MAPping affective dimensions of behavior: Methodologic and pragmatic advancement of the Multidimensional Assessment Profiles scales

Abstract

Rigorous validation of the full developmentally sensitive normal:abnormal spectrum, including evaluating the incremental value of age‐specific behaviors, is necessary for nuanced characterization of dimensional features of psychopathology. To maximize the clinical utility of transdiagnostic approaches to risk identification, derivation of psychometrically sound, pragmatic versions with empirically derived cutoffs is also key. This special section has a central focus on rigorous, developmentally‐based measurement of irritability as an exemplar of this theory‐ and pragmatically‐based approach. Elevated irritability is a robust transdiagnostic predictor of the common psychopathologies of childhood. The Multidimensional Assessment Profiles Temper Loss (MAPS‐TL) Scales are the only irritability tool specifically designed to capture the normal:abnormal dimensional spectrum. These have been extensively investigated in preschool age but lack rigorous modeling at older and younger ages. In this special issue, (with three independent—and one longitudinal—set of samples), we test and improve measurement of irritability as a transdiagnostic phenotype of psychopathology risk as it unfolds across development, expanding the MAPS‐TL scale in three important ways: (1) extending irritability dimensional modeling and the developmental specification approach to older ages, (2) advancing science to practice translation by generating pragmatic irritability screening tools across ages, and (3) extending the dimensional, developmental specification approach to other dimensions of behavior, that is, internalizing. Collectively, the special issue operationalizes and advances application of a neurodevelopmental, dimensional and transdiagnostic approach to psychopathology.

1. INTRODUCTION

A fundamental methodologic principle of mental health fields is that the closer to “reality” the measurement of the construct, the better the utility of the model, for example, for decision‐making (Ferguson et al., 2012). Both in neurodevelopmental frameworks of psychopathology and from a clinical perspective, the “reality” is that psychiatric phenomena do not fall neatly into categorical boxes as discrete entities. Rather, clinical phenomena are developmentally unfolding and fall across dimensional spectra that span a normal:abnormal continuum (Mittal & Wakschlag, 2017). Measurement of dimensional spectra goes beyond symptom counts as dimensions are ordered in severity and include normative expression, and thus, can capture neurodevelopmental vulnerability prior to clinical manifestation. Yet historically and for ease of clinical decision‐making, categorical frameworks have dominated psychiatric nosology. Neurodevelopmental conceptualizations make it imperative to advance dimensional measurement, which is closer to “reality”, with an urgent need for truly dimensional measures (Cuthbert, 2020; Mittal & Wakschlag, 2017). Yet, this nuanced theoretical conceptualization of the dimensional nature of unfolding behavior and the importance of transdiagnostic processes has lagged in its methodologic applications, limiting translation of dimensional, developmentally based constructs to measurable tools with practical application.

2. DEVELOPMENTAL CHARACTERIZATION OF DIMENSIONS

Current nosologies do not sufficiently capture varied developmental phenotypes of psychiatric syndromes and their unfolding patterns. Both developmental and neuroscientific theories have underscored the need to define abnormality as deviation from normal patterns within a developmental period (Marquand et al., 2019; Wakschlag et al., 2010). Thus, dimensional measures, capturing the range of normative to severe behaviors, must be both lifespan coherent and developmentally sensitive. This approach has been extensively applied for young children (Krogh‐Jespersen et al., 2021; Wakschlag et al., 2012). This sets the stage for explicit testing of this approach for older youth, including testing the added value of developmentally specific behaviors at particular age periods, key for advancing this approach in a clinically meaningful way.

3. REAL WORLD TRANSLATION

This nuanced characterization is crucial for research aimed at understanding mechanistic linkages and developmental trajectories. Yet, in the real world, clinical settings have limited time, resources and, in primary care settings, specialization in in‐depth psychiatric resources. Administration of a full, “deep phenotype” measure can be impractical and is likely not necessary if the needed outcome is a “thumbs up or down” decision about whether a child is at risk and in need of further evaluation and or intervention. Transdiagnostic, broad brush psychiatric risk determination with tools that are pragmatic for clinical use are key for this (Lewis et al., 2021; Stanick et al., 2021). Pragmatic tools are brief, easily administered, freely available measurement that can be implemented in the clinic and a variety of settings (Glasgow, 2013). Extant measures often have shortened forms (i.e., fewer items but still psychometrically acceptable) that can measure the psychopathological construct in resource‐limited settings. Similarly, screening tools may focus on items that are most related to impairment and/or mental disorder and are meant to flag risk. Yet, clinically optimized forms combine these elements (brevity, focus on flagging risk) and moreover facilitate pragmatic, real‐world application by providing an empirically‐derived cutoff score for clinical decision‐making. To maximize the clinical utility of developmentally sensitive, dimensional measurement, derivation from these developmentally based scales to clinically optimized screeners with empirically derived cutoffs to flag risk is of high value.

4. ASSESSMENT OF IRRITABILITY VIA THE MULTIDIMENSIONAL ASSESSMENT PROFILES (MAPS) SCALES: APPLICATION OF THE DEVELOPMENTAL SPECIFICATION APPROACH

Irritability, that is, a dispositional tendency to react with anger and frustration to blocked goal attainment, is a highly robust indicator of broad psychopathology risk, measurable from infancy with distribution covering normative to severe expressions (Evans et al., 2017). Elevated irritability in early childhood is a robust transdiagnostic predictor and feature of internalizing and externalizing disorders, the most common and modifiable psychological problems of childhood (Finlay‐Jones et al., 2023; Hirsch et al., 2022; Wiggins et al., 2018). Of course, there are other well‐validated measures of psychopathology risk (e.g., Achenbach & Ruffle, 2000; Goodman, 1997; Reynolds & Kamphaus, 2004; Stringaris et al., 2012). However, the Multidimensional Assessment Profiles Temper Loss (MAPS‐TL) Scale is the only irritability tool specifically designed to capture the normal:abnormal dimensional spectrum of irritability in a developmentally specified manner (Wakschlag et al., 2012), and is freely available for non‐commercial use by contacting Dr. Wakschlag. (Note: MAPS was formerly the Multidimensional Assessment Profiles—Disruptive Behavior [MAP‐DB] Scales, subsequently renamed the MAPS because of its expanded transdiagnostic focus and expansion to include attention dysregulation (Nili et al., 2023) and internalizing behaviors [Wakschlag et al., 2023, this issue].) The MAPS‐TL broke new measurement ground by generating and then empirically testing the novel developmental specification method, which posited that behavioral atypicality should be defined as deviation from normative patterns within a developmental period (Wakschlag et al., 2010), an approach which has subsequently been put forth for trajectories of neural development (Marquand et al., 2019). Operationalization of this method was carried out through (a) generating a broad range of behaviors reflecting developmental expression of irritable mood and tantrums in early childhood, spanning normative misbehaviors that most children do all the way to severe, rarely occurring behaviors; (b) taking interactional and motivational context into account (c) employing an objective frequency scale (i.e., how many times per week/day); to generate thresholds of concern that have reduced bias because they do not rely on subjective judgment (e.g. “often”); and (d) validating across multiple large community samples (Krogh‐Jespersen et al., 2021; Wakschlag et al., 2018). When the scales were first developed, this dimensional, developmentally based approach was particularly crucial for defining a normal:abnormal distinction in young children because of the difficulty differentiating the normative misbehaviors of early childhood (of which tantrums are hallmarks) versus clinical red flags.

Thus, original validation of the MAPS‐TL was at preschool age (Wakschlag et al., 2012). To model the irritability dimensional spectrum for the MAPS measure, item response theory (IRT) methods were deployed (Reise & Waller, 2009). IRT is especially useful for dimensional modeling because IRT maps the location of both items and respondents on an underlying latent trait continuum from mild, commonly occurring behaviors to severe, rarely occurring behaviors. Within this IRT framework, behaviors are psychometrically defined as “abnormal” or severe when they are rarely occurring (i.e., in less than 5% of the population) (Wakschlag et al., 2014). In preschool age, previous work has provided robust psychometric evidence for validity of MAPS‐TL (Wakschlag et al., 2012), and demonstrated predictive, clinical and mechanistic utility of this measure (e.g., Wakschlag et al., 2012; Wakschlag et al., 2018; Wiggins et al., 2018). In order to enable dimensional characterization of irritability in a manner that combines developmental coherence with specificity—a focus of this special issue—the MAPS‐TL has since been extended across the pediatric period: downward to the first years of life (12 months of age) and upward through adolescence (17 years). For versions that span this broader set of age periods, the MAPS‐TL includes the core set of 22 irritable mood and behavior items believed to be applicable across age periods (with minor modifications) and an additional set of age‐specific items designed to investigate whether these add incremental utility beyond the core validated set. However, the scale at younger and older ages lacks the extensive body of modeling and validity evidence that has been established for preschool age. Thus, advancing developmentally sensitive dimensional measurement of irritability, with pragmatic tools for application, will be crucial for coherent characterization of irritability across the pediatric lifespan, mechanistic linkages and, ultimately, for clinical applications.

5. MOVING FORWARD: THIS SPECIAL ISSUE

With this suite of scales and the foundational validation of the utility of the MAPS‐TL for characterization of the normal:abnormal spectrum of irritability, the field may benefit from broader testing and application, which is the goal of this special issue. In this particular collection of papers, we test and improve measurement of irritability as a transdiagnostic developmental phenotype of psychopathology risk as it unfolds across development and extend it to other key dimensional features of psychopathology, expanding the MAPS‐TL in three important ways (Figure 1):

• (1) Extending irritability dimensional modeling and the developmental specification approach to older ages. Although the normal:abnormal spectrum of irritability on the MAPS‐TL has been modeled from ages 1–5 (Krogh‐Jespersen et al., 2021; Wakschlag et al., 2012); it has not been dimensionally characterized for school age through adolescence. In this special issue, Hirsch et al. (2023), Alam et al. (2023), and Kirk et al. (2023), conduct rigorous application of the developmental specification model by testing the utility of age‐specific behaviors and dimensional modeling of the MAPS‐TL from school‐age through adolescence.
• (2) Translating science to practice by developing pragmatic tools for irritability screening. Despite the identification of irritability as a “top problem” clinically (Evans et al., 2022), little work has been done to translate the robust irritability science base to clinically feasible tools in a manner that reliably differentiates normative misbehavior from behaviors of clinical concern. Our prior efforts to achieve this with the MAPS‐TL preschool version has yielded promising results (Wiggins et al., 2018). In this special issue, we expand this pragmatic approach across the pediatric range, by developing clinically optimized screeners for infancy/toddlerhood through adolescence (Hirsch et al., 2023; Kirk et al., 2023; Wiggins, Ureña Rosario, MacNeill et al., 2023; Wiggins, Ureña Rosario, Zhang et al., 2023). Strikingly, across developmental periods, all the parent‐report clinically optimized screeners include the combination of low frustration tolerance (a normative misbehavior that occurs in most children regularly but not daily) as well as severe, pathognomonic behavior (uncommon at any frequency) as indicative of impairment. See Table 1 for summary of clinically optimized screeners across developmental periods.
• (3) Expanding the dimensional, developmental specification approach. Although our central focus is on the MAPS‐TL as an assessment of irritability, the original MAP‐DB measure captures multiple dimensions of externalizing behavior (noncompliance, aggression, punishment insensitivity and low concern for others; Nichols et al., 2015; Wakschlag et al., 2014) and has recently been expanded to cover attention dysregulation (Nili et al., under review). However, application of this approach to other key affective aspects of psychopathology is lacking. This special issue includes a paper by Wakschlag et al. (2023) expanding this novel, dimensional, developmentally specified approach to internalizing (anxious and depressive) behaviors in early childhood. This now provides provide nuanced dimensional and developmentally based characterization across the full internalizing/externalizing spectrum via the MAPS scales.

FIGURE 1.

Overview of special issue papers within extant literature. Blue outline indicates papers in this special issue. Check marks indicate work that has already been done. ¹Krogh‐Jespersen et al. (2021); ²Wakschlag et al. (2012); ³Wakschlag et al. (2015).

TABLE 1.

Contrasting MAPS‐TL clinically optimized screeners by informant and developmental period.

Developmental period	Informant	Items	Cutoff score	Special issue paper
Transition to toddlerhood (1–2 years)	Parent	Frustrated about small things	5	Wiggins, Ureña Rosario, Zhang et al., this issue
		Hit/bite/kick during tantrum
		Trouble cheering up when grumpy
		Act grumpy during fun activities
		Temper tantrum in public
Preschool age (3–6 years)	Parent	Frustrated easily	3	Wiggins et al., 2018; Wiggins, Ureña Rosario, MacNeill et al., this issue
		Break/destroy during a tantrum
Early school age (6–9 years)	Parent	Frustrated easily	4	Hirsch et al., this issue
		Act irritable
Preadolescence (8–11 years)	Parent	Frustrated easily	4	Alam et al., this issue
		Yell angrily at someone
	Self	Difficulty calming down when angry	3
		Lose temper with parent
Adolescence (12–17 years)	Parent	Frustrated easily	4	Kirk et al., this issue
		Angry/irritable/grouchy throughout day
		Difficulty calming down when angry
	Self	Trouble calming down when angry	2
		Hit/shove/kick when lose temper

Note: Items are scored 0–5 based on frequency in the past month (see citations for objective anchors, which are the same across scales) and summed. Cutoff scores indicate empirically derived thresholds related to impairment.

5.1. Data sources for the special issue

Data from three overarching studies were used to conduct this work (Table 2). First, the MAPS Study recruited a diverse U.S. Midwestern sample of preschoolers to elucidate dimensions of disruptive behavior that mark a persistent clinical course. The MAPS Study first psychometrically validated the MAPS scales via two, large, independent community samples (Preschool Age Samples 1 and 2) of 3‐ to 5‐year‐olds from pediatric clinics, stratified for race/ethnicity and poverty, for psychometric validation of the MAPS measure. A longitudinal cohort was drawn from the first of these community psychometric samples (Longitudinal Subset of Preschool Age Sample 1) enriched for psychopathology risk via oversampling for family domestic violence and child disruptive behaviors. This longitudinal subset included three key waves: baseline at preschool age (3–5 years), early school age (6–8 years), and pre‐adolescence (9–10 years). In this issue, Preschool Age Samples 1 and 2, including the Longitudinal Subset, were used in Wiggins, Ureña Rosario, MacNeill et al. (2023) and Alam et al. (2023) and Hirsch et al. (2023) used waves of the Longitudinal Subset of Preschool Age Sample 1. Next, the When to Worry (W2W) Study (Transition to Toddlerhood Sample) was also a diverse Midwestern sample designed to extend this work on early identification of dimensional phenotypes that portend clinical risk by beginning at 12 months of age, with risk enriched via irritability oversampling (Krogh‐Jespersen et al., 2021; Zhang et al., 2023). Key W2W waves were baseline, which was the transition to toddlerhood (∼12–14 months), bi‐monthly follow‐ups to toddlerhood (∼24–26 months), and another visit at preschool age (∼36–54 months). The W2W sample was used in Wiggins, Urena Rosario, MacNeill et al. (2023) and a sub‐sample of W2W was used in Wakschlag et al. (2023). Finally, the Temper Outburst Follow Up Study (Adolescent Sample) was an East Coast‐based sample of adolescents (ages 12–17 years) who were originally recruited at school age (ages 5–9 years) and, at baseline, included three groups: children with clinically significant irritability characterized by severe temper outbursts, children with attention‐deficit/hyperactivity disorder (ADHD) without such outbursts, and healthy comparisons. The data at the adolescent follow‐up were used in Kirk et al. (2023). For details on these samples and their methods of ascertainment, please see, for MAPS Study: Dirks et al. (2019) and Wakschlag et al. (2014); for W2W Study: Krogh‐Jespersen et al. (2021); Zhang et al., 2023‐a; and for Temper Outburst Follow‐Up Study: Roy et al., 2013; Hirsch et al., 2022.

TABLE 2.

Overview of samples contributing to special issue papers.

		Enrichment for clinical risk	Recruitment sources	Sociodemographics		Longitudinal waves (N, ages a )	Special issue papers drawing on this sample
				Race/ethnicity	Poverty
MAPS study	Preschool age sample 1	None	Community pediatric clinics	White non‐Hispanic: 23% Black/African American: 40% Hispanic: 24% Other: 13%	Not poor: 55% Poor: 44%	N = 1857 Ages 3–5 years	Wiggins, Ureña Rosario, MacNeill et al.
	(Dirks et al., 2019; Wakschlag et al., 2014)
	Longitudinal subset of preschool age sample 1	Child disruptive behaviors Family domestic violence	Community pediatric clinics	White non‐Hispanic: 19% Black/African American: 50% Hispanic: 30% Other: 2%	Not poor: 50% Poor: 49%	Preschool age N = 425 Ages 3–6 years	Wiggins, Ureña Rosario, MacNeill et al.
						Early school age N = 315 Ages 6–9 years	Hirsch et al.; Wiggins, Ureña Rosario, MacNeill et al.
	(Wakschlag et al., 2015; Wiggins et al., 2018)					Preadolescence N = 296 Ages 8–11 years	Alam et al.; Wiggins, Ureña Rosario, MacNeill et al.
	Preschool age sample 2	None	Community pediatric clinics	White non‐Hispanic: 26% Black/African American: 33% Hispanic: 30% Other: 11%	Not poor: 55% Poor: 41%	N = 1490 Ages 3–5 years	Wiggins, Ureña Rosario, MacNeill et al.
	(Dirks et al., 2019; Wakschlag et al., 2012)
When 2 Worry study	Transition to toddlerhood sample	Child irritability	Community pediatric clinics, flyers and social media	White non‐Hispanic: 42% Black/African American: 23% Hispanic: 15% Other: 20%	Not poor: 69% Near poor: 12% Poor: 19%	Transition to toddlerhood N = 310 Age 14 months	Wiggins, Ureña Rosario, Zhang et al.; Wakschlag et al.
						Bimonthly online Follow‐ups b N = 269, 16 months N = 250, 18 months N = 236, 20 months N = 220, 22 months N = 205, 24 months
	(Krogh‐Jespersen et al., 2021)					In‐lab toddlerhood Follow‐up b N = 169, 27 months
Temper outburst follow‐up study	Adolescent sample	Longitudinal follow up study of young children including those with clinically significant outbursts	Adolescents recalled from previous study of pediatric irritability b	Race White: 45% Black/African American: 24% Other: 19% Missing: 13% Ethnicity Hispanic: 28%	Not poor: 78% Poor: 22%	N = 80 Ages 12–17 years	Kirk et al.
	NIH R15MH115356

Note: Percentages that do not add to 100 indicate missing data. Poverty status was determined using income and lack of services in the preschool age samples and longitudinal follow‐ups, using income to needs ratio in the transition to toddlerhood sample, and using income in the adolescent sample.

^a Ages approximate.

^b Note that COVID‐19 online‐only/in‐person restrictions that went into effect after baseline sample recruitment affected sample size for follow‐ups.

As Table 2 highlights, these samples are well suited for the purpose of the special issue, as the samples are reasonably large, socio‐demographically and racially/ethnically diverse, span relevant age periods and are irritability enriched. Future extension of this work in population‐based samples will further extend generalizability, and application in clinical samples will also amplify impact.

In summary, this special issue brings applications of dimensional, transdiagnostic methodology full circle, now capturing the periods from the first year of life through adolescence, with developmental nuance and pragmatic approaches. Taken together, we hope these papers will illuminate the state of dimensional methodology with irritability as exemplar, highlight pragmatic applications and predictive utility and catalyze interest in their scientific and real‐world utility.

AUTHOR CONTRIBUTIONS

Jillian Lee Wiggins wrote the manuscript, and Lauren S. Wakschlag and Amy K. Roy provided edits and feedback.

CONFLICT OF INTEREST STATEMENT

All authors declare no conflicts.

Wiggins, J. L. , Roy, A. K. , & Wakschlag, L. S. (2023). MAPping affective dimensions of behavior: Methodologic and pragmatic advancement of the Multidimensional Assessment Profiles scales. International Journal of Methods in Psychiatric Research, 32(S1), e1990. 10.1002/mpr.1990

DATA AVAILABILITY STATEMENT

No data were used to create this manuscript.