linked to child externalizing behaviors (e.g., Keown, 2012; Yan et al., 2019), much of the current research has ignored longitudinal associations and/or has focused on maternal negative affect or harshness (e.g., McCullough & Shaffer, 2014). Because of this, the association between maternal intrusiveness and child externalizing behaviors is not fully understood. Second, much of the work regarding temperamental self-regulation as it relates to externalizing behavior has been conducted with older child samples (e.g., Choe et al., 2013; Eisenberg et al., 2015), despite the evidence that effortful control emerges in toddlerhood and begins to stabilize in early childhood (Rothbart et al., 1994). Furthermore, disagreement is apparent in the developmental literature regarding the role individual differences in nervous system functioning play in child behavior problems. For instance, researchers have reported that lower baseline RSA is predictive of a) more externalizing problems (Hinnant & El-Sheikh, 2009), b) greater externalizing problems only in adolescence (Beauchaine et al., 2007), or c) increased externalizing problems for some children based on their home environment (Ugarte et al., 2021).

To our knowledge, no previous work has assessed the combined effects of RSA, effortful control, and maternal intrusiveness on externalizing behaviors in early childhood. It would, however, be beneficial to evaluate the roles of these early biopsychosocial factors to get a fuller understanding of whether these infant and maternal factors operate in tandem to predict child externalizing behaviors. Our study aimed to address these gaps in the literature by examining these relations among a younger sample with a longitudinal study design. We utilized path modeling to examine whether maternal intrusiveness (5-months) exerted an indirect effect on child externalizing behavior (36-months) via child effortful control (24-months), and whether this anticipated indirect effect was conditional upon infant baseline RSA level (5-months) after controlling for infant orienting (5-months). Given that children with low baseline RSA are already at an increased risk for exhibiting behavioral problems later in development (e.g., Hinnant & El-Sheikh, 2009), we hypothesized that infants with a high baseline RSA would be more sensitive to the effects of intrusive parenting behavior during their first postnatal year with respect to the development of self-regulatory skills (effortful control) and later behavioral problems (externalizing behaviors). In other words, children with a higher baseline RSA are typically regarded as being more temperamentally reactive (Calkins, 1996), which could lead to more interactions between the child and their mother. If mothers are intrusive during these interactions, they could be expected to hinder the child’s development of effortful control (e.g., Graziano et al., 2010). Specifically, we predicted that the indirect pathway between greater maternal intrusive parenting during infancy and increased childhood externalizing behaviors through lower levels of toddler effortful control would be contingent upon infants having a higher baseline RSA level.

Method

Participants

Participants for the current study included 410 mother-child dyads who came from a larger longitudinal study examining cognitive and emotional development from infancy through middle childhood across three cohorts: cohort 1 (n = 106), cohort 2 (n =105), and cohort 3 (n = 199). Across our current study’s three data collection waves, children were approximately 5-months (M = 5.42, SD = 0.26), 24-months (M = 25.26, SD = 0.80), and 36-months (M = 36.64, SD = 1.05) of age at each lab visit. There were 201 boys and 209 girls. The majority of the children were identified by their mothers as White (77.6%), with smaller proportions identified as Black or African American (13.7%), Multi-Racial or other (8.3%) and Asian (0.4%). In terms of ethnicity, 7.1% of children were identified as Hispanic. Maternal age ranged from 16 – 44 years when the children were 24-months old (M = 31.68 years, SD = 6.01 years). Regarding maternal education, 2.2% of mothers, did not complete high school, 32.2% completed high school or had a technical degree, and 63.2% had completed college or graduate school. The remaining 2.4% of mothers did not report education status. The cohorts were recruited when the children were infants using mailing lists, media advertisements, flyers, and word of mouth by two research locations, a rural college town (Blacksburg, VA) and a mid-sized city (Greensboro, NC), both in the mid-Atlantic region of the United States. Each location recruited half of the participants and the demographics of participants from each location reflected the demographics of the area in which each research site was located. Participants at each location did not differ with respect to sex [χ² (1, N = 410) = 2.16, p = .14]. However, the college town location had mothers with higher levels of education, on average [χ² (3, N = 410) = 12.94, p = .01].

Of the 410 participants in this study, 405 mother-child dyads (98.8%) participated in the 5-month lab visit and contributed usable 5-month data to the study. That is, 350 dyads completed all three of the 5-month measures, and the remaining 55 dyads completed at least one of the three 5-month measures [i.e., only infant RSA (n = 3), only infant orienting (n = 8), infant RSA and orienting (n = 18), infant RSA and maternal intrusiveness (n = 6), or infant orienting and maternal intrusiveness (n = 20)]. An additional five dyads are missing 5-month data as they joined the study at the 10-month lab visit (10-month data not reported here). Out of the 410 children originally recruited in infancy, 337 mother-child dyads (82.2%) provided 24-month, child effortful control data. Finally, 226 dyads (55.1%) provided 36-month child externalizing behavior data. Cohort 1 was not administered the Child Behavior Checklist (CBCL) as this cohort is three years older than cohorts 2 and 3, and this measure had not yet been added to the study protocol at the time of cohort 1’s 36-month lab visit. This means that 106 of the 184 children (57.6%) who did not provide 36-month externalizing behavior data are missing this data because they were not administered the CBCL. In sum, 195 mother-child dyads in this study have complete study data (i.e., no missing data), and the remaining 215 dyads completed at least one measure across the three timepoints measured. Little’s Missing Completely at Random (MCAR) test was used to confirm that the MCAR assumption was met [χ²(46) = 44.36, p = .541]. We used full information maximum likelihood (FIML) estimation to handle missing data (see Data Analytic Plan section below for more information) and participants with complete or partial data were included in the analyses. Using Chi-square tests for our categorical variables, there were no differences between families who did or did not participate at 24-months in terms of child sex [χ² (1, N = 410) = 0.518, p = .47] or child race [χ² (3, N = 410) = 1.873, p = .60]. Families who did or did not participate in the study at 36-months did not differ with respect to child age [χ² (1, N = 410) = 0.192, p = .66], but a difference was detected in terms of child race [χ² (3, N = 410) = 10.971, p = .01]. That is, there was a greater number of Black or African American children who did participate in the study at 36-months than did not. Using an independent samples t-test, we examined whether there were any differences between families who did or did not participate at 24-months as a function of 5-month infant RSA, maternal intrusiveness, infant orienting, or maternal education. No differences emerged (all ps > .05). We further evaluated whether there were any differences between families who did or did not participate at 36-months as a function of 5-month infant RSA, maternal intrusiveness, infant orienting, maternal education, or 24-month effortful control. Once again, no differences were detected (all ps > .05).

Procedure

Data were collected at both research locations using identical protocols. Research assistants from each location were trained together by the project’s principal investigator (final author) on protocol administration, data collection, and psychophysiological coding. To ensure that identical protocol administration was maintained between the two laboratories, the Blacksburg team periodically viewed video recordings and raw ECG files collected by the Greensboro team. To ensure that identical ECG processing criteria were maintained between laboratories, the Blacksburg team provided verification of artifact screening of processed ECG data collected by the Greensboro team. All maternal behavior coding was done by the Blacksburg team.

Upon arrival at the laboratory, research assistants greeted mothers and children. Mothers completed consent to participate forms and children provided verbal assent starting at the 24-month visit. Mothers had been mailed questionnaires about their child before each lab visit and brought completed forms with them to the lab. At all three lab visits, children completed a battery of cognitive and affective tasks (not included in this current study) while wearing ECG electrodes. Mothers and children also participated in several interactive tasks. The laboratory visit was video recorded for later behavioral coding by research assistants. For their participation at all three visits, mothers received monetary compensation and children received a small gift.

Measures

Maternal Intrusiveness at 5 Months

Mothers and infants completed two interaction tasks that lasted 2 minutes each at the 5-month visit. The first task was a toy play task, during which the experimenter asked mothers to play with their infants as they normally would with two age appropriate toys (rattle, keys) that were provided. The second task was a peek-a-boo task, during which experimenters asked mothers to play peek-a-boo with their infants as they normally would, using their hands or a provided cloth. The child was in an infant seat and mother was seated facing her infant. Both tasks were recorded and coded offline by trained research assistants. Examples of maternal intrusive behaviors in both tasks included failing to let infants set the pace of play or overwhelming/overstimulating the infants. Intrusive behavior was coded in 30-second epochs using a coding scheme adapted from previous work (Calkins et al., 2004) and was rated on a 4-point scale with a 1 indicating no evidence of intrusiveness and a 4 being consistent, high levels of intrusive behaviors across the epoch. Reliability coding was accomplished on 29.5% of the sample for the toy play task and on 28.8% of the sample for the peek-a-boo task. Interclass correlations (ICCs) were acceptable for both tasks (ICC toys = .85; ICC peek-a-boo = .88). Scores were summed across epochs and divided by the total number of epochs to create an average maternal intrusiveness score. Both tasks were then averaged together to get one maternal intrusiveness score.

Baseline RSA at 5 Months

RSA data was collected during a baseline period at the 5-month visit. RSA was extracted from continuous ECG data that were recorded for 1-minute while infants sat in their mother’s lap and watched a research assistant manipulate a toy containing brightly colored balls on a testing table 1.1 m in front of them. During the baseline period, mothers were instructed not to talk to or interact with their infant.

ECG data were recorded from two neonatal disposable electrodes using modified lead II alignment (right collarbone and lower left rib; Stern et al., 2001) and grounded on the scalp near electrode Fz. The cardiac electrical activity was amplified using a James Long Company Bioamp (Caroga Lake, NY) and band-passed from 0.1 to 100 Hz. The QRS complex was displayed on the acquisition computer monitor and digitized at 512 samples per second. The acquisition software was Snapshot-Snapstream (HEM Corporation, Southfield, MI) and the raw data were stored for later R-wave detection and analyses.

ECG data were examined and analyzed using IBI Analysis System software developed by the James Long Company. First, R waves were detected offline with a four-pass peak detection algorithm, resulting in a data file with onset times for each detected R wave. To edit ECG artifact, the ECG signal was viewed alongside tick marks representing the times of software-detected R waves. For undetected visible and obscured R waves, the tick marks were inserted manually. Movement artifact was designated by the absence of at least three consecutive R waves. These artifact-scored epochs were eliminated from all calculations.

The edited R wave was converted to heart period (i.e., time between heart beats). Spectral analysis was used to calculate high-frequency variability (i.e., RSA) in the heart period data, using a discrete Fourier transform with a 16-s Hanning window and 50% overlap. The frequency band for quantification of RSA at 5-months was 0.24-1.04 Hz (Bar-Haim et al., 2000). The RSA data were transformed using natural log to normalize the distribution. Data were considered missing if RSA values fell outside of an expected range of 1-10 (n = 1). See Table 1 for RSA descriptive statistics.

Table 1.

Descriptive Statistics and Correlation Matrix for the Primary Study Variables

	1	2	3	4	5	M (SD)	N	Min	Max
5-Months
1. Maternal Intrusiveness	--					1.338 (0.38)	376	1.00	3.25
2. Infant RSA	−.051	--				3.942 (1.14)	377	1.06	9.36
3. Infant Orienting	.059	.031	--			5.051 (0.55)	396	3.57	6.60
24-Months
4. Child Effortful Control	−.046	.009	.349**	--		4.437 (0.51)	337	2.89	5.85
36-Months
5. Child Externalizing Behavior	−.037	.008	−.099	−.386**	--	12.597 (8.09)	226	0.00	43.00

Note.

^*= p < .05,

^**= p < .01

Orienting Regulation at 5 Months

Infant orienting regulation was assessed via maternal report on the Infant Behavior Questionnaire- Revised (IBQ- R; Gartstein & Rothbart, 2003) at the 5-month visit. The IBQ- R contains 191 items assessing 14 dimensions of child temperament across a variety of behavioral and emotional contexts. Mothers reported on specific behaviors using a 7-point Likert scale with response options ranging from 1 (never) to 7 (always).

Of primary interest to the present study was the orienting regulation factor, which consists of subscales comprised of 12 items of the IBQ- R. These subscales reflected the infant’s tendency to exhibit pleasure during low-stimulating activities, enjoy being held by a caregiver, hold attention to objects for long periods of time, and be easily soothed by a caregiver. In our data, the orienting regulation factor had acceptable internal consistency (α = .81). See Table 1 for descriptive statistics on infant orienting sensitivity

Effortful Control at 24 Months

Child effortful control was assessed via maternal report on the Early Childhood Behavior Questionnaire (ECBQ; Putnam et al., 2006) at the 24-month visit. The ECBQ contains 201-items assessing 18 dimensions of child temperament across a variety of behavioral and emotional contexts. Mothers reported on specific behaviors using a 7-point Likert scale with response options ranging from 1 (never) to 7 (always).

Of primary interest to the present study was the effortful control factor, which consists of subscales comprised of 59 items of the ECBQ. These subscales reflected the child’s tendency to focus attention and inhibition to suppress dominant responses in favor of subdominant ones. In our data, the 59-item effortful control factor had acceptable internal consistency (α = .89). See Table 1 for descriptive statistics on child effortful control.

Externalizing Behaviors at 36 Months

Child externalizing behaviors were assessed via maternal report on the Child Behavior Checklist (CBCL; Achenbach & Rescorla, 2001) at the 36-month visit. The CBCL contains 99 items regarding child behaviors that mothers rated on a 3-point Likert scale with response options ranging from 0 (not true) to 2 (very true).

Of primary interest to the present study was the externalizing behaviors scale of the CBCL, which contains 24 items that assess a variety of externalizing behavior problems such as aggression and attention problems. The raw scores of the externalizing behaviors scale were used in our analyses at the recommendation of the authors of the measure (Achenbach & Rescorla, 2001). In our sample, the externalizing behaviors scale had excellent internal consistency (α = .92). See Table 1 for descriptive statistics on the externalizing behaviors scale of the CBCL.

Analytic Plan

Prior to analysis, descriptive statistics were examined and the data were inspected for outliers and normality of the distributions using IBM SPSS software (Version 27). No evidence of nonnormality was detected in the dataset as both skewness and kurtosis fell within an acceptable level (i.e., skewness ≤ 3 and kurtosis ≤ 10; Kline, 2011). Outliers were defined as values that were ±3SD of the mean, and the number of outliers present in the dataset are as follows: 5-month RSA (5 cases ≥ +3SD), 5-month maternal intrusiveness (6 cases ≥ +3SD), 24-month effortful control (1 cases ≤ −3SD ), 36-month externalizing behavior (3 cases ≥ +3SD), Our analyses were conducted with the outliers retained and with the outliers handled via Winsorization, which is an outlier management technique that replaces outliers with the next closest score (Salkind, 2010). Because no significant differences emerged in our findings as a function of outlier treatment, we present our results below with the outlier values retained in the dataset. Bivariate and biserial correlations were conducted to determine whether sample demographic factors (i.e., child sex, child race, and maternal education) were related to the primary variables in the current study. Child sex (coded as 0 = girl, 1 = boy) was associated with 24-month effortful control (r = −.19, p < .001) and 36-month externalizing behavior (r = .20, p = .003). Thus, girls in this sample were reported to have greater effortful control at 24-months and fewer externalizing problems at 36-months in comparison to boys. Maternal education was associated with 5-month RSA (r = −.12, p = .017), 5-month maternal intrusiveness (r = −.12, p = .018), and 36-month externalizing behavior (r = −.20, p = .004). This pattern of relations illustrates that more highly educated mothers in this sample had infants with lower baseline RSA and children with fewer externalizing problems. More highly educated mothers were also less intrusive on average during interactions with their infant. As a result, both child sex and maternal education were included as covariates in the main analyses. We also controlled for infant orienting at 5-months in our analyses to evaluate whether our primary infancy variables accounted for unique variance in 24-month effortful control, above and beyond any stability in children’s temperament-based self-regulation skills.

The hypothesized associations were tested using path modeling in Mplus (Version 8; Muthén & Muthén, 2017), following Hayes’ (2013) recommendations for testing moderated mediation. FIML was used to account for missing cases. This estimation method utilizes all available data and has been shown to produce less biased estimates and standard errors in comparison to other missing data techniques (e.g., listwise deletion or single imputation; Enders, 2001; Widaman, 2006). The following inferential and descriptive indices of model fit were measured in our study: Chi-squared (χ²) goodness of fit test, root mean square of approximation (RMSEA; values ≤ .08 are reflective of acceptable model fit; MacCallum et al., 1996), standardized root mean squared residual (SRMR; values ≤ .08 are reflective of acceptable model fit; Hu & Bentler, 1999), and confirmatory fit index (CFI; values ≥ .90 are reflective of acceptable model fit; Hu & Bentler, 1999). Bootstrapping was applied (n = 10,000) and the indirect effect was examined using bias-corrected 95% confidence intervals (CI; MacKinnon et al., 2004). Significant indirect effects are indicated by the absence of zero in the 95% confidence interval (Hayes, 2013). Maternal intrusiveness and infant RSA values were mean centered prior to creating the interaction term. Moderation was probed by estimating the conditional direct and indirect effects of maternal intrusiveness on effortful control and externalizing behavior at values of infant RSA that correspond to the mean (i.e., average baseline RSA), mean +1SD (i.e., high baseline RSA), and mean −1SD (i.e., low baseline RSA) to determine whether the observed effects are different from zero across levels of infant baseline RSA.

Results

Descriptive statistics and bivariate correlations between the primary variables using the observed dataset can be found in Table 1.

Moderated Mediation Model

We estimated a moderated mediation model to examine whether the hypothesized indirect effect of maternal intrusiveness (5-months) on early childhood externalizing behavior (36-months) through toddler effortful control (24-months) was conditional on levels of infant baseline RSA (5-months). Acceptable model fit was achieved, χ²(9) = 15.384, p = .081; RMSEA = .045 (90% CI [0.00, 0.08]); SRMR = .045; CFI = .931 (Figure 1).

Figure 1. Longitudinal Moderated Mediation Model.

Note. * = p < .05, ** = p < .01. Standardized estimates are presented. Solid lines represent statistically significant paths and dashed lines represent non-significant paths.

Infant RSA Moderates the Association Between Maternal Intrusiveness and Child Effortful Control

As expected, a significant direct effect was detected between child sex and 24-month effortful control (β = .184, SE = .052, p < .001), and between 5-month infant orienting and 24-month effortful control (β = .379 SE = .046, p < .001). In other words, girls exhibited greater effortful control at 24-months in comparison to boys, and infants with greater orienting/regulation skills demonstrated greater effortful control in toddlerhood. Neither 5-month infant RSA nor maternal intrusiveness directly predicted 24-month effortful control (ps > .05). However, the infant RSA by maternal intrusiveness interaction term was significant (β = −.131, SE = .060, p = .03). Together, these infant predictors and control variables accounted for 20.6% of the variance in child effortful control. To probe this interaction, we estimated the simple slope of maternal intrusiveness at values of infant RSA that correspond to average (mean), high (one standard deviation (SD) above the mean), and low (one standard deviation below the mean) baseline RSA levels. Among infants who exhibited high baseline RSA, maternal intrusiveness at 5-months was negatively associated with effortful control at 24-months (B = −0.11, SE = 0.05, p = .04; Figure 2). By contrast, the relation between maternal intrusiveness and child effortful control was not statistically significant for infants with average (B = −0.04, SE = 0.03, p = .22) or low (B = 0.03, SE = 0.04, p = .52) baseline RSA levels. Thus, the strength of the association between maternal intrusiveness and child effortful control was conditional upon level of infant baseline RSA.

Figure 2. Interaction of 5-Month Maternal Intrusiveness by Infant RSA Predicting 24-Month Child Effortful Control.

Note. * = p < .05.

Infant Baseline RSA Moderates the Indirect Effect of Maternal Intrusiveness on Child Externalizing Problems

Next, we examined the direct paths between the infant/toddler variables and children’s externalizing behavior problems. Child sex (β = −.136, SE = .063, p = .03), maternal education (β = .161, SE = .063, p = .01), and 24-month effortful control (β = −.372, SE = .070, p < .001; Figure 1) significantly predicted 36-month externalizing problems, accounting for 20.5% of the total variance. Maternal intrusiveness at 5-months did not directly predict 36-month child externalizing behavior (p > .05).

Finally, evidence of a conditional indirect effect was detected in the current model, [β = .403, SE = .225, 95% CI (0.065, 0.960)]. Once again, simple slopes were generated by estimating the indirect effect of maternal intrusiveness on early childhood externalizing behaviors via toddler effortful control at low, average, and high infant baseline RSA levels. The indirect effect of maternal intrusiveness on early childhood externalizing behavior via toddler effortful control was statistically significant for infants with high baseline RSA (β = .642, SE = .334, 95% CI [0.135, 1.448]). By contrast, the indirect effect was not significant for infants with average baseline RSA (β = .240, SE = .199, 95% CI [−0.126, 0.657]) nor for infants with low baseline RSA (β = −0.163, SE = .264, 95% CI [−0.723, 0.333]). The moderated mediation model was then visually depicted using a Johnson-Neyman plot, which best illustrates the conditional indirect effect across multiple values of infant RSA with a 95% CI (Figure 3). In support of our hypothesis, these findings demonstrate that the indirect effect of maternal intrusiveness on early childhood externalizing behaviors via toddler effortful control was significant only for infants with high baseline RSA values.

Figure 3. Johnson-Neyman Plot of the Conditional Indirect Effect.

Note. The vertical blue line represents the boundary of significance, whereby the indirect effect through child effortful control was significant only for children with infant baseline RSA values to the right of the vertical blue line.

Discussion

The purpose of our study was to identify child and maternal factors in infancy and toddlerhood that relate to externalizing behavior problems in early childhood. We demonstrated that there was a significant, indirect pathway from maternal intrusiveness during infancy to early childhood externalizing behaviors through toddler effortful control. However, the magnitude of this indirect effect varied as a function of infant parasympathetic nervous system functioning, as the indirect pathway from maternal intrusiveness to externalizing behaviors through effortful control was only significant for children who had high levels of baseline RSA. For infants with low or average baseline RSA, the indirect effect was not significant. These results were significant even after controlling for infant orienting sensitivity. Together, these findings highlight the importance of considering biological, psychological, and social factors longitudinally when examining the development of child outcomes, such as externalizing behaviors. These results are further discussed in light of previous work.

Baseline RSA as a Moderator of the Indirect Effect Between Maternal Intrusiveness and Externalizing Behavior in Early Childhood

Understanding the mechanism(s) that can account for the association between caregiving factors and externalizing behavior problems in early childhood is important to the development of early interventions aimed at improving outcomes for at-risk child populations. The role of toddler effortful control in accounting for the association between maternal intrusiveness during infancy and child externalizing behaviors among children with high baseline RSA was not surprising in our study. Indeed, previous work has demonstrated similar findings between other aspects of self-regulation, maternal parenting, and child outcomes across development. For instance, Smith & Day (2018) reported that higher levels of effortful control in toddler-aged boys predicted fewer externalizing behaviors in these children, and Gartstein et al. (2012) reported that 24-month-olds’ who demonstrated lower levels of effortful control exhibited more externalizing behaviors at 36-months (Gartstein et al., 2012). Furthermore, the association between warm parenting behaviors and fewer adolescent externalizing behaviors appears to be mediated by higher levels of late childhood effortful control (Eisenberg et al., 2005). We were therefore able to replicate this indirect pathway using a more complex developmental model with a sample of much younger children and their mothers.

Specifically, our results showed that the strength of the association between maternal intrusiveness and child effortful control differed as a function of infant baseline RSA, which supports the notion of differential susceptibility (Belsky, 2007; Belsky & Pleuss, 2009). Although maternal intrusiveness negatively related to child effortful control among infants with high RSA, this association was not significant for infants with average or low RSA. Consequently, the indirect effect of maternal intrusiveness on child externalizing behaviors via effortful control was only significant for children who exhibited high baseline RSA during infancy. It is important to note that infant baseline RSA was not correlated with toddler effortful control or child externalizing behaviors, and was weakly correlated with maternal intrusiveness. Taken together, this result suggests that adverse maternal behaviors do not uniformly relate to children’s self-regulation abilities after taking into consideration biologically-based individual differences in infancy; rather, maternal intrusiveness seems to be especially detrimental to child outcomes among infants with high levels of RSA at baseline. This means that refraining from engaging in highly controlling and/or disruptive behaviors during parent-child interactions may serve as a protective factor against the development of externalizing behaviors for infants with high baseline RSA.

Situating this finding in the early childhood literature is difficult due to the relative absence of literature examining RSA and intrusive parenting interaction effects in predicting children’s later self-regulation skills or externalizing behavior; however, our finding does align with previous work reporting that RSA moderates the association between caregiver factors more broadly and early childhood behavior/cognition. For example, Bagner and colleagues (2012) demonstrated that baseline RSA moderated the association between family-based interventions and child externalizing behaviors. That is, children with low levels of baseline RSA exhibited a significantly greater reduction in disruptive behavior following a parent-training intervention in comparison to children with high RSA levels. Likewise, Gueron-Sela et al. (2017) reported an interactive effect between child RSA and maternal sensitivity, such that sensitive parenting positively predicted children’s later executive functioning performance among children with low baseline RSA. As previously mentioned, the absence of intrusive behavior does not inherently imply that mothers are interacting with their child in a manner that is warm or responsive. As such, a complex relation may exist between caregiving behavior and child RSA whereby sensitive parenting may positively predict outcomes primarily for children with low baseline RSA, while intrusive parenting may negatively predict outcomes primarily for children with high baseline RSA. Additional research examining an array of maternal parenting behaviors in early childhood is needed to formally investigate this interpretation.

Finally, in the context of younger children and temperamental risk for developing externalizing behaviors, Morales and colleagues (2015) reported that lower baseline RSA moderated the association between higher levels of exuberance and more externalizing behaviors. These authors suggest that a low baseline RSA is maladaptive to children who might be already biologically primed to develop externalizing behaviors (e.g., have high levels of exuberance). As such, a high baseline RSA might be especially adaptive to children who are biologically and socially protected (e.g., have high levels of effortful control and less intrusive mothers) from developing externalizing behaviors. To summarize, our results demonstrate that baseline RSA and maternal intrusiveness during infancy interact with respect for optimal developmental outcomes, and that effortful control serves as a mechanism of the predictive association between maternal intrusive parenting and child externalizing behavior for infants with high baseline RSA.

The Role of Infant Orienting Regulation

As expected, infant orienting at 5-months was directedly related to toddler effortful control at 24-months, which replicates the existing developmental literature (Bridgett et al, 2011; Rothbart et al., 2011). In other words, children who scored high in temperament-based self-regulation skills in infancy were reported to have greater self-regulation abilities in toddlerhood, thereby demonstrating a degree of stability over time in early childhood. Of particular interest to our study, it’s important to note that the moderated mediation findings held when controlling for infant orienting at age 5-months. This means that even after accounting for children’s rudimentary self-regulation abilities in infancy, maternal intrusiveness still accounted for unique variance in toddler effortful control, which in turn predicted externalizing behavior among children with high baseline RSA. An in-depth examination of the direct and/or indirect effects of infant orienting on children’s externalizing problems was beyond the scope of the current study. Yet, as previous work has demonstrated conflicting findings regarding the association between infant orienting regulation and later behavioral problems (Gartstein et al,, 2012; Toffol et al., 2019), future research is necessary to further quantify this association as well as to explore whether the trajectory of early self-regulation is related to childhood socioemotional problems.

Strengths and Limitations

In addition to contributing to the literature base on the biosocial predictors of the development of child externalizing behaviors, our study had several strengths. First, we had a large community sample which allows our findings to be generalized to the larger population. Second, our study design was longitudinal in nature and utilized sophisticated modeling techniques, allowing us to examine the effects of social and self-regulatory predictors of childhood externalizing behaviors over time after controlling for child sex. Specifically, we were able to capture the effects of our predictors during infancy and toddlerhood on externalizing behaviors as children transitioned into a new developmental era. Past work has suggested that children make strides in self-regulation and social skills at this transitional time (Barnes et al., 2013; Bornstein et al., 2010), and thus our results help to identify some of these gains at this age. Our study also helps to contribute to the limited literature on the direct and indirect associations between parenting, child self-regulation, and child externalizing behaviors by providing evidence that baseline RSA moderates the indirect effect of maternal intrusiveness on child externalizing behaviors through child effortful control.

Despite clear strengths, our study also had some weaknesses. First, although we noted that we had a large community-based sample, our sample was lacking in racial and ethnic diversity. Our sample demographics are representative of the geographic regions from which mother-child dyads were recruited; however, this is a limitation in terms of generalizability as it is unclear whether these findings apply to more diverse mother-child samples. Individual studies in the past have pointed towards cultural and racial differences in maternal intrusiveness (e.g., Carlson & Harwood, 2003; Ispa et al., 2004; Whiteside et al., 2003). However, a recent meta-analysis suggested that there are more ethnic and regional/cultural similarities than there are differences with respect to parenting behaviors and child outcomes (Pinquart & Kauser, 2018). Related, we also did not account for developmental context in the maternal intrusiveness parenting behavior used in the current study. In our own work with these same mother-child dyads, for example, we have reported that maternal intrusiveness at age 5-months predicts more optimal executive function outcomes in toddlerhood, whereas maternal intrusiveness at age 10-months predicts less optimal executive function outcomes (Broomell et al., 2019). Given this research, replication of our current study is certainly warranted. Second, we only had data on externalizing behaviors at age 36-months. The CBCL preschool version is valid for use in children aged 1.5-5 years (Achenbach & Rescorla, 2001), so it would have been advantageous to have externalizing behaviors to control for at ages 18-months and 24-months, given previous work demonstrating the changes in the trajectory of such across early childhood (e.g., Yan et al., 2019). The CBCL did not enter the protocol for our study, however, until the 36-month visit. Third, our study focused on maternal characteristics and excluded the use of father characteristics as the larger longitudinal data from which this study was drawn only collected data on mothers. As previous research has suggested that paternal intrusiveness may also play a role in predicting child outcomes (Karreman et al., 2008), it would be beneficial to examine an array of early childhood social factors in future work. Fourth, our study was focused on the direct and indirect effects of intrusiveness caregiving measured in toddlerhood on early childhood externalizing behavior. These findings cannot speak to the association between these factors earlier in development, although it is possible that social factors present even in infancy may have a cascade effect on childhood behavioral outcomes. Finally, although we had a range of maternal intrusiveness in our sample, we did not have high levels overall. This may be due to our study’s restricted demographic diversity and/or our use of a short laboratory measure of maternal intrusiveness. In other words, because we used a laboratory-based task to assess maternal intrusiveness, it is possible that we were not able to capture each individual mother’s level of intrusiveness outside of an unfamiliar and observed setting, as it is unlikely that mothers and children are doing the same task on a day-to-day basis. Other measures, such as an unstructured free-play period or home observations, could might have yielded more of the day-to-day intrusive behaviors in the mothers in our sample.

Future Directions and Conclusion

Future research should first consider addressing the above limitations by recruiting a more diverse sample, collecting data on paternal intrusiveness, and incorporating additional measures of maternal intrusiveness into the protocol. Additionally, future work might consider incorporating measures of maternal personality into the model. Recent research has demonstrated that negative maternal personality traits appear to moderate the associations between child temperament and externalizing behavior problems in adolescents (Thartori et al., 2018). Others have suggested maternal Big-5 personality traits (conscientiousness, agreeableness, neuroticism, openness, and extraversion) can impact maternal parenting behaviors such as intrusiveness and positivity (Smith et al., 2007). Thus, it may be informative to incorporate measures of maternal personality into models with toddlers and young children, not only to see if findings are similar to those found in previous works, but also to see if the results of our study are replicated. Additionally, it may be useful for future research to examine how maternal intrusiveness, effortful control, and baseline RSA continue to work together to predict externalizing behaviors into later childhood and adolescence. Doing so would allow for an examination of the growth trajectories of externalizing behaviors across development and give us a better understanding of whether early social and biological factors continue to influence externalizing behavior problems beyond the early childhood years. A final avenue for future work might be to further examine the associations between infant orienting regulation and later regulatory capacities in predicting the trajectory of behavioral problems in childhood.

Overall, our findings demonstrate the importance of considering biological, psychological, and social factors in examining the development of externalizing behaviors in early childhood. Our study makes an important contribution to the literature, as it examines how maternal intrusiveness, toddler effortful control, and infant baseline RSA predict the onset of externalizing behaviors in early childhood.