Parent stress and trauma, autonomic responses, and negative child behaviors

Abstract

Cumulative stress and trauma in parents may alter autonomic function. Both may negatively impact child behaviors, however these links have not been well established. We tested hypotheses that parent stress and trauma are associated with and interact with altered autonomic function during the Toy Wait Task (TWT), an acute parent-child interaction challenge, to predict greater negative child behaviors. Sixty-eight parents and their 2-5 year old children were enrolled. More parent major and traumatic life events, and more parent recent life events coupled with increased heart rate and decreased heart rate variability (HRV), each related to more child disruptive/aggressive behavior. More major life and traumatic life events coupled with greater HRV predicted more child attention seeking behavior. Our novel approach to assessing parental life stress offers a unique perspective. Interventions mitigating parent stress and regulating physiological coping during parent-child interactions may both promote better parent health and improve child behavioral outcomes.

Introduction

Parenting specific stress, such as being overwhelmed with parenting responsibilities and an inability to provide emotional support, has been consistently associated with child developmental issues [1–3]. Evidence suggests that non-parenting related stressful life events may also contribute to problems surrounding child development, by prompting more parenting stress and maladaptive parenting behaviors [4, 5]. However, this link is understudied and, when it is, stress is often defined very broadly to either include any and all life events [6] or only one specific life stressor (e.g., marital strain [7, 8]). Stress and trauma throughout people’s lives are very common, with nearly one-half of Americans reporting frequent stress throughout their day, and parents experiencing multiple types of stressful life events [9, 10]. Therefore, it is important to determine the effects of overall cumulative stress as well as the impact of specific stressors on parenting stress response during parent-child interactions and on child behaviors.

Research on life stress has categorized stressful life events by: 1) major life events (MLE; non-life threatening but highly stressful experiences;), 2) traumatic life events (TLE; life threatening or violent stressors), 3) recent life events (RLE; highly stressful and/or traumatic within the past year), and 4) chronic subjective sense of being overwhelmed by specific negative (chronic) life events (CLE) [11–13]. CLE, such as work stress and lack of sleep, is associated with less attentive parent-child interactions and less positive parenting [14, 15]. MLE and TLE, such as interpersonal betrayal and war trauma, are similarly linked to parenting issues with a greater likelihood of insecure attachment style, poorer child well-being, and less parental responsiveness [16, 17]. Additionally, RLE, such as changes to romantic partnerships and ongoing consequences of parental alcohol use disorder, are correlated with more reported maternal stress, stricter parenting style [18], and greater internalizing symptoms including, higher negative emotions, sadness, depression, and anxiety, in the child [19]. Given the associations between these stressors, maladaptive parenting behaviors, and negative child consequences, examining the impact of parent life stress may provide one pathway to preventing negative outcomes.

The effects of stress and trauma have also been studied in the context of physiological responses and adaptations. Specifically, heightened heart rate and decreased heart rate variability (HRV), or the time and frequency between beats [20], have been associated with greater cumulative stress [21]. Time domains of HRV include root mean square successive differences between beats (RMSSD) and the standard deviation between NN intervals (SDNN). Daily, chronic, and total life stress have also been associated with lower daily SDNN and RMSSD values [22, 23]. Of note, acute (momentary) or provoked stress has also been related to HRV [21]. For instance, acute stress has been linked to lower RMSSD and higher heart rate, during both a group based social stressor for healthy individuals, [24] and in an acute trauma script paradigm amongst those with past trauma and trauma-related psychopathology [25]. Additionally, more RLE have been linked to lower RMSSD and higher heart rate during acute social stress [26]. This research has primarily been conducted in adults, and not specifically in parents, nor on the impact of stress on interactions with their young children. Stress dysregulation of the sympathetic nervous system has been associated with lower functioning of medial prefrontal brain regions involved in emotion regulation and behavioral control [27–29] and cardiovascular disease [30–32]. Therefore, heart rate and HRV stress reactivity may be important markers for emotion and behavioral regulation as well as later health consequences.

Stress and trauma in parents are negative experiences in a child’s environment that may contribute to factors underlying negative behavioral outcomes in young children [33–35]. Parents’ experiences of cumulative stress and trauma may impact acute and chronic physiological states (i.e., parent heart rate and HRV basal levels and acute responses) and serve as markers of emotional and behavioral dysregulation, as suggested by Thayer et al [27, 29], that may influence negative child behaviors. To examine this conceptual framework of the effects of parent stress and trauma on their physiological state and on child behavioral outcomes, and to address gaps in the literature surrounding the effects of non-parenting related stress, we recruited 68 parents and their young children ages 2-5 years old. These parent-child dyads participated in the Toy-Wait Task (TWT), a previously validated parent-child mildly challenging interaction task [36–38]. We assessed parent physiological autonomic responses prior to, during, and right after the task and conducted structured behavioral coding of video-taped recordings of observed child attention-seeking and aggressive behaviors during the TWT. Prior to the TWT, cumulative stressful and traumatic life events were assessed in parents using the well validated Cumulative Adversity Interview (CAI) [12]. This scale provides a lifetime index of stressful and traumatic events, specific subscales for MLE, TLE, and RLE, and a CLE subscale focused on subjective response of being overwhelmed by sustained chronic negative life events.

We hypothesized that greater parent cumulative stress will predict child attention seeking and aggressive behaviors and parent HR and HRV levels during the TWT. We further hypothesized that higher cumulative stress and trauma in parents will interact with altered parent HR and HRV levels as an average index of parents’ stress responsivity in this challenge task, to predict greater attention-seeking and aggressive child behaviors during the TWT. We did not have specific hypotheses for the 4 cumulative stress subscales (MLE, TLE, RLE, CLE) as previous literature has not tested the effects of different forms of stress in parents on child outcomes. Therefore, this work takes a novel approach by measuring cumulative stress and trauma, physiologic response in parents, and their acute impact on child behaviors.

Methods

Participants

The sample included 68 parent-child dyads recruited from a mid-size New England urban area and the surrounding communities via flyers and online advertising. The parent-child dyads included a diverse group of parents (34.69 ± 5.99 years old, 55.9% Caucasian) and their preschool aged children (22-69 months, 57.4% male). For additional demographic information see Table 1. Parents were recruited as part of a larger stress, nutrition, and healthy parenting intervention study for parents who were overweight or with obesity. Inclusion criteria were: (a) being 18+ years old, (b) having a BMI ≥ 27 kg/m² and (c) having children 2 to 5 years old. Exclusion criteria were: (a) Parents’ child had a confirmed diagnosis of developmental disorder (e.g. autism), (b) the parent was pregnant or had an acute, current medical illness or psychiatric illness precluding participation, or (c) parent was non-English speaking. All procedures were approved by the institutional IRB.

Table 1.

Demographics.

	Total Sample N = 68
Parent Gender (Female)	67 (98.5%)
Parent Age (Years)	34.69 ± 5.99 (21-49)
Parent Racea (Caucasian/African American/Other)	38 (55.9%)/17 (25%)/ 13 (19.1%)
Parent Ethnicity (Hispanic)	20 (29.4%)
Parent BMIb (kg/m2)	35.96 ± 5.88 (28.46-51.28)
Child Gender (Female)	29 (42.6%)
Child Age (Months)	42.1 ± 13.8 (22-69)
Chronic Stress Checklist
Major Life Events	2.97 ± 2.29 (0-8)
Traumatic Life Events	7.5 ± 4.38 (0-21)
Recent Life Events	2.65 ± 2.8 (0-12)
Chronic Stressors	14.74 ± 7.47 (0-33)
Heart Rate Variability Measures
Heart Rate	85.04 ± 10.54 (59.52 – 107)
SDNN	36.53 ± 14.19 (4.82-90.74)
RMSSD	28.52 ± 16.52 (2.8-111.59)

^aOther includes 1 who identified as American Indian or Alaska Native, 2 who identified as Asian, 1 who identified as Native Hawaiian or Other Pacific Islander, and 9 who did not identify with any of these racial options

^bData available on 65 participants.

Continuous variables have mean ± standard deviation and range. Categorical variables have a count (% of total). Heart rate variability measures are presented on average over the 5 timepoints. HR Baseline: 86.23 ± 11.48 TWT: 85.11 ± 11.29 Recovery 1: 85.24 ± 10.53 Recovery 2: 84.32 ± 10.96 Recovery 3: 83.31 ± 10.39; SDNN Baseline: 37.94 ± 14.6 TWT: 37.34 ± 16.09 Recovery 1: 37.54 ± 14.64 Recovery 2: 34.37 ± 14.33 Recovery 3: 34.57 ± 14.9; RMSSD Baseline: 29.54 ± 16.47 TWT: 29.18 ± 18.01 Recovery 1: 29.11 ± 16.76 Recovery 2: 26.99 ± 16.77 Recovery 3: 27.02 ± 17.46; SDNN = standard deviation of the N to N interval; RMSSD = root mean square of successive differences between normal heart beats.

Procedures

Potential participants were screened by phone to ensure initial eligibility criteria were met. Participants then came in for 2-3 intake appointments. After signing informed consent, study measures were captured, including perceived parent stress and stress history. Parents brought in their children to complete the TWT, an acute child challenge paradigm.

Materials

Parent Cumulative Stress and Trauma

Parent stress was measured using the Cumulative Adversity Index (CAI [12]). The CAI is a 140-item structured interview that is comprised of 4 subscales: Major Life Events (MLE), Traumatic Life Events (TLE), Recent Life Events (RLE), and Chronic Life Events (CLE). The MLE subscale contains 11 items across the lifespan (e.g., growing up in an orphanage). The TLE subscale includes 41 items across the lifespan (e.g., being sexually assaulted or witnessing a shooting). The RLE subscale contains 34 items (e.g., serious illness and major financial issues happening to oneself, partner, parent, or close person within the past 12 months). Items in these three subscales were asked as “yes/no” questions and endorsed items were summed per scale. Finally, the CLE subscale contained 63 items with specific events in life for which parents reported a perceived sense of being overwhelmed by the event as rated on a 3-point scale (rated as “not true,” “somewhat true,” or “very true”). Items included feeling increased pressure and expectations as well dissatisfaction in relationships. Items endorsed at “somewhat” or “very” true were summed to create the final score. Overall, the yes/no subscales had good reliability (total α: 0.85, individual α’s = 0.68-0.75) as did the CLE subscale (α = 0.84).

Toy Wait Task

Parents and their children were taken to a room with 2 discrete video cameras set up to record dyad interactions during the task. Parents were outfitted with a Firstbeat BodyGuard 2 (BG2) device to record heart rate and HRV. A 5-minute baseline period of passive physiology was assessed. The researcher then showed the child a new toy and told them, “This is a surprise for you, but you have to wait 5 minutes until you can open it.” The toy was then put on a shelf, out of reach though visible to the child. The child was told they could play with an old, broken toy in the meantime. Parents were informed of the task during the consent process and told they would be filling out paperwork during this time. Parents were instructed that if the child asks about them about the new toy to tell them, “There are other toys to play with” and to continue to focus on filling out the forms. The researcher left the room and the task proceeded for 5 minutes. At the end, the researcher returned to the room and gave the child the new toy. Parent heart rate and HRV were monitored throughout the task and for 15 more minutes post-task as a “recovery” period. This task has been previously validated as a mildly challenging interaction task for preschool aged children [36–38]. See Figure 1 for an overview of the task.

Figure 1. Schematic of Toy Wait Task.

Heart rate variability was captured for a 25 minute period, broken into 5 minute segments. The segments are Baseline (pre task), Reactivity (during the task), and 3 points of recovery. Parent Child Interaction Coded - observed behavioral coding of child aggressive disruptive behavior and bids to toy (coders were oblivious to stress and trauma levels of parents).

Child Aggressive and Attention-Seeking Behaviors.

Parent and child interactions during the task were video-recorded for later coding of child behaviors. Observed frequency with which the child exhibited Disruptive Aggressive Behavior and a Bid to the Toy (a measure of attention-seeking behavior) were coded using TWT behavioral coding manual (based on:[39–42]) by 2 research assistants and a supervisory coordinator. All individuals underwent specific training (by TMC, who developed the behavioral coding system) in coding of these observed child behaviors.

Bids to Toy (the measure of attention-seeking behavior) was defined by the child looking at, gesturing to, talking to the parent about, or attempting to reach the new toy. The number of bids made across the 5-minute task period was measured as a count variable. Child Disruptive Aggressive Behavior was defined as speaking or acting aggressively (e.g., physically trying to destroy a piece of furniture, kicking, hitting, saying rude or mean things) or in a disruptive manner (e.g., engaging in behaviors that an adult would typically stop, such as grabbing the physiological equipment). The frequency of Child Disruptive Aggressive behavior was rated on a scale from 1 (none) to high (5) across the 5-minute task period. Interrater reliability between behavior coders on Child Disruptive Aggressive Behavior was assessed for 36 videos (53%) and was good (intraclass correlation coefficient = 0.79). Bids to Toy was a count variable determined by one well-trained research assistant coder. Any questions on either coding scale were reviewed by the supervisory coordinator.

Heart Rate Variability

Parents were outfitted with a Firstbeat Body Guard 2 heart rate variability (HRV) monitoring device prior to the TWT. Both heart rate and HRV measures (including standard deviation between NN intervals [SDNN] and root mean square successive differences between beats [RMSSD]) were automatically recorded and provided by Firstbeat software. Recordings took place beginning from the 5-minute period prior to the task until 15 minutes after the task ended. This 25-minute timeframe was broken into five segments (baseline, reactivity during the task, and 3 recovery periods of 5 minutes each). We calculated the average heart rate and HRV for each measure in each of the time blocks based on the values provided by Firstbeat. Measuring HRV during the task allowed us to “time lock” parent HRV stress responsivity in an acutely challenging situation.

Data Analysis

All data analysis was conducted in R v. 3.6.1 [43]. Separate linear mixed effects models with a random intercept (lmerTest [44]) examined if heart rate and the HRV measures changed over the course of the task and if there were significant effects of parent stress as measured by the four CAI subscales (MLE, TLE, RLE, CLE) on heart rate and HRV. Parent stress was z-scored to remove “micro” multicollinearity [45] and the 5 within-subjects timepoints were tested as a categorical variable to allow for non-linear effects of time. The stress by HRV models predicting child behavior were assessed using ordinary least squares regression. In these interactive models, HRV was taken as the average heart rate, SDNN, and RMSSD (separate models) over the 25-minute timeframe. The continuous stress and HRV measures were z-scored. Child age (in months), child sex, parent race, and parent ethnicity were included as covariates in all models.

Results

Parent Stress and Child Aggressive Behaviors

MLE.

Models containing MLE and MLE interactions with HR and HRV measures significantly predicted child aggressive behaviors. This was driven by the MLE main effect (F(1,59) = 4.96, p < 0.03; F(1,59) = 4.17, p < 0.046; F(1,59) = 4.77, p < 0.033, in HR, SDNN, and RMSSD models respectively), indicating that greater MLE in the parent’s life was associated with greater disruptive and aggressive behavior by the child during the TWT. Figure 2a displays these main effects. No MLE by HR or HRV interactions were significant (p’s > 0.14).

Figure 2. Major and Traumatic life Events Predict Greater Child Disruptive Aggressive Behavior.

Main effects are shown for demonstration based on heart rate variability being removed from the models. *p<0.05 Panel A: Greater number of Parent Major Life Events is associated with more Child Disruptive Aggressive Behavior during an acute challenge task (b = 0.251, t = 2.25, p < 0.028). The minimum stress score (0) was associated with score of 0.96 while a maximum score endorsed (8) was associated with a predicted score of 1.84. Panel B: Greater number of Parent Traumatic Life Events is associated with more Child Disruptive Aggressive Behavior during an acute challenge task (b = 0.239, t = 2.2, p < 0.032). The minimum stress score (0) was associated with score of 0.88 while a maximum score endorsed (21) was associated with a predicted score of 2.03.

TLE.

Similar effects were seen for TLE in models containing TLE and TLE interactions with HR and HRV models. The TLE main effect was significant (F(1,59) = 4.75, p < 0.034; F(1,59) = 4.73, p < 0.034; F(1,59) = 4.52, p < 0.038, in HR, SDNN, and RMSSD models respectively), suggesting that greater TLE in the parent’s life was related to higher disruptive and aggressive behavior by the child during the TWT. Figure 2b displays these main effects. No TLE by HR or HRV interactions were significant (p’s > 0.07).

Parent Stress by HRV Interactions Predicting Child Disruptive Aggressive Behavior

RLE.

The interaction between RLE x HR and HRV measures significantly predicted Child Disruptive Aggressive Behavior (F(1,59) = 4.49, p < 0.039; F(1,59) = 7.16, p < 0.01; F(1,59) = 7.89, p < 0.007, in HR, SDNN, and RMSSD models respectively). Further breakdown of the interactions revealed that greater RLE was most strongly related to more Child Disruptive Aggressive Behavior at heightened heart rate (+1 SD heart rate: b = 0.4, p < 0.038; Mean heart rate: b = 0.17, p > 0.14; −1 SD heart rate: b = −0.06, p > 0.61). An opposite pattern emerged with the RLE by SDNN interactions (+1 SD SDNN: b = 0.01, p > 0.89; Mean SDNN: b = 0.22, p > 0.06; −1 SD SDNN: b = 0.43, p < 0.014) and RLE by RMSSD interactions (+1 SD RMSSD: b = 0.04, p > 0.71; Mean RMSSD: b = 0.23, p > 0.059; −1 SD RMSSD: b = 0.41, p < 0.012) suggesting that greater RLE was related to more Child Disruptive Aggressive Behavior when SDNN and RMSSD values were low. Figure 3 displays these interaction effects.

Figure 3. Recent Life Events Interact with Heart Rate Variability Measures to Predict Child Disruptive Aggressive Behavior.

*p<0.05 Panel A: There is a significant Recent Life Events by parent heart rate interaction (b = 0.23, t = 2.12, p < 0.039). Greater parent Recent Life Events is associated with more Child Disruptive Aggressive Behavior when the parent heart rate is approximately +1 SD (b = 0.40, t = 2.12, p < 0.039). This effect is non-significant at approximately the mean (b = 0.17, t = 1.46, p > 0.14) and −1 SD (b = −0.06, t = −0.51, p > 0.61) heart rate. Panel B: There is a significant Recent Life Events by SDNN interaction (b = −0.21, t = −2.68, p < 0.01). Greater parent Recent Life Events is associated with more Child Disruptive Aggressive Behavior when the parent has experienced below the average (approximately −1 SD) (b = 0.43, t = 2.56, p < 0.014) and trending at mean SDNN (b = 0.22, t = 1.89, p <0.064). This effect is non-significant at +1 SD SDNN (b = 0.01, t = 0.14, p > 0.89). Panel C: There is a significant Recent Life Events by RMSSD interaction (b = −0.19, t = −2.81, p < 0.007). Greater parent Recent Life Events is associated with more Child Disruptive Aggressive Behavior when the parent below the average (approximately −1 SD) RMSSD (b = 0.41, t = 2.60, p < 0.012) and trending at the mean (b = 0.23, t = 1.92, p < 0.06). This effect is non-significant at +1 SD RMSSD (b = 0.04, t = 0.37, p > 0.71).

All CLE by HRV models predicting Child Aggressive Behaviors were non-significant (model p’s > 0.06).

Parent Stress and Trauma by HRV Interactions Predicting Child Bids to Toy

MLE.

The MLE by SDNN (F(1,59) = 6.53, p < 0.014) and MLE by RMSSD (F(1,59) = 4.42, p < 0.04) models, but not MLE by heart rate model (p > 0.78) showed interactions predicting the number of Child Bids to Toy. Breaking down the MLE by SDNN interaction revealed that greater MLE was related to fewer Bids to Toy at low SDNN levels (+1 SD SDNN: b = 0.27, p > 0.17; Mean SDNN: b = −0.12, p > 0.45; −1 SD SDNN: b = −0.5, p < 0.038). A similar pattern emerged for the MLE by RMSSD interaction (+1 SD RMSSD: b = 0.31, p > 0.16; Mean RMSSD: b = −0.05, p > 0.74; −1 SD RMSSD: b = −0.41, p > 0.08) (Figure 4A–4B).

Figure 4. Major and Traumatic Life Events Interact with Heart Rate Variability Measures to Predict Child’s Bids to Toy.

*p<0.05 Panel A: A significant Major Life Events by SDNN interaction was found predicting Child Bids to Toy (b = 0.38, t = 2.56, p < 0.014). Greater parent Major Life Events is associated with fewer Child Bids to Toy when the parent has below the average (approximately −1 SD) SDNN (b = −0.50, t = −2.13, p < 0.038). This effect is non-significant at approximately mean (b = −0.12, t = −0.76, p > 0.45) and +1 SD SDNN (b = 0.27, t = 1.39, p > 0.17). Panel B: Greater parent Major Life Events is associated with non-significant Child Bids to Toy at SDNN levels approximately +1 SD (b = 0.31, t = 1.39, p > 0.16), approximately mean (b = −0.05, t = −0.33, p > 0.74) and −1 SD (b = −0.41, t = −1.75, p > 0.08). However, there is a significant interaction (b = 0.36, t = 2.1, p < 0.04). Panel C: A significant Traumatic Life Events by SDNN interaction was found predicting Child Bids to Toy (b = 0.46, t = 2.72, p < 0.009). Greater parent Traumatic Life Events is associated with more Child Bids to Toy when the parent has experienced above the average (approximately +1 SD) SDNN (b = 0.54, t = 2.42, p < 0.019). This effect is non-significant at approximately mean (b = 0.09, t = 0.59, p > 0.55) and −1 SD SDNN (b = −0.37, t = −1.69, p > 0.09).

TLE.

The TLE by SDNN (F(1,59) = 7.38, p < 0.009) model also showed a significant interaction predicting the number of Child Bids to Toy. Breaking down the TLE by SDNN interaction revealed that greater TLE was linked to more Child Bids to Toy at high SDNN levels (+1 SD SDNN: b = 0.54, p < 0.019; Mean SDNN: b = 0.09, p > 0.55; −1 SD SDNN: b = −0.37, p > 0.097) (Figure 4C). TLE by RMSSD model was significant overall (F(8,59) = 2.23, p < 0.038). However, no TLE, RMSSD, or interaction terms were significant predictors (p > 0.22) in that model. TLE by heart rate did not reach significance (p > 0.055).

RLE and CLE.

All Recent Life Event and Chronic Stress by HRV models predicting Bids to Toy were either non-significant or did not have significant stress, HRV, or interaction terms.

HR and HRV During the TWT.

HR, SDNN, and RMSSD were significantly different over the course of the TWT (HR: F(4, 261) = 4.27, p < 0.003; SDNN: F(4, 261) = 4.93, p < 0.001; RMSSD: F(4, 261 = 3.03, p < 0.019). Further examination revealed that for all 3 measures, baseline and TWT time periods did not differ from one another (p’s > 0.06), but the 2 latter recovery timepoints were significantly less than baseline and the task period (p’s < 0.04, except the second recovery timepoint did not significantly differ from the task for HR (p > 0.52). None of the four stress measures had direct effects on heart rate, SDNN, or RMSSD either on average across the Toy Wait Task period (main effects p’s > 0.07) nor differentially per time period (p’s > 0.28).

Discussion

Current findings indicate that parents’ cumulative experiences of stressors and trauma directly impact negative child behaviors. Additionally, parents’ stress and trauma impact negative child behaviors differently depending on the parents’ autonomic responses during an acute parent-child challenge task. We found that more major and traumatic life events in the parents’ lives were each significantly associated with greater child aggressive behaviors. More recent life events paired with greater heart rate or lower HRV (SDNN or RMSSD) was associated with greater child aggressive behaviors as well. More major life and traumatic life events also interacted with greater HRV (SDNN and RMSSD, and only SDNN, respectively) to predict more child attention seeking (as measured by bids to toy).

The finding that greater major and traumatic life events previously in the parent’s life were associated with more child aggressive behavior matched our hypotheses and previous literature on the impact of maternal trauma and family stress [4, 46]. Furthermore, these findings are supported by the Process Model of parenting, suggesting that parental environment and stress impacts child behavior [47]. While recent parent stressors did not reflect more aggressive child behavior directly, recent stress coupled with heightened heart rate or lower HRV response did significantly affect more child aggressive behavior during the task. To our knowledge, no previous research has examined whether acute parent HRV (in contrast to the child’s own HRV [48]) is predictive of child aggression. However, given the previous reported association between parent stress and greater negative child behavior [49, 50], the reflection of limited HRV on the experience of stress [21], and theories surrounding sympathetic nervous system response and cognitions [27], our finding fits with the larger literature and our hypotheses.

Somewhat surprisingly, more major life events coupled with more of either measure of HRV (RMSSD or SDNN) was associated with more attention seeking behavior. Traumatic events coupled with greater SDNN reflected a similar pattern. While this effect may initially seem somewhat counterintuitive as greater HRV is not typically associated with negative outcomes, it is important to note that greater parent HRV may suggest that the child’s distress is not engaging the parent [51]. Such a lack of engagement paired with the negative effects of parent stress may be sufficient to encourage this type of “acting out” behavior in the child. This is particularly concerning as attention seeking behavior can be a precursor to child anger [39, 52].

Embedded in our work is the assumption that the Toy Wait Task elicited an acute physiological response from the parents throughout the task. Indeed, our findings confirmed this with heightened heart rate and lower HRV at baseline (in anticipation of the task) and during the task, relative to latter recovery timepoints. This finding is congruent with previous work using the TWT, thereby further validating the acute mildly challenging aspect of this paradigm [36–38].

While we found a number of significant effects to support our main hypotheses, the hypotheses that parent stress and trauma would be directly associated with parent heart rate and HRV during the TWT were not supported. Prior work suggests that stress and trauma would be associated with increased heart rate and more limited HRV [53, 54]. The incongruence between our findings and those of some others may stem from differences in how stress was defined. For instance, previous work often looks at HRV reactivity as a reflection of an acute stressor such as cognitively challenging [53] or socially stressful [54] tasks. When examining life stress and HRV, one meta-analysis saw no such direct association unless psychopathology was present [55]. Additionally, acute stress reactivity may not always be associated with negative outcomes. One study found that an increase in heart rate in parents to an acute parent-child stressor was associated with more positive parenting [56], suggesting a link between physiology and positive outcomes, that was not in the purview of our analyses. We sought to draw conclusions about how past or ongoing stress, separate from the challenge task, may contribute to HRV response. This expanded measurement technique to capture life stress adds a unique perspective. Therefore, even though we had a lack of significant findings between parent life stress and trauma and HRV, our work represents an important contribution to the literature as it offers an additional perspective amongst a generally healthy parent population.

Although this study had many strengths, including a unique experimental design to test how parent cumulative stress and trauma may affect young children’s behaviors in an acutely challenging situation, there were a few limitations as well. First, our parent sample was predominantly mothers. Known sex differences exist in how men and women experience and report on stressful life events [57, 58]. Therefore, future work should study fathers and male caretakers of children as well to see if the impact of stress and trauma on parent-child interactions varies. Second, our sample specifically focused on parents with obesity or high overweight BMIs due to the subsequent intervention they were enrolled in. While not of primary interest here, it is worth nothing that increased BMI has been associated with dysregulation in some measures of HRV [59, 60]. Future work should examine what impact this might have in interaction with parental stress with a cohort that includes the full BMI spectrum. Finally, while our goal was to measure the influence of stress and trauma that the parent experienced on child behavior, this does not preclude the effects that the child’s stress may have had on their own behavior. For instance, some of the events experienced by the parents may have been experienced by the child as well, shaping child reaction in an acutely challenging situation.

Despite these limitations, the current study reports on novel findings that cumulative major and traumatic life events in the parent’s life were associated with more negative child behaviors observed during a mildly challenging task. Specifically, more major and traumatic life events was associated with greater child disruptive/aggressive behavior outright, and, coupled with changes to HRV, was associated with more child attention seeking behavior. Recent life events combined with greater heart rate and limited HRV was also linked to more child disruptive/aggressive behavior. These findings suggest that managing and intervening with parents’ coping of their cumulative stress and trauma experiences may be an important pathway to preventing negative child responses when the child is faced with a challenge themselves. Providing coping techniques, such as increasing mindfulness and establishing networks of parental support, to improve both negative feelings of stress and diminish autonomic response when watching their child in potential distress may also be beneficial. Future work would benefit from additional exploration of these pathways in mitigating ill effects on child development.

Summary

Despite links between parenting stress and subsequent child developmental problems, the impact of non-parenting related stress on children has been understudied. We have taken a novel approach by assessing major, traumatic, recent, and chronic cumulative life stressors in parents and connecting it both to parents’ autonomic activity and to child behavior during a parent-child interaction task. We recruited 68 parent-child dyads (2-5 year old children). Children were asked to wait briefly for a new toy (Toy Wait Task; TWT) with their parents present. Parent autonomic function (heart rate and heart rate variability; HRV) and negative child behaviors during the task were recorded. Stressful life events for the parents were recorded via interview in the Cumulative Adversity Index (CAI). Results indicated that more major and traumatic life events for the parents were associated with more child disruptive/aggressive behavior during the task. Additionally, more recent life events and an increase in heart rate or decrease in HRV in parents was also associated with child disruptive/aggressive behavior. More major and traumatic life events for the parents was associated with more child attention seeking behavior. No direct significant relationship between parent stress and autonomic activity was observed. Our findings suggest that many different parent stressors, not just those that are parenting related, can impact child behavior. This is particularly salient as an increase in negative child behaviors can manifest into later developmental disorders. Therefore, interventions that aid parents in coping with stress and improve physiological duress, such as increasing mindfulness and establishing networks of interpersonal support, may ameliorate some of the negative impacts on both parents and their children.