Short Report: Parasympathetic Withdrawal Indexes Risk for Emotion Dysregulation in Children with Autism Spectrum Disorder

Abstract

Children with autism spectrum disorder (ASD) experience significant difficulties with emotion regulation. Theory and empirical evidence suggest substantial biological contributions to regulatory challenges, which may be related to core ASD symptoms. Respiratory sinus arrythmia (RSA) is a measure of parasympathetic nervous system activity that serves as a putative biomarker for emotion regulation. Higher baseline RSA and more RSA reactivity (parasympathetic withdrawal; RSA-R) in response to challenge appear adaptive in non-clinical populations, but existing evidence for children with ASD remains inconclusive. The current study examined correlates of observed emotion dysregulation in 61 children with ASD between the ages of 6 and 10 years, including ASD symptom levels as well as both baseline RSA and concurrent RSA reactivity. Consistent with previous research, ASD symptom level was significantly correlated with observed dysregulation whereas additional factors such as child IQ were not. Baseline RSA was unrelated to observed dysregulation, but higher RSA reactivity predicted concurrent dysregulation above and beyond the contribution of child ASD symptoms. Findings contribute to an emerging understanding of dysregulation in these children, raise questions about the utility of traditional baseline RSA measures for this population, and clarify the functional significance of RSA reactivity as a risk factor for emotion dysregulation in children with ASD.

Lay Summary:

Children with autism spectrum disorder (ASD) are at high risk for emotion dysregulation. This study identifies core ASD symptom level as important to understanding regulatory challenges, and suggests that certain biological arousal processes (e.g., the “freeing” of regulatory control to meet a challenge) may operate in a different way for this population as compared to what is generally observed for most children with neurotypical development.

Emotion regulation, or the ability to monitor, evaluate, and modulate emotion states to achieve a goal (Gross, 2008; Thompson, 1994), is a challenge for many children with autism spectrum disorder (ASD).¹ Heightened dysregulation, which can impair functioning and increase risk for co-occurring disorders (Mazefsky et al., 2012; White et al., 2014), may stem in part from underlying biological differences related to ASD (Mazefsky et al., 2012; White et al., 2014). Indeed, Fenning et al. (2018) identified core ASD symptom level as the strongest correlate of observed emotion dysregulation in 4- to 10-year-old children with ASD among select factors including child age, IQ, sympathetic arousal, and parental scaffolding. In addition to experiencing heightened dysregulation, children with ASD appear at a disadvantage for acquiring skills to assist with emotion management, including evidence of a marked delay in the internalization of parental co-regulatory supports (Baker et al., 2019; Fenning et al., 2018). Even when children with ASD utilize adaptive emotion regulation strategies, these efforts appear to function less effectively than they do for other children (Jahromi et al., 2012).

Autonomic nervous system measures can provide insight into arousal and regulatory processes (Benevides & Lane, 2015). The parasympathetic branch, which slows heart rate and reduces arousal through increased vagal output (Benevides & Lane, 2015), is commonly indexed by high-frequency heart-rate variability related to respiration, or respiratory sinus arrythmia (RSA). Baseline RSA (RSA-B) is often conceptualized as a peripheral marker of emotion regulation capacity (Beauchaine, 2015). Interpretations of withdrawal of RSA in response to challenge (RSA reactivity; RSA-R) are more complex and may depend upon population and measurement context (e.g., Beauchaine et al., 2019; Graziano & Derefinko, 2013).

In children with neurotypical development, higher resting RSA-B is generally associated with better emotion regulation (Beauchaine, 2015). Although some studies have demonstrated comparatively lower RSA-B in children with ASD, others have not (for reviews see Arora et al., 2021; Barbier et al., 2022; Benevides & Lane, 2015). Among children with ASD, higher RSA-B has been linked to better social-emotional skills (e.g., Arora et al., 2021; Bal et al., 2010) and fewer anxiety symptoms (Guy et al., 2014), but not parent-reported emotional control (Guy et al., 2014) or externalizing problems (Baker et al., 2020; Neuhaus et al., 2014). In adults with ASD (IQ > 80), greater high frequency heart rate variability, a measure analogous to RSA-B, has been associated with more adaptive self-reported regulatory strategies (Cai et al., 2019).

In community samples, greater RSA-R (more parasympathetic withdrawal) is often theorized to reflect autonomic flexibility, permitting enhanced engagement to meet contextual demands (Porges, 2007). However, in clinical populations at risk for dysregulation, higher RSA-R may represent a loss of regulatory control, allowing for an increased and potentially problematic role of sympathetic “fight or flight” responding (Beauchaine et al., 2012; 2015; Fenning et al., 2019). Emerging evidence suggests each of these functions in ASD, with high RSA-R indexing differential susceptibility to certain environmental and internal factors. For example, Baker et al. (2020) found that higher RSA-R in 6- to 10-year-old children with ASD predicted more externalizing problems in the context of high parental negativity, but fewer problems when parental negativity was low. Similarly, Fenning et al. (2019) found that higher RSA-R predicted externalizing problems differentially based upon children’s sympathetic nervous system responding. The relatively distal nature of the externalizing outcome used in these studies and evidence of moderated effects raise questions about how RSA-R relates to in-the-moment regulatory behavior.

Concurrent direct assessment of RSA and emotion regulation is needed to understand the functional significance of RSA in children with ASD. Utilizing the Baker et al. (2020) sample, we examined RSA-B and RSA-R in relation to emotion dysregulation observed during a challenge task from which RSA-R was also derived. First, we attempted to replicate Fenning et al. (2018) by testing whether ASD symptom level would emerge as the strongest correlate of dysregulation. Second, we explored RSA-B in relation to children’s dysregulation, though we anticipated potential non-significance given a lack of association between RSA-B and externalizing problems in this sample (Baker et al., 2020). Finally, we examined RSA-R in relation to concurrent, observed emotion dysregulation. Based upon the proposition that higher RSA-R in clinical groups is reflective of greater underlying dysregulation (Beauchaine, 2012; 2015), we hypothesized that greater RSA-R would correlate with more observed emotion dysregulation.

Method

Participants

Families of children with ASD aged 6 to 10 years were recruited from the community and local service providers. From an initial 77 children, 61 (74% male) provided sufficient RSA data. Missing RSA data occurred more frequently for males than females, and for children with greater ASD symptoms (details in Baker et al., 2020). The sample was diverse in intellectual ability (IQ range 47–121), ASD symptom levels, race/ethnicity (47% Hispanic, 33% White, non-Hispanic), and annual family income (median = US$50,000–70,000; scale: 0 = $0-$15,000/yr, 7 = >$95,000/yr; Table 1). Most primary caregivers were married (71%) and 3% were fathers.

Table 1.

Descriptive Statistics and Correlations among Variables of Interest (n=61)

	1	2	3	4	5	6	M (SD)
1. Age in Years	--						7.95 (1.48)
2. IQ	−.23+	--					80.02 (20.68)
3. ASD Symptom Level	−.06	−.10	--				7.21 (2.06)
4. RSA Baseline (RSA-B)	.36**	−.25+	.04	--			5.96 (1.17)
5. RSA Reactivity (RSA-R)	.02	.00	.06	.00	--		0.00a (0.73)
6. Family Income	.09	.09	.02	−.09	.10	--	4.95b (1.96)
7. Emotion Dysregulation	−.13	−.06	.28*	.00	.29*	.27*	1.39 (1.05)

Note: ASD = Autism spectrum disorder; RSA = Respiratory sinus arrhythmia

^aScores are regression residuals.

^bLikert scale: 0=$0–$15,000/yr, 7=>$95,000/yr.

⁺p<.10

^*p<.05

^**p<.01

Procedures

The institutional review board approved procedures. Parents consented and children assented. Electrodes were placed on the child’s lower ribs and right clavicle. An adjustment period was followed by a 3-minute baseline involving viewing of nature slides (Erath et al., 2016). The child then engaged in a 3-minute challenge task designed to elicit physiological arousal related to negative emotion (e.g., El-Sheikh, 2005). The task required tracing a star using an indirect, mirrored image; all children engaged in basic tracing. Parents reported demographic information.

Measures

ASD symptom level and IQ.

Community ASD diagnosis was confirmed by clinical best estimate procedures (see Baker et al., 2020), which included administration of the Autism Diagnostic Observation Schedule-2 (ADOS-2; Lord et al., 2012). The ADOS-2 comparison score indexed ASD symptom levels from 1 (minimal to no evidence) to 10 (high). Child IQ was estimated using the Stanford-Binet 5 ABIQ (Matthews et al., 2015; Roid, 2003).

Respiratory sinus arrhythmia (RSA).

As detailed in Baker et al. (2020), RSA was measured using a MindWare data acquisition system (MindWare Technologies, Inc., Gahanna, OH). Data were sampled at 500 Hz and RSA scores were quantified using spectral analysis (Berntson et al., 1997) as the natural log of the variance in heart period within age-adjusted respiratory frequency bands (Shader et al., 2018). RSA was expressed in units of ln(ms²). Possible artifacts were flagged by an algorithm that detected improbable interbeat intervals (Berntson et al., 1997); very few R-peaks required manual insertion or correction. RSA reactivity (RSA-R) was calculated as the residual of the regression of RSA during the challenge task on RSA during baseline (Burt & Obradović, 2013). Residualized change scores were multiplied by −1, with higher RSA-R scores indicating greater reduction (withdrawal) in RSA from baseline to challenge.

Emotion Dysregulation.

Child emotion dysregulation was coded by blind raters from the videotaped star-tracing challenge task using the Dysregulation Coding System (Baker et al., 2007), which considers overall lability and soothability as well as the appropriateness of the type, duration, and intensity of emotional expressions. The behavioral manifestation of poor regulatory control that impedes task engagement is also included. Dysregulation scores range from 0 (no evidence) to 4 (significant dysregulation). Psychometric support has been established in ASD (Baker et al., 2019; Fenning et al., 2018). Interrater reliability based upon a randomly selected 31% of cases was excellent (ICC = .91).

Results

Family income, ASD symptom level, and RSA-R demonstrated positive bivariate correlations with children’s observed dysregulation (Table 1). No other demographic or child variables considered (e.g., age, gender, race/ethnicity, IQ, medication status) were associated with dysregulation, nor was RSA-B. In a regression including family income, ASD symptom level, and RSA-R, the latter two remained uniquely, positively related to dysregulation, with the entire regression accounting for approximately one-fifth of the variance in emotion dysregulation, F(3,57)=4.95, p=.004 (Table 2).

Table 2.

Linear Regression Predicting Emotion Dysregulation

	Emotion Dysregulation
	B	SE	ß	t	p	R 2
Family Income	.13	.06	.23	1.95	.056	.21
ASD Symptom Level	.13	.06	.25	2.06	.044
RSA Reactivity (RSA-R)	.38	.17	.27	2.23	.030

Note: ASD = Autism spectrum disorder; RSA = Respiratory sinus arrhythmia.

Discussion

Current findings replicate Fenning et al. (2018) regarding the role of ASD symptoms in emotion dysregulation in children with ASD, augment evidence suggesting that traditionally-measured RSA-B may be a less useful dysregulation index in this population, and clarify that greater RSA-R in response to challenge likely indexes regulatory risk rather than beneficial autonomic flexibility in these children.

As was the case in the somewhat younger and smaller Fenning et al. (2018) sample, ASD symptom level was a significant predictor of children’s emotion dysregulation whereas child age and IQ were not. Whereas Fenning et al. (2018) did not find sympathetic arousal to correlate with concurrent dysregulation, the present study focused on RSA, which has been more closely tied to emotion regulation, and identified a significant association between RSA-R and observed dysregulation. In conjunction with evidence from Baker et al. (2019) indicating a substantial delay in the internalization of parental co-regulatory support for children with ASD, the current study underscores the apparent primacy of child-driven regulatory processes in this population.

Evidence that higher RSA-R in response to challenge may be a risk rather than a beneficial factor for regulation in children with ASD contrasts with findings from the general population. However, this is consistent with data from clinical samples, particularly those with externalizing problems, (Beauchaine, 2015), and aspects of emerging models in ASD (e.g., Condy et al., 2019; Patriquin et al., 2019). Moreover, interactions with other physiologic systems (e.g., sympathetic; Fenning et al., 2019) and environmental factors (e.g., parenting; Baker et al., 2020) may ultimately determine whether RSA-R vulnerability eventuates in functional impairments such as externalizing problems. Further consideration of these multi-system processes and additional child (e.g., temperament, executive functions) and environmental contributions will facilitate a more comprehensive characterization of emotion regulation in ASD.

The null RSA-B findings contrast with data for children without ASD but parallel previous results with this sample. As discussed in Baker et al. (2020), the ability to establish a true physiological baseline for children with ASD is challenging in a laboratory setting, as children may be emotionally activated resulting in baseline measures capturing a snapshot of an already dynamic process. Thus, certain “baseline” RSA measurements in this population may reflect some degree of reactivity, which may contribute to mixed findings regarding status group differences. Identifying avenues for differentiating physiological baseline from reactivity is an important future research endeavor.

Study limitations include those addressed in Baker et al. (2020) such as limited covariates and lack of respiration measurement (though age-adjusted ranges were applied; Shader et al., 2018). Our sample was comparatively diverse and relatively large for physiological studies in ASD, but sample size nonetheless prevented subgroup analyses. Generalization considerations remain given missing data were more common for children with greater ASD symptoms. Additionally, the unexpected positive correlation between family income and dysregulation is deserving of further investigation.

Findings suggest that heightened RSA-R represents a biological process corresponding to, and perhaps underlying, observed emotion dysregulation in children with ASD. Prior evidence that certain forms of external support such as low negative parenting may buffer risk for externalizing problems conferred by higher RSA-R (Baker et al., 2020) suggests the intriguing possibility of personalizing environmental support to manage the sequelae of psychophysiological risk in children with ASD.