It has long been recognized that children with attention deficit hyperactivity disorder (ADHD) have difficulty regulating their emotions. In particular, research has shown that children with this disorder exhibit greater emotional reactivity [1
], higher levels of negative affect [2
] and lower levels of emotional awareness [3
]. Emotion regulation is defined as an individual’s ability to modify an emotional state so as to promote adaptive, goal-oriented behaviours [4
]. Emotion dysregulation arises when these adaptive processes are impaired, leading to behaviour that defeats the individual’s interests [5
]. Although prevalence rates of emotion dysregulation in ADHD are high [6
], the clinical significance of these findings and how specific they are to ADHD remain unclear. It has not yet been established, for example, whether deficits in emotion regulation are evident in all children with ADHD or perhaps only in a subgroup of children with this disorder.
In the early conceptualisation of ADHD, emotion dysregulation was considered a cardinal symptom [7
]. It was only with the introduction of the DSM-III [8
] that emotion regulation became an associated feature rather than a diagnostic criterion. The current conceptualization of ADHD is made up of two age-inappropriate behavioural dimensions, these being inattention and hyperactivity-impulsivity (ADHD; DSM 5; [9
]). However, many argue that emotion dysregulation should take more of a consideration in the assessment of ADHD due to its impact on psychological, physical and social outcomes [10
]. Now, conceptual theories of emotion regulation and ADHD can generally be characterised by three separate models; emotion dysregulation as a core feature of ADHD, emotion dysregulation as a distinct, but correlated dimension to ADHD or the addition of emotion dysregulation and ADHD as a distinct entity [5
] argues that emotion dysregulation is a core feature of ADHD and stems from executive functioning difficulties at the neurological level. Specifically, the inability to inhibit responses causes difficulties with selective attention, hyperactivity and impulsivity inherent in ADHD, as well as an impaired ability to inhibit strong emotional responses. However, emotion dysregulation is a dimensional trait that undercuts the traditional divide between internalizing and externalizing diagnoses, and it is not unique to ADHD [13
]. Regulation of emotions is compromised in children with disruptive behavioural disorders (DBDs), like conduct disorder (CD) and oppositional defiant disorder (ODD), as well as mood disorders. A recent study by Factor et al.
] suggests that ADHD alone is not sufficient for children to display significantly impaired emotional regulation, but it is only in the presence of a comorbid disorder that this pattern of deficiency begins to emerge.
Between 30% and 50% of children with ADHD meet the criteria for conduct disorder [15
], and this subgroup shows greater ADHD symptom severity than those with ADHD alone and worse outcomes [16
]. This group also appears to have higher familial and genetic loading for ADHD [17
], especially those with aggressive CD symptoms [19
]. However, previous research on emotion regulation in children with ADHD has often not considered the effects of comorbid CD [20
]. It is difficult therefore to know whether it is the core features of ADHD that are linked to emotion dysregulation or whether the relationship is explained by associated CD.
Studies on emotion regulation have primarily used retrospective, self-report questionnaires. Experimental studies (e.g., [23
]) have often used frustration eliciting tasks to assess emotion regulation; for example, by asking participants to hide their emotions from a confederate competitor. In this type of paradigm, however, the participants have no real motive to regulate their emotion apart from complying with the experimenter’s demands. Economic decision-making games, such as the Ultimatum Game (UG), provide another way of measuring emotion regulation by assessing effects on decision making [25
]. These paradigms involve two players interacting to decide how to divide a sum of money. One player (the proposer) offers a portion of the money to the second player (the responder). The responder can either accept the offer (in which case, both players split the money as proposed) or reject the offer (in which case, both players get nothing). Traditional economic theories, which view decision-making as a rational, cognitive process (e.g., [29
]) state that all offers, regardless of their fairness, should be accepted. Previous studies, however, have found that offers made to the responder that are comparatively small, and therefore deemed as unfair (20% of the total), have a 50% chance of being rejected by most individuals [25
Most individuals experience a negative emotional response and increased arousal when receiving unfair offers [31
], and a number of studies provide evidence that emotion regulation processes are a critical component in the UG. Negative emotions, such as anger and frustration, provoke participants to penalise their opponent rather than to make a utilitarian choice [32
], and the rejection of unfair offers increases when feelings of sadness are induced [34
]. The percentage of accepted unfair offers is influenced by the use of specific emotion regulation strategies, such as reappraisal [35
], and when participants are asked to “stay calm”, they accept more unfair offers [37
], suggesting that the ability to regulate negative emotions is necessary for the (rational) acceptance of unfair offers.
The rejection of unfair offers has been found to be associated with activity in neural substrates involved in negative emotions, such as the amygdala [38
] and anterior insula [39
]. Ventromedial prefrontal cortex (VMPC) damage is reliably associated with poorly-controlled emotional responses. In response to relatively minor provocation or frustration, patients with such damage are often irritable, angry, argumentative and even abusive [40
], yet generally show shallow affect. Similarities have been observed between patients with VMPC damage and patients with psychopathy [26
], and Koenigs et al.
] found similarly high rejection rates to the UG in VMPC damage patients and prisoners with low-anxiety psychopathy. In a community sample, Viera [47
] showed that the rejection rate of unfair offers was associated with VMPC activity in those with high psychopathy scores compared to those with low psychopathy scores; they interpreted this as reflecting an angry reaction to the frustration of not obtaining the desired outcome.
The results of studies in children and adults assessing emotion regulation suggest that age is an important factor. It is consistently found that adolescents reject more unfair offers than younger children and adults [48
], suggesting that there is a U-shaped developmental trajectory. This is consistent with the conceptualization of a peak in emotional reactivity during adolescence [51
]. However, until now no studies have examined emotion regulation using the UG in a clinical sample of adolescents with ADHD.
When assessing the contributory effects of comorbid externalizing disorders, it is important to consider the clinical and aetiological heterogeneity of disorders, such as CD and ODD [6
]; not all children who engage in antisocial behaviour will display emotion regulation problems. Frick and Morris [52
] argue deficits in emotion regulation are likely to underlie conduct problems that involve the angry and overt confrontation of others (e.g., fighting and assault), but are less likely to be associated with conduct problems that are not associated with confrontation or negative affect (e.g., stealing, vandalism). Burt and Donnellan [53
] also argue that there are unique personality correlates of different forms of antisocial behaviour. They found that aggression was uniquely predicted by high stress reaction (e.g., easily upset, has unaccountable mood changes), but this was not related to non-aggressive rulebreaking behaviour. This suggests that adolescents who display high aggressive CD symptoms might have more difficulty regulating their negative emotions during the UG and reject more unfair offers.
Callous-unemotional traits (CU) are another potentially important source of heterogeneity when looking at externalizing disorders. Such personality traits identify those at greater risk for severe antisocial behaviour [54
] and reduced responsiveness to treatment [55
]. The importance of such traits has been acknowledged by including limited prosocial emotions as a specifier for CD in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association; [9
]). CU traits are characterised by low prosocial emotions and behaviours, including shallow or blunted affect, lack of guilt or remorse, physiological under arousal and low empathy. Individuals who lack empathy and are not concerned about the emotions of others may be less likely to be driven by anger and the motivation to punish the proposer. This was supported in a non-clinical sample, which found that students scoring high on psychopathic traits rejected fewer unfair offers, interpreted as favouring self-interest [57
]. However, this contradicts Koenigs et al
] study of psychopathic inmates and also contrasts the results of more recent studies, which found no differences in rejection rates between high and low psychopathy scorers in community adults [47
] and adolescents [58
] or between healthy and (high psychopathy scoring) incarcerated individuals [59
Only a few studies have examined emotion regulation using the UG in adolescents [48
], and we are not aware of any study that has done so in a clinical sample of adolescents with ADHD. This study compared the decision-making of those with ADHD against those with ADHD and CD; with respect to the latter group, we distinguished between those with low aggressive CD symptoms and high aggressive CD symptoms. Within these groups, we also looked at the effect of additional CU traits. We included a sample of typically-developing adolescent males for comparison.
This study sought to examine whether ADHD adolescents in general have a problem with emotion regulation or whether this is a specific problem in those with conduct disorder, especially those with predominantly aggressive symptoms. No study until now has examined economic decision-making using the UG in a clinical sample of youths with ADHD, with or without CD. This study supports previous work [14
] suggesting that children with ADHD show significantly higher levels of emotional dysregulation than control children only in the presence of a comorbid disorder.
Unsurprisingly, the vast majority of the adolescents accepted the fair offers and rejected the unfair offers. Generally our adolescent male groups accepted fewer unfair offers than those reported for adults [26
]. This supports the suggestion of a peak in emotional reactivity during adolescence [51
]. There were no differences between the four adolescent groups’ acceptance rates for fair (5/5) and seriously unfair offers (9/1 and 8/2); however, a significant group effect was found for the moderately unfair offers (6/4 and 7/3), suggesting that problems in emotion regulation become more apparent under ambiguous conditions.
Follow-up tests showed that the ADHD with aggressive CD group rejected significantly more moderately unfair offers than any other group. Previous studies claim that the rejection of unfair offers is due to anger and a desire to punish the opponent, and the responder’s ability to regulate anger and frustration therefore plays a critical role in task performance [32
]. All three clinical groups reported the same amount of internalising emotionality in the SDQ. However, when faced with being treated unfairly, group differences in the ability to regulate externalising emotions became clear. Our results suggest that emotion regulation difficulties are not found in adolescents with ADHD alone, but rather only in those who have additional aggressive behaviour. This reflects their clinical presentation: being unable to control their aggressive behaviour [52
The results suggest that ADHD alone is not associated with emotion dysregulation during the UG compared to normal adolescents, supporting the view that emotion dysregulation is not a core feature of ADHD. The fact that the biggest difference between groups was between the two CD groups highlights the importance of treating CD as a heterogeneous disorder. The results showed that aggressive symptoms predicted performance on the UG better than overall CD severity, supporting the idea that aggressive antisocial behaviour has a different aetiology than non-aggressive behaviour [19
In the present study, participants were told that the aim of the game was to gain as many points as possible. Apart from this, there was no other incentive for them to win. The use of real reward incentives might, therefore, have a large impact on the rate of offers accepted. Further research is needed to help determine this in order to facilitate the development of more specific interventions for CD. For example, intervention programmes may be more beneficial by focusing on emotion regulation management in individuals with aggressive CD, whilst working with incentive-based goals in individuals with low-aggressive CD.
Unlike aggression, CU traits did not influence the acceptance of offers, supporting previous studies, which found no significant difference between individuals scoring high or low in psychopathy [47
], but not others [28
]. Previous studies have found that aggressive behaviour is positively correlated with negative emotionality and dysfunction [53
], whereas CU traits are negatively correlated with these same traits [72
]. In a modified version of the UG, Radke et al.
] found that offenders high in psychopathy, like controls, took the context of the offer into consideration (i.e.
, whether the proposer had a fair or unfair alternative offer to choose from), whereas offenders low in psychopathy did not, suggesting stronger impairments in social decision-making. However, even if similar behaviour patterns are shown in high and low psychopathy scorers, these might represent different motivations, as suggested by recent imaging studies [47
], and this now needs to be tested further in clinical samples. Koenigs et al.
] observed poorer regulation during the UG in low-anxious psychopathic offenders in comparison to high-anxious psychopathic and non-psychopathic offenders. However, due to the small sample sizes (n
= 6) and lack of a non-ASB control sample, further research is needed. Our groups did not differ significantly in internalising emotionality (measured by the SDQ), and it may be the case that the low anxiety component of psychopathy drives regulation problems. Further research is needed using a clinical sample of children with disruptive behaviour problems, and improvements could be made by using a combination of parent, teacher and behavioural observation to assess CU traits, as suggested by the new CU specifier for the CD diagnosis in the DSM-5.
An issue that needs further exploration is the assumption that the acceptance of unfair offers is the rational decision. From an economic perspective, the rejection of offers is irrational, because it results in a personal loss. However, from a social perspective, rejection of unfair offers can be seen as a rational, altruistic action to preserve social norms. Rather than maximizing self-interest, the participant chooses to punish the socially-inappropriate action from the proposer for the good of the general population [73
]. This would explain why similar rejection rates are found in a modified version of the UG in which the participants play on behalf of a third party, compared to one played by themselves [75
]. We would argue that it is unlikely that boys high in aggressive CD symptoms rejected offers for the “good of the general population”, and this is supported by a recent imaging study, which found differences in response to the UG between severely antisocial adolescents and controls [76
]. That study found decreased right inferior frontal gyrus (rIFG) activity in antisocial youngsters during the UG and no correlation between rIFG activity and behavioural responses, as was found in the controls. These results complement previous studies that suggest that juvenile antisocial behaviour is associated with difficulties in engaging the regulatory processes associated with the frontal cortex [77
], in particular the rIFG, which is associated with response inhibition [79
]. This supports the notion that the rejection of unfair offers in antisocial populations is due to deficient self-regulatory processes. Further research should now investigate more thoroughly participants’ reasoning behind the rejection of offers in order to support this.
Another limitation of the study is that during the UG task, no direct measure of participants’ emotional responses, such as psychophysiological recordings or subjective ratings, were obtained. Because we did not measure participants’ emotional response to offers, we do not know how much participants needed to self-regulate. One would assume that the more intense the shift in emotion, the more regulatory resources would be needed in order to modify that emotion. It is difficult, however, to determine from this study whether the boys with aggressive CD had deficient regulatory resources or experienced a more intense emotional reaction. Previous findings of reduced psychophysiological responding to aversive stimuli in the sample group would suggest the former [81
]. Future research including such additional measures should help untangle the various potential factors affecting performance on the UG. The simplicity of the paradigm is also well suited for testing this sample using brain imaging techniques.
Emotion dysregulation is a dimensional trait that is not unique to ADHD. It is important to uncover to what extent individuals with ADHD and comorbid CD develop emotion regulation deficits for reasons that are different from those with CD alone by testing psychopathology in non-ADHD samples. Longitudinal studies are needed in order to define how the developmental trajectories interact with one another, to see, for example, whether emotion regulation difficulties bridge the development of aggressive behaviour in children and adolescents with ADHD or whether factors underlying both ADHD and comorbid aggressive CD (i.e.
, temperamental or biological factors) lead children to demonstrate impairing levels of emotional dysregulation. Furthermore, existing treatments need to be modified to address the role of emotional regulation in children with ADHD. They should incorporate cognitive-behavioural techniques to teach emotion recognition and physiological relaxation exercises for negative emotions [83
] and encourage problem-solving techniques to help children adjust and self-regulate when their expectations are not met [85