Infant frontal EEG asymmetry moderates the association between maternal behavior and toddler negative affectivity

Abstract

Relatively little work has examined potential interactions between child intrinsic factors and extrinsic environmental factors in the development of negative affect in early life. This work is important because high levels of early negative affectivity have been associated with difficulties in later childhood adjustment. We examined associations between infant frontal electroencephalogram (EEG), maternal parenting behaviors, and children’s negative affect across the first two years of life. Infant baseline frontal EEG asymmetry was measured at 5 months; maternal sensitivity and intrusiveness were observed during mother-child interaction at 5 and 24 months; and mothers provided reports of toddler negative affect at 24 months. Results indicated that maternal sensitive behaviors at 5 months were associated with less negative affect at 24 months, but only for infants with left frontal EEG asymmetry. Similarly, maternal sensitive behaviors at 24 months were associated with less toddler negative affect at 24 months, but only for infants with left frontal EEG asymmetry. In contrast, maternal intrusive behaviors at 5- and 24-months were associated with greater toddler negative affect, but only for infants with right frontal EEG asymmetry at 5-months. Findings suggest that levels of negative affect in toddlers may be at least partially a result of interactions between children’s own early neurophysiological functioning and maternal behavior during everyday interactions with children in the first two years of life.

1. Introduction

Temperament plays an important role in social interactions and social functioning as well as in a wide range of child outcomes, including early school-related success and both maladaptive and adaptive patterns of behavior (Diaz et al., 2017; Eisenberg, et al., 1996; Hernandez et al., 2017; Keogh, 2003; Rothbart, Ahadi, & Hershey, 1994; Rothbart & Bates, 2006). One aspect of temperament in particular, negative affectivity measured in toddlerhood, has been consistently associated with difficulties in later childhood adjustment (Lemery, Essex, & Smider, 2002). Negative affectivity is a temperament trait characterized by a predisposition toward anger, frustration, fear, and sadness (Rothbart et al., 2006). Negative affect, in general, has been shown to peak in infancy and early toddlerhood and then gradually decrease across early childhood as children become better at regulating their own emotions. Indeed, an important goal of early development is control over one’s own level of affective arousal and an increasing ability to use strategies for modulation of arousal in a wide range of novel and potentially threatening stimuli or environments typically achieved in the second year (Fox, Schmidt, Calkins, Rubin, & Coplan, 1996). Children who are not successful in these early stages may be embarking on a pathway to more negative or maladaptive outcomes, such as internalizing and externalizing disorders (Calkins, Dedmon, Gill, Lomax, & Johnson, 2002; Murphy, Shephard, Eisenberg, & Fabes, 2004). Thus, understanding factors that predict higher levels of negative emotion in toddlers is vital for identifying children at risk for maladaptive pathways that have the potential to interfere with future success and well-being. We focus here on intrinsic factors in infancy and extrinsic factors in infancy and toddlerhood as predictors of toddler negative affectivity.

1.1. Maternal Caregiving

Caregiving behavior has been strongly implicated in the development of the regulation of emotion, and particularly negative emotion, in infancy and early toddlerhood (see Calkins & Hill, 2007 for a review). The development of regulating capacities for arousal and affect occurs early in development and begins via a reliance on caregivers for regulation guidance then becomes increasingly more independent to achieve self-regulation (Kopp, 1982; Swingler, Perry, Calkins & Bell, 2014). Given that infants’ first achievements of regulated states occur almost exclusively in the context of social interactions and routines with their caregiver (Calkins, 2008), it is in the context of these caregiver-infant interactions that children initially begin to experience emotion regulation as well as learn about regulating their own emotional experiences (Kopp, 1989; Swingler et al., 2014).

Current models of caregiving assert that sensitive caregivers who recognize infant distress and coping signals (e.g., fussing, crying, or looking away), and respond to them in a consistent and appropriate manner, have infants who develop a sense that they can help to regulate the dyadic exchange and, eventually, themselves (Tronick, 2007). Maternal sensitivity is the extent to which mothers are aware of and can read their children’s verbal and nonverbal cues and, more importantly, appropriately respond to those cues (Spieker, Oxford, Kelly, Nelson, & Fleming, 2012). When infants are able to elicit caregiver interaction through the use of affective signals, and experience appropriate and sensitive responding to these signals, they begin to develop self-regulatory skills needed to cope with longer periods of frustrating or emotionally challenging episodes later in development (Tronick, 2007). A body of research has demonstrated that mothers who are more sensitive have infants who show more attention seeking behaviors (e.g., looking to her, smiling, reaching) and display greater positive affect, and less negative affect, across a variety of contexts (Kogan & Carter, 1996; Mesman, van Ijzendoorn, & Bakermans-Kranenburg, 2009; Mills-Koonce et al., 2007; Perry, Mackler, Calkins, & Keane, 2014). In contrast, negative maternal behavior, over-control, and intrusiveness across a variety of contexts is negatively related to toddlers’ self-regulatory skills during stressful or frustrating situations (Calkins, Smith, Gill, & Johnson, 1998) and predicts higher overall levels of child negativity (Paulussen-Hoogeboom, Stams, Hermanns, & Peetsma, 2007). Taken together, these data suggest that children of more sensitive mothers may experience less distress and negativity early in development and develop more effective self-regulatory skills in later toddlerhood and early childhood. The result of this may be a general overall pattern of greater negative affect, distress, and frustration present across development for children experiencing less sensitive parenting.

1.2. Infant Neurophysiology

In spite of the growing evidence for the role of sensitive parenting behaviors for shaping infants’ experiences, regulation of arousal, and affective behavior (Calkins & Hill, 2007), the infant’s own biological functioning also contributes to this process (Swingler et al., 2014). Specifically, developmental psychophysiological work highlights the significance of physiological processes and functioning for the development of early patterns of affective and regulatory behavior and for the prediction of later emotional functioning (Fox et al., 1995; Smith & Bell, 2010). Activation within the frontal cortex in particular, as measured by electroencephalogram (EEG) at rest, has been associated with predictable patterns of emotional responding and affective tendencies (see Coan & Allen, 2004; Reznik, & Allen, 2018 for a review), and has been proposed to underlie the more sophisticated and effortful regulation of emotion that children engage in when emotionally and behaviorally aroused (Davidson, 1992; Davidson & Fox, 1982; Fox, 1991, 1994). Fox’s (1994) model of differential activation of the left and right frontal cortices states that frontal EEG asymmetry can index individual differences in emotion reactivity and current emotional states (Davidson & Fox,1989; Dawson et al, 1992). Classifying children’s frontal asymmetry based on positive or negative asymmetry scores has demonstrated high internal consistency for indicating individual differences in infant emotion reactivity (Allen, Coan, & Nazarian, 2004; Coan & Allen, 2004; Fox et al., 1992; Schmidt, 2008; Smith & Bell 2010).

In addition, the categorization into right frontal or left frontal asymmetry is thought to reflect an individual’s inherent bias to behaviorally respond in certain ways (Fox, Henderson, Rubin, Calkins, & Schmidt, 2001). For instance, right frontal EEG asymmetry during a rest or baseline period has been consistently associated with a greater tendency to exhibit behaviors facilitating withdrawal from novel or stressful stimuli or situations such as negative affect (e.g., sadness, fear), as well as the expression of withdrawal types of behaviors (Buss et al., 2003; Calkins, Fox, & Marshall, 1996). In contrast, left frontal EEG Asymmetry during rest is associated with a tendency to demonstrate behaviors facilitating approach behaviors, such as approach motivated anger (Harmon-Jones, 2004) as well as the expression of certain positive emotions (Fox, 1991). These patterns of emotional behavior associations with EEG asymmetry group membership are evident as early as the first year of life (Buss, et al., 2003; Diaz & Bell, 2012; Fox, 1994; Smith & Bell, 2010; Smith, Diaz, Day, & Bell, 2016).

One possible explanation proposed for frontal EEG asymmetry associations with emotional responding and affective tendencies is that intrinsic approach-withdrawal tendencies act as a potential bias for an individual’s interactions with their environment, leading them to preferentially respond in specific ways to stressful or unfamiliar situations (Henderson, Fox, & Rubin, 2001). These inherent biases may then influence later behavioral functioning across a variety of contexts such that infants exhibiting right frontal EEG asymmetry at rest compared to left frontal asymmetry may have fewer adaptive strategies available and, thus, exhibit patterns of behavior which undermine the control of negative affect (Henderson et al., 2001; Smith et al., 2016).

1.3. Interactions between Maternal Behaviors and Infant Frontal EEG Asymmetry

Depending on individual patterns of neurophysiological functioning, environmental factors such as maternal behavior that influence the development of emotion expression and regulation may not function in the same way for all children. Recent developmental theory and resiliency literature has suggested that children’s intrinsic biological characteristics may impact the way in which environmental influences relate to developmental outcomes (Belsky & Pluess, 2009) and emphasize the importance of bioregulatory mechanisms in the relation between the home environment and child outcomes (Luthar, Cicchetti, and Becker, 2000). Previous theoretical work suggests that the development of self-regulation of negative emotion seems to hinge on both the infant’s own capacity for utilizing necessary regulatory strategies and the parent’s sensitivity in meeting the regulatory needs of the infant (Calkins, 1994; Swingler et al., 2014). Thus, examining both maternal behavior and infant neurophysiology may allow for a greater understanding of extrinsic factors from the home environment that interact with children’s intrinsic physiological characteristics to shape emotional development.

However, relatively little work has examined potential interactions between child intrinsic and extrinsic environmental factors on the development of negativity and negative affect in early life. Previous work examining functioning in other physiological systems, such as cardiac regulation, suggests that individual differences in physiological regulation may serve as a buffer against non-supportive parenting in preschool aged children (Perry, Calkins, Nelson, Leerkes, & Marcovitch, 2012). Integral to this study, there is evidence that young infants’ EEG asymmetry interacts with maternal behaviors to predict observed infant regulation strategy use in response to a challenging task at 5-months (Swingler et al., 2014). Swingler and colleagues (2014) reported that greater maternal sensitivity during a mother-child interaction task predicted greater infant use of distraction during an arm restraint procedure for infants with left frontal EEG asymmetry at rest. Moreover, greater maternal sensitivity predicted greater observed infant negative reactions during the arm restraint procedure, but only for infants with resting right frontal asymmetry. These findings provide support for the idea that maternal sensitivity and infant neurophysiological functioning may interact to predict regulatory behavior and reactivity in the first year of life.

Moreover, when examining interactions between intrinsic and extrinsic factors influencing the development and regulation of negative affectivity in toddlerhood, it may be especially important to examine the period before the rapid change in regulatory abilities occurs (from about 6 to 12-months; Kopp, 1982 & 1989; Posner & Rothbart, 1998; Stifter, & Braungart, 1995) as well as concurrently in toddlerhood, when most children have achieved a degree of independent self-regulation of emotion (Feldman, 2009). It may be that early infant physiology has a stronger influence on relations between maternal behavior and toddler negative affect development due to the rapid development in emotional responding and regulation, as well as the predominant dependence on caregivers as external sources of regulation of negative affect, that both occur in infancy. By 24 months of age, however, toddlers no longer rely as heavily on external sources of regulation of negative affect. Therefore, the moderating role of infant physiology on maternal behaviors predicting negative affectivity in toddlers may be distinct and smaller in magnitude. Thus, measuring resting EEG during early infancy may yield important information regarding the infant’s capacity for regulatory strategies and how this may moderate the influence of maternal behaviors on the pattern of negative affectivity that emerges in later toddlerhood.

The current study examined the influence of maternal behaviors at 5 months and 24 months and infant baseline EEG asymmetry at 5 months, as well as interactions between these two factors, on toddler negative affectivity. Specifically, we hypothesized that infants with right frontal EEG asymmetry exposed to early intrusive maternal behaviors would demonstrate greater negative affect at the toddler time point. Right frontal infants, however would benefit from early sensitive maternal interactions and demonstrate less negativity in toddlerhood compared to infants with left frontal asymmetry. Since left frontal activation is associated with more positive behavior and tendency to approach and engage (Fox, 1991; Fox & Davidson 1994), it was hypothesized that infants who exhibited left frontal EEG asymmetry would exhibit less negative reactivity irrespective of early maternal behaviors. We sought to extend the current literature on the development of negative affectivity by considering the role of both early and concurrent maternal behavior on toddler negative affectivity, as well as the possible moderating role of early infant EEG asymmetry on this relationship. Lastly, we expected to find similar relations among infant EEG asymmetry, toddler negative affect, and maternal sensitivity and intrusiveness observed in toddlerhood.

2. Method

2.1. Participants

Infants and mothers were part of an ongoing longitudinal examination of cognition and emotion from infancy through childhood. This report focuses on the research lab visits when children were 5 and 24 months of age. As part of our longitudinal study, 410 infants (209 girls, 201 boys; 27 Hispanic, 383 non-Hispanic; 318 Caucasian, 56 African American, 34 Multi-Racial, two Asian) were recruited by two research locations (BLINDED FOR REVIEW, BLINDED FOR REVIEW), with each location recruiting approximately half of the total sample. For parents who reported educational information (404 mothers, 392 fathers), 99% of mothers and 97% of fathers graduated from high school (6% and 7% had a technical degree, 42% and 31% had a bachelor’s degree, and 22% and 24% had a graduate degree, respectively). Mothers and fathers were approximately 29 and 32 years old (SD = 5.59 and 6.78 years), respectively, at the child’s birth. Demographics of the sample reflect those of the populations where the study was conducted. Infants at both sites were recruited via commercial mailing lists, newspaper birth announcements, and word of mouth. The same families were contacted when the children were 24 months old and all analyses were conducted with the sample of 320 families contributing data when the child was 24 months. Attrition was mainly due to families moving out of the local regions. Mothers were paid $50 for each visit and children received a small toy.

Data were collected in both research locations using identical protocols. Research teams were trained together by the last author on protocol administration, as well as on behavioral and psychophysiological coding. To ensure that identical protocol administration was maintained between the labs, the (BLINDED FOR REVIEW) site team periodically viewed video recordings and psychophysiology files collected by the (BLINDED FOR REVIEW) lab and provided verification of artifact screening for psychophysiology data collected and coded by the (BLINDED FOR REVIEW) lab. The (BLINDED FOR REVIEW) team coded all mother behavioral data collected by both labs.

2.2. Procedures

Upon arrival to the research lab for the 5-month visit, mother and infant were greeted by a research assistant who explained the study procedures and obtained signed consent. After a brief warm-up period, the infant sat on the mother’s lap and was distracted with toys while research assistants placed an EEG cap on the infant’s head. Baseline EEG was recorded for 60 seconds while the infant sat and watched an experimenter manipulate a toy with colorful bouncing balls on top of the testing table 1.1m away from the infant. This procedure has been shown to quiet infants and reduce both eye and gross motor movements thus allowing infants to better tolerate the EEG cap and minimize muscle artifact for EEG recording (Bell, 2001, 2012; Bernier, Calkins, & Bell, 2016; Cuevas & Bell, 2011). Mothers were instructed not to move or talk to infants during the baseline recording After the baseline EEG recording, a series of cognitive, emotion, and mother-infant interaction tasks were administered, with EEG recorded continuously throughout. This report focuses exclusively on data from the baseline EEG recording and the mother-infant interaction tasks.

2.3. Five-Month EEG Recordings

EEG recordings were made from 16 left and right scalp sites [frontal pole (Fp1, Fp2), frontal (F3, F4, F7, F8), central (C3, C4), temporal (T7, T8), parietal (P3, P4, P7, P8), and occipital (O1, O2)]. All electrodes were referenced to Cz during the recordings. The recordings were obtained using a stretch cap (Electro-Cap, Inc.; Eaton, OH; E1-series cap) with tin electrodes in the 10/20 system pattern. Our EEG methodology has been published elsewhere (e.g., BLINDED FOR REVIEW).

Power was computed for the dominant frequency for infants, the 6–9 Hz developmental alpha frequency band, and expressed as mean square microvolts (Bell & Fox, 1994; Marshall, Bar-Haim, & Fox, 2002). Data were then transformed using the natural log (ln) to normalize the distribution. Frontal EEG asymmetry scores at 6–9 Hz during a baseline recording, especially those associated with F3/F4 medial frontal scalp locations, have been used as evidence of the associations between brain electrical activity and later observed emotion-related behaviors (Fox, Henderson, Rubin, Calkins, & Schmidt, 2001; Davidson & Fox, 1989). Data were analyzed with a discrete Fourier transform (DFT) using a Hanning window of 1-sec width and 50% overlap. Across infants, the mean number of DFT windows was 76.97 (SD = 23.01), which is approximately 38.49 sec of artifact-free EEG data. Researchers have determined that 20 sec of artifact-free resting baseline EEG is sufficient for the reliable calculation of EEG power in 10- to 13-year-old children (Gasser, Bächer, & Steinberg, 1985). In a recent study, Chow et al. (under review) reported that Cronbach’s alpha for EEG power at various electrodes ranges between .95 and .98 for 60 sec of baseline EEG in 8- to 13-year-old children. We are unaware of any reports of reliability assessments in infant resting EEG. It is typical for infant EEG researchers to obtain approximately 60 seconds of resting baseline recording (e.g., Marshall et al., 2002; Orekhova, Stroganova, & Posikera, 2001; Thatcher, Walker, & Giudice, 1987).

We calculated baseline frontal EEG asymmetry scores by subtracting ln left frontal power at the F3 scalp electrode from ln right frontal power at the F4 scalp electrode (ln right – ln left = frontal EEG asymmetry). In the emotion EEG literature, brain activation is indicated by lower EEG power values in the alpha frequency band. Thus, negative asymmetry scores reflect greater relative activation in the right hemisphere compared to the left and vice versa. Because classifying children’s frontal asymmetry based on positive or negative baseline asymmetry scores has been shown to be a reliable and valid measure of asymmetry (Fox et al., 2001; Smith & Bell, 2010), asymmetry groups were formed based on frontal baseline asymmetry scores. No infants had symmetry and thus all were classifiable into either a left frontal asymmetry group (n = 172) or a right frontal asymmetry group (n = 125). Thus, 297 of the 320 toddlers contributing data at 24 months had baseline EEG data from F3 and F4 scalp electrodes at the 5-month lab visit. Reasons for missing 5-month EEG data (n = 23) included not accepting the EEG cap, low-quality EEG recordings, or EEG equipment failure.

2.4. Maternal Behaviors

2.4.1. Five months.

For the mother-child interaction task at 5-months infants were placed in an infant seat and mothers were instructed to interact with their infants as they normally would at home using two simple infant toys (keys, rattle) during a 2-minute free-play session. Mothers were asked not to remove their infants from the infant seat if possible. After giving the mother instructions, the experimenter left the room during the interaction. Sessions were videotaped and later coded.

Maternal behaviors were scored using a coding scheme previously reported (BLINDED FOR REVIEW) and adapted from Calkins and colleagues (Calkins, Hungerford, & Dedmon, 2004) for assessing maternal interactive style during toy play. Maternal interactive style is a multifaceted construct that includes sensitivity to infant’s signals, facilitation of attention, lack of intrusive behavior, appropriate physical stimulation, appropriate vocal stimulation/excitement, appropriate pacing of activities, and praise. Two dimensions of maternal behaviors are used in the current study. The first dimension, maternal sensitivity, included the extent to which the mother’s interactions were appropriate and coordinate with the infant’s behavior. This variable included such behaviors as well timed and synchronous responses, acknowledgment of infant’s affect, appropriate levels of stimulation, soothing, picking up on the infant’s interests, and attention-directing behaviors. The second dimension, maternal intrusiveness, was the extent to which the mother displayed over-controlling behavior or was focused on her own agenda, ignoring the infant’s behavior or cues. Maternal intrusiveness included such behaviors as failing to modulate the pace or intensity of her behavior when the infant withdrew or turned away, appearing to force toys or self on infant, and intrusive physical interactions (e.g. grabbing or manipulating the infant). Maternal behaviors were coded every 30 seconds using a 4-point scale (1=none, 4=high) for each dimension. Inter-rater reliabilities (intraclass correlations (ICC); over 25% coding overlap) were acceptable for both sensitivity (.74) and intrusiveness (.87). Table 1 provides summary scores for maternal behavior variables at 5 months.

Table 1.

Descriptive Statistics and Correlations among Key Variables

	1	2	3	4	5	6	7	8	9	10
1. child sex	-
2. maternal education	.06	-
3. maternal age	.03	.44**	-
4.5-mo maternal sensitivity	.05	.10	.12*	-
5.5-mo maternal intrusiveness	−.01	−.16**	−.19**	−.73**	-
6.24-mo maternal sensitivity	−.15**	23**	.18**	.19**	−.13*	-
7.24-mo maternal intrusiveness	.06	−.19**	−.18**	−.12*	.28**	−.57**	-
8.5-mo infant negative affect	.01	−.11**	−.27**	−.06	.16**	−.05	.15*	-
9. 24-mo toddler negative affect	.05	−.21**	−.21**	−.11*	.16**	−.20**	29**	.38**	-
10.5-mo infant EEG asymmetry	.03.	−.04	.002	.01	−.06	.05	−.06	−.05	.08	-
N	320	313	317	305	305	294	294	307	320	297
M	.49	2.64	29.72	3.42	1.39	3.32	1.40	2.97	3.07	.58
SD	.50	1.14	5.46	.55	.46	.50	.48	.64	.54	.49

Note: Child sex was coded as Male = 1 and Female = 0; 5-mo EEG asymmetry was coded Right Frontal = 0 and Left Frontal = 1; maternal education was coded from 0= Did not complete High School to 4 = Graduate School.

^*p < .05,

^**p < .01

2.4.2. Twenty-four months.

When the dyad returned at 24-months, mothers and toddlers sat at a child-sized table and mothers were asked to play with a set of puzzles of increasing difficulty with their child for 5 minutes. She was asked to play with the puzzles as she typically would at home. After giving the mother instructions, the experimenter left the room during the interaction.

As during infancy, maternal behaviors were coded during this interaction task (Smith, Calkins, Keane, Anastopoulos, & Shelton, 2004). Maternal behaviors, maternal sensitivity (promptly and appropriately responding to the child’s bids to her) and maternal intrusiveness (being too strict, demanding, or harsh considering the child’s behavior; exerting influence toward completion of the child’s activity; displaying a no-nonsense attitude; constantly guiding the child and enforcing a structured environment) were coded every 30 seconds using a 4-point scale (1=none, 4=high) for each dimension. Inter-rater reliabilities (ICC; over 20% coding overlap) were acceptable for both sensitivity (.92) and intrusiveness (.95). Table 1 provides summary scores for maternal behavior variables at 24 months. Maternal intrusiveness means demonstrate that this group of mothers showed relatively low levels of intrusiveness. The sample range, however, shows that there was variability in the intensity of intrusive behaviors (Table 1).

2.5. Temperament Questionnaires

2.5.1. Five months.

Prior to the laboratory visit, mothers were asked to complete the IBQ-revised (Gartstein & Rothbart, 2003), which measures 14 domains of child temperament. Mothers rated each item on a 7-point, Likert scale (1 = never, 4 = half the time, 7 = always). This questionnaire is a highly differentiated assessment of emotional and behavioral responses across a number of situations for ages 3- to-12-months. Although all IBQ-r temperament subscales were collected, the Negative Affect factor was of particular interest to the current study. As per Gartstein and Rothbart (2003), the subscales of sadness (e.g., When it was time for bed or a nap and your baby did not want to go, how often did s/he: whimper or sob?), distress (e.g., When placed on his/her back, how often did the baby fuss or protest?), fear (e.g., How often during the last week did the baby startle to a sudden or loud noise?), and loading negatively, falling reactivity (e.g., When rocking your baby, how often did s/he soothe immediately?) were combined to create a negative affectivity score (α = .72). Table 1 provides summary scores for Negative Affect at 5 months.

2.5.2. Twenty-four months.

Prior to the laboratory visit, mothers were asked to complete the ECBQ (Putnam, Gartstein, & Rothbart, 2006), which measures 18 domains of temperament in children 18- to 36-months of age. Following Putnam et al., (2006), a negative affectivity factor was created by combining 8 temperament scales (α = .78); discomfort (e.g., During everyday activities how often did your child seem to be disturbed by loud sounds?), fear (e.g., While in a public place, how often did your child, show fear when the caregiver stepped out of sight?), sadness (e.g., While having trouble completing a task, how often did your child become sad?), frustration (e.g., When s/he couldn’t find something to play with, how often did your child get angry?), motor activation (e.g., When playing alone, how often did your child, chew his/her lower lip?), perceptual sensitivity (e.g., During everyday activities, how often did your child notice low-pitched noises such as the air-conditioner, heater, or refrigerator running or starting up?) and shyness (e.g., When approached by an unfamiliar person in a public place, how often did your child cling to a parent) and loading negatively soothability (e.g., After getting a bump or scrape, how often did your child forget about it in a few minutes?). Table 1 provides summary scores for Negative Affect at 24 months.

3. Results

3.1. Preliminary Analyses

Descriptive statistics and preliminary correlations between all study variables are displayed in Table 1. Mothers displayed no difference in sensitivity or intrusiveness towards their male or female infants at 5-months; no differences in intrusiveness at 24-months; nor did mothers rate infants differently on negative affect at 24-months by sex (ps >.24). There were, however, sex differences in 24-month maternal sensitivity, with girls (M = 3.40, SD = .48) observed to receive more sensitivity from mothers than boys (M = 3.24, SD = .51; t(292) = 2.66, p = .008). EEG asymmetry group membership (left frontal, right frontal) at 5-months did not vary by infant sex, p > .46. However, parents of boys had higher education, p < .05.

Maternal education was negatively correlated with 5- and 24-month maternal intrusiveness and child 5- and 24-month negative affect, as well as positively correlated with 24-month maternal sensitivity (see Table 1). Mother’s age at her infant’s birth was positively correlated with 5- and 24-month maternal sensitivity and negatively correlated with 5- and 24-month maternal intrusiveness and the child’s 5- and 24-month negative affect (see Table 1).

Based on these correlations, infant sex and maternal education and age at infant birth were used as covariates in all subsequent analyses. Maternal ratings of infant 5-month negative affect were positively correlated with her ratings of toddler 24-month negative affect. Maternal behaviors were correlated with one another in the expected directions (i.e., negative correlations between sensitivity and intrusiveness; positive correlations between 5- and 24-month behaviors). Infant EEG asymmetry at 5 months was not correlated with maternal ratings of child negative affect at either 5 or 24 months.

3.2. Moderation Analyses

Regression analyses were performed in Mplus using full information maximum likelihood estimation with robust standard errors (MLR). Continuous variables were centered prior to conducting analyses. Centering was accomplished by subtracting the sample mean from all individual scores on the variable; this produced a revised sample mean of 0. This procedure reduced the multicollinearity between predictors and any interaction terms among them and facilitated the testing of simple slopes (Holmbeck, 2002). Guidelines from Holmbeck (1997) were used in the moderation analyses.

To test the hypotheses that the association between maternal behaviors (sensitivity, intrusiveness) and toddler negative affect is moderated by infant frontal EEG asymmetry (left frontal asymmetry, right frontal asymmetry), we computed two regression analyses: the first focused on maternal behaviors at the 5-month lab visit and the second focused on maternal behaviors at the 24-month lab visit. In each regression, we controlled for infant sex, maternal education, maternal age, and 5-month infant negative affect. Because Mplus does not provide an F test for regression analyses, regression models were compared against a baseline model in which all paths from the predictors to toddler negative affect were fixed to zero. As is the case for an F test, a significant χ2 for the baseline model indicates that the combined contribution of predictors is statistically significant (Muthén, & Muthén, 1998–2010). Significant interactions were examined in Mplus in accordance with the procedures outlined by Aiken and West (1991). A meta-analytic review of over 100 studies’ independent moderation effect sizes reported a median effect size value of only 0.2% suggesting that Cohen’s conventions (i.e. 2%, 18%, and 54% small, medium, and large effects sizes respectively) may need adjustment in behavioral research (Aguinis, Beaty, Boik, & Pierce, 2005; Rogers, 2002). Recently, Durand (2013) justified several scenarios in which a study’s effect size and sample size support the increase in Type 1 error if in conjunction with priori reasoning, theory, or previous research. Given the difficulty in detecting significant interaction terms in the social sciences where statistical power may be an issue, and given previous theoretical and empirical support, we followed recommendations from McClelland & Judd (1993) and Whisman & McClelland (2005) and probed interaction terms at p-values of .10 and lower in order to interpret interactions.

3.2.1. Main effect.

Regarding main effects, maternal intrusiveness positively predicted toddler negative affect in each regression analysis. The main effect of 5-month maternal intrusiveness was qualified by the interaction of maternal intrusiveness and infant EEG asymmetry (see Table 2 and described below); the main effect of 24-month maternal intrusiveness was also qualified by an interaction with infant EEG asymmetry (see Table 3). There was no main effect of either 5-month maternal sensitivity (see Table 2) or 24-month maternal sensitivity (see Table 3) on toddler negative affect. Finally, infant EEG asymmetry was only a significant (on trend level) predictor of toddler negative affect in the analysis with 5-month maternal behaviors, and this main effect was also qualified by the interaction with maternal behaviors (see Tables 2 & 3 and described below).

Table 2.

Predicting 24-mo Negative Affect from 5-mo Maternal Behaviors and 5-mo Infant Frontal EEG Asymmetry

	B (SE)	t	CI 95%
child sex	.05 (.05)	1.02	−.05 – .15
maternal education	−.15 (.06)	−2.50*	−.26 – −.03
maternal age at birth	−.05 (.06)	−.84	−.17 – .07
5-mo negative affect	.34 (.05)	6.40**	.24 – .45
5-mo EEG asymmetry	.09 (.05)	1.76++	−.01 – .20
5-mo maternal sensitivity	.21 (.13)	1.59	−.05 – .47
5-mo maternal intrusiveness	.24 (.12)	2.06*	.01 – .48
5-mo maternal sensitivity × 5-mo EEG asymmetry	−.28 (.12)	−2.36*	−.51 – −.05
5-mo maternal intrusiveness × 5-mo EEG asymmetry	−.21 (.11)	−1.92+	−.42 – .01

Notes: χ² (9) = 71.69^**, R² = .21^**; N = 320. Coefficients are standardized. Child sex was coded as Male = 1 and Female = 0; 5-mo EEG asymmetry was coded Right Frontal = 0 and Left Frontal = 1; maternal education was coded from 0= Did not complete High School to 4 = Graduate School

⁺⁺p = .078,

⁺p = .056,

^*p < .05,

^**p < .01

Table 3.

Predicting 24-mo Negative Affect from 24-mo Maternal Behaviors and 5-mo Infant Frontal EEG Asymmetry

	B (SE)	t	CI95%
child sex	.02 (.05)	.48	−.07 – .12
maternal education	−.12 (.06)	−2.06*	−.24 – −.01
maternal age at birth	−.04 (.06)	−.75	−.15 – .07
5-mo negative affect	.33 (.05)	6.51**	.23 – .43
5-mo EEG asymmetry	.09 (.05)	1.75++	−.01 – .19
24-mo maternal sensitivity	.16 (.14)	1.20	−.10 – .43
24-mo maternal intrusiveness	.35 (.12)	2.87**	.11 – .58
24-mo maternal sensitivity × 5-mo EEG asymmetry	−.26 (.12)	−2.09*	−.49 – −.02
24-mo maternal intrusiveness × 5-mo EEG asymmetry	−.20 (.11)	−1.77+	−.42 – .02

Notes: χ² (9) = 77.91^**, R2 = .25^**; N = 320. Coefficients are standardized. Child sex was coded as Male = 1 and Female = 0; 5-mo EEG asymmetry was coded Right Frontal = 0 and Left Frontal = 1; maternal education was coded from 0= Did not complete High School to 4 = Graduate School

⁺⁺p = .079,

⁺p = .077,

^*p < .05,

^**p < .01

3.2.2. Interactions.

The interaction effect in the model with 5-month maternal behaviors showed that infant EEG asymmetry significantly interacted with 5-month maternal sensitivity (see Table 2 and Figure 1, graphs A & B). Higher levels of 5-month maternal sensitivity was related to less toddler negative affect for children with left frontal asymmetry, B = −.14, p = .053, but not for children with right frontal asymmetry, B = .20, ns. The interaction of infant EEG asymmetry and 5-month maternal intrusiveness was significant at a trend level (p=.056). Higher levels of 5-month maternal intrusiveness were associated with more toddler negative affect for children with right frontal asymmetry, B = .28, p < .05, but not for children with left frontal asymmetry, B = −.05, ns.

Figure 1.

Maternal behavior predicting 24-mo Negative Affect as moderated by infant EEG asymmetry. Significant simple slopes are represented by an *. Infant EEG Asymmetry was examined at right and left frontal asymmetry.

The interaction effect in the model with 24-month maternal behaviors showed that infant EEG asymmetry interacted with 24-month maternal sensitivity (See Table 3 and Figure 1, graphs C & D). Just as with the model using 5-month maternal sensitivity, interaction of infant EEG asymmetry and 24-month maternal intrusiveness was significant at a trend level (p = .077). Higher levels of 24-month maternal sensitivity was related to less toddler negative affect for children with left frontal asymmetry, B = −.17, p < .05, but not for children with right frontal asymmetry, B = .18, ns. Additionally, higher levels of 5-month maternal intrusiveness was associated with more toddler negative affect for children with right frontal asymmetry, B = .39, p < .01, but not for children with left frontal asymmetry, B = .09, ns.¹

4. Discussion

To determine whether the relations between toddler negative affect at 24-months and maternal behaviors were associated with psychophysiological functioning early in development, we examined mother-child interactions at both 5 months and 24 months, with particular interest in the role of maternal sensitivity and intrusiveness on toddler negative affectivity and the moderating role of infant EEG asymmetry. Examining both maternal behavior and infant neurophysiology allows greater understanding for the way in which extrinsic factors from the environment, such as caregiver behaviors, may interact with children’s intrinsic biological characteristics, in this case EEG asymmetry, to shape children’s emotional development.

We expected that the association between maternal parenting behaviors and toddler’s negative affect would be moderated by children’s frontal EEG asymmetry measured at 5 months, which may provide a measure of child susceptibility toward greater negative emotionality. We found that children of mothers who demonstrated more sensitive behaviors towards their infants at 5 months were rated as exhibiting lower negative affect at 24 months; however, this was only true for infants who exhibited left frontal EEG asymmetry during a baseline recording at 5 months. The same interaction was found for sensitive behaviors during mother-child interaction at 24 months, with only left frontal asymmetry infants demonstrating a negative relation between toddler negative affect and maternal sensitivity at 24 months. We hypothesized that regardless of parenting, left frontal children would exhibit less negative reactivity. However, results only partially supported our hypothesis, as lower negativity was evident in the context of sensitive maternal behaviors. Responsive and stimulating parenting is considered a vital component of a child’s social environment that promotes optimal psychological and behavioral functioning (O’Brien, Johnson, & Anderson-Goetz, 1989). Our findings support that notion but also indicate that it is especially true for children with left frontal asymmetry who may be utilizing greater self-regulation of negative behaviors compared to children with right frontal asymmetry.

Although at a trend level, we found some evidence to support the hypothesis that children of mothers who demonstrated more intrusive behavior at 5 months and who also had right frontal EEG asymmetry at baseline exhibited greater negative affect as a toddler compared to those whose mother were less intrusive and had right frontal EEG asymmetry. Similarly, greater intrusive behaviors towards children at 24 months was also related to greater negative affect at 24 months only in infants who exhibited right frontal EEG asymmetry; however, again, this effect was only at a trend level and therefore should be interpreted with caution. Although we cannot offer conclusive interpretations for these effects, Calkins and Johnson (1998) reported that mothers who had a tendency to take control of tasks and do things for their toddlers had children who were more distressed and more likely to act out negatively. Indeed, intrusive mothers are characterized as not reading children’s cues and interests well and over-controlling child behavior during tasks and play situations. Moreover, Smith et al. (2015) provided evidence that infants with right frontal asymmetry were less able to engage in regulatory strategies that facilitate the expression of more appropriate behavior during social interactions and emotionally challenging or frustrating situations. These findings lend some support that this may be especially true for right frontal infants with intrusive mothers as their over controlled interactions with their mothers may further prevent opportunities for the development of their self-regulation. Repeated exposure to these interactions may lead children to experience and express more negative feelings as well as develop more negative interpersonal interaction styles (Kahen, Katz, & Goffman, 1994; Pettit et al., 1991; Tronick, 1989). Thus, intrusive mothers may set up a pattern of interaction that impedes children’s emotional development over time by creating a context in which children are unable to develop, use, or practice the ability to modulate their arousal (Kopp, 1989); thereby heightening their child’s risk for negative outcomes (Nachmias, Gunnar, Mangelsdorf, Parritz, & Buss, 1996; Rubin et al., 2009). However, further research is warranted.

It was also hypothesized that infants who exhibited right frontal asymmetry at 5 months would show less negative affect at 24 months when their mothers were more sensitive. Sensitive caregivers who limit their use of negative or controlling behavior might be more successful in helping children acquire and successfully use regulatory skills to self-regulate and manage negative emotions (Braungart-Rieker, Garwood, Powers, & Notaro, 1998; van den Boom, 1994). Although biological sensitivity to context theories (see Ellis & Boyce, 2008 for a review) would suggest that right frontal infants should gain the most from sensitive interactions, Swingler and colleagues (2014) reported that maternal sensitivity predicted more negative affect during an arm restraint procedure in infants with right frontal EEG asymmetry. It may be that within a context of sensitive caregiving a child’s frontal asymmetry may influence the access and the application of adaptive strategies for regulating negative affect in infancy and toddlerhood. It may be that right frontal asymmetry infants need more help and input from a sensitive caregiver to regulate their behaviors than left frontal infants and signaled this need with negative affect to their sensitive caregiver presumably experiencing her as an effective regulator (Swingler et al., 2014). Previous and current results, suggest that children’s underlying physiological processes may play an important role in the etiology and development of affective behavior (Fox et al., 1995; Smith & Bell, 2010).

The findings of our report must be interpreted in light of the limitations of our study. Although the IBQ-r and ECBQ are well-validated measures, only parent-report data were used to assess infant and toddler negative affect. Much of the developmental work examining temperament and temperament traits has historically used parent report to assess temperament characteristics in children. Parent reports have the advantage of utilizing the parent’s extensive observations of the child over a wide variety of situations. Furthermore, parent reports assess child temperament as it is exhibited in the context of child-caregiver interaction (Garstein & Rothbart, 2003), which is central to the focus of our study. However, future research could obtain both direct observation and parent report measures of negative affect, as multiple methods of measuring a construct can provide greater confidence and understanding of results. It is also unclear with the current data if the children are a product of the parent’s behavior or the parent’s behavior is a product of children’s negative behavior. Future studies should utilize methods that allow the examination of the unfolding of reciprocal behaviors between parent and child over more than two time points. Furthermore, although this study utilized infant asymmetry and its interaction with maternal behavior to provide longitudinal predictions of child emotion reactivity, it is quite possible that children’s asymmetry patterns may have changed from 5- to 24-months of age. Previous developmental work has suggested that frontal EEG asymmetry is a relatively stable trait measure (Allen, et al., 2004; Schmidt, 2008). Fox and colleagues (2001) demonstrated empirically that children who were highly inhibited (demonstrating internalizing behaviors such as high anxiety and social withdrawal) across the first four years of life displayed right frontal EEG asymmetry at 9-months and also at 14-months. Nevertheless, stability in early physiological tendencies as a predictor of later behavior is an important and critical next step in research on frontal EEG asymmetry, given the conceivable maturation and development of more complex approach and withdrawal behaviors over time (Achenbach & Rescorla, 2004; Smith et. al., 2016).

This study also focused on negative affect. High negative emotionality in early childhood has been associated with an array of later negative outcomes including greater risk for behavioral problems and poorer academic engagement (Diaz et al., 2017). However, neuroplasticity research in adults suggests that positive emotional states may initiate changes in the structure and function of the brain to promote more adaptive thoughts and behaviors (Garland & Howard, 2009). Future research examining associations between children’s psychophysiology and maternal parenting behaviors should also investigate both child and maternal positive affect, as positive affect not only plays a role in important child outcomes (Diaz et al., 2017; Eisenberg et al., 1993; 1996), but also may act as a buffer against the effects of stressful contexts as well as the effect of parenting stress on maternal sensitivity (Garland et al., 2010; Smith & Stephens, 2018). Lastly, the children in this sample were predominately Caucasian; thus, ethnic and cultural differences could not be thoroughly examined. This is an important avenue of research as studies have indicated differences in temperament across distinct ethnic group in various Eastern and Western cultures (Ahadi, Rothbart, & Ye, 1993).

In sum, developmental theory suggests that children’s intrinsic biological characteristics may impact the manner that environmental influences are related to developmental outcomes. Our data show that maternal behavior during interaction with children in infancy and toddlerhood interacts with infant frontal EEG asymmetry to predict ratings of toddler negative affect. This suggests that toddler negative emotion hinges on the child’s neurophysiological capacity for utilizing necessary regulatory strategies and the maternal sensitivity in meeting the regulatory needs of the infant. This work joins a growing body of researching demonstrating that work that considers both biological and environmental correlates of child outcomes gives a more informative picture of development than either factor considered in isolation. Funding: This work was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) [HD049878 and HD043057] to Martha Ann Bell. The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the NICHD or the National Institutes of Health.

Highlights.

• Infant EEG asymmetry, maternal parenting & children’s negative affect were examined

• Early maternal behaviors & EEG interacted when predicating later toddler affect

• Sensitive behaviors were related to less negative affect at left frontal asymmetry

• Intrusive behaviors were related to more negative affect at right frontal asymmetry

• Extrinsic and intrinsic factors shape children’s emotional development