Characterizing the spectrum of irritability in preadolescence: Dimensional and pragmatic applications

Abstract

Objectives

Characterize the dimensional spectrum of preadolescent (PA) irritability, a robust transdiagnostic vulnerability marker, using the youth version of the Multidimensional Assessment Profiles Temper Loss (MAPS‐TL‐Youth) scale including common and with developmentally specific items. Based on this, derive and validate a clinically optimized irritability screener to flag psychopathology risk in preadolescents.

Methods

The normal:abnormal irritability spectrum was modeled using MAPS‐TL‐Youth data from the Multidimensional Assessment of Preschoolers Study (MAPS) Study PA wave (n = 340) via item response theory. Both cross‐cutting core items from the MAPS scales and developmentally specific items were used to generate this dimension. Stepwise logistic regression was then used to optimize MAPS‐TL‐Youth irritability items in relation to Kiddie Schedule of Affective Disorders and Schizophrenia impairment to generate a clinically optimized irritability screener. Receiver operator characteristic analysis identified the irritability threshold for the screener. For the first time, youth self‐report of their own irritability on the MAPS‐TL was also modeled via the MAPS‐TL‐Youth‐Self‐Report (MAPS‐TL‐Youth‐SR).

Results

Irritability was unidimensional and ranged from mild and common to severe and rare behaviors. Developmentally specific items allowed detection of more severe irritability. Items for the screener were identified in relation to concurrent impairment. These included low frustration tolerance and pathognomonic severe behaviors. The clinically optimized screener demonstrated very good sensitively (87%) and specificity (81%) in regard to concurrent irritability‐related DSM disorders. Modeling of the MAPS‐TL‐Youth‐SR yielded similar results.

Conclusion

Characterizing the normal: abnormal spectrum of irritability in preadolescence advances application of Research Domain Criteria methods to this developmental period. This foundational work yielded two developmentally specified tools for irritability characterization in preadolescence: a nuanced dimensional scale to precisely characterize the full normal‐abnormal irritability spectrum, and a pragmatic, clinically optimized screener suitable for real world use. Future application in mechanistic and clinical studies will be important for establishing validity and incremental utility.

1. INTRODUCTION

Irritability is the relative tendency to experience anger and frustration in response to blocked goal attainment and exists along a spectrum from developmentally normative to clinically concerning (Brotman, Kircanski, Stringaris, et al., 2017; Wakschlag et al., 2015). When elevated and/or dysregulated, irritability is one of the most robust and developmentally meaningful behavioral predictors of transdiagnostic psychopathology (Wiggins, Urena Rosario, Zhang, et al., 2023; Wiggins Urena Rosario, MacNeill et al., 2023). Burgeoning work on the measurement of irritability has been crucial for operationalizing this concept for scientific purposes, for example, characterizing the full dimensional spectrum, as well as its translation for clinical use, for example, developing screeners that quickly and efficiently identify youth with clinically concerning irritability (Wakschlag et al., 2023; Wiggins et al., 2018). Such work has been driven by developmental specification theory to differentiate atypical irritability in regard to normative variation within a developmental period (Wakschlag et al., 2010). To date, this measurement work has primarily been conducted with preschool age samples, as the challenge of typical vs. atypical differentiation is most pronounced due to overlap with normative misbehavior during this developmental period (Finlay‐Jones et al., 2023; Wakschlag et al., 2010; Wiggins et al., 2018). However, elevated/dysregulated irritability continues to be a transdiagnostic marker of risk for psychopathology beyond early childhood; indeed, moodiness and temper outbursts are a normative hallmark of adolescence (Evans et al., 2017; Vidal‐Ribas et al., 2016). Thus, identifying youth (ages ∼9–12 years) during the pre‐clinical phase who are at risk for impairing psychopathology prior to the dramatic uptick of mental disorders during (Merikangas et al., 2010) will be crucial for timely and effective prevention. Further bolstering the argument for identifying atypical irritability within the developmental context of preadolescence, variable irritability patterns have differential predictive utility (Damme et al., 2022), and evidence suggests that preadolescence is a particularly vulnerable turning point when trajectories of irritability can shift and change direction (Yu et al., 2023). Here, we sought to characterize the normal:abnormal spectrum of irritable behavior as developmentally specified within preadolescence. Guided by a “translational mindset” (Wakschlag et al., 2022), we then derived a parsimonious clinically optimized irritability screener for psychopathology risk in preadolescence, based on the most psychometrically discriminative irritable behaviors.

Irritability can manifest as temper outbursts and/or angry mood, and its phenotypic expression falls along a normal:abnormal spectrum that shifts over the course of development. For example, temper tantrums are typically considered normative responses in preschoolers and are only atypical when very frequent, dysregulated, or disproportionate to context. By contrast, tantrums are not frequently observed in older youth and adolescents (Brotman, Kircanski, & Leibenluft, 2017; Wakschlag et al., 2012). From early preschool age through adolescence, youth typically experience increases in brain maturation, self‐regulation capacities, and expectations for behavioral control. Accordingly, irritability normatively decreases with age, following a well‐defined trajectory that decreases into later childhood and further into adulthood (Brotman, Kircanski, & Leibenluft, 2017; Brotman, Kircanski, Stringaris, et al., 2017; Leibenluft & Stoddard, 2013; Sorcher et al., 2022; Wiggins et al., 2014). However, there exists a subset of youth for whom irritability remains chronically elevated and/or dysregulated across childhood and adolescence (Pagliaccio et al., 2018; Sorcher et al., 2022; Wiggins et al., 2014; Yu et al., 2023; Zhang et al., 2023). This highlights the need for developmentally meaningful assessment over time.

Longitudinal work suggests that the transition period between childhood and early adolescence (i.e., preadolescence) may be a particularly critical time when persistent irritability is associated with emerging psychopathology. For example, Savage et al. (2015) found that the predictive relationship from irritability to internalizing symptoms was strongest during the transition from late childhood (ages 8–9) to early adolescence (ages 13–14). Also, in a longitudinal study following youth from ages 3 to 15, Yu et al. (2023) found around age 9 is an inflection point in the trajectories of irritability, with irritability decreasing in some youth and increasing in others from age 9 through 15. This highlights the salience of modeling the normal:abnormal irritability spectrum, as well as a companion screener for real world use. There are a dearth of methods specifically developed for preadolescence that apply the NIMH Research Domain Criteria (RDoC) neurodevelopmental, dimensional and transdiagnostic approach (Cuthbert & Insel, 2013).

During preadolescence, youth are simultaneously experiencing physical, cognitive, emotional, and social transitions (Mascia et al., 2023), including prefrontal regions underpinning executive function gaining peak gray matter volume (Sung et al., 2021), as well as the onset of increased risk‐taking behaviors and mental health problems (Sung et al., 2021). Preadolescence also marks increased importance of peers compared to parents (Casey et al., 2010) and the drive to build their own identity and redefine their relationships with reference figures (i.e., parents and teachers) (Mascia et al., 2023). Given the multitude of developmental changes, assessment at preadolescence must take into account developmental context, characterizing the spectrum of normal:abnormal irritability based on knowledge of expectable developmental patterns in the preadolescent period (Wakschlag et al., 2010).

The Multidimensional Assessment Profiles Temper Loss (MAPS‐TL) Scale was specifically designed to characterize irritability severity within a developmental framework (Wakschlag et al., 2012, 2014). [Note: Originally termed the Multidimensional Assessment Profile of Disruptive Behavior, the scales have since been renamed the Multidimensional Assessment of Preschoolers Scales (MAPS) reflecting the expanded coverage of internalizing and other externalizing domains (see Wakschlag et al., 2023; Nili et al., 2023)]. [Correction added on 19 October 2023, after first online publication: In the preceding sentence, the term ‘Multidimensional Assessment of Preschoolers Study (MAPS) scales’ has been changed to ‘Multidimensional Assessment of Preschoolers Scales (MAPS)’.] The MAPS‐TL scale has been validated for early childhood in >6000 young youth across the infant‐toddler (Krogh‐Jespersen et al., 2021) and preschool periods (Wakschlag et al., 2012, 2014, 2018), and now at early school age and in adolescence (Hirsch et al., 2023; Kirk et al., 2023). Wiggins et al. have also generated MAPS‐TL clinically optimized screeners for early childhood ‐for preschool age (2018) and infant‐toddlers (Wiggins, Urena Rosario, MacNeill et al., 2023). These have demonstrated predictive utility included 4 to 7 times greater odds of subsequent impairment and psychopathology (Wiggins, Urena Rosario, Zhang et al., 2023; Wiggins, Urena Rosario, MacNeill et al., 2023; Wiggins et al., 2018). However, the normal:abnormal spectrum of irritability has not yet been characterized in preadolescent youth and a pragmatic screener drawing on this spectrum is lacking.

Thus, we sought to distinguish normal from abnormal expressions of irritability in preadolescent youth and generate a developmentally specific, clinically optimized irritability screener, with the aim of improving identification of those who may be at risk and/or in need of intervention. We do so via psychometric validation of the MAPS‐TL‐Youth scales. The theoretical framework that undergirds the MAPS scales emphasizes lifespan developmental coherence of measurement (balancing specificity with continuity over time) (Wakschlag et al., 2010). Thus, the youth scales contain core items validated for early childhood together with additional developmental specific items theorized to capture important features of irritability from school age to adolescence. Here we validate the MAPS‐TL‐Youth for preadolescents. Other papers in this issue validate it for early school age (Hirsch et al., 2023) and adolescents (Kirk et al., 2023). Moreover, as preadolescence marks a period when youth can report on their own internal states and behaviors, we additionally explore the properties of a self‐report version of the MAPS‐TL‐Youth (MAPS‐TL‐Youth‐SR) (see also Kirk et al., 2023). Goals of this paper are validation in a diverse sample of PA youth: (1) Assess the unidimensionality of the MAPS‐TL‐Youth irritability factor assessed via parent report. This includes testing the added value of developmentally specific items over and above core items; (2) Model this irritability factor along a severity spectrum, from mild, common to severe, rare behaviors; (3) Derive a clinically optimized screener for MAPS‐TL‐Youth for preadolescents via differential association to concurrent impairment; and (4) Optimize this screener for maximal sensitivity and specificity for DSM‐5 irritability‐related disorders. We additionally explored Aims 1–4 with parallel analyses based on youth report via the MAPS‐TL‐Youth‐SR.

2. METHODS

2.1. Participants

Data were derived from the preadolescent wave of the MAPS (N = 497), which included intensive assessment and survey only participants (for details, see Damme et al., 2022; Wakschlag et al., 2012; Wakschlag et al., 2015). The central longitudinal sample for the MAPS Study were the n = 425 families who participated in intensive lab‐based assessment at the preschool baseline visit. Eighty‐two percent (n = 348) of these families participated at the MAPS preadolescent wave. This preadolescent sample (n = 348) did not differ sociodemographically from the preschool age baseline sample (N = 497) [child gender (χ ² = 0.179, df = 1, p = 0.672), race/ethnicity (χ ² = 0.328, df = 3, p = 0.955), poverty status (χ ² = 0.108, df = 1, p = 0.742), or maternal education (χ ² = 3.516, df = 5, p = 0.621)]. Consistent with the initial sample, the preadolescent sub‐sample was diverse and fairly evenly distributed by child gender, race/ethnicity, poverty status, and mother's educational level (see Table 1).

TABLE 1.

Sample demographics for MAPS‐TL‐Youth sample (n = 340).

	N	Percentage
Youth gender female	183	52.5
Race/ethnicity
Hispanic	102	29.3
Non‐Hispanic African‐American	171	49.1
Non‐Hispanic White/Caucasian	70	20.1
Non‐Hispanic other	5	1.4
Poverty status
Below poverty level	167	47.9
Maternal education
Less than high school	15	4.3
High school or GED	67	19.2
Associate's degree/trade school	112	32.1
Some college (no degree)	49	14.0
Bachelor's degree	51	14.7
Graduate degree (MA, PhD, MD, JD)	48	13.8

Of the N = 348 families who participated in the preadolescent wave of MAPS, n = 8 were excluded from the parent report data analyses due to missing MAPS‐TL‐Youth data at this wave. Thus, the total analytic sample for modeling of parent reported data was n = 340 (M _age = 9.31 years, SD_age = 0.77 years; see Table 1 for demographics). Consistent with the initial sample, this sub‐sample was fairly evenly distributed by child sex, race/ethnicity, poverty status, and mother's educational level. See Table 1 for participant demographic information. Three hundred and seven youth participated in the preadolescent wave via self‐report surveys. However, 5 youth did not complete the MAPS‐TL‐Youth‐SR. Thus, the analytic sample consisted of n = 302 youths (M _age = 9.18 years, SD_age = 0.66 years).

2.2. Measures

2.2.1. Irritability

Irritability was assessed using the MAPS‐TL‐Youth scale by parent report and by youth self report (MAPS‐TL‐Youth‐SR). Of the 40 items on the MAPS‐TL‐Youth, 22 were core MAPS‐TL items developed through theoretical delineation of the core facet of irritability, preliminary validation utilizing secondary data, pilot testing, focus groups, and review of extant measures (Wakschlag et al., 2014). The MAPS‐TL‐Youth scale introduces 18 additional items that were theorized to capture facets and contexts salient to the normal:abnormal spectrum of irritability beyond preschool age, i.e., phenotypic expression from early school age to adolescence. The MAPS‐TL‐Youth covers irritability from normative expressions to dysregulated manifestations, and its items assessed irritability behaviors that emerge from contextual antecedents (“when tired, hungry or sick,” “when frustrated, angry or upset,” “during daily routines,” “to get something s/he wanted,” “for no reason or out of the blue”) and that occur across interactional contexts (i.e., with parents, with other adults, with siblings, with other youth). Each participant’s parent rated the occurrence of items on a 6‐point objective frequency scale (0 = Never; 1 = Rarely [less than once per week]; 2 = Some [1–3] days of the week; 3 = Most [4–6] days of the week; 4 = Every day of the week; 5 = Many times each day). Due to a low endorsement of the rating “5 = Many times each day”, we combined the scaling responses of “4 = Every day of the week” and “5 = Many times each day” to “4 = Every day of the week AND/OR Many times each day”, for both parent‐ and self‐reports. For the preadolescent vs. preschool version (see also Kirk et al., 2023), the word “temper tantrum” was replaced with “temper outburst,” due to considerations of face validity in this older age.

Youth‐reported irritability was also measured using the MAPS‐TL‐Youth‐SR. Thirty‐six items from the parent‐reported MAPS‐TL‐Youth were retained and were slightly reworded to generate a self‐report version to administer to youths (See Table S1 for contrast of MAPS‐TL parent‐ and youth‐report items and core vs. developmentally specific items).

2.2.2. Clinical diagnoses

Participant diagnoses were assessed via parent interview using the Kiddie Schedule for Affective Disorders and Schizophrenia–Present and Lifetime Version (K‐SADS‐PL; Kaufman et al., 1997). The K‐SADS‐PL captures diagnoses based on the Diagnostic and Statistical Manual of Mental Disorders (DSM–5; American Psychiatric Association, 2013). DSM irritability‐related syndromes (here defined as those with explicit irritability symptoms, i.e., oppositional defiant [ODD], disruptive mood dysregulation [DMDD], major depressive [MDD] and persistent depressive [PDD] disorders). Concurrent presence of any of these disorders served as the validation outcome for deriving the cutoff score for the screener (see Statistical Approach). Reliability among interviewers was acceptable, with kappa = 0.79–1.0 across diagnoses, based on 20% of interviews independently re‐rated.

2.2.3. Impairment

Concurrent impairment was measured using the Developmentally and Contextually‐Based Rating Impairment Scale, a supplement to the K‐SADS‐PL. The K‐SADS‐PL has a pre‐existing impairment rating scale which measures impairment with peers, family, and school. The Developmentally and Contextually‐Based Rating Impairment Scale expands these areas of functioning to include eight distinct contexts of impairment: life at home, “out and about”, at school; and relationships with parents, siblings, peers, nonparental adults, and teachers. Impairment was assessed at the end of each K‐SADS interview section as it relates to a specific set of symptoms assessed in each section (e.g., Depression symptoms, ODD symptoms, etc.), and can be measured as “no impairment”, “minimal impairment”, “moderate impairment”, or “severe impairment”. Cross‐domain impairment was defined as present if there was moderate or severe impairment in two or more of the eight different areas of functioning. (See Supporting Information S1 for details.).

2.3. Statistical approach

Aim 1: Assess the unidimensionality of the parent‐report MAPS‐TL‐Youth irritability factor. A confirmatory factor analysis (CFA) was conducted using Mplus 8 (Muthén & Muthén, 1998‐2017) to evaluate unidimensionality of the MAPS‐TL‐Youth items. As the data are ordinal and do not follow a continuous and multivariate normal distribution, a weighted least square mean and variance adjusted (WLSMV) estimator was used. Data were specified as categorical (which in Mplus, includes ordinal data). Root‐mean‐square error of approximation (RMSEA), Tucker‐Lewis index (TLI), and the comparative fit index (CFI) were used to assess an acceptable model fit for CFA, as they tend to perform well with respect to detecting model misspecification and do not depend on sample size (Jackson et al., 2009). The criteria set for the preceding fit measures were based on Hu and Bentler (1999), who recommend that the assumption of a good fit between the hypothesized model and the observed data should have a cutoff value approximating 0.06 for RMSEA and a cutoff value approximating 0.95 for TLI and CFI. A second CFA was run to assess the unidimensionality of the irritability factor for the MAPS‐TL‐Youth‐SR, using the same approach.

Aim 2: Model developmentally specific and core items contributing to the irritability factor along a severity spectrum, from mild, common to severe, rare behaviors. IRTPRO 5 (Cai et al., 2020) was used to apply item response theory methods to estimate each item parameter. A graded response model was applied, due to the ordered categorical nature of the response categories. An expected a posteriori (EAP) estimation was used with a normal distribution to reflect the population mean (μ = 0) and standard deviation (σ = 1). These parameter outputs (i.e., item difficulty, item discrimination, latent trait threshold) were equipped to estimate the “severity” of each item and align them along the dimension from “least severe” to “most severe.” The developmentally specific items in the MAP‐TL‐Youth scale were calibrated along with the core items, using the item parameters of the preschool version of the MAPS‐TL (Wakschlag, et al., 2012). The 95th percentile of the trait score values, i.e., the EAP scores, which identified each individual's trait level along the Θ scale based on their response to a set of items, was calculated and used as a cutoff to indicate severity; using this parameter, items with a location parameter above the 95th percentile value were considered to be “clinically severe.” The category threshold (b1–b4) for each frequency in each item represented the estimated level on the irritability spectrum at which there is a >50% probability that the respondent would choose the keyed response category or higher. An item location was calculated by taking the average of each category threshold or difficulty parameter for the item. Item locations that exceeded the 95th percentile of the trait score were considered “severe” or “problematic,” regardless of the frequency at which they occurred (i.e., “Keep on having a temper outburst even when you tried to help him/her calm down”). Furthermore, if an item location did not exceed the 95th percentile, but had its category threshold for a particular frequency category exceed the 95th percentile value, that item was considered “problematic” or “severe” only if the behaviors measured by the items were manifested at that certain frequency or higher (i.e., “lose temper or have a temper outburst during daily routines, such as bedtime, mealtime, or getting dressed” is considered to be severe only when it occurs most days of the week or higher).

A second IRT was applied to the 36 items for the MAPS‐TL‐Youth‐SR to estimate each item parameter. Item parameters were obtained with the application of a Graded Response Model via IRTPRO 4.2 (Cai et al., 2011). In contrast to the parent‐report data analyses, all parameters for the self‐report were freely estimated and were not calibrated along with the previously established item parameters, due to the expected difference between parent‐report and self‐report. The 95th percentile of the trait value scores were also calculated and used as a cutoff value to indicate severity. Similar to the parent‐report, the item location was calculated and considered “severe” or “problematic,” regardless of the frequency at which they occurred if they exceeded the 95th percentile of the trait score value.

Aim 3: Derive a pragmatic irritability screener based on differential association to impairment. Here, we derived the individual irritability behaviors from the MAPS‐TL‐Youth most strongly related to cross‐domain K‐SADS impairment ratings. A stepwise logistic regression, using forward entry with likelihood ratio comparison of models, in IBM SPSS statistical software (IBM Corp., Armonk, NY) was used to identify the predictive value of each irritability item from the MAPS‐TL‐Youth scale using K‐SADS cross‐domain impairment as an indicator. Stepwise logistic regression tested items both individually and in combination to reduce the model to the items yielding the strongest relationship with impairment. The core 22 items of the MAPS‐TL are well validated items (Wakschlag et al., 2014; Wiggins et al., 2018), and thus were input into Block 1 of the stepwise logistic regression. The 18 developmentally specific items were added in Block 2 (these items are also validated for early school age by Hirsch et al., 2023, and adolescents by Kirk et al., 2023 in this special issue). These analytic steps were repeated for parent‐ and self‐reports separately.

Aim 4: Optimize irritability screener for maximal sensitivity and specificity in relation to DSM irritability‐related disorders. Clinically optimized cutoff scores for both parent and youth‐report were derived using the summed score of the reduced irritability item set generated via cross‐domain impairment from Aim 3. Implementing a receiver operating characteristics (ROC) analysis, the cutoff scores balanced sensitivity and specificity considering ODD, DMDD, MDD, and/or PDD diagnoses from the K‐SADS. Diagnoses for these four disorders were combined into one variable (present or absent) for youth who were diagnosed with any of these irritability related syndromes to provide a transdiagnostic cutoff point for irritability related disorders. The pragmatic score of the parent and self‐report irritability items derived in Aim 3 served as the criterion variable, and presence or absence of an irritability‐related diagnosis was the classification variable. To identify the cutoff scores, Youden index values were utilized to optimize the highest level of sensitivity (i.e., true positive rate) and specificity (i.e., true negative rate). Analyses were implemented in R with package Optimal Cutpoints (López‐Ratón et al., 2014). These steps were repeated for parent‐ and youth‐report.

3. RESULTS

Aim 1: Assess the unidimensionality of the MAPS‐TL‐Youth irritability factor (parent report). The results indicated that the scale measured one underlying latent construct, as the one‐factor model fit well statistically (χ ² [740, N = 340] = 1242.087, p < 0.001) and descriptively (CFI = 0.979, TLI = 0.978, RMSEA = 0.045). All standardized factor loadings were generally large and statistically significant (ranging from 0.680 to 0.917) (Cheung et al., 2023), indicating that all 40 items (core plus developmentally specific) were strongly representative of irritability. The factor structure of the one‐factor model of the MAP‐TL‐Youth parent‐report is presented in Table 2.

TABLE 2.

Standardized regression weights of MAPS‐TL‐youth items (parent report).

Items	Standardized regression weights
Lose temper or have a temper outburst when frustrated, angry, or upset	0.917
Have a temper outburst	0.901
Have a hot or explosive temper	0.899
Act angry, irritable, or grouchy throughout most of the day	0.891
Lose temper easily	0.882
Lose temper or have a temper outburst with you or other parent	0.88
Have a temper outburst until exhausted	0.879
Get extremely angry	0.877
Lose temper or have a temper outburst to get something he or she wanted	0.876
Stay angry for a long time	0.874
Keep on having a temper outburst even when you tried to help him/her calm down	0.872
Have a temper outburst that lasted longer than 5 min	0.864
Act angry, irritable, or grouchy no matter what you do	0.86
Lose temper or have a temper outburst “out of the blue” or for no reason	0.853
Have a “short fuse” (become angry quickly)	0.851
Make you feel that you have to “walk on eggshells” to avoid setting him/her off	0.851
Become frustrated easily	0.844
Lose temper or have a temper outburst with a teacher	0.839
Have difficulty calming down when angry	0.837
Lose temper or have a temper outburst during daily routines, such as bedtime, mealtime, or getting dressed	0.837
Act irritable	0.833
Cause family outings/activities to revolve around preventing him/her from having outbursts	0.828
Act grumpy or grouchy	0.826
Get in a bad mood even during fun activities	0.826
Have less expected of him/her than other youth in the family because of angry mood	0.826
Lose temper or have a temper outburst when tired, hungry, or sick	0.825
Get annoyed easily	0.816
Hit, bite, or kick during a temper outburst	0.814
Lose temper or have a temper outburst with other adults (e.g., teacher, babysitter, family member)	0.8
Lose temper or have a temper outburst with a friend a	0.793
Act angry all day long	0.792
Break or destroy things during a temper outburst	0.782
Yell angrily at someone	0.772
Act grouchy most of the day	0.771
Seem sullen	0.758
Complain about others, activities; not be satisfied by anything	0.758
Spoil something for your family because of his/her bad mood	0.752
Have temper outbursts that get in the way of getting along with other youth	0.745
Stamp feet or hold breath during a temper outburst	0.734
Lose temper or have a temper outburst when doing schoolwork	0.68

Note: Developmentally specific youth items, as specified in Hirsch et al., 2023 indicated in italics.

^a Items added for this preadolescent sample.

Assess the unidimensionality of the MAPS‐TL‐Youth‐SR irritability factor (youth self‐report). The results indicated that the items in the self report of the MAPS‐TL‐Youth scale measured one underlying latent construct as well, with good model fit (χ ² [740, N = 594] = 1025.588, p < 0.001) and descriptively (CFI = 0.946, TLI = 0.943, RMSEA = 0.049). All standardized factor loadings were statistically significant (Cheung et al., 2023) and presented in Table 3.

TABLE 3.

Standardized regression weights of MAPS‐TL‐youth items (self report).

Items	Standardized regression weights
Lose your temper easily?	0.762
Get extremely angry?	0.754
Lose your temper when you were frustrated, angry, or upset?	0.73
Have a temper outburst?	0.717
Feel angry?	0.708
Continue to have a temper outburst even when your parent tried to calm you down?	0.689
Have a hot or explosive temper?	0.671
Have a temper outburst that lasted longer than 5 min?	0.665
Break or destroy things during a temper outburst?	0.659
Lose your temper with your parent?	0.651
Have an attitude a	0.651
Get annoyed easily?	0.647
Hit, shove, or kick someone when you lost your temper?	0.641
Ruin a family activity because of your bad mood?	0.64
Have a temper outburst until you were exhausted?	0.618
Stay angry for a long time?	0.607
Lose your temper for no reason?	0.607
Feel angry or grouchy no matter what others do to try to make you feel better?	0.606
Lose your temper with other kids?	0.604
Feel grouchy?	0.599
Feel angry or grouchy most of the time?	0.594
Get angry over small things?	0.593
Lose your temper with a friend a	0.593
Lose your temper with your brother or sister	0.592
Yell at someone because you were angry?	0.591
Lose your temper to get something you wanted?	0.58
Get out of doing something because you were angry or grouchy?	0.58
Get rejected by friends because you were losing your temper?	0.579
Stamp your feet when you lost your temper?	0.572
Lose your temper with a teacher?	0.572
Complain about everything?	0.569
Get in a bad mood even during fun activities?	0.562
Have trouble calming down when you were angry?	0.541
Lose your temper when doing schoolwork?	0.518
Become frustrated easily?	0.443
Lose your temper when you were tired, hungry, or sick?	0.434

Note: Italics indicate items that were above the 95th percentile threshold.

^a Items added for this preadolescent sample.

Aim 2: Model the irritability normal:abnormal severity spectrum (parent report). In the IRT analysis, across the scale which indicated the range and severity of the irritability items, item locations ranged from as low as 0.88 (“get annoyed easily”) to as high as 2.71 (“act angry all day long”). The 95th percentile value was calculated to be 1.07, with 20% of all items scoring ≤95th percentile. Table 4 displays the item location mean and category threshold of each item. Figure 1 displays all items scaled from highest to lowest severity.

TABLE 4.

Difficulty parameters of MAPS‐TL‐youth items (parent report).

	Category thresholds
	Item location mean (b)	Rarely or higher (b1)	Some days of the week or higher (b2)	Most days of the week or higher (b3)	Every day/many times a day or higher (b4)
Items (95th percentile threshold = 1.07)
Stay angry for a long time	2.10	0.61	1.97	2.69	3.13
Hit, bite, or kick during a temper outburst	1.83	0.81	1.62	2.16	2.74
Have a temper outburst until exhausted	1.73	0.69	1.51	2.16	2.57
Lose temper or have a temper outburst with other adults (e.g., teacher, babysitter, family member)	1.70	0.38	1.47	2.28	2.67
Keep on having a temper outburst even when you tried to help him/her calm down	1.55	0.43	1.33	2.02	2.41
Have a temper outburst that lasted longer than 5 min	1.53	0.23	1.19	2.13	2.59
Break or destroy things during a temper outburst	1.45	0.63	1.53	2.17
Have a hot or explosive temper	1.45	0.45	1.26	1.84	2.25
Get extremely angry	1.44	0.33	1.21	1.92	2.31
Stamp feet or hold breath during a temper outburst	1.42	0.11	1.07	1.94	2.57
Lose temper or have a temper outburst “out of the blue” or for no reason	1.42	−0.67	1.53	2.21	2.60
Act irritable	1.34	−0.25	0.97	1.99	2.66
Have difficulty calming down when angry	1.30	−0.09	1.00	1.86	2.45
Have a “short fuse” (become angry quickly)	1.19	−0.07	0.92	1.69	2.22
Yell angrily at someone	1.15	−0.41	0.85	1.79	2.38
Lose temper or have a temper outburst with you or other parent	1.07	−0.15	0.83	1.59	2.00
Lose temper or have a temper outburst during daily routines, such as bedtime, mealtime, or getting dressed	1.06	−0.24	0.74	1.66	2.09
Lose temper or have a temper outburst to get something he or she wanted	1.04	−0.24	0.77	1.53	2.09
Have a temper outburst	1.04	−0.15	0.80	1.53	1.98
Lose temper or have a temper outburst when tired, hungry, or sick	1.03	−0.29	0.71	1.56	2.16
Become frustrated easily	1.02	−0.52	0.68	1.66	2.27
Lose temper or have a temper outburst when frustrated, angry, or upset	0.98	−0.28	0.75	1.45	2.00
Act angry all day long a	2.71	1.51	2.92	3.72
Have temper outbursts that get in the way of getting along with other youth a	2.44	1.05	1.79	3.07	3.84
Cause family outings/activities to revolve around preventing him/her from having outbursts a	2.33	1.08	1.89	2.77	3.59
Lose temper or have a temper outburst with a friend a	2.16	0.92	1.88	2.17	3.66
Have less expected of him/her than other youth in the family because of angry mood a	2.12	0.80	1.52	2.65	3.51
Seem sullen a	1.99	0.93	1.96	2.25	2.83
Lose temper or have a temper outburst with a teacher a	1.97	0.80	1.38	2.16	3.54
Act angry, irritable, or grouchy throughout most of the day a	1.94	0.62	1.51	2.33	3.28
Act grouchy most of the day a	1.94	0.81	1.88	2.39	2.67
Get in a bad mood even during fun activities a	1.92	0.49	1.39	2.26	3.55
Complain about others, activities; not be satisfied by anything a	1.84	0.63	1.39	2.57	2.78
Make you feel that you have to “walk on eggshells” to avoid setting him/her off a	1.71	0.92	1.52	2.10	2.31
Spoil something for your family because of his/her bad mood*	1.69	0.47	1.68	2.92
Lose temper or have a temper outburst when doing schoolwork a	1.60	0.17	1.07	2.22	2.95
Act angry, irritable, or grouchy no matter what you do a	1.49	0.41	1.43	1.83	2.28
Act grumpy or grouchy a	1.26	−0.24	0.95	1.96	2.39
Lose temper easily a	1.10	0.04	0.86	1.58	1.91
Get annoyed easily a	0.88	−0.54	0.63	1.56	1.88

Note: Each bolded value represents an item location mean or category threshold that exceeds the 95th percentile threshold.

^a One of the 18 additional items modeled for the youth version.

FIGURE 1.

Parent‐report items along severity dimension. Italicized items represent the 18 newly developed items.

Model the irritability normal:abnormal severity spectrum (self‐report). Across the irritability severity scale by youth report, item locations ranged from as low as 0.13 (e.g., “lose your temper with your brother or sister”) to as high as 2.54 (e.g., “ruin a family activity because of your bad mood?”). The 95th percentile value was 1.48, with 61.1% of all 36 items locating below the 95th percentile threshold value. Table 5 displays the item location mean category threshold of each item. Figure 2 displays all items scaled from highest to lowest severity.

TABLE 5.

Difficulty parameters of MAPS‐TL‐youth items (self report).

	Category thresholds
	Item location mean (b)	Rarely or higher (b1)	Some days of the week or higher (b2)	Most days of the week or higher (b3)	Every day/many times a day or higher (b4)
Items (95th percentile threshold = 1.48)
Ruin a family activity because of your bad mood?	2.54	1.2	2.11	3.06	3.78
Get in a bad mood even during fun activities?	2.18	0.82	1.83	2.78	3.32
Lose your temper for no reason?	2.06	0.94	1.67	2.68	2.95
Break or destroy things during a temper outburst?	1.97	1.15	1.66	2.37	2.72
Get rejected by friends because you were losing your temper?	1.91	1.05	1.81	2.29	2.5
Lose your temper when doing schoolwork?	1.87	0.46	1.55	2.46	3.01
Lose your temper with a teacher?	1.87	1.1	1.71	2.27	2.4
Complain about everything?	1.71	0.53	1.51	2.23	2.59
Feel angry or grouchy no matter what others do to try to make you feel better?	1.7	0.76	1.71	2.05	2.29
Hit, shove, or kick someone when you lost your temper?	1.64	0.65	1.46	2	2.46
Get out of doing something because you were angry or grouchy?	1.63	0.56	1.35	2.02	2.61
Feel angry or grouchy most of the time?	1.6	0.22	1.41	2.1	2.68
Have a temper outburst until you were exhausted?	1.53	0.41	1.44	1.88	2.37
Lose your temper with a friend	1.49	0.32	1.28	1.95	2.4
Stay angry for a long time?	1.48	0.05	1.25	2.06	2.57
Continue to have a temper outburst even when your parent tried to calm you down?	1.48	0.57	1.31	1.95	2.08
Lose your temper to get something you wanted?	1.45	0.18	1.24	1.9	2.47
Lose your temper when you were tired, hungry, or sick?	1.42	−0.3	1	2.23	2.75
Lose your temper with your parent?	1.36	0.23	1.13	1.86	2.22
Have a hot or explosive temper?	1.36	0.58	1.25	1.68	1.95
Stamp your feet when you lost your temper?	1.26	0.2	1.02	1.74	2.07
Have a temper outburst that lasted longer than 5 min?	1.16	0.12	0.82	1.6	2.11
Have a temper outburst?	1.1	−0.02	0.8	1.52	2.11
Lose your temper with other kids?	1.1	−0.11	0.85	1.65	2.02
Feel grouchy?	1.04	−0.58	0.74	1.82	2.18
Get angry over small things?	0.91	−0.42	0.51	1.5	2.04
Lose your temper easily?	0.88	0.03	0.75	1.2	1.57
Get extremely angry?	0.74	−0.29	0.53	1.17	1.55
Have an attitude	0.71	−0.65	0.48	1.31	1.7
Have trouble calming down when you were angry?	0.59	−1.05	0.34	1.29	1.76
Yell at someone because you were angry?	0.56	−0.82	0.19	1.05	1.82
Feel angry?	0.5	−0.84	0.24	1.1	1.48
Become frustrated easily?	0.39	−1.75	−0.02	1.24	2.08
Get annoyed easily?	0.37	−0.91	0.17	0.87	1.35
Lose your temper when you were frustrated, angry, or upset?	0.36	−0.88	0.11	0.86	1.34
Lose your temper with your brother or sister	0.13	−1.06	−0.16	0.59	1.16

Note: Each bolded value represents an item location mean or category threshold that exceeds the 95th percentile threshold.

FIGURE 2.

Self‐report items along severity dimension.

Aim 3: Derive a parsimonious subset of irritability items most strongly associated with impairment (parent report). The final stepwise logistic regression model for the pragmatic screener identified 4 out of the 40 MAPS‐TL‐Youth items: “become frustrated easily” (p = 0.003), “yell angrily at someone” (p = 0.008), “lose temper or have a temper outburst to get something s/he wanted” (p = 0.023) and “spoil something for your family because of his/her bad mood” (p = 0.022). These items explained 41.3% of the variance in cross‐domain impairment, and the final model was a good fit to our data (Hosmer & Lemeshow χ ² = 5.986, df = 6, p = 0.425). Notably, three of these items (“become frustrated easily,” “yell angrily at someone,” “lose temper or have a temper outburst to get something s/he wanted”) were from the core MAPS‐TL items, and one was from the developmentally specific items from the MAPS‐TL‐Youth version. As the frequency of “become frustrated easily” increased, the odds of observing cross‐domain impairment increased by 1.88 times (95% CI [1.236, 2.842]); odds ratio for “yell angrily at someone” was 1.76 (95% CI [1.161, 2.671]); for “lose temper or have a temper outburst to get something s/he wanted” was 2.11 (95% CI [1.110, 3.991]); and for “spoil something for your family because of his/her bad mood” was 2.39 (95% CI [1.166, 7.338]). The final model performed significantly better than the baseline model (Δχ ² = 107.925, df = 4, p < 0.001) as well as the block 1 model containing only the core 22 MAPS‐TL items plus baseline (Δχ2 = 101.511, df = 3, p < 0.001). These four items were then summed as the basis for the screener.

Derive a parsimonious subset of irritability items most strongly associated with impairment (self ‐report). The final stepwise logistic regression model identified 2 out of the 36 of the self‐report MAPS‐TL‐Youth items: “difficulty calming down when angry” (p < 0.001) and “lose temper with parent” (p = 0.007). These items explained 12.0% of the variance in cross‐domain impairment, and the final model was a good fit to our data (Hosmer & Lemeshow χ ² = 4.602, df = 7, p = 0.708). As the frequency of “difficulty calming down when angry” increased, the odds of observing cross‐domain impairment increased by 1.36 times (95% CI [1.148, 1.608]); for “lose temper with parent,” the odds increased by 1.33 (95% CI [1.082, 1.640]). The final model performed significantly better than the baseline model (Δχ ² = 26.705, df = 2, p < 0.001). Items were then summed for further analysis.

Aim 4: Clinically optimize the irritability screener via maximizing sensitivity and specificity in regard to DSM irritability‐related disorders (parent report). Using the parsimonious irritability screener summed scores (mean = 2.56, SD = 2.91, range = 0–20), ROC analyses indicated an optimal screener cutoff score of 4 optimally balanced sensitivity (87%) and specificity (81%) in relation to DSM irritability‐related syndromes (peak Youden Index: 0.69). In other words, 87% of youth who were diagnosed with an irritability‐related syndrome were correctly identified as meeting the cutoff criteria, and 81% of youth who did not meet diagnoses of irritability related syndromes were correctly identified as not meeting the cutoff criteria. Area under the curve for DSM irritability‐related syndromes (AUC = 0.91, p < 0.001) indicated excellent classification accuracy of the cutoff score.

In this MAPS preadolescent sample, 28.2% scored above and 68.7% scored below the cutoff of 4 (Figure S1a). Of youths who met the irritability cutoff by parent report, nearly all were reported to “become frustrated easily,” with a large portion (82.6%) reporting the behavior occurring at least weekly. Youth who had positive screens varied more substantially in frequency in dysregulated/developmentally inappropriate behaviors, i.e., “lose temper or have a temper outburst to get something s/he wanted” (20.4% never, 40.8% less than once per week 31.6% 1–3 days per week, and 7.1% most days of the week or greater), and “spoil something for your family because of his/her bad mood” (37.8% never, 42.9% less than once per week, 15.3% 1–3 days per week, and 3.1% every day or more). Nearly two‐thirds (61.2%) of youth who scored above the cutoff never or infrequently lost their temper or had a temper outburst to get something they wanted in the past month, and 81.7% of irritable youth with impairment never or infrequently spoil something for their family because of their bad mood. In contrast, for youth who did not meet irritability criteria, less than half (43.5%) endorsed mild frequencies of “becom[ing] frustrated easily” (score of 1 or 2), but the large majority did not “lose their temper to get something they wanted” (89.1% with a score of 0) or “spoil something for their family because of their bad mood” (94.6% with a score of 0). Youths above and below the cutoff did not differ on age, gender, race/ethnicity, or poverty status (for details, see Table S1).

Determine a clinically optimized irritability score on the MAPS‐TL‐Youth‐SR. Using the summed scores of the parsimonious youth self report items (mean = 2.54, SD = 2.30, range = 0–10), ROC analyses showed that an optimal cutoff score of 3 (peak Youden Index: 0.29) balanced sensitivity (67.5%) and specificity (61%), meaning that 67.5% of youth who were diagnosed with an irritability‐related syndrome were correctly identified as meeting the cutoff criteria, and 61% of youth who did not meet diagnoses of irritability related syndromes were correctly identified as not meeting the cutoff criteria. Area under the curve for DSM irritability‐related syndromes (AUC = 0.65, p < 0.001) indicated adequate classification accuracy of the cutoff score.

36.0% of youths scored above and 42.2% scored below the cutoff of 3 (Figure S1b). Nearly all youths who met the cutoff score of 3 reported having “difficulty calming down when angry” at least weekly, with 86.7% of youths reporting that the behavior occurs weekly or more. In contrast, almost all youths who did not meet cutoff reported never or rarely having “difficulty calming down when angry” with a large majority (88.5%) reporting the behavior never happening or occurring less than weekly. For “lose temper with parent,” over three‐quarters (78.2%) of youths who did not meet the irritability cutoff score reported that they never “lost temper with parent” in the past month, and an additional 20% reported only losing their temper with their parent less than weekly. However, for youths who met the irritability cutoff, over half (51.2%) reported never or rarely losing their temper with their parent, while the remaining 48.9% reporting they “lost temper with parent” once a week or more, suggesting youths may view this item as a more dysregulated behavior in pre‐adolescence (See Supporting Information S1 for item frequency comparisons for participants who met and did not meet each cutoff score.).

4. DISCUSSION

To chart the trajectory of transdiagnostic neurodevelopmental vulnerability to pediatric psychopathology, indexed by irritability, requires a lifespan coherent approach that incorporates developmental specificity with consistency across developmental periods (Wakschlag et al., 2010). The current paper advances this RDoC informed approach by characterizing the normal:abnormal spectrum of irritability within the PA period, one which is a crucial developmental inflection point (Yu et al., 2023). We have provided initial evidence of the reliability and validity of the MAPS‐TL‐Youth scale as a dimensional irritability measure in preadolescents. We also applied a “translational mindset” (Wakschlag et al., 2022) to increase the practical applications of this work by empirically deriving a clinically optimized screener for the MAPS‐TL‐Youth in preadolescents. This is to advance the goal of providing a pragmatic transdiagnostic mental health screening tool. This has promise for screening in pediatric primary care, which is often the primary point of entry for youths at risk (Wakschlag et al., 2019). Nuanced characterization of the full spectrum of irritability is important to “look under the hood” for its dimensional patterning and for identifying where along the spectrum youths fall. However, clinicians, who are in resource‐limited environments, also need brief, easily administered screeners to flag youths with clinically significant irritability who may be in need of more in‐depth assessment and intervention. This work is an important step toward that goal.

Developmental psychopathology theory posits that abnormality may be conceptualized as deviations from normal patterns within developmental context (Mittal & Wakschlag, 2017). Yet rigorous application of this beyond early childhood has lagged, in part due to the lack of validated measures to characterize the full spectrum of irritable behavior dimensions. Building on our prior work extensively validating the severity spectrum of irritability in early childhood (Krogh‐Jespersen et al., 2021; Wakschlag et al., 2012), we and others in this special issue (Hirsch et al., 2023; Kirk et al., 2023) now advance validity of the MAPS scales approach to youth. This dimensional spectrum approach goes beyond “symptom counts” to provide an ordered characterization of different facets of irritable behavior from mild normatively occurring to severe and clinically salient. The MAPS scales approach uses an objective frequency rating scale, providing concrete frequency thresholds at which each behavior—even behaviors qualitatively deemed mild—becomes severe. For example, whereas simply having a temper outburst was fairly common and was not one of the more severe behaviors, this behavior was psychometrically severe (occurring in less than 5% of the population) when it occurred most days of the week. By contrast, destructive temper outbursts (e.g., break/destroy things during an outburst; outbursts in unexpectable developmental contexts (e.g., out of the blue); and pervasive irritable mood (e.g., bad mood even during fun activities), protracted irritable mood (e.g., irritable no matter what) that led to family impairment (e.g., cause family outings to revolve around preventing outbursts) were uncommon (<5% prevalence) and thus severe even at lower frequencies (rarely or some days of the week). Moreover, we demonstrated the added value of developmentally specific items for youth, as these provide more comprehensive coverage of the upper range of the severity spectrum.

Moving beyond a research tool for nuanced characterization of the irritability dimension, we also developed a practical screener to be deployed as a transdiagnostic indicator of irritability that is associated with increased psychopathology risk in preadolescents. Together with prior work at preschool age and older and younger ages in this special issue (Hirsch et al., 2023; Kirk et al., 2023; Wiggins Urena Rosario, MacNeill et al., 2023; Wiggins, Urena Rosario, Zhang et al., 2023), we have now validated developmentally specific versions of clinically optimized irritability screener of the MAPS‐TL from ages 12 months‐adolescence. Interestingly, we found “easily frustrated” to be a foundational trait key for flagging clinically significant irritability across age periods. The robustness and consistency with which low frustration tolerance appears as an important item for screening suggests that the low threshold for dysregulation may be a necessary (albeit not sufficient) component of clinically significant irritability. This association with clinically significant irritability is particularly remarkable because “easily frustrated” is a relatively mild indicator along the dimensional severity spectrum both in the current paper and in IRT analyses of other ages (Krogh‐Jespersen et al., 2021; Wakschlag et al., 2012; Hirsch et al., 2023). It is important to note, however, that the mere presence of occasional low frustration tolerance was not sufficient to raise red flags. It either exhibiting low frustration tolerance daily or more OR having low frustration tolerance in combination with other, more severe behaviors (“yell angrily at someone”, “lose temper or have a temper outburst to get something s/he wanted”, and “spoil something for your family because of his/her bad mood”) that marked clinically significant irritability here and at other age periods. Thus, frequent low frustration tolerance, in combination with a small set of developmentally derived indicators, i.e., crucial elements of pathognomonic severe behaviors, can be a pragmatic screening tool for clinicians.

To our knowledge, this work is the first to validate a dimensional irritability self‐report scale for youth. We showed that preadolescents could reliably report on their own irritability behaviors, indeed, with a broader range of severity than their parents, adding to a fuller characterization and measurement of irritability. Future directions should include methods for combining this multi‐informant approach for clinical decision making. While preadolescents could also report on the key irritable behaviors that led to impairment and risk for psychopathology, the screener was somewhat more precise by parent report. Replication and extension in clinical populations will be useful, including efforts to increase sensitivity of this measure. In addition, the parent report screener may have had greater precision due to shared variance with the parent reported outcomes on the K‐SADs. This should be systematically studied in future studies that include youth report of symptoms as well as clinician ratings.

This study is not without limitations. First, while this study includes a richly characterized sample, we relied on concurrent assessment for validity analyses. Future studies that employ longitudinal follow‐up would be optimal. Second, while we leveraged parent and child report, objective measures (e.g., performance‐based measures) and cross‐informant and/or clinician rated outcomes would be ideal as an optimal test of predictive validity. Third, while both poor and non‐poor participants from historically marginalized groups were represented, a population‐based follow‐up study would be ideal. To further ensure the generalizability of this study's results, future studies will need to norm and replicate our findings across multiple populations. Finally, we here optimized the screener in relation to disorders that explicit irritability symptoms. However, common psychopathologies, particularly anxiety and ADHD, also have irritability as a key feature or sub‐type which guides behavioral and psychopharmacologic treatment, even if not explicitly included within symptom criteria (Evans et al., 2017).

The results of this study enable a more precise characterization and pragmatic screening of irritability during preadolescence—a critical period marked by significant changes in cognitive, social, and emotional development. Given the deleterious consequences of pediatric irritability, especially at the critical inflection point of preadolescence (Yu et al., 2023), it is crucial to develop proper tools and frameworks that can effectively capture irritable behaviors that may require intervention. To date, there are a dearth of transdiagnostic indicators of psychopathology risk that capture the full spectrum of irritability within developmental context. The findings of this paper move toward providing two such tools, a nuanced, developmentally‐specified, dimensional research measure to allow precise characterization of the irritability spectrum, and a clinically optimized screener for clinicians for use in real world settings to identify preadolecents with broad risk for irritability‐related psychopathology. The longer‐term vision for these measures is that the dimensional scale can advance mechanistic work on irritability‐related pathways to psychopathology that will inform new, targeted prevention efforts. The efficiency of a transdiagnostic irritability screener can also advance broad‐based earlier identification of youth risk for psychopathology, which is of great importance in light of the youth mental health crisis. The advancement of these tools may open avenues to promote youth's well‐being and use these behavioral markers to alter youths' trajectories toward improved mental health.

AUTHOR CONTRIBUTIONS

Drs. Jillian Lee Wiggins and Amy Krain Roy conceptualized and designed the study, interpreted the data, and revised the manuscript critically for intellectual content. Ms. Tasmia Alam and Mr. Nathan Kirk analyzed the data and wrote portions of the initial draft of the manuscript. Ms. Emily Hirsch analyzed the data and revised the manuscript critically for intellectual content. Drs. Margaret Briggs‐Gowan and Lauren S. Wakschlag obtained funding, interpreted the data, and revised the manuscript critically for intellectual content. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

CONFLICT OF INTEREST STATEMENT

All authors report no conflicts of interest.

Supporting information

Supporting Information S1

Alam, T. , Kirk, N. , Hirsch, E. , Briggs‐Gowan, M. , Wakschlag, L. S. , Roy, A. K. , & Wiggins, J. L. (2023). Characterizing the spectrum of irritability in preadolescence: Dimensional and pragmatic applications. International Journal of Methods in Psychiatric Research, 32(S1), e1988. 10.1002/mpr.1988

DATA AVAILABILITY STATEMENT

Deidentified data are available upon reasonable written request to Dr. Wakschlag for non‐commercial uses.