The Neurodevelopmental Basis of Early Childhood Disruptive Behavior: Irritable and Callous Phenotypes as Exemplars

Our own work generates a model of disruptive behavior with four dimensions: 1) aggression, 2) noncompliance, 3) irritability, and 4) callous behavior (75). The current review focuses on irritability and callous behavior for two key reasons: First, these two dimensions possess a strong neurodevelopmental science base (76, 77). Second, a comparison of their phenomenology highlights heterogeneous developmental expressions. For example, expression of irritability is exemplified by behaviors such as temper tantrums, which occur normatively in most young children and are only pathognomonic when very frequent and dysregulated (65, 78). In contrast, callous traits, as reflected in diminished response to the distress of others, do not normatively occur in development. When they do, however, they are developmental risk markers for severe antisocial syndromes (79–81). Concern for others’ feelings manifests in rudimentary form as early as infancy, making identification of callous tendencies from very early in life plausible (82–84). Together, these two dimensions predict many common childhood- and adult-onset mental disorders (5, 9, 11, 85). Thus, characterizing their neurodevelopment within a mental disorder framework could impact potentially chronic problems at a much earlier phase of the clinical sequence, which is of great import for prevention (4).

Our in-depth focus on these two phenotypes is intended as an exemplar of the neurodevelopment of disruptive behavior, leveraging the specificity and rich science base of irritability and callous behaviors. Aggression, another core disruptive behavior dimension, also emerges early and predicts later antisocial pathways (5, 6). However, it lacks clinical specificity as its varied forms are associated with both irritability (reactive aggression) and callous behaviors (proactive aggression) (76, 77). Further, relative to these first three dimensions, noncompliance has received far less attention in neuroscience.

Irritability reflects a relative dispositional tendency to respond with anger to blocked goal attainment (65, 94, 95). There are substantial individual differences in the intensity, ease of elicitation, and persistence of irritability and the extent to which irritability impairs functioning. These individual differences are shaped by developmental assets and vulnerabilities as well as bidirectional associations with experience (96–101). Long considered a complicating factor in other developmental psychopathologies, irritability has more recently been considered a problem in its own right (93). Clinical levels of irritability occur in 3%−20% of children, with higher estimates in early childhood and developmentally varying heritability estimates (i.e., 0.3–0.8) (61, 97, 102). Irritability has traditionally been classified as a feature of oppositional defiant disorder. In DSM-5, it is also the defining feature of the new diagnosis disruptive mood dysregulation disorder. Oppositional defiant disorder and disruptive mood dysregulation disorder are overlapping in their characterization of irritability, with most children meeting criteria for both disorders (103). In young children, irritability is currently best captured as a disruptive syndrome, as being younger than age 6 is an exclusion criterion for disruptive mood dysregulation disorder in DSM-5. Further, given the arbitrary adevelopmental threshold for irritability severity in disruptive mood dysregulation disorder, its stability and incremental utility remain in question (103, 104). Irritable mood may also be an indicator of depressed mood in young children. However, predominant features of young children’s depression are anhedonia and sad/withdrawn affect, and disruptive behavior is much more common than depression at preschool age (105). Irritability also commonly co-occurs in many other childhood-onset disorders (98–100, 106–108), and it increases the risk of adult mental disorder, particularly depression and anxiety, even with concurrent irritability symptoms controlled (62, 109). Pervasive infant irritability is also associated with nearly triple the risk of disruptive behavior through middle childhood (101). As clinical investigations of irritability have been largely disorder-specific, work on the transdiagnostic import of its early emergence is needed (73, 98, 99, 110, 111). Since the transdiagnostic approach to irritability is relatively new, it is also not yet known whether there is an adolescent- or adult-onset phenotype. However, clinically salient, persistent childhood irritability is typically presaged by an irritable disposition in the first years of life (112).

Clinical patterns of early childhood irritability are exaggerations of normative patterns. Irritable mood and temper tantrums are normative misbehaviors that occur in most children (113, 114) (Figure 2A). For example, 83.7% of preschoolers have regular temper tantrums (78). As a result, when oppositional defiant disorder irritability symptom thresholds for older children are used for preschoolers, this triples the prevalence compared to developmentally specified thresholds (51). Thus, delineating the boundaries between normal and abnormal irritability in early childhood is particularly complex. Figure 2 highlights the application of the developmental specification framework to irritability.

Irritable mood and behavior are normatively heightened during early childhood as increased capacity leads to enhanced frustration with environmental limits (95). Emotion regulation, which involves both top-down arousal processes and bottom-up control processes, is the key developmental capacity undergirding modulation of irritability (25, 115, 116). Early executive function capabilities support self-regulation of anger. These capabilities become increasingly sophisticated across the preschool period (117–119), most likely due to growth during early childhood of cortical structures mediating anger regulation, with the most rapid growth from 0 to 2 years and 90% of adult brain volume achieved by age 6 (87).

We have established the normal–abnormal spectrum of irritability using the MAP-DB temper loss scale in the two large community samples of preschoolers in the MAPS study (see Table S1 in the data supplement accompanying the online version of this article) (65, 75). Relatively mild expressions are common (online Table S1; Figure 3 delineates severity of individual behaviors). In contrast, abnormal expressions are captured in items that are above a psychometrically derived clinical threshold, representing the severe end of a latent irritability trait. Thus, abnormality is defined in terms of frequency, dysregulation (e.g., destructive tantrums), persistence (e.g., longer than 5 minutes), and developmental expectability in context (e.g., frustration versus “out of the blue”) (78) (Figure 2B). Similar patterns in multiple community and clinical samples demonstrate replicability (online Table S1) (65, 78, 120). Thus, most preschoolers exhibit some irritable behavior, but frequent, highly dysregulated, and long-lasting behaviors are not endorsed for most children. These patterns appear to be parallel in infancy with some variation in frequency thresholds (121).

The central emphasis of pathophysiologic investigations of irritability has identified atypical responses to frustrative nonreward, reflecting abnormalities in intensity, duration, and ease of elicitation (65, 77). This mirrors clinical features of irritability in young children (78, 113). Dysfunctions may involve atypical reward processing, aberrant subcortical activation, and perturbed prefrontal functions (Figure 2C). More recently, research on adolescents also points to dysfunctional threat processing in the pathophysiology of irritability (77, 122). However, this work has not yet been systematically extended to younger children. In early explorations in the MAPS study, perturbed threat processing was implicated in preschool anxiety (123) but not irritability.

Executive function involves top-down control of processes that guide behavior and undergird emotion, self-regulation, and goal-oriented achievement (115, 118). Deficits in executive function impede children’s ability to regulate negative emotions and execute adaptive responses (118). Developmental perturbations of the prefrontal cortex may impair the regulation of emotional arousal in irritable children (124–126). In the only investigation to date of the neural processes underlying executive function in early irritability, Li and colleagues found that children who are dispositionally, but not clinically, irritable display increased lateral prefrontal cortex activation during a task requiring cognitive flexibility relative to nonirritable children (72). The relative capacity of irritable children to recruit this region may serve an adaptive regulatory function and prevent impairment.

From a behavioral perspective, young children with deficits in response reversal are less effective at regulating mood and behavior across varied social contexts, while those who can shift attention flexibly during frustration/demand tasks are at reduced risk of externalizing problems (116, 127, 128). In terms of predicting clinical problems, irritable preschoolers with a blunted error-related negativity response manifest signs of deficient conflict monitoring (73). Such children face risk for later disruptive behavior, unlike irritable preschoolers with an enhanced error-related negativity, who face risk for anxious/depressed symptoms (73). These findings reflect the context-sensitive deployment of executive function processes (39). For example, irritable infants with left frontal electroencephalographic (EEG) asymmetry exhibit poorer inhibitory control (129). Parallels at older ages include the presence of cognitive flexibility deficits along with underlying prefrontal cortex and striatal functional deficits in irritable adolescents (130, 131). Of course, executive function is both a common and dissociable mechanism of a range of psychopathologies (132). For example, although executive dysfunction is not the hallmark of callousness, there is evidence that it serves as a moderator of clinical escalation from both childhood irritability and callous behavior (133, 134).

Individual differences in irritability are associated with aberrant reward processing (93). Irritable young children have relatively high reward orientation and approach tendencies, putting them at risk for frequent and intense frustration (111). These patterns may differentiate irritable young children with adaptive versus maladaptive functioning (12, 135). Variation in severity across the irritability spectrum is associated with differential activation in the dorsolateral prefrontal cortex (DLPFC), a region that also serves executive function, manifesting as an inverted U-shaped association (12). Specifically, along the ascending arm of the curve, irritability and DLPFC activation are positively associated at frustration onset, i.e., higher irritability is correlated with greater DLPFC activation. In contrast, in the descending arm of the inverted U curve, irritability and DLPFC activation are negatively associated at frustration activation, i.e., higher irritability is associated with weaker DLPFC activation (12). Children within the first subgroup (ascending arm of inverted U) had variable irritability levels, but all fell within the normative range of irritability on the MAP-DB temper loss scale, whereas the latter subgroup (descending arm of inverted U) had subclinical- to clinical-level MAP-DB temper loss scores. This inverted U pattern points to the possibility that developmentally typical versus atypical reward processing may be a pathophysiologic mechanism associated with the transition from vulnerability to clinical disorder in children at risk for irritability-related pathology. Of note, in this study the neural threshold marking the transition of the inverted U (i.e., the point at which the association between DLPFC activation and irritability switched from positive to negative) closely approximates the psychometrically derived severity on the MAP-DB scale. Other salient work relies on event-related potentials (ERPs). Here, clinically salient patterns of irritability at preschool age, as captured via symptoms of disruptive mood dysregulation disorder, predicted preadolescent enhanced neural sensitivity to reward (ERP reward positivity), possibly contributing to exaggerated stimulus response learning and perseveration (136).

This model of irritability provides a map to clinical patterns and progression (Figure 2D). At least 10 studies have demonstrated specific clinical and predictive utility of early childhood irritability. These studies support five major conclusions (Table 1):

Callous/unemotional traits reflect diminished responsiveness to the distress of others (67, 79, 145). Their early emergence is a marker of antisocial behavior persistence and severity, and they may serve as a developmental substrate of psychopathy (146, 147). Prevalence of conduct disorder with callous/unemotional traits in youth is estimated at 2−4% (about 25%−35% of children with conduct disorder), with heritability of 40%−78% (145, 148). In DSM-5, callous/unemotional behavior is the core element of the “limited prosocial emotions” specifier for conduct disorder (149). However, as DSM callous/unemotional symptoms are not delineated in a developmentally meaningful manner for young children, this has precluded their clinical evaluation in young children (53). Callous/unemotional behavior probabilistically increases the risk of severe antisocial behavior and psychopathy, but most youths with these traits do not develop psychopathic tendencies. Thus, early callous/unemotional behavior represents a pattern of insensitivity to others that predisposes to severe antisocial behavior, but whether or not severe psychopathology develops depends on a combination of genetic risk, environment, learning, and opportunity (146, 150–152). (Note: We focus on callous behaviors, which have received far more attention in early childhood research than unemotional features.)

Processes that go awry in callous syndromes are the development of empathic concern for others and development of the moral emotions (e.g., shame) that undergird rule internalization (e.g., guilt) (Figure 4A) (75, 83, 148, 153, 154). While empathic concern was originally thought to be a late-developing process, developmental science now suggests that rudimentary concern for others is present at birth (83). The implication is that top-down control is required to directly express concern; however, the capacity to feel for others is more automatic (83). For example, neonates are negatively aroused by the distress of others, reflecting overlap in the neural circuitry underlying perception of one’s own versus others’ emotions (155). Infants in the first year of life also demonstrate prosocial versus antisocial preferences and empathic concern (83, 156–159). Moreover, toddlers exhibit stable dispositional empathy and internalization of rules, and they perform prosocial acts to ameliorate others’ distress (157). Thus, sensitivity to others’ feelings and the capacity to follow social norms manifest in the first years of life (83, 158). Moral emotions (e.g., guilt) are also identifiable in nascent forms as exhibited through bodily tension and gaze aversion following transgression in toddlers, and through expressions of remorse and reparation at preschool age (159). Both concern for others and guilt inhibit antisocial behavior that will cause others distress (160, 161).

Applying the psychometric approach, we have established the severity spectrum of callous behavior by means of the MAP-DB scale “low concern for others” (Figure 4B) (75). Items range from indifference (e.g., “not seem to care about others’ feelings”) to goal-oriented distress evocation (e.g., “enjoy making others mad”). In contrast to the common display of early irritable behaviors described above (and shown in the online Table S1), the severity spectrum of the “low concern for others” scale indicates that these are not normative misbehaviors. None of the callous behaviors are common at preschool age (i.e., less than one-third of children are reported to have engaged in any these behaviors over the past month; see online Table S2). Of note, all of these behaviors were identified as psychometrically severe across two community samples of preschoolers (i.e., above the 95th percentile severity threshold; Figure 5), regardless of socioeconomic context. Despite the greater severity of these callous-type behaviors in general, there is still dimensional variation. The least severe item reflects insensitivity to others’ feelings in developmentally expectable contexts (e.g., “not caring about others’ feelings when upset”), whereas the most severe item reflects intentional provocation of distress in others (e.g., “doing things to humiliate others”) (Figure 5). This severity spectrum is consistent with findings that toddlers’ active evocation of others’ distress predicts adolescent conduct disorder symptoms, whereas lower levels of prosocial concern do not (15).

In adolescents and adults, specific affective and cognitive atypicalities have been robustly established as neural substrates of callousness (Figure 4C), particularly amygdala dysfunction (162–164). At least 10 studies have demonstrated specific, rather than general, deficits in emotion processing in callous youth (80, 81, 165). In particular, deficits in distress processing, especially fear processing, have been identified (166, 167). A recent fMRI meta-analysis also suggests specificity in patterns of neural disruptions in disruptive youth with and without psychopathic traits (which roughly parallels our irritable versus callous distinction) (67).

Neurodevelopmentally, such socioemotional deficits impair responsiveness to socialization. This is because attunement to others’ distress and displeasure inhibits other-directed negative behavior by transgression- and empathy-related arousal, and undergirds the development of conscience (84, 165). In terms of emotion processing deficits, we have demonstrated a specific association between the MAP-DB scale for low concern for others and decrements in fear processing at preschool age, even with impulsivity, aggression, and irritability controlled (81). While this requires replication, the specificity of these findings parallel to patterns in older youth is striking. Other pathophysiologic indicators of early callous behavior include impaired eye contact (during performance-based and social interaction paradigms) (168), reduced cardiac and behavioral arousal in response to distress-eliciting stimuli, and increased fearlessness in contrast to irritable infants (169). The progression of early childhood callous behaviors to later externalizing problems is moderated by theory of mind deficits, underscoring the role of social processing deficits in these pathways (170). Evidence of heritability of callous traits in an adoption sample is also consistent with biologic plausibility (171). To our knowledge, developmentally sensitive neuroimaging techniques have not yet been employed to examine neural underpinnings of callous behavior in very young children, an important area for future research.

This model of the early callous phenotype points to clinical patterns and progression (Figure 4D). There has been a veritable explosion of research on callous behavior in early childhood (Table 2). Drawing on both direct observation and maternal report, at least 20 independent studies (beginning as young as 14 months of age) have led to several conclusions.