Understanding irritability through the lens of self-regulatory control processes in children and adolescents: a systematic review

The aim of this systematic review is to synthesize and assess existing research on the connection between socio-psycho-physiological SRC processes and irritability during childhood and adolescence. In particular, we aim to systematically review existing studies that investigate irritability through the lens of at least one of the SRC processes. This will serve to identify gaps in the present literature and highlight opportunities for future research. By identifying SRC processes that are associated with irritability, this systematic review may provide knowledge to serve as a basis for determining preventive or therapeutic approaches to address irritability.

The JBI Manual for Evidence Synthesis, chapter 7: systematic reviews of etiology and risk [32] guided the realization of the review. Also, the preferred reporting items for systematic review and meta-analysis (PRISMA) 2020 [33, 34] were followed for reporting. The review protocol is available on PROSPERO (#CRD42022370390). We incorporated studies that (a) sampled children and adolescents (0–17 years of age); (b) assessed at least one psychophysiological self-regulatory process (e.g., cognitive control, emotion regulation, autonomic regulation or social regulation); (c) measured irritability (e.g., anger proneness, low tolerance to frustration, outbursts); and (d) were published in English, German or French. All study designs were included (i.e., observational and case studies as well as qualitative and quantitative methodology). We excluded studies that (a) were not peer-reviewed or referred to conference acts, (b) focused on psychometric properties of instruments assessing irritability, (c) reported interventions or (d) mainly focused on children and adolescents with an autism spectrum disorder or intellectual disabilities as it refers to neurodevelopmental disorders which imply different developmental pathways, underlying mechanisms and specific assessments [35,36]. Consequently, the SRC processes and their interactions with irritability in these populations may differ significantly from those observed in more typical developmental contexts. Including studies on these populations in the review may therefore introduce excessive heterogeneity, potentially complicating the synthesis of findings and the identification of broader patterns.

Two blind and independent reviewers (SU and AOW) conducted study selection (abstract and title screening as well as full text selection) and data extraction. Choices that differed between reviewers were discussed to achieve a consensus.

In collaboration with a medical librarian (JRA), a literature search was conducted in October 2023 in six bibliographic databases: Embase.com, Medline ALL Ovid, APA PsycInfo Ovid, Web of Science Core Collection, the Cochrane Database of Systematic Reviews Wiley and ProQuest Dissertations & Theses A&I. The searches were performed without language or date restrictions. Manually (backward search) and through the use of Web of Science Core Collection (forward search), further records were discovered by tracing citations of studies that were included.

Figure 1 displays the PRISMA 2020 flow diagram [33]. The supplementary File 1 provides details regarding the search syntax, keywords and index terms used.

Bild vergrößern

We screened the titles and abstracts of identified studies for possible inclusion (k = 2612), which led to k = 106 studies selected for full text screening. Among them, 74 studies (and 82 reports) met the inclusion criteria defined in the study protocol (see Table S1). Below, we indicate whether we described the studies (i.e., general project) or the reports (i.e., published article).

We chose the appraisal tool for cross-sectional studies (AXIS) [37], as it seems to be the most appropriate tool for analysing the quality of observational studies. Specifically, the AXIS is a 20-item instrument and was applied by SU and AOW to evaluate the quality of the retrieved reports by assessing the main bias that may be observed in observation studies (e.g., selection of the participants, sample size justification, drop out analyses, statistical method, ethical consideration and role of funding sources; for details, see Table S2 and Figure S2). However, it should be noted that this instrument lacks, for example, aspects such as study protocol and planned analysis publication, blinded analyses or non-published results.

Number of participants (for both males and females), sample origin and study design were extracted from the retrieved studies. Table 1 describes how irritability was operationalized (e.g., frustration, anger proneness, aggressivity, anger dysregulation, grumpy mood, tantrum). Finally, we characterized the specific process of SRC (i.e., cognition [either “purely” or combined with “affective/motivation” dimensions], emotion, physiology and social) for each study (see Figure S1 for overlaps between the studies regarding which processes were examined).

The included reports were published between 1981 and 2023. However, only three reports were published before 2000, and k = 70 reports (85.4%) were published after 2017 (within the last 6 years), from which k = 14 reports (17.1%) were published in 2023, highlighting the topicality of this subject. The vast majority of studies were performed in Western countries, particularly in the United States (k = 48, 64.8%) and in European countries (k = 18, 24.3%). Less than 7% of the studies (k = 5) originated from Asia and South America. Most studies adopted a cross-sectional design (k = 31, 41.3%). Further, k = 29 studies (39.1%) and k = 14 studies (18.9%) adopted a longitudinal design (with follow-up ranging from 6 months to 17 years) and a case–control design, respectively. A total number of n = 26,764 participants (n = 12,384 girls and n = 12,905 boys, n = 1475 no information) were present in the selected studies. The study samples referred mainly to community samples (k = 42, 56.8%) or clinical samples (k = 32, 43.2%). In thirteen studies, more than 60% of those sampled were boys [38‐46], whereas in two studies, more than 75% of those sampled were boys [47, 48]. In four studies, more than 60% of those sampled were girls [49‐52], whereas one study sampled girls, exclusively [52]. Finally, four studies did not specify the gender in their samples [31, 53‐56]. The remaining reports included an equivalent proportion of boys and girls. The average age of all who were sampled was 8.08 years (SD = 5.26). Children younger than 5 years old were included in k = 29 studies (39.2%), and 26 studies (35.1%) sampled adolescents above 12 years old. Only parents were included in the samples of twenty-one studies (28.4%), from which k = 14 recruited only mothers and k = 7 included both mothers and fathers. The majority of reports assessed both tonic and phasic irritability components (80.5%) without distinguishing them. The reason for doing so is that the instruments did not allow for a distinction to be made between components. For terminology, the majority of reports (69.5%) used “irritability” without specification. In 8.4% of reports, “tantrum” was used alongside other terms such as anger proneness, frustration intolerance, aggression or grumpy mood.

Regarding the SRC components, approximately one quarter (25.7%) of the reports examined physiological aspects. Conversely, either “purely” cognitive aspects or “affective/motivation” combined with cognitive aspects were examined in about 20.8% of reports. Finally, emotional components (i.e., emotion regulation) were considered in almost 14% of the reports.

The AXIS assessment for each specific report is presented in the supplement (see supplementary Table S2 and Figure S2). Only k = 2 (2.4%) reports justified the sample size, either by a priori or a posteriori power analysis. Moreover, the question of drop out was addressed in k = 13 (15.9%) studies, of which seven reports described drop out bias (the vast majority of reports [k = 66; 79.5%] did not address this aspect). Patients with the most severe forms of irritability might have been more prone to drop out. Therefore, important information on the full range of irritability might have been overlooked by these studies. Overall, the included studies were of high quality, with none demonstrating a particularly strong bias, as measured with this instrument.

In the following subsections, we provide a narrative summary of the included studies according to the perspectives of the main SRC processes (cognitive, emotional, physiological and social) on irritability. We illustrate the main themes and subthemes examined in the included reports (Fig. 2). Also, we highlight the main characteristics and findings for each report (Table S1).

Bild vergrößern

One of the most clinically relevant aspects of the research on SRC and irritability is how the associations between them unfold over the long-term, especially from infancy to childhood, adolescence and young adulthood. Notably, knowledge on the developmental pathways of irritability may guide prevention and intervention strategies. Further, the manner in which these associations predict clinical outcomes, such as internalizing or externalizing psychopathology, is of crucial importance. A considerable number of identified studies have used a longitudinal design, but few have focused on irritability and SRC as predictors of later psychopathology [31, 53, 58, 59, 67, 69, 81, 106, 107, 110‐112], and very few have used clinical samples [67, 112].

Establishing comparability due to variations in both the definition and measurement of the construct of irritability is one challenge across the studies identified in this review [113]. Recent research has reached a consensus on defining irritability as an increased proneness to anger that may lead to aggression but often does not [20, 114, 115]. Anger is described as a transient negative emotional state ranging from mild irritation to intense rage, with physiological, cognitive and behavioural components. Aggression refers to intentional behaviours that cause harm, including reactive, impulsive or proactive types. However, distinguishing between irritability, anger and aggression proves challenging, as the constructs are closely intertwined and because rating scale measures of the constructs correlate with medium to large effect sizes [116‐118], raising doubts about their distinctiveness. The reliability of assessments of irritability (as well as the other constructs) is further complicated by differences in informants, with small to medium correlations between reports from different sources [80, 116, 118]. It is likely that irritability manifests differently at different ages as well as to various degrees in different contexts (e.g., at home vs. in school), and informant perspectives differ not just due to measurement error. In the studies, emotional SRC has primarily been analysed via parent report or physiological measures during challenging situations, offering insight into either the perception that the parent has of their child’s general functioning or real-time observation of emotion regulation reactivity and recovery phases. However, there remains a lack of child and youth self-reporting in this domain.

Few studies have examined peripheral physiological SRC processes such as ANS or HPA-axis functioning in relation to irritability, suggesting a need for more detailed examination in future research. While many reports have outlined aberrant neural patterns associated with irritability or related constructs, findings are mixed, and a recent meta-analysis showed no common functional or structural substrates underlying irritability [119].

In exploring the role of social interaction in regulating irritability, numerous studies have shed light on the reciprocal relationship between irritability and parenting, predominantly within a core family and primarily focusing on mothers. However, limited attention has been given to the social components of SRC outside the family context, particularly in that of peer interactions or support. Considering the importance of peer relationships in middle childhood and especially in adolescence, investigating their influence on irritability would be exceptionally valuable.

A further research gap that has emerged from our systematic review is the role of gender in the association between SRC and irritability, which has been reported in only five studies. We found only one study in which stronger relationships between inhibitory control and, later, irritability were observed in girls than in boys [59]. More generally, this topic has received little attention in research on irritability [120]. However, some evidence from population-based studies suggests that developmental patterns differed between males and females, with irritability being more common during childhood (decreasing with age) in boys and during adolescence (with levels increasing with age) in girls [121, 122]. The authors suggested two “types” of irritability based on gender: an early-onset type more common in boys (a pattern typical of neurodevelopmental problems) and a later-onset “type” that starts in adolescence and is more common in girls (a pattern typical of mood problems) [121]. In this perspective, one longitudinal study reports early differential associations between the sexes and in parental symptoms in predicting phasic and tonic irritability in adolescence [123]. Thus, future studies that investigate SRC processes of irritability should be designed with a specific focus on gender.

The areas that were explored in previous studies and identified in this systematic review delineate the potential developmental pathways of SRC for irritability (see Fig. 3). More specifically, irritability may arise from a problematic temperamental profile. Then, it is increased through delay intolerance, low cognitive control skills and attention bias towards anger or threat. Mutual negative influence is eventually observed through irritability and parenting. Finally, emotion dysregulation and physiological regulation mediate the link between irritability and specific psychopathology. The proposed developmental pathways entail certain gaps that need to be substantiated in future studies.

Bild vergrößern

In previous research, irritability was mainly studied as a unitary clinical phenomenon [124, 125]. This resulted in a poor definition of irritability and an unclear understanding of its clinical correlates and pathophysiology. To overcome this shortcoming, tonic and phasic irritability are distinguished in more recent conceptualization [126‐128], the usefulness of which should be tested in future studies. Similarly, the majority of studies have focused on single components of SRC. However, in daily life, all SRC processes work together to allow the individual to adapt to the ever-changing environment. While certain studies have incorporated multiple SRC components, a need exists for their integration in future studies.

Moreover, research on irritability may benefit from a more ecological or naturalistic approach. Taking into account both a between-person and a within-person perspective as opposed to the standard or static approach may offer an improvement in the understanding of the nature and variability of irritability over time (e.g., [129‐131]). In addition, this approach will allow for an assessment of the temporal sequences among irritability and SRC [132, 133]. Such knowledge is of crucial importance for developing ecological momentary interventions or just-in-time adapted interventions (see [134]). Furthermore, adopting an integrative approach in ambulatory assessment, which involves not only repeated prompting of subjective experiences but also cognitive assessments and physiological indices, may enhance the current understanding of the dynamic interplay between psychological and physiological factors within the everyday contexts that individuals encounter.

The studies included in this review are international in scope; however, the cultural conceptualizations of irritability (and the implication of SRC processes) have not been specifically addressed in prior research or in this systematic review. The perception, expression, and regulation of emotions—such as anger or irritability—vary significantly across cultures [135, 136] , shaped by culturally shared “decoding rules” [137], which influence how individuals interpret and express emotions within their specific cultural context. Moreover, the threshold for what is considered irritable behavior can differ according to cultural norms and expectations [8]. Consequently, cross-cultural studies are essential for examining both the similarities and differences in the phenotyping of irritability and its association with SRC processes. In this context, the Cross-Cultural Consortium on Irritability (C3I, https://medicine.yale.edu/childstudy/research/collaborative-labs/cross-cultural-consortium-irritability/) is a noteworthy collaborative initiative dedicated to investigating irritability across different cultures. The C3I aims to establish an international network of researchers to advance understanding of the cross-cultural similarities and differences in irritability, with a focus on, though not limited to, the pediatric population.

It has been reported that the prevalence of irritability is rising substantially in community samples [4]. Hence, obtaining a better understanding of irritability through the lens of SRC may represent the foundation for intervening early. Childhood represents an early window of opportunity for the prevention of negative long-term outcomes for the affected individuals [18] and might inform the development of therapeutic interventions on specific SRC processes. For instance, cognitive remediation targeting specific self-control deficits [139, 140] and biofeedback [141, 142] or virtual reality techniques [143] targeting specific psychophysiological SRC may represent an interesting therapeutic approach.

We limited this systematic review to children and adolescents. Thus, further reviews should focus on adults. Moreover, this systematic review was limited to literature published in English, German and French. We may, therefore, have missed some information published in other languages. However, this is unlikely to taint the global observed picture, as we did not limit the search based on language in the first place. Due to the heterogeneity in study designs and operationalizations of the associations between irritability and SRC, we did not conduct a quantitative meta-analysis, which will be helpful in the future when methodologies are more harmonious and, thus, comparable.