Do Individual Differences in Early Affective and Cognitive Self-Regulation Predict Developmental Change in ADHD Symptoms From Preschool to Adolescence?

Emotion regulation

Findings from several recent studies support the long-standing theoretical perspective of poor emotion regulation as an important associated feature of ADHD (Karalunas et al., 2014; Maedgen & Carlson, 2000; Musser et al., 2013; Walcott & Landau, 2004). Importantly, poor emotion regulation skills have been demonstrated to be associated with ADHD symptoms independently of impairments in cool aspects of EF, perhaps constituting a separate self-regulatory pathway to the disorder (Forslund et al., 2016; Nigg et al., 2004; Sjowall et al., 2013). Of late, biological evidence for a role of an emotion regulatory deficit underlying the heterogeneity in ADHD was demonstrated in terms of a deviant parasympathetic mechanism during an emotion regulation task (Musser et al., 2011; Musser et al., 2013), results which were not explained by comorbid conduct problems. Although recent behavioral findings (Sjowall, Backman, & Thorell, 2015) coincide with Musser and colleagues’ (2013) conclusion, there are also studies suggesting the opposite (Maedgen & Carlson, 2000; Melnick & Hinshaw, 2000).

Moreover, there is support for particular emotions and their respective regulation to be differentially important for externalizing behavior problems (Martel et al., 2009; Nigg et al., 2004). Regulation of intense positive emotionality, which is associated with high approach tendencies and impulsivity, has been found important in relation to externalizing behavior problems (Nigg et al., 2004; Rydell, Berlin, & Bohlin, 2003; Rydell, Thorell, & Bohlin, 2007; Stifter, Putnam, & Jahromi, 2008) and as a specific risk factor for ADHD (Martel, 2009; Musser et al., 2013; Nigg et al., 2004). Importantly, this type of positive emotionality/exuberance does not refer to happiness or joy in general, but rather to high temperamental approach tendencies and unrestrained excitement. Rydell et al. (2003) speculate that the distinction between adaptive and problematic positive emotionality rests in the child’s capacity to regulate his or her emotions to achieve social acceptance.

Emotion regulation can be understood as composed of two related and interactive developmental processes: emotional reactivity and regulation. Although there is a discussion about the extent to which emotional reactivity and emotion regulation are separable, research supports considering them as dissociable constructs at the empirical level (Cole, Martin, & Dennis, 2004; Martel et al., 2009). It has been suggested that while high negative emotionality or reactivity may be linked to disruptive disorders in general, possibly accounting for their comorbidity, a deficiency in emotion regulation may be specific for ADHD (Martel, 2009). Thus, studying emotion regulation without accounting for individual differences in emotional reactivity might mask a specific link between the regulation-related process and ADHD.

Developmental Change in ADHD Symptoms

It has been demonstrated, although by relatively few studies, that the manifestations of ADHD symptoms change with age. Typically, levels of hyperactivity and impulsivity decline with development, whereas difficulties with inattention remain more stable over time (Hart, Lahey, Loeber, Applegate, & Frick, 1995; Molina et al., 2009; Willoughby, 2003). Despite this knowledge, very few studies have so far investigated symptom change in relation to early self-regulatory skills, which of course is particularly true for studies simultaneously examining different aspects of self-regulation in relation to ADHD-symptoms across long time periods.

Participants and Procedures

The sample has previously been described by (Brocki et al., 2007). A total of 72 children (60 boys and 12 girls) entered the longitudinal study. To obtain a sample of children scoring across the full range of ADHD symptom severity, we recruited one third of the children from local Child Health Care Centers. These children had been identified as being at risk for developing a disruptive disorder (i.e., ADHD and/or ODD) based on clinical assessment by trained psychologists specialized in screening children for developmental disorders. The remaining sample consisted of a group of typically developing children chosen to be proportionally matched in terms of sex and age. These children were randomly selected and recruited through the local birth register of Uppsala County, Sweden. No children in the study were at the time of testing receiving psychostimulant medication for ADHD. Parental educational status was high, with 59% of the mothers and 44% of the fathers having a university degree. The majority of parents were Swedish in origin (mothers 94% and fathers 86%), and the majority of children (90%) lived with both parents at T1. Furthermore, 21% of the children were initially rated by parents and teachers as having levels of ADHD symptoms that was considered as being “at-risk.” This means that these children were rated as having at least a score of 2 (i.e., often occurring symptoms) on six symptoms of either hyperactivity/impulsivity or inattention (see Table 2 for descriptive data on ADHD symptoms at T1 and T2). At follow up (T2), 20% of the children remaining in the study were reported by the parents to have received an ADHD diagnosis and 70% out of the diagnosed children had either received or were receiving some sort of behavioral or medical intervention.

At T1 (M = 5 years, 5 months; SD = .68), the children were tested individually at the Department of Psychology, Uppsala University. The tasks being analyzed in the current article were part of a larger battery designed to tap various EFs. To control for possible fatigue effects, the tasks were administered during two sessions over a period of 1 to 2 weeks. The tasks within each session were administered in a randomized order. Each session lasted approximately 1 hr with a break half-way through. At the end of each session, the children received a gift worth approximately €10. At T1, questionnaire data from the children’s parents and teachers were collected for 72 and 68 of the children, respectively. Participant attrition was due to lack of parental consent to contact the teacher (N = 1) and an unwillingness to complete the questionnaire by the teacher (N = 3).

At T2 (M = 13 years, 5 months; SD = .82), 63 children remained in the study (longitudinal gap M = 8 years, 0 months, SD = .73). At T2, the data collection consisted of parents and teachers completing questionnaires for 63 and 47 of the children, respectively. Participant attrition was due to that families no longer wished to participate in the study (N = 9) and lack of parental consent to contact the teacher (N = 5), and an unwillingness to complete the questionnaire by the teacher (N = 11).

Measures

Statistical Analyses

Data were screened for extreme values, and data points that met criteria for > 3 SD were defined as outliers. We found three outliers for cool inhibition and one for emotion regulation and replaced them with the value next most extreme (Tabachnick & Fidell, 2013). As some of the performance variables and ratings of ADHD and ODD symptoms were skewed (skewness ranging from .03 to 2.89, kurtosis from .20 to 3.21), logarithmic transformations were conducted. Analyses with these more adequately distributed variables did not change the level of significance in any of the analyses. For ease of interpretation, all results presented below are based on raw scores. We calculated a standardized residual for total emotion regulation by partialling out emotional reactivity in line with the proposed importance of isolating the effect of regulation from reactivity (Cole et al., 2004; Rydell et al., 2003). The residuals were calculated in separate regression analyses, with the total emotion regulation measure as outcome variable and the emotionality measure as the predictor. Similarly, for analysis of regulation of the four distinct emotions, a standardized residual was created for each of the separate emotions in separate analyses, with the distinct emotion regulation variables as outcomes and emotionality-matching variables as predictors. To account for comorbidity, symptoms of ODD at T1, which were significantly related to all of the predictors and outcome variables (r = .26 –.80), were controlled for in the analyses specified below.

First, Pearson’s correlations were conducted to examine bivariate relations between T1 and T2. Then, two repeated-measures ANOVAs were performed, with time as independent variable (T1 and T2) and symptoms of inattention and hyperactivity as dependent variables, respectively. To evaluate the independent effects of the three self-regulatory processes on developmental change in in the two symptom domains between T1 and T2, two ANCOVAs were performed with the three self-regulatory processes (cool inhibition, hot inhibition, and emotion regulation) as covariates. Finally, early symptoms of ODD were added as covariate in the analyses. As neither sex nor social background (indexed by a mean value of the mother’s and father’s highest educational level) influenced the patterns of significance for any of the analyses, we did not control for these variables in the final analyses. Independent samples t tests showed that there were no significant differences in any of the study variables at T1 for those children who did not remain in the study at T2 (all ps > .05).

Effects of Cool and Hot Inhibition on Developmental Change in ADHD Symptoms

We hypothesized that those with poor self-regulation in preschool would show relatively higher levels of ADHD symptoms in adolescence in comparison with those with good self-regulation skills. Although the levels of hyperactivity/impulsivity were higher in adolescence for children with poor cool inhibition as shown in the correlational data, the interactive effect did not support the prediction of a developmental lag in the expected decrease of symptoms. The result appeared somewhat counterintuitive in that the poor inhibitors showed a steeper decline in hyperactivity/impulsivity across time compared with the good inhibitors. Importantly, the reason for this larger symptom decline was that the poor inhibitors started off with significantly higher levels of hyperactivity/impulsivity in preschool. Thus, it appears that poor cool inhibition is primarily important for increased levels of hyperactivity/impulsivity in the preschool age. And although these children continue to show a high symptom level during the school years (Hart et al., 1995; Molina et al., 2009; Willoughby, 2003), they are catching up somewhat with the good inhibitors in hyperactivity/impulsivity in adolescence.

The results for hyperactivity/impulsivity are in line with previous studies demonstrating poor preschool inhibition in the cool domain to be an important associate of later ADHD symptoms (Berlin & Bohlin, 2002; Bohlin, Eninger, Brocki, & Thorell, 2012; Brocki et al., 2010). Importantly, however, this study adds to these findings in significant ways. First, our results indicated that lower levels of cool inhibition in preschool over time become a less important predictor of higher levels of hyperactivity/impulsivity, whereas its relation to levels of inattention symptoms is more stable as demonstrated in the correlational data. Second, the longitudinal relations between hot inhibition and symptoms of inattention (and the marginal significant relation with hyperactivity/impulsivity) disappeared when controlling for cool inhibition. This finding suggests that it was the variance shared with cool inhibition that was captured in the association between hot inhibition and ADHD symptoms. This result is in line with Zelazo and Mueller’s (2002) theoretical proposition that ADHD is viewed primarily as a disorder of cool EF, with hot EF being relatively, but not completely intact. Furthermore, our hot EF task involves compliance and resisting norm breaking behavior. This type of oppositional behavioral problems may well be a part of, although not limited to, the manifestation of ADHD. As there is no clear consensus on the definition of hot EF, variations in operationalizing this process may well have contributed to differences in results. Future longitudinal studies with a focus on hot EF in relation to disruptive behaviors specific for ADHD and ODD/CD, preferably comparing various existing “hot” EF tasks in different age groups, are needed to establish a clear definition of hot EF and its contribution to these disorders.

Effects of Emotion Regulation on Developmental Change in ADHD Symptoms

Early individual differences in emotion regulation predicted ADHD symptoms across time in that those with poor preschool emotion regulation not only had higher levels of inattention at both time points as can be seen in the correlational data, but they also showed a slight increase in inattention symptoms over time, whereas those with good emotion regulation showed a decrease in inattention symptoms. This result adds to recent cross-sectional reports emphasizing emotional functioning as an important associated feature of ADHD (Sjowall, Backman, & Thorell, 2015). Our result also extends recent longitudinal findings (Sjowall, Bohlin, Rydell, & Thorell, 2015) that demonstrated a predictive relation between preschool emotion regulation and symptoms of inattention in adolescence by showing that poor emotion regulation actually may indicate a risk for increased levels of inattention across childhood over time. Our results also corroborate recently reported biological evidence of disrupted emotional processing in relation to ADHD in middle childhood (Karalunas et al., 2014; Musser et al., 2011; Musser et al., 2013).

Interestingly, further analysis on the regulation of separate emotions revealed that, although there were bivariate longitudinal relations between all of the emotions (anger, fear, sadness, and happiness/exuberance) and ADHD symptoms (see Table 1), it was poor regulation of positive emotion (i.e., happiness/exuberance), that independently explained the longitudinal effect of the total measure of emotion regulation on the relative increase in inattention symptoms across time. At a behavioral level, this type of poor regulation of positive emotionality can be observed in situations in which excessive levels of excitement are not appropriate (e.g., happiness over one’s own winning and a friend’s losing in games or in failure to dampen play activities involving high laughter and shouting even when asked to do so by and adult). This finding supports proposals of a temperamental pathway to ADHD symptoms through unrestrained positive approach tendencies and difficulties regulating such emotions (Karalunas et al., 2014), proposed to be independent of deficits in executive functioning (Nigg et al., 2004; Sjowall et al., 2013). This finding is also in line with previous research using the same rating scale for emotion regulation reporting poor regulation of positive emotion in relation to externalizing behavior problems (Rydell et al., 2003) and in relation to ADHD symptoms (Forslund et al., 2016) in nonclinical samples and between regulation of positive emotion and ADHD in a clinical sample.

A key question concerns the actual mechanism involved in the dysregulation of positive emotion that drives the developmental progression of symptoms of inattention. The recent biological evidence published by Musser and colleagues (Musser et al., 2011; Musser et al., 2013) suggests abnormal parasympathetic mechanisms contributing to emotion regulation deficits, and specifically, regulation of positive emotion in childhood ADHD. It is possible that a biologically based difficulty in modulating strong positive affect leads to a series of additive social adversities propelling the child from one step to the next along a specific pathway in the progression of ADHD symptoms. Thus, the importance of emotion dysregulation as an early risk marker for ADHD may allow for new intervention methods with a focus on teaching children with ADHD regulatory skills to manage extreme levels of emotionality and elevated approach tendencies. Indeed, aspects of early family environment and parenting as significant contributing factors in the development of cognitive self-regulation have been emphasized in recent research (Bernier, Carlson, & Whipple, 2010; Bibok, Carpendale, & Muller, 2009).

The finding that dysregulation of negative emotionality, particularly anger, did not contribute to change in ADHD symptoms across time may seem counterintuitive given consistent reports of dysregulation of negative emotion in ADHD (Shaw et al., 2014). However, the aim of the present study was examination of specificity in relation to ADHD symptoms, and we therefore controlled for comorbid ODD symptoms, which are also strongly associated with dysregulation of negative emotionality. Indeed, the present findings give support to theoretical propositions (Martel & Nigg, 2006; Martel et al., 2009), and extend recent empirical findings (Forslund et al., 2016; Sjowall et al., 2013), suggesting that dysregulation of negative emotionality in ADHD may depend on the presence of comorbid ODD symptoms, whereas dysregulation of positive emotionality may constitute a specific risk factor for ADHD symptoms. It is, however, important to remember that ADHD symptoms and ODD symptoms are highly comorbid (Nock, Kazdin, Hiripi, & Kessler, 2007; Waschbusch, 2002), and dysregulation of negative emotion is therefore an important and common associated feature in children and adolescents high in ADHD symptoms.

A possible limitation in the current study was the use of parental ratings of emotion regulation, instead of using objective tasks similarly to the hot and cool EF measures. Using partly the same informants for the predictor and outcome measure involves a possible risk of inflated relations. However, to confirm that the predictive effect of poor regulation of positive emotion on inattention symptoms across time was not due to shared source variance, we conducted an analysis using parental ratings for emotion regulation and teacher ratings only for ADHD symptoms and the conclusion did not change. This means that the unique predictive relation between poor regulation of positive emotion and adolescent ADHD symptoms is robust and not a result of a reporter effect.