Irritability in Youths: A Critical Integrative Review

Abstract

Irritability, defined as proneness to anger that may impair an individual’s functioning, is common in youths. There has been a recent upsurge in relevant research. The authors combine systematic and narrative review approaches to integrate the latest clinical and translational findings and provide suggestions for addressing research gaps. Clinicians and researchers should assess irritability routinely, and specific assessment tools are now available. Informant effects are prominent, are stable, and vary by age and gender. The prevalence of irritability is particularly high among individuals with attention deficit hyperactivity disorder, autism spectrum disorder, and mood and anxiety disorders. Irritability is associated with impairment and suicidality risk independent of co-occurring diagnoses. Developmental trajectories of irritability (which may begin early in life) have been identified and are differentially associated with clinical outcomes. Youth irritability is associated with increased risk of anxiety, depression, behavioral problems, and suicidality later in life. Irritability is moderately heritable, and genetic associations differ based on age and comorbid illnesses. Parent management training is effective for treating psychological problems related to irritability, but its efficacy in treating irritability should be tested rigorously, as should novel mechanism-informed interventions (e.g., those targeting exposure to frustration). Associations between irritability and suicidality and the impact of cultural context are important, underresearched topics. Analyses of large, diverse longitudinal samples that extend into adulthood are needed. Data from both animal and human research indicate that aberrant responses to frustration and threat are central to the pathophysiology of irritability, revealing important translational opportunities.

Irritability can be defined as proneness to anger that may impair an individual’s functioning. It is one of the most common reasons that youths present for mental health evaluation and care (1). The most salient feature of clinically relevant irritability is frequent temper outbursts that are developmentally inappropriate and out of proportion with a precipitating event (phasic irritability) (2). Such outbursts are typically accompanied by chronic grouchiness and angry mood (tonic irritability) (3). Severe irritability is associated with impairment at home, in school, and with peers (2). Irritability is a central symptom of three DSM-5 diagnoses, as well as a transdiagnostic, dimensional construct that is especially common in attention deficit hyperactivity disorder (ADHD), autism spectrum disorder (ASD), and anxiety and mood disorders (4-6).

Irritability is a phenotype characterized by negatively valenced emotions and approach behavior (7). Etiologically, irritability is moderately heritable, is associated with environmental adversities such as peer victimization (8) and adverse childhood experiences (9, 10), and can follow several different developmental trajectories. One model of irritability proposes two potential mechanisms: difficulty tolerating frustration (i.e., the emotional and behavioral response to omission of an expected or desired reward; frustrative nonreward in the Research Domain Criteria [RDoC]) and aggressive responses to threat (7). This model has been elaborated to specify intervention targets (11) and can be studied in animals (7).

The number of publications focused on irritability published per year has quadrupled since 2000 (12). Nonetheless, irritability remains understudied relative to its clinical importance. In this review, we present the latest clinical and translational findings and identify research gaps. Depending on the content of each section of the review, we performed systematic reviews, relied on previous reviews and meta analyses, or used narrative methodology (see the online supplement). Of the 11 sections of this review, four focus on foundational topics, three on specific developmental periods or comorbidities, one on treatment, and three on methods and mechanisms. Each section concludes with prioritized suggestions for future research (Table 1).

TABLE 1.

Priorities in different areas of irritability research^a

Research Area	Research Priorities
Conceptualization	Conduct cross-cultural studies of the external validity of irritability as a construct and its differentiation from other constructs (incremental validity) and use these data to refine its definition and operationalizationb
Measurement	Investigate how assessment measures and differences across informants relate to external validators, for example, cognitive and psychophysiological processes, longitudinal course, treatment response, passive sensor data, and in vivo ecological momentary assessment datab Identify mechanisms mediating informant effects Develop self-report measures to differentiate tonic versus phasic irritability Develop developmentally and culturally sensitive idiographic and qualitative tools to better reflect the perspective of young people experiencing irritability
Epidemiology	Conduct studies of the mechanisms underlying developmental trajectories of irritability, including trajectories into adulthoodb Include multiple informants, including adolescents providing self-reports Identify environmental and cultural stress and resilience factors across development that influence the expression of irritability
DSM-5 and beyond	Conduct cross-cultural epidemiological studies to derive data-driven thresholds for irritability and enable harmonization of intermittent explosive disorder, oppositional defiant disorder, and disruptive mood dysregulation disorderb Integrate a dimensional framework into the diagnosis of irritability-related disorders
Early childhood	Develop tools combining behavioral patterns, neural biomarkers, and parent-child interactional processes to identify highly irritable young children who will develop clinical problems and malleable protective processes that can be leveraged for preventionb
Middle childhood to adolescence	Identify mechanisms mediating associations between irritability and suicidality, including testing whether interventions targeting irritability cause reductions in suicidality, and understanding mechanistic links between depression and irritabilityb Conduct mechanistic studies of tonic versus phasic irritability to explain their longitudinal associations Examine whether exposure to minority stressors is associated with increased irritability
Comorbidity with neurodevelopmental conditions	Conduct mechanistic developmental studies testing whether irritability is a feature of ADHD or ASD or whether it is an early indicator or a risk factor for a secondary conditionb Examine mechanistic associations among domains of executive function and self-regulation (e.g., impulsivity, deficient inhibitory control, and poor emotion regulation, including irritability) Characterize longitudinal course of irritability into adulthood in ADHD and ASD, including impact of treatment and of stress associated with the neurodevelopmental conditions themselves
Treatment	Conduct rigorous psychotherapeutic and psychopharmacological RCTs for children with severe irritability that does not improve when the primary disorder has been treatedb Test hypothesized mechanisms of treatment response to guide development of novel interventions Conduct RCTs to test whether irritability and emotion dysregulation show differential responses to treatment Clarify within-session psychological mechanisms (e.g., habituation or extinction) Identify predictors of treatment response
Neuroimaging	Increase sample sizes and reliability, for example, by using valid assessment measures, longitudinal approaches, valid and reliable tasks evoking potent and relevant emotional responses, multivariate measures identified using machine learning, and multimodal approachesb
Genetics	Use longitudinal genetic approaches in large, diverse samples to test whether irritability represents the same construct by gender, at different ages, and in the context of different disorders (e.g., major depressive disorder vs. ADHD)b Test whether irritability indexes biological or genetic heterogeneity Examine the genetic architecture of phasic versus tonic irritability
Translational research	Identify interindividual differences in animal responses to frustrative nonreward and the mechanisms mediating these differencesb Increase discussion and collaboration between clinical researchers using neuroimaging to study irritability and animal researchers doing related work

^aADHD, attention deficit hyperactivity disorder; ASD, autism spectrum disorder; RCT, randomized controlled trial.

^bHighest-priority research aim.

DEFINITION AND CONCEPTUALIZATION

Irritability has both mood and behavioral components. It can be differentiated from related constructs, specifically emotion dysregulation (2), which is a broader construct encompassing multiple emotional phenotypes including surgency, anxiety, mania, depression, and others (5). Questionnaire data indicate that irritability can be differentiated psychometrically from emotion dysregulation (13), anger, and aggression (14). Irritability is also unique in that it is associated with both externalizing (15) and internalizing problems. Although this review focuses on the literature on irritability, in discussing comorbid neurodevelopmental conditions and the treatment of irritability, we also include the literature on emotion dysregulation and externalizing disorders because irritability is usually not differentiated from these other constructs in these subfields.

Beyond defining irritability as distinct from other constructs, it is important to test its external validity, that is, its ability to reliably provide relevant clinical information (16, 17). Kendler (16) suggested assessing external validity using familial aggregation, illness course, stability, and treatment response. As detailed below, the main indicators of external validity for irritability are its moderate heritability (18-20), data on its illness course (i.e., longitudinal associations with anxiety, depressive disorders, behavioral problems, and suicidality) (5, 21, 22), and its ability to predict future impairment and psychopathology beyond that of comorbid disorders (4, 23-25).

Irritability waxes and wanes, and research on diagnostic criteria for bipolar disorder in youths demonstrates that such temporal heterogeneity is clinically important (26). There are two complementary approaches to characterizing the temporal course of irritability. The first differentiates episodic versus chronic (trait) irritability over a relatively long period of time. Episodic irritability is present for a delimited period, and its onset and offset are contemporaneous with changes in other characteristic symptoms (e.g., mania and changes in sleep or activity). In contrast, chronic irritability is relatively constant. The second subtyping approach differentiates acute temper outbursts (phasic irritability) and chronic angry mood (tonic irritability) over shorter periods of time. Phasic and tonic irritability are highly correlated but may differ in comorbid conditions, genetic mechanisms, longitudinal course, and treatment response (27, 28). Research into these associations is ongoing, as are studies of acute versus sustained emotional responses in paradigms inducing anger or frustration (29, 30); these studies could elucidate mechanisms underlying phasic versus tonic irritability.

Finally, there are individual differences in proneness to experiencing and exhibiting anger. For example, research suggests that angry, aggressive responses are influenced by interindividual differences in the extent to which an environmental stimulus (e.g., another young person) is perceived as threatening (31). However, such differences in the experience of anger and subsequent behaviors have not been explored from a first-person perspective with a deep qualitative understanding of the impact of earlier experiences and the present context. Regarding context, evidence suggests that individuals’ anger expression is causally influenced by environmental affordances; that is, when the latter support fighting, anger is more likely to be expressed (32). This finding supports the importance of considering cultural and social demands. Indeed, the efficacy of parent management training for clinical problems related to irritability demonstrates that these demands are relevant (5, 33, 34). The mechanisms of these different effects could be studied with real-time digital phenotyping techniques (e.g., ecological momentary assessment and actigraphy) and functional MRI (fMRI) paradigms evoking frustration and anger in a social context.

The highest-priority research questions regarding the conceptualization of irritability involve extending the work on the external and incremental validity of the construct and refining its definition based on these data. These efforts should be cross-cultural, and they overlap significantly with other research suggestions below.

MEASUREMENT

Initial advances in irritability research relied on secondary analyses of symptom data from large data sets (35). Post hoc measures of irritability have been derived from structured (e.g., the Development and Well-Being Assessment [36]) and semistructured (e.g., the Schedule for Affective Disorders and Schizophrenia for School-Age Children [37]) DSM-based interviews, parent- and youth-report checklists (e.g., the Child Behavior Checklist and the Youth Self-Report), and temperament rating scales, but these measures have psychometric limitations. Irritability has also been measured by using items designed to assess specific diagnoses (e.g., disruptive behavior and mood, anxiety, or stress disorders) (see Table S1 in the online supplement).

It is essential to have instruments that measure irritability specifically. Notable examples include the Affective Reactivity Index, the Clinician Affective Reactivity Index, the Multidimensional Assessment of Preschool Disruptive Behavior, the Disruptive Behavior Diagnostic Observation Schedule, the Emotion Dysregulation Inventory reactivity scale, and the Emotional Outburst Inventory (5, 38, 39) (see Table S1 in the online supplement). When measuring irritability, it is important to consider near-neighbor constructs (e.g., emotion dysregulation, anger, and aggression) and co-occurring symptoms and disorders (e.g., depression, ADHD, traumatic stress, and sleep disturbance) (5, 40).

Multi-informant irritability studies show evidence of validity, reliability, and stability for both parent- and self-report measures; however, interrater agreement is modest, and item- and scale-level differences are apparent (14, 41). Such low agreement is common in youth psychopathology and likely reflects differences in perspective rather than measurement error (42). This evidence argues against averaging scores across informants or using a single informant. Researchers examining data from multiple informants should parse informant variance (e.g., with bifactor or informant-specific models). Clinicians should also view discrepancies between informants as potentially meaningful, explore them in detail, and consider parent and child reports as equally important sources of information (5, 38). Finally, it is not uncommon for irritability to be reported by one informant, typically a parent, in the absence of corroborating reports. In such cases, the treating clinician should assess the reliability of the informant, the overall clinical impairment, and available intervention options.

Inconsistent results have been reported in the neuroimaging and genetics literatures on irritability, possibly because of different assessment instruments. Given the relatively nascent state of irritability research, it is premature to designate gold-standard measures. However, a reasonable assessment might include some of the instruments high-lighted in Table S1 in the online supplement, such as an irritability-specific measure (e.g., the Affective Reactivity Index), an outburst-specific measure (e.g., the Emotional Outburst Inventory), and a broad measure of psychopathology and regulation (e.g., the Child Behavior Checklist and its dysregulation profile [43, 44]). As data emerge on the reliability and external validity of different measures, the field could adopt consistent measures.

Future studies should investigate how symptom measures and their variation among informants relate to external validators, including cognitive and psychophysiological measures, longitudinal course, treatment response (5, 10, 14), passive sensor data, and in vivo real-time ecological momentary assessment data from parents and children. Self-report measures differentiating tonic versus phasic irritability, along with developmentally and culturally sensitive idiographic and qualitative tools, would enable young people’s perspectives to be better incorporated in care and research and help elucidate factors underlying observed informant effects (1). The literature contains only scant descriptions of how vulnerable populations (e.g., marginalized populations or those with neurodevelopmental conditions) experience or express irritability. Research combining qualitative and quantitative methods is key to addressing such issues of epistemic injustice.

EPIDEMIOLOGY

The frequency and intensity of normative irritable mood and temper outbursts varies developmentally, generally peaking in early childhood and declining into adulthood. International data suggest that 44% to 82% of individuals show low, stable levels of irritability from early childhood to early adolescence, and 2% to 5% of children show elevated and persistent or increasing levels of irritability (34, 45-49) (Table 2). Nonstable trajectories also exist; in one study, 5% of children had a curvilinear trajectory peaking in late childhood (47). Vulnerable populations may show greater variability in developmental trajectories.

TABLE 2.

Developmental trajectories of irritability in community-based samples

Study	Cohort (Country)	N	Age (years)	Assessment	Irritability Trajectories
Ezpeleta et al., 2016 (49)	Barcelona cohort (Spain)	622	3–5	Diagnostic Interview of Children and Adolescents for Parents of Preschool Children; oppositional defiant disorder irritability symptoms	Very low persistent, 46%; low persistent, 44%; high persistent, 4%; high decreasing, 4%; low increasing, 3%
Boylan et al., 2017 (48)	Pittsburgh Girls Study (United States)	2,450	5–13	Child Symptom Inventory–4; oppositional defiant disorder irritability symptoms	Low persistent, 39%; moderate persistent, 56%, high increasing, 5%
Riglin et al., 2019 (47)	Avon Longitudinal Study of Parents and Children (United Kingdom)	7,924	7–15	Development and Well-Being Assessment; oppositional defiant disorder irritability symptoms	Low persistent, 81%; high decreasing, 6%; low increasing, 6%; limited to late childhood (inverted U), 5%; high persistent, 3%
Orri et al., 2019 (45)	Quebec Longitudinal Study of Child Development (Canada)	1,393	6–12	Social Behavior Questionnaire; Child Behavior Checklist; Ontario Child Health Study Scales; Preschool Behavior Questionnaire	Low persistent, 75%; high decreasing, 7%; low increasing, 13%; high persistent, 5%
Wiggins et al., 2014 (46)	Fragile Families and Child Wellbeing Study (United States)	4,712	3–9	Child Behavior Checklist	Low decreasing, 61%; moderate decreasing, 21%; high persistent, 11%; high decreasing, 5%; high increasing, 2%
Yu et al., 2023 (34)	Fragile Families and Child Wellbeing Study (United States)	4,462	3–15	Child Behavior Checklist	Reverse-J shaped, 6.67%; high stable, 11.27%; medium stable, 19.96%; high inverted-U shaped, 2.33%; low U shaped, 60.76%

Evidence suggests that there are complex interactions among age, gender, and informant in rating a child’s irritability (41, 47). Comparisons across studies are difficult because informants vary by subject age (e.g., parent ratings of young children vs. self-ratings, ratings with or without the parent present, and ratings in adolescence) and individual studies generally do not cover broad developmental periods.

When evaluated cross-sectionally, chronic irritability clusters with depression, anxiety, oppositional behavior, conduct disorder, ADHD, and ASD (50). However, a meta-analysis of longitudinal studies found that irritability has the strongest links with depression (odds ratio=1.80), anxiety (odds ratio=1.72), and oppositional defiant disorder (odds ratio=2.62); the relationship with oppositional defiant disorder may be inflated by item overlap (21). Irritability in youths is associated with suicidal thoughts and behaviors in community cross-sectional samples (odds ratio=1.08–4.40) and longitudinal samples, especially when irritability is accompanied by depressive symptoms (odds ratio=1.15–2.22) (51).

Recent research has identified developmental trajectories of irritability. The highest priority for epidemiological research on irritability is to conduct related mechanistic studies and an examination of the continuity of irritability into adulthood. Studies should have multiple informants, including self-report measures in adolescents (especially in girls) to avoid missing less overt symptoms. Finally, few studies have included underrepresented populations (34, 52).

DSM-5 AND BEYOND

“Irritability” or “anger” appears as a symptom in 20 conditions in DSM-5, including disorders characterized by aggression. Episodic irritability occurs in manic episodes, depressive episodes in youths, withdrawal syndromes, and premenstrual dysphoric disorder. Chronic irritability manifests in multiple disorders, including generalized anxiety disorder and posttraumatic stress disorder; it is a cardinal symptom of intermittent explosive disorder, oppositional defiant disorder, and disruptive mood dysregulation disorder (DMDD); and it may be a major feature of neurodevelopmental disorders.

Here, we discuss three questions regarding irritability in DSM-5. One concerns the boundary between DMDD and mania. In the early 2000s, an emerging school of thought maintained that bipolar disorder did not present in youths with distinct manic episodes, as it does in adults, but instead with nonepisodic irritability and hyperarousal (53). Because the latter phenotype overlaps considerably with ADHD, there was an upsurge in the rate at which youths were given the diagnosis of bipolar disorder (54). However, subsequent research revealed that chronic, nonepisodic irritability in youths does not confer increased risk for mania later in life, indicating that the diagnosis of bipolar disorder should be reserved for youths with distinct manic episodes (21, 26). DMDD was included in DSM-5 to provide a diagnosis for children receiving the diagnosis of bipolar disorder inappropriately. ICD-11 did not adopt DMDD, opting instead for introducing a chronic irritability subtype of oppositional defiant disorder (35). DMDD has had rapid uptake in clinical practice and is associated with reduced rates of bipolar disorder diagnosis (55). However, patients with DMDD, like those with bipolar disorder, have high rates of antipsychotic and polypharmacy prescriptions (55), perhaps reflecting the lack of specific, evidence-based therapeutic options.

A second question relates to boundaries among the three DSM-5 diagnoses in which chronic irritability is a core symptom—intermittent explosive disorder, oppositional defiant disorder, and DMDD. DMDD is classified as a depressive disorder, whereas oppositional defiant disorder and intermittent explosive disorder are classified as disruptive and impulse-control disorders. Questions exist about the uniqueness of these syndromes, particularly given item (i.e., symptom) overlap. Few studies have compared them using external validators, such as familial aggregation, outcomes, and treatment response.

Third, the optimal threshold for differentiating pathological and subclinical irritability is unknown and is likely to vary developmentally (56). Confusion about the threshold and the use of different assessment instruments likely contribute to the greater than 10-fold difference in reported prevalence rates of DMDD across studies (ranging from lower estimates of <0.5% to upper estimates of >8%) (4, 50, 56-64) (Table 3); notably, such variation is also found in more established psychiatric disorders (65).

TABLE 3.

Prevalence of disruptive mood dysregulation disorder (DMDD), tantrums, and irritable mood in community-based samples and comorbidity rates in youths with DMDD

Study	Cohort (Country)	N	Female (%)	Age (years)	Assessment	Prevalence	Comorbidity
Brotman et al. 2006 (57)	Great Smoky Mountains Study (United States)	1,420	49	9–19	Child and Adolescent Psychiatric Assessment	Severe mood dysregulation: lifetime rate, 3.3%; severity criteria met, 1.8%	ADHD, 27%; conduct disorder, 26%; oppositional defiant disorder, 25%; anxiety disorder, 15%; depressive disorder, 13%
Copeland et al., 2013 (50)	Duke Preschool Anxiety Study (United States)	918	52	2–5	Preschool Age Psychiatric Assessment	DMDD: 3-month rate, 3.3%; 3-month rate if irritable mood and tantrums “every day”, 1.7%	ADHD, 31%; conduct disorder, 22%; oppositional defiant disorder, 68%; anxiety disorder, 73%, depressive disorder, 12%
Copeland et al., 2013 (50)	Great Smoky Mountains Study (United States)	1,420	49	9–17	Child and Adolescent Psychiatric Assessment	DMDD: 3-month rate, 1.1%; cumulative rate by age 16, 4.4%	ADHD, 6%; conduct disorder, 23%; oppositional defiant disorder, 57%; anxiety disorder, 9%; depressive disorder, 33%
Copeland et al., 2014 (50)	Caring for Children in the Community Study (United States)	920	50	9–17	Child and Adolescent Psychiatric Assessment	DMDD: 3-month rate, 0.8%	ADHD, 9%; conduct disorder, 19%; oppositional defiant disorder, 71%; anxiety disorder, 8%; depressive disorder, 36%
Dougherty et al., 2014 (58), 2016 (59)	Stony Brook University (United States)	462	46	6	Preschool Age Psychiatric Assessment	DMDD: 3-month rate, 8.2%	Internalizing disorder, 26%; externalizing disorder, 55%; either disorder, 61%; both disorders, 21%; ADHD, 33%; conduct disorder or oppositional defiant disorder, 22%; anxiety disorder, 22%; depressive disorder, 5.6%
Dougherty et al., 2016 (59)	Stony Brook University (United States)	473	46	9	Schedule for Affective Disorders and Schizophrenia	DMDD: 3-month rate, 1.3%
Althoff et al., 2016 (60)	National Comorbidity Survey–Adolescent Supplement (United States)	6,483	51	13–18	Composite International Diagnostic Interview	DMDD: single-point rates: broad criteria, 5.26%; without hypomania or mania, 1.71%; with outburst frequency criteria, 0.56%, without hypomania or mania and with frequency criteria, 0.12%	Any, 93%; ADHD, 32%; conduct disorder or oppositional defiant disorder, 68%; anxiety disorder, 46%; mood disorder, 58%; substance use disorder, 43%; suicidality, 48%
Munhoz et al., 2017 (61); Laporte et al., 2021 (4)	2004 Pelotas Birth Cohort Study (Brazil)	3,490	48	Mean=11, SD=0.4	Development and Well-Being Assessment	DMDD: single-point rate, 2.5%	Any, 63%; ADHD, 44%; conduct disorder or oppositional defiant disorder, 21%; anxiety disorder, 23%; mood disorder, 12%
Chen et al., 2019 (62); Lin et al., 2021 (63)	Taiwan’s National Epidemiological Study of Child Mental Disorders (Taiwan)	4,816	48	7–15	Schedule for Affective Disorders and Schizophrenia, epidemiological version	DMDD: lifetime rate, 0.3%; 6-month rate, 0.2%	ADHD, 63%; conduct disorder, 7%; oppositional defiant disorder, 37%; anxiety disorder, 40%; depressive disorder, 13%
Wiggins et al., 2021 (56)	Multidimensional Assessment of Preschoolers Study enriched with clinical cases (United States)	388	52	Mean=4.7, SD=0.9	Preschool Age Psychiatric Assessment	DMDD: 3-month rate, 6.7%
Herzoff and Tackett, 2016 (64)	Community samples from southern Ontario and from Texas (Canada and United States)	439, 291	51	9–10, 8–10	Computerized Diagnostic Interview Schedule for Children	Oppositional defiant disorder–irritability: 14.7% and 3.1%

The most important research priority within the DSM framework is for cross-cultural epidemiological studies to derive data-driven thresholds for irritability (4) and to facilitate harmonizing intermittent explosive disorder, oppositional defiant disorder, and DMDD. Longitudinal studies could identify criteria for nonnormative irritability based on outcomes. External validity could be tested using family measures, physiological measures, and treatment response.

Outside the DSM framework, a high-priority research goal is integrating a dimensional perspective into diagnosis because the boundary between normative and pathological irritability is unlikely to be sharp. Dimensionalizing the symptoms comprising or frequently associated with irritability (e.g., temper outbursts and reactive aggression) might help characterize the heterogeneous presentations of irritability and facilitate mechanistic studies. Dimensional approaches (i.e., the Hierarchical Taxonomy of Psychopathology [66] and the RDoC [67]) presume that irritability-related symptoms are expressions of a shared liability or are hierarchically organized, with general psychopathology factors and lower-level domains and subfactors.

EARLY CHILDHOOD

Although all young children exhibit temper outbursts, individual differences in outbursts can be a transdiagnostic risk marker for later psychopathology. Studying such differences can aid in prevention. A recent meta-analysis found small associations between infant irritability and later internalizing and externalizing symptoms; for preschool children, the associations with irritability were small to moderate (68). Specific markers of atypical irritability, including dysregulation (e.g., tantrums until exhausted), occurrence in unexpected contexts (e.g., grumpy during fun activities), and high frequency (daily or more), can be identified as early as 12 months of age (69). In one longitudinal study of children recruited at pediatric visits and oversampled for irritability (N=425; mean baseline age was 4.7 years and mean follow-up time was 2.9 years; 51% were girls), low frustration tolerance and destructive tantrums at baseline were associated with impairment and DSM-5 diagnoses in both cross-sectional and longitudinal analyses (25). Repeated assessment was important because single assessments yielded frequent false positives but few false negatives (56). Secondary data analyses suggest that early irritability responds to interventions designed to promote early self-regulation (70, 71).

Studying brain-behavior substrates of early irritability is challenging, but methodological advances have yielded interesting pilot data. One study using functional near-infrared spectroscopy linked irritability to interdyadic differences in neural and behavioral synchrony between parents and their children (72). Another study found that irritability was inversely correlated with striatal and anterior cingulate activity during frustration (73).

The highest-priority research aim is to develop tools that combine behavioral patterns, neural biomarkers, and parent-child interactional processes to identify which highly irritable young children will develop clinical problems and what malleable protective processes can be leveraged for prevention. Such tools could allow the integration of personalized prediction of risk into pediatric primary care (74). Improvements in early identification will enable the development of interventions designed to increase adaptive parent-child coregulation.

MIDDLE CHILDHOOD AND ADOLESCENCE

Multiple community-based studies have described longitudinal developmental trajectories of irritability (Table 2). One large study (N=4,462) followed a community sample of youths, oversampled for children of unmarried parents, from ages 3 to 15 (34). Results showed that preadolescence and adolescence are important inflection points in several trajectories and that groups were differentiated based on harsh or neglectful parenting and child internalizing symptoms.

Irritability in middle childhood and adolescence is often associated with co-occurring symptoms that complicate clinical care. Irritability can obscure internalizing psychopathology, raising questions about the primary focus of treatment and whether treating the co-occurring problem could worsen the irritability. When irritability co-occurs with another syndrome or diagnosis, it is associated with increased impairment and a worse course (23, 75).

Irritability confers risk for suicidality and nonsuicidal self-injury to a greater extent than do major depressive disorder and other risk factors. Increased irritability often precedes suicidal behavior (45, 76), particularly in the context of high irritability during middle childhood (77), and increases in irritability may be useful in detecting near-term risk of suicide. Mechanisms mediating associations between irritability and suicidality are unknown, although several models have been proposed (e.g., irritability as an endophenotype of suicidality, a risk factor for suicidality because of associations with other forms of psychopathology, or a risk factor for the transition from suicidal ideation to suicide attempt) (76).

In adolescence, irritability may play a role in increased mood problems and risk-taking behaviors, including substance abuse. For instance, irritability mediates the association between ADHD and alcohol use problems (78), and temper outbursts (i.e., phasic irritability) predict substance use and risky sexual behaviors (28).

The highest-priority research questions in this area revolve around suicidality. It is important to test whether interventions that target irritability reduce suicidality and other clinical problems (e.g., substance abuse) and to study mechanistic links between irritability and depression. Further research on longitudinal outcomes of tonic and phasic irritability (28, 79) is warranted and could include studying the mechanisms underlying these relationships. Given high correlations between tonic and phasic irritability, large samples will be needed. Finally, researchers should examine whether minority stressors are associated with increased youth irritability and suicidality and thereby contribute to mental health disparities in sexual and gender minority youths and other marginalized populations.

COMORBIDITY WITH NEURODEVELOPMENTAL CONDITIONS

Youths with ADHD or ASD have high rates of irritability, often occurring in the context of more generalized emotion dysregulation—that is, excitability, anxiety, and lability. Over half of youths with ADHD or ASD have emotion dysregulation (6, 15), and approximately 40% of 7- to 12-year-olds with ADHD experience extreme irritability; a similar proportion experience extreme surgency with more moderate irritability (15). Emotion dysregulation, including irritability, is related to genetic liability for ADHD (15). Thus, emotion dysregulation may be a central feature of ADHD, complicating both assessment and treatment. Data suggest that variation in emotion dysregulation may form a basis for ADHD subtypes (15, 80).

The clinical significance of irritability in ASD is illustrated by 45 clinical trials with irritability as the outcome (81). However, in the context of these clinical trials, irritability was often defined to include self-injury and aggression, with an insufficient focus on a propensity for anger. There are few studies focused on the mechanisms underlying irritability in ASD. Some studies have used tasks from ADHD research to examine responses to frustration and associated physiological arousal in ASD (82), but it is unknown whether factors influencing irritability are shared across these disorders. Irritability may stem from core symptoms of these disorders: for example, inattention may contribute to blocked goal attainment and frustration for an individual with ADHD or to violations of preference for sameness for an individual with ASD. However, given broad emotion dysregulation in both populations, cross-domain self-regulation difficulties (e.g., difficulties in top-down regulatory mechanisms) may play a larger role than diagnosis-specific features.

ADHD, ASD, and irritability all have an early age at onset (83). The highest-priority research question in this area is whether irritability is a feature of those conditions or an early indicator of or risk factor for a secondary condition. Do alterations in the development of top-down control in early life influence both impulsivity and irritability? It is also important to study the longitudinal course of irritability into adulthood for individuals with ADHD and ASD, including the impact of treatment and the impact of the stress associated with the neurodevelopmental conditions themselves.

TREATMENT

Evidence for the effective treatment of irritability is increasing, yet few randomized controlled trials (RCTs) use irritability as a primary or secondary outcome. Our systematic review found 14 studies evaluating a medication or a psychological intervention in the past 5 years in the treatment of irritability (excluding studies focused on ASD) (see Figure S1 and Table S2 in the online supplement). However, most of these studies are at high risk of bias by either including no control group or including a control group where unblinding is likely (see Figure S2 in the online supplement). Here we also include information from articles published before 2018; for these studies, we used a narrative review approach (see the online supplement).

Parent management training teaches parents reinforcement-based methods, reduces externalizing problems in youths under age 13 (an effect size of approximately d=0.5) (33), and is scalable. Indeed, the efficacy of parent management training is substantial and has been shown in many trials of good quality. Considering the efficacy of this approach across methods, cultures, and treatment settings (84), it should be studied explicitly in young people with irritability.

Cognitive-behavioral therapy (CBT) shows medium effect sizes in children and adolescents with externalizing symptoms, disruptive and antisocial behavior, and anger-related problems; however, the relevant studies were small and often included inactive comparators (85). Transdiagnostic CBT programs (e.g., MATCH and Unified Protocol) show promise (86), as do behavioral analysis and skills enhancement programs (87, 88). Preliminary data support the potential efficacy of dialectical behavior therapy (81, 89), interpersonal psychotherapy (82, 90), and exposure-based treatments for irritability (83, 91). Irritability may also be reduced as a secondary outcome with a treatment whose primary target is another condition (e.g., children who received CBT with exposure and response prevention for their primary diagnosis of obsessive-compulsive disorder [92]).

When irritability does not respond to behavioral interventions, medication augmentation is common. Risperidone and aripiprazole have a U.S. Food and Drug Administration indication for treating irritability in individuals with ASD. Although findings about antipsychotic treatment of patients with ASD are often extended to those without ASD, practitioners should do so with caution and ensure that non-pharmacological interventions have been optimized. There are a number of concerns regarding generalizing findings from the ASD literature. First, the irritability measure in these trials was designed for individuals with developmental disabilities (e.g., measures include items related to self-harm [93]). Second, medication side effects (e.g., rapid weight gain) can lead to unblinding and inflated effect sizes, or they can result in differences from placebo because of nonspecific effects (e.g., sedation). Finally, antipsychotic efficacy should be balanced against the high frequency and severity of side effects (94).

Secondary data analyses show that stimulants may reduce irritability in youths with ADHD (95). Thus, in children with ADHD and irritability, ADHD medication should be optimized first (5). Antidepressants may also be useful. A small trial in youths with DMDD and ADHD found that citalopram plus methylphenidate reduced irritability more than placebo plus methylphenidate (96).

Clinically, the top priority is to conduct rigorous RCTs in children with severe irritability that does not improve when the primary disorder has been treated. Careful assessment of both efficacy and side effects is important in both pharmacological and psychological trials. Increased mechanistic research is essential. For example, studies testing whether irritability and emotion dysregulation respond differently to treatment would be clinically important and informative regarding the external validity of both constructs. As another example, an intervention that exposes youths to anger-inducing events (91, 97) would build on work showing predictive associations between irritability and aberrant brain reorganization after frustration (29), but the relevance of these mechanistic findings to treatment response has not been tested. Mechanisms that operate during the session (e.g., enhancing distress tolerance, habituation, or extinction) should also be studied.

NEUROIMAGING

Neuroimaging research on irritability has increased but faces the challenges of reproducibility and small effect sizes seen in clinical neuroscience broadly (98) (see Tables S2-S4 in the online supplement). A meta-analysis of 30 task-based fMRI studies showed no spatial convergence of neural activation associated with irritability across or within neurocognitive domains, including emotional reactivity, cognitive control, and reward processing (99). Studies of irritability and functional connectivity of brain regions during a task or during rest have yielded inconsistent results (100). The amygdala has been the focus of the largest number of functional connectivity studies, which have had mixed results. These inconsistent findings may reflect methodological variability, heterogeneity in associations between irritability and brain function, or both. The literature on structural MRI is much smaller but has also shown inconsistent findings in relation to differences in gray or white matter volume or microstructure (99).

A few studies of irritability in preschoolers have used event-related potentials as metrics of executive function, reward, and error processing (101). Further work should test constructs and event-related potentials more consistently across development. A few investigations of frustration and cognitive control with functional near-infrared spectroscopy found that preschoolers with irritability have lower prefrontal activity, suggesting greater difficulty regulating frustration (102).

In summary, the neuroimaging literature on irritability is small, young, and inconclusive. Cross-sectional studies have not yielded neurobiological markers. Methodological adaptations could yield more promising results. First, neuroimaging researchers should employ developmentally sensitive, multi-informant, and valid measures of irritability. Second, existing cross-sectional studies can be used to inform longitudinal studies of neural mechanisms underlying intraindividual changes in irritability resulting from development or treatment. Third, for fMRI studies, task designs that evoke relevant affective states (i.e., frustration) have greater ecological validity and evoke stronger emotional responses, and multivariate measures identified using machine learning are more reliable than those identified using univariate or single-region approaches, improving rigor and reproducibility. For example, a recent pilot study assessing resting-state scans before and after an experience of frustration found associations between irritability and postfrustration connectivity across limbic, reward, and sensorimotor networks (29), a finding that merits follow-up in future studies. Fourth, future work should capitalize on technological and analytic advances that integrate multiple modalities (e.g., EEG and fMRI) to probe interactions between underlying structural and functional connectivity and task-based responses at multiple timescales. Multisite collaborations will enable large sample sizes, internal and external validation, and more robust investigation of socioenvironmental factors, including race/ethnicity, cultural context, and psychiatric comorbidity. Such robust investigations could identify mediators and moderators and may differentiate more homogeneous subgroups. These methodological advances could yield an enhanced understanding of specific neurobiological mechanisms underlying irritability and guide development and testing of treatments with tractable neural targets (e.g., real-time fMRI neurofeedback or transcranial magnetic stimulation).

GENETICS

Twin studies suggest that irritability is moderately heritable, with estimates ranging from 22% to 51% (18-20, 103, 104) (Table 4). Genetic contributions to irritability appear to be dynamic, with stable genetic influences from childhood to adolescence followed by novel genetic and environmental influences later in life (20). What is distinctive about genetic and environmental influences on irritability, and how do they inform our conceptualization?

TABLE 4.

Key findings from twin studies of irritability

Study	Country	N	Age (years)	Study Design	Irritability Measurea	Key Findings
Stringaris et al., 2012 (18)	United Kingdom	2,652 pairs	12–19	Cross-sectional and longitudinal	ASEBA, youth and adult self-report	Irritability predicts and shares genetic variance with depression. Headstrong or hurtful behaviors share genetic variance with and predict delinquency.
Merwood et al., 2014 (19)	United Kingdom	1,920 pairs	5–18	Cross-sectional	Conners 10-item measure of emotional lability; parent report	Emotional lability shows phenotypic and genetic overlap with ADHD symptoms.
Savage et al., 2015 (103)	Sweden	1,348 pairs	8/9–19/29	Longitudinal	ASEBA, youth self-report and parent report	Child irritability predicts early adolescent anxiety or depression. Most variance and covariance over time is explained by genetic factors.
Mikolajewski et al., 2017 (104)	United States	1,225 pairs	11 and 17	Longitudinal	Diagnostic Interview for Children and Adolescents–Revised; parent report, interview	Irritability predicts overall but not specific internalizing disorders. Headstrong or hurtful behaviors predict substance use. Both predict antisocial and externalizing behavior.
Roberson Nay et al., 2015 (20)	Sweden	2,490 children	8/9, 13/14, 16/17, 19/20	Longitudinal	ASEBA (Child Behavior Checklist and Adult Behavior Checklist)	Male participants: heritability increases (36% to 89%) from early childhood to adulthood. Female participants: early genetic influences decline with age.

^aASEBA, Achenbach System of Empirically Based Assessment.

First, the magnitude of heritability appears to differ across sexes, which is unusual. Furthermore, heritability increases for males and decreases for females from childhood to adolescence (20).

Second, a central question at the genetic level is whether irritability is best conceptualized as a mood problem (as in DSM-5), an oppositional defiant disorder subtype (as in ICD-11), or closely related to ADHD and other neurodevelopmental disorders. An early twin study (18) showed that irritability predicted later depression and, unlike headstrong or hurtful oppositional defiant disorder symptoms, showed a genetic correlation with depression (r=0.70, 95% CI=0.59–0.82). This finding has been replicated and highlights the distinction of irritable versus headstrong or hurtful oppositional defiant disorder symptoms. Nonshared environmental influences make a substantial contribution to irritability but shared environmental factors do not (18), whereas shared environment factors make a substantial contribution to behavioral problems (105); these findings again suggest a distinction between behavioral problems and irritability. However, a twin study of emotional lability, a construct comprising irritability and mood volatility, showed substantial genetic overlap with ADHD (19), whereas other studies suggest that irritability predicts and shows genetic overlap with antisocial behavior (104). Therefore, the available data are mixed, and further research is needed.

More recent genetic investigations have used composite measures of common genetic risk variants (polygenic scores [PGSs]). In a longitudinal population-based study, an early-onset, persistent form of irritability was associated with an ADHD PGS, male preponderance, and ADHD (47). A later, adolescent-onset form of irritability with an increasing trajectory was associated with a depression PGS, female preponderance, and adolescent depression. In addition, a large cross-sectional study found that an ADHD PGS was associated with irritability, particularly global emotion dysregulation, in youths with ADHD (106). These findings, together with twin studies, suggest that irritability is related to ADHD genetic liability and that early-onset male irritability may be more neurodevelopmental in origin and later-onset irritability may be more akin to other mood problems, although the groups overlap.

Finally, gene-environment correlations—for example, when genetically encoded child characteristics evoke parental irritability (evocative correlation) or when a child inherits not only genes but also the environment of irritable parents (passive correlation)—are probably abundant in developmental psychology but are difficult to study (107). However, both clinical experience and developmental theory suggest that person-environment evocative correlations exist such that parent-child dyads engage in mutually coercive, angry cycles (108). Indeed, the efficacy of parent management training is thought to be predicated on breaking this cyclic reinforcement of aberrant behaviors, including irritability.

The highest-priority goal in this area would be to use longitudinal genetic designs to test whether irritability represents the same construct by gender, across ages, and in the context of different disorders (e.g., major depressive disorder vs. ADHD). A related question is whether irritability indexes biological or genetic heterogeneity and could be useful for stratifying treatment and prognosis. A third topic involves examining the genetic architecture of phasic versus tonic irritability, given evidence from twin studies suggesting differences (109). Detailed clinical inquiry among diverse patient groups and large-scale genetic data can be used to address clinically important questions on how best to conceptualize and treat irritability.

TRANSLATIONAL RESEARCH

Translational models that are guiding work to bridge pre-clinical and clinical research on irritability center on two constructs: frustrative nonreward and reactive aggression.

Frustrative nonreward was first described by Amsel in rodents and has since been observed in multiple other species, including humans (7, 110). Frustrative nonreward is the normative response to frustration, or blocked goal attainment; this response can be viewed as a form of negative prediction error (11). The frustrative nonreward response is characterized by increased activity, increased aggression (typically measured with the resident intruder test [111]), and resistance to extinction of conditioned responses. Data suggest that aberrant behavioral and neural responses to frustration are central to pediatric irritability (29, 112), supporting the clinical relevance of frustrative nonreward paradigms in animals.

Rodent studies induce frustrative nonreward using operant conditioning followed by omission or diminution of the expected reward. Studies of the frustrative nonreward response in adult animals implicate the central amygdala and dopaminergic pathways in the frontal cortex and ventral striatum (97, 98, 113, 114). However, these paradigms often require lengthy training and mature motor skills, making them less applicable to juvenile animals (115, 116). One new paradigm addresses these limitations by capitalizing on the mouse’s tendency to explore two places alternately. This paradigm elicits expected frustrative nonreward responses in juvenile mice without affecting behavior on tasks modeling anxiety, depression, or nonaggressive social behavior (117). Using this paradigm and measuring c-fos expression, investigators found frustrative nonreward–induced activation in multiple regions and a brain-wide shift toward a more integrated, network organization, broadly consistent with one human study (29). Future frustrative nonreward studies in rodent models using techniques such as fiber photometry optogenetics (118) could guide human fMRI frustrative nonreward studies.

In certain contexts, reactive aggression, or aggressive responses to threat, are adaptive (119). However, youths with irritability often have maladaptive or excessive aggressive responses to perceived threat and other negative stimuli (120-124). Reactive aggression is defensive, characterized by impulsivity, and associated with emotional and physiological hyperarousal. In contrast, proactive or instrumental aggression is goal-driven, rewarding, and associated with hypoarousal (125); it is not a model for irritability. A rodent model of hyperaroused reactive aggression can be created using alcohol or anabolic steroid exposure, inducing frustration, or priming aggression through exposure to an opponent (119). In nonhuman primates, the human intruder paradigm can be used to model reactive aggression (126). Human data finding associations between irritability and autonomic hyperarousal at baseline and after frustration demonstrate the translational potential of such models (82, 97). Human research could be extended by using other arousal measures, including actigraphy and pupillometry. Rodent studies of reactive aggression implicate the amygdala and possibly the medial prefrontal cortex, findings that are generally consistent with human studies in the literature (119, 122).

In irritability and other clinical domains, one barrier to clinically relevant cross-species research is that clinical pathophysiological studies focus on interindividual differences—that is, asking why some people are more irritable than others. However, such interindividual differences remain understudied in animals; exceptions include differences in aggressive or anxiety-related threat responses (127, 128). Given the central role of frustration in irritability, the highest-priority studies could examine interindividual differences in frustrative nonreward in animals. Human experimental medicine studies that pair a therapeutic intervention with mechanistic approaches parallel animal studies of mean intraindividual changes in response to an experimental (vs. a control) intervention; these types of studies provide another possible translational bridge. More discussion and collaboration between clinical researchers using neuroimaging to study irritability and animal researchers interested in the topic could foster the development of novel interventions.

FUTURE RESEARCH DIRECTIONS

Research on irritability has progressed. Significant advances have occurred in defining the scope and nature of the clinical problem, creating developmentally sensitive tools that measure irritability specifically, identifying developmental trajectories, confronting diagnostic challenges, leveraging clear clinical connections between frustration and irritability into mechanistic studies, obtaining preliminary treatment data, and elucidating genetic associations.

However, given the relatively recent focus on irritability, it is unsurprising that much work remains to be done. Five overarching research priorities have emerged (Table 1). The first is the need for studies in large, diverse samples of the impact of cultural and social context on the presentation of irritability, parental responses, and children’s experience. There has been related research on the broader construct of emotion dysregulation (129), and international efforts are being organized to undertake this type of research in irritability (https://medicine.yale.edu/childstudy/research/collaborative-labs/cross-cultural-consortium-irritability/).

A second major research need concerns the external validity of the irritability construct and the measurement tools used to assess it. This requires going beyond self- and parent-report measures and confronting the general lack of agreement in psychology between subject reports and biomarkers such as brain imaging and physiology (130). However, because self- and parent-report measures will always be essential to the clinical endeavor, a third research need is for greater understanding of the factors driving differences in informant reports (29, 41).

A fourth major research priority encompasses understanding the transdiagnostic nature of irritability. One important and tractable question is whether treatments targeting irritability also diminish co-occurring symptoms associated with, for example, ADHD, depression, or anxiety. An underlying mechanistic question regards the extent to which irritability manifesting in different clinical contexts is mediated by similar brain mechanisms.

Finally and importantly, the evidence base for treatment must be expanded through rigorously conducted, well-powered RCTs that either repurpose existing treatments or test novel interventions. More mechanistic work is needed to guide the development of novel interventions. Regarding mechanisms, neuroimaging studies would benefit from the adoption of multiple methodological advances, and increased discussion and collaboration is needed between clinical researchers and those doing relevant basic and translational research (110, 113). Although this lays out an ambitious research agenda, the scope and importance of the clinical problem, coupled with recent advances laying the groundwork for future studies, warrant the investment.

Continuing Medical Education

You can earn CME credits by reading this article. Three articles in every American Journal of Psychiatry issue comprise a short course for up to 1 AMA PRA Category 1 Credit^™ each. The course consists of reading the article and answering three multiple-choice questions with a single correct answer. CME credit is issued only online. Readers who want credit must subscribe to the AJP Continuing Medical Education Course Program (psychiatryonline.org/cme), select The American Journal of Psychiatry at that site, take the course(s) of their choosing, complete an evaluation form, and submit their answers for CME credit. A certificate for each course will be generated upon successful completion. This activity is sponsored by the American Psychiatric Association.

Examination Questions for “Irritability in Youths: A Critical Integrative Review”

1. Chronic, as compared to episodic irritability, is a characteristic symptom of:

   1. Pediatric bipolar disorder
   2. Oppositional defiant disorder
   3. Enuresis
   4. Arachnophobia
2. Regarding the genetics of irritability, which of the following statements is true:

   1. It is highly (>70%) heritable
   2. It has a substantial component of non-shared environmental influences
   3. Early onset irritability does not share genetic risk with ADHD
   4. Early onset irritability is mainly linked with depression genes
3. Which of the following statements about the treatment of irritability are true:

   1. Irritability in ADHD frequently improves with the administration of stimulant medication
   2. Irritability in disruptive mood dysregulation disorder does not respond to citalopram added to stimulant medication
   3. Antipsychotics should be the first line treatment for non-autistic children with irritability
   4. Parent management training is not a promising treatment for irritability