Emotional dysregulation (ED), defined as a deficit in the ability to monitor and modulate the valence, intensity, and expression of emotions, is typically expressed with mood instability, irritability, tantrums, and self-harm in young children with autism spectrum disorder (ASD) (1–4). Although ED is not among the diagnostic criteria of ASD, its manifestations seem to be an important contributor to functional impairment and clinical referral (5); furthermore it appears to be related to ASD core symptoms (5). Indeed, children with ASD often have problems in recognizing and managing emotions, inhibiting emotional reactions, delaying gratification, and tolerating transitions (6).

Emotional dysregulation represents a transdiagnostic risk factor for mental health disturbances both in the general population (7, 8) and in ASD (9, 10), and could have relevance as a possible target for developing specific therapeutic interventions (11). ED occurs when the ability to control the intensity, duration, and type of the emotions experienced is impaired, resulting in negative affectivity or irritability. Emotion regulation typically occurs in an adaptive way with the capacity to upregulate positive emotions and down regulate negative ones. There are many strategies that can be used to down regulate negative emotions: some of them, such as avoidance, denial, and negative rumination, are maladaptive and associated with poorer outcome in term of anxiety and depression, while others, such as cognitive reappraisal, problem solving, and acceptance, are more adaptive (12). In particular, the flexibility in being able to adopt more than one strategy seems to be especially adaptive (12).

When emotional regulation is not effective, intense emotions interfere with social interaction. Children with such difficulties tend to engage in peer conflicts and, consequently, to be less accepted by others and to experience negative feed-back in relationships, with consequent feelings of exclusions (13). Indeed, in longitudinal studies of children with ASD, emotion regulation abilities seem to be associated with subsequent prosocial behaviors (14). Ineffective emotional regulation impairs also the ability to perform goal oriented activities and are therefore involved in learning processes (15).

The ability to regulate emotions grows with age, being less effective in preschoolers than in older children, is influenced by puberty with decreased regulation during adolescence (16), and becomes more effective in adulthood (17). Parenting style is a key factor for developing good self emotion strategies during childhood (18), as is also the social context in which the child grows (19). Sex differences in emotion regulation strategies have been reported across different phases of development in the general pupulation (19). For example, a study reported that females engaged in more social support seeking strategies and dysfunctional rumination, whilst males employed more passivity, avoidance, and suppression strategies (20).

People with ASD seem to be at higher risk of having impaired emotional regulation because of poor emotional awareness and low competence in emotional language, poor flexibility, high sensitivity to change and to environmental stimulation (21). Due to these deficits, emotions become difficult to manage, negative emotions tend to last longer and can lead to prolonged ruminations after social setbacks and to intense reactions to social rejection, with consequent need of intervention from others to externally regulate emotions. In younger kids, ED can be expressed with externalizing behavior, such as aggression, irritability, tantrums, self harm, and “meltdowns” or “shutdowns.” If left untreated, ED in children with ASD tends to persist over time (22).

As ED underlies several behavioral problems in ASD, therapeutic interventions aimed at empowering the ability to recognize and regulate emotions might be an effective way of addressing behavioral problems (21). The social adaptive domain (SAD) has been found to be more impaired than other domains in ASD even after accounting for sex, age, and cognitive development (23, 24). Although there is wide heterogeneity in the level of functioning among people with ASD, cross sectional studies have documented an inverse correlation between autism symptom severity and adaptive functioning (25). Moreover in ASD people there is often a pronounced discrepancy between cognitive abilities and adaptive functioning profile, and this seems to be more pronounced in individuals with older age, higher IQ and higher social-communication symptoms (26). Compared with individuals with the same age and IQ, people with ASD have a lower overall adaptive functioning profile (27). Results indicated that impairment in adaptive behavior within the domain of socialization skills remains a distinctive factor of ASD (24). Adaptive functioning in children with ASD is positively associated with a family’s quality of life (28).

More in detail, the ABAS-II Social Adaptive Domain score is derived from the scores of the two subscales: play and social abilities. Since ABAS-II examines “real-world” functions, it provides an index of the child’s competency in everyday contexts, such as the ability to engage in leisure activities by its own, to interact with others while playing, to ask for adult’s involvement while playing, to share emotions, and to exhibit social engagement, stages of according to different developmental age.

To our knowledge no study has analyzed the relative contribution of ED to adaptive functioning in ASD in very young children.

This study aimed to examine the relationship between ED and adaptive functioning in preschoolers clinically referred for ASD or other neurodevelopmental disorders, while taking into account autism symptoms, sex, and presence of global developmental/cognitive delay.

The sample consisted of 100 patients, of whom 74 were males (74%). Mean age at first evaluation was 39.43 months (SD 12.25, range 18–70 months) (Table 1). Sixty-five children (65%) met the specified criteria for inclusion in the ASD group. Sixty children (60%) scored under the 3rd percentile at cognitive or developmental tests and were therefore categorized as having intellectual disability/developmental delay (GDD).

On the CBCL, mean scores were: 56.51 ± 11.01 on the internalizing scale and 54.20 ± 9.74 on the internalizing scale, with a total score of 54.92 ± 11.20. Detailed subscales of CBCL are reported in Table 1, with comparison between the ASD and non-ASD groups.

The mean CBCL-AAA T-score of the total sample was 173.11 ± 19.85), with no statistically significant difference between the ASD (175.30 ± 19.32) and the non-ASD group (169.05 ± 20.47) (p = 0.054). No sex difference was found in our sample with regard to adaptive functioning, ADOS CSS, CBCL AAA, and GDD (Table 2).

The ABAS mean General Adaptive Composite score (GAC) of the sample was 71.45 ± 18.66, conceptual adaptive domain (CAD) mean score was 69.40 ± 18.38, social adaptive domain (SAD) mean score was 78.56 ± 18.95, and practical adaptive domain mean score (PAD) was 76.82 ± 19.08.

The mean GAC score of patients who received a diagnosis of ASD (65.96 ± 16.76) was lower that in patients without ASD (81.62 ± 17.91; p < 0.0001). Scores for DAS (74.26 ± 16.04 in ASD vs. 86.54 ± 21.46 in non-ASD, p = 0.0004) and DAP (71.30 ± 16.05 in ASD vs. 87.05 ± 20.25 in non-ASD, p < 0.0001) were also lower in children with ASD. No significant difference was found between the ASD and non-ASD with regard to CAD (p = 0.05, Table 1).

The mean ADOS CSS score of the whole sample was 4.95 ± 3.16). As expected, the mean ADOS CSS in ASD (6.75 ± 2.37) was higher than in non-ASD (1.60 ± 0.88; p < 0.0001).

The association between ED and overall adaptive functioning was then evaluated. The model was adjusted by ADOS CSS and GDD. Higher ED was associated with lower overall adaptive functioning (p = 0.003). Significant associations between ADOS CSS and adaptive functioning (p < 0.001) and GDD and adaptive functioning (p = 0.023) were also revealed (Table 3).

To study the impact of ED on the diverse domains of adaptive functioning, we analyzed a univariate multivariable model for each adaptive functioning domain (Table 3). Lower CAD levels were predicted by the presence of GDD (estimate mean difference −11.34, p = 0.008) and by higher ED (estimate −0.18, adjusted p value = 0.046), while both higher ED (estimate −0.28, adjusted p value = 0.002) and ADOS CSS (estimate −1.34, adjusted p value 0.026) were associated with lower SAD. Higher levels of ADOS CSS (estimate −1.85, adjusted p value = 0.001) and ED (estimate −0.32, adjusted p value < 0.001), and the presence of GDD (estimated mean difference 9.43, adjusted p value = 0.011) were all significant predictors of worse PAD.

Likewise, when considering only children with ASD, higher ED was associated with lower adaptive functioning levels. Also the ADOS CSS was associated with overall adaptive functioning. With regard to the three adaptive functioning domains (Table 3), CAD was only predicted by GDD (estimated mean difference 10.52, adjusted p value = 0.016). ADOS CSS (estimate −1.88, p value 0.032) and ED (estimate −0.28, adjusted p value 0.002) were both found to be associated with worse SAD. ADOS CSS (estimate −2.30, adjusted p value 0.011) and CBCL AAA (estimate −0.24, adjusted p value = 0.015) predicted lower PAD.

Prior research has suggested that ED in children with ASD is a significant problem that is likely to underlie behavioral difficulties, such as tantrums, self-harm and irritability, and thus to contribute to functional impairment. The pathogenetic relevance and the mechanisms of the relationship between ED and clinical symptoms are still a matter of study, and investigation into ED in ASD children is ongoing. This study aimed to contribute to the research in this area by examining the relationship between ED and functional impairment, and by measuring different domains of adaptive functioning in preschoolers clinically referred for suspected ASD. To our knowledge, this study has been the first to evaluate autism symptoms, emotional dysregulation and adaptive functioning in preschoolers with ASD and other neurodevelopmental disorders. The clinical characteristics of the study sample are consistent with those expected in this population, with a predominance of the male sex and about two-third of the referred children meeting criteria for ASD or being at high risk for ASD (48).

Most studies have reported prevalence rates of ED between 50 and 60% in ASD (49–52), which are higher than those reported in typically developing children (22, 53–55) or in children with other neurodevelopmental disorders (54). However, a recent large study in clinically referred children aged 6–18 years reported higher rates of ED among non-ASD compared to ASD children, using the CBCL-AAA as a measure of ED (56). In our sample, no significant differences were found between ASD and non-ASD regarding ED, even if ED was numerically higher in ASD, with a p = 0.05 (Table 1) that closely missed the pre-specified alpha threshold (p < 0.05). Clinically referred children, like the preschoolers included in our sample, have been generally found to have high levels of ED. Among samples of clinically referred children for a suspicion of ASD, about 60% (57) receive an ASD diagnosis, while the others are very often diagnosed with language disorders, global developmental delays or having inattention, hyperactivity, and disruptive behaviors. The relationship between ADHD (58–61), developmental delays (62) or other disorders (63) and emotional has been extensively reported. Moreover boys with specific language impairment, have been reported to have lower emotional regulation abilities than typically developing children (64). However, the current literature on ED in clinically referred preschoolers remains rather limited.

The close relationship between ED and core autism symptoms has led some to suggest that ED should be considered among the ASD features (50). But, ED is also a well known transdiagnostic risk factor for many mental disorders in children and adolescents, such as, among the others, ADHD, mood disorders, posttraumatic stress disorder, non-suicidal self-injury, and disruptive mood dysregulation disorder (65). In fact, the CBCL-AAA index was initially studied as a measure of high risk for bipolar disorder in adolescence (66–69), and subsequently as a risk factor for other psychopathology associated with suicidality in adults (68, 70). CBCL-AAA has been applied also to the preschooler population, using the 1 and 1/2 –5 years old version of the CBCL (39, 40). In non-clinical preschoolers population there was an association between higher scores on CBCL AAA and psychiatric symptomatology, lower adaptive functioning, temperament and maladaptive parenting (39).

A specific impairment in the social adaptive functioning domain (SAD) had been reported to be distinctive of ASD compared with other neurodevelopmental disorders (24). In our sample, however, we found that ASD was accompanied by a general impairment in multiple domains of functioning, and not just SAD (Table 1). This is consistent with what was recently reported in an Italian preschooler sample (71). Previous reports have confirmed that toddlers (younger than 2 years old) with ASD have a lower adaptive functioning compared to children with developmental delay (72).

In our study, considering both the overall sample and also ASD sample, higher ED was associated with lower adaptive functioning levels (Table 3). Social and practical domains are significantly influenced by both autism symptoms severity, as expected (25), and by ED, as hypothesized for this study. The relationship between emotional regulation strategies and subsequent social competences is well documented. Berkovits et al. analyzed a group of 4–7 years old children with ASD and found that ED was strongly related to social and behavioral functioning but was largely independent of IQ (5). Reyes et al. compared two samples of ASD children and typical developing peers (3–7 years), and found that in ASD better social skills were associated with greater emotional regulation and increased expression of emotions. Better emotional regulation was linked to fewer peer problems and more prosocial behaviors (22).

We didn’t find any sex difference in ED or in other clinical variables. This could be due to the rather small representation of females in the study sample and/or to referral bias whereby females with high ED might have been more likely to be clinically referred. Few data on sex differences in emotional regulation among ASD people are available, but scarcely comparable to our population (55, 73). Further studies on sex differences in preschoolers with ASD are needed to evaluate different trajectories of development, and so identify vulnerable children.

The results of this study are clinically important because they document the relevance of ED to adaptive functioning in young children and thus provide support to efforts to develop emotional regulation strategies as a possible means of ultimately improving functioning. In typical development, good parental scaffolding helps children overcome frustration and improve emotion regulation. ASD children too can benefit from parental scaffolding but they seem to have difficulty in internalizing and generalizing co-regulatory support. This is consistent with the clinical experience that children with ASD benefit from the presence of one-to-one support for regulatory management in education settings well into middle childhood and beyond (74). Parent-mediated early interventions targeted to ED in preschool years could therefore be considered as a means of improving functioning. More research is needed to better understand the relationship between ASD symptoms, ED, and parenting style within the social context.

This study has several limitations. First, the cross sectional design precludes the possibility of assessing prospectively the interplay between ED and adaptive functioning or to evaluate the developmental trajectories of these constructs over time. Second, the study sample included only clinically referred children without a comparison group of typically developing children. Third, the male preponderance, while being consistent with the epidemiology of the disorder, limits inferences to female patients. Fourth, both the ED and adaptive functioning ratings were provided by the same informant (main caregiver), thus limiting somewhat the external validity of the associations. Finally, no systematic information on the social context was available. Further studies with larger samples and prospective design are needed to confirm and better clarify our findings.

Higher levels of ED were associated with lower adaptive functioning in all the domains evaluated, regardless of the severity of the autistic symptoms or the presence of GDD. Decreasing ED in preschoolers with ASD or other neurodevelopmental disorders may be an effective way of improving their overall adaptive functioning, including prosocial behavior, ability to focus on goal oriented activities, and academic achievement. In particular, since a child’s ability to regulate emotions is modeled by parental scaffolding, parent-mediated early interventions may be an effective approach to improving the developmental trajectory of children with high levels of ED.

The raw data supporting the conclusions of this article can be made available by the authors upon qualified request.

The study was approved by the institution review committee (Comitato Etico Interaziendale A.O.U. Città della Salute e della Scienza di Torino). Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

CD and DM conceived and designed the study protocol with input from FA, FR, and BV. CF, AG, LA, and BS carried out literature searches. BS, CF, VFC, and LA collected the data. DM designed and carried out the statistical analysis. CD, AG, DM, and BV interpreted the data and drafted the manuscript. FR and FA supervised the writing of the manuscript. All authors critically reviewed and contributed to the final version of the manuscript.

In the last 2 years, BV has received consultant fees or honoraria from Medice, Menarini, and Angelini Pharmaceuticals. FR has received sponsorship or fee for advisory board from Roche, Novartis, Biogen, PTC Therapeutics, Sanofi Genzyme, Sarepta/biogen and CD has received consultant fee from Roche and Lundbeck, and DM has received consultant fee from Ethos Ltd. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.