Social competence

Social competence has been defined as the level of ability a child possesses to successfully coordinate and implement the processes and resources available to meet social demands and achieve social goals (e.g., peer interactions) [8–10]. Environmental factors, and family influences in particular (e.g., parents’ attitudes, beliefs, social role models, parent-child interactions), contribute significantly to the development of social competence through childhood [11]. With a handful of recent exceptions, social competence is assessed using parent- or teacher-rated, ‘broad-band’ questionnaires [12] (e.g., Social Skills Improvement System [SSIS] [13]; Child Behavior Checklist [14]; Strengths and Difficulties Questionnaire [SDQ] [15], School Social Behavior Scale (SSBS-2) [16] or diagnosis-specific tools (e.g., Autism Spectrum Rating Scales [17]). These measures provide insight into everyday function, but not to the socio-cognitive skills that underpin social competence.

Social cognition

Social cognition refers to “those aspects of higher cognitive function which underlie smooth social interactions by understanding and processing interpersonal cues and planning appropriate [social] responses.” [18. p.559]. These processes underpin the ability to perceive and process social information and include recognizing others’ thoughts and emotions, attributing mental states to oneself and others, understanding that others have mental states different from one’s own and attributing causes or intent to others’ behaviors [11]. Social cognition skills emerge gradually through infancy into adolescence [19], beginning with infant smile and imitations of others’ actions [20]. The ability to recognize and understand emotions and theory of mind emerge later, around three to four years [21–23] and develops through childhood. Complex social information processing, social communication and intent attribution mature later and, by the end of adolescence, most young people are capable of high-level social decision making [24, 25].

Targeted assessment of social cognition, using tools that precisely characterize the child’s ‘socio-cognitive profile’ is critical for precise, effective intervention. With a few exceptions (e.g., NEPSY Affect Recognition and Theory of Mind [26]), tools evaluating children’s social cognition have inadequate normative data, are designed primarily for adults and have limited appeal for children (e.g., Ekman faces [27]; Mind in the Eyes Test, [28]; Trolley Task, [29]). Optimally, tools assessing children’s social cognition should take account of developmental factors and incorporate real-life situations, viewed from a first-person perspective, to maximize ecological validity [7, 12, 30].

Disruptions to socio-emotional development

Maturation of social competence and social cognition is dependent upon a healthy environment, strong social models, and a healthy brain [11]. Disruptions to development, due to biological or environmental factors, increase the risk of derailing these skills, though the magnitude of the impact may be difficult to accurately quantify due to imprecise definition and assessment [1, 3, 4].

To date, most research has focussed on social competence, employing either dimensional approaches such as observer ratings, which place the child’s social competence along a continuum, or diagnostic tools (e.g., Diagnostic Statistical Manual of Mental Disorders–V [31], which classify a child’s function in terms of symptom clusters. Based on such approaches, global estimates suggest 50–60% of children with neurodevelopmental disorders, (NDDs) [32, 33] and 10% of children in the general population experience impaired social competence [34]. Elevated risk of impaired social competence is also reported in children with acquired brain injury [35–39], epilepsy [40], chronic illness [41], learning disabilities [42], attention-deficit/hyperactivity disorder (ADHD) [43], and anxiety [44]. Consequences of deficits in social competence include social isolation and exclusion, bullying, lack of satisfying friendships, behavior problems and poor quality of life.

In contrast, apart from ASD research, little attention has been paid to the social cognitive mechanisms underpinning impaired social competence or to whether specific socio- cognitive ‘phenotypes’ can be identified for NDDs and child mental health conditions which can then be used to guide, targeted, effective treatments to optimize socio-emotional development. Further, the question remains as to whether these problems represent a primary social cognitive impairment or a secondary consequence of other conditions and comorbidities.

Our study objective was to comprehensively assess and characterize social competence and social cognition in typically developing children (TDC) and children with neurodevelopmental and mental health (‘clinical’) diagnoses (ASD, ADHD, anxiety [ANX]) using novel tools able to distinguish diagnosis-specific socio-emotional profiles. Specifically, we predicted that: i) compared to TDC, children with clinical diagnoses would be differ on both parent ratings and child-direct assessment; ii) standardised parent questionnaires would not differentiate between the clinical groups and instead yield a global, non-specific profile of socio-emotional impairment common to all groups; and iii) child-direct assessment would yield evidence for distinct, disorder-specific “social cognition profiles,” which could represent novel and potentially modifiable targets for personalised treatment in these high-risk populations.

Design

This study employed a multi-site, cross-sectional design and recruited a community-based sample of TDC, and a clinical sample of children with NDDs and mental health diagnoses. Participants in the TDC sample and clinical groups are mutually exclusive.

Participants

Measures

Social cognition.

Paediatric Evaluation of Emotions Relationships and Socialisation (PEERS). PEERS [51, 52] is a child-direct, app-based, interactive digital health tool delivered via iPad providing a sex and age-normed assessment of social cognition, social communication and attention/executive skills, using real-life situations, viewed from a first-person perspective, to maximize ecological validity. PEERS includes three sex and age-adjusted composite scales (Cognition, Primary Social Processing, Complex Social Processing, M = 100, SD = 15), a Total score (M = 100, SD = 15) and nine subtests (M = 10, SD = 3) and three Supplementary subtests which do not contribute to the Total Score. (Table 1, Fig 1). Each subtest begins with a practice item, followed by the scored test. PEERS software automatically calculates sex and age-normed scores based on error and completion times scores from the Community-based TDC sample (referred to as the ‘normative sample’).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1.

Examples from the PEERS subscales: A. Matching Emo, B: Odd One, C. Get This, D. Multiple Morals.

https://doi.org/10.1371/journal.pone.0291929.g001

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Pediatric evaluation of emotions, relationships and socialization.

https://doi.org/10.1371/journal.pone.0291929.t001

Procedure

This study was approved by the Human Research Ethics Committees of the Victorian Department of Education and Early Childhood (002318), Catholic Education Melbourne (2166) and The RCH (34046), which conform to the Declaration of Helsinki. The individuals depicted in Fig 1 of this manuscript have given written informed consent to publish these images. Schools selected to provide a demographically representative sample, were approached and provided with a study overview. Those agreeing to participate distributed study information and consent forms to parents of all enrolled children (electronic/ hardcopy). Consenting parents returned forms either to the school in hardcopy or electronically via a secure research database (REDCap) [57]. If questionnaires were not received on the day of the assessment, parents were sent daily reminders for four days. If questionnaires were still outstanding, the research team made two reminder phone calls, after which the data were considered missing.

Once parents returned consent forms, an assessment was scheduled either at the child’s school, or outpatient clinic at The RCH. Children 12 years and older provided written consent, and children younger than 12 years provided verbal assent. Assessments were conducted individually by trained researchers beginning with the IQ assessment followed by PEERS and took on average 1–1.25 hours. Parent ratings and child-direct assessment data were available for all children in the sample.

Statistical analysis

Analyses were conducted using Stata v15.1 [58]. Means and standard deviations were generated for continuous variables, frequencies and percentages for categorical variables. Planned contrasts of clinical groups compared to the TDC group were conducted, with independent t-tests comparing continuous measures, and chi-squared tests for categorical data.

Group means for the TDC and Clinical groups were compared for intellectual ability, social cognition, social competence alongside 95% confidence intervals (95% CI), allowing comparison of measurement and accuracy of clinical group scores to normed values. Planned contrasts between TDC and each clinical group were explored using independent t-tests. For these analyses, a value of p<0.05 was deemed statistically significant.

Sample characteristics

Participant demographics, FSIQ data, and parent-ratings are provided in Tables 2 and 3. For the TDC sample, mean age was 8.9 years (SD = 2.4), 357 (69.6%) were of Caucasian origin, 79 (15.4%) Asian origin, with the remainder including small numbers of African, Indigenous, and Hispanic children. In the clinical sample, 116 (85.3%) of children were of Caucasian background, with 8 (5.9%) of Asian heritage, 4 (2.9%) were Indigenous, 1 (0.7%) African descent, with race unspecified for 7 (5.1%). SES was comparable across groups, with the exception that the ADHD group recorded lower SES (p = 0.006). For FSIQ, all group means fell within the average range, but for ASD and ADHD groups, mean FSIQs were lower than the TDC (ps <0.001).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Sample demographics.

https://doi.org/10.1371/journal.pone.0291929.t002

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Comparison of intellectual ability, behavior and social competence across TDC and clinical groups.

https://doi.org/10.1371/journal.pone.0291929.t003

For clinical groups, SDQ Total Difficulties scores were similar, and in the ‘clinically significant’ range for all clinical groups (ASD: M = 21.4; ADHD: M = 21.0; ANX: M = 20.7) (see Table 3). In keeping with expectations, on the SDQ, children with ASD and ANX were rated as having internalizing behaviors in the clinical range, with borderline-clinical externalizing symptoms. In contrast, the ADHD group had clinical level externalizing behaviors and borderline-clinical internalizing symptoms. Prosocial and peer problems were only detected in the ASD and ADHD groups. Of note, all clinical group mean ratings fell in the clinical range for the hyperactive/ inattentive subscale, and only the ANX group had mean scores in the clinical range for Emotional Symptoms.

Social competence: Parent ratings

On broad band social competence questionnaires (SSIS: Social skills, PEERS-Q), parents endorsed significantly more problems across all clinical groups compared to TDC (all t-test comparison ps<0.001) (See Table 3). Mean parent ratings on the SSIS Social Skills (ASD = 75.8, ADHD = 77.4, ANX = 80.9) fell in the borderline-clinical range for all clinical groups. In contrast, PEERS-Q provided a more fine-grained description of everyday social competence. PEERS-Q total scores were in the clinically significant range for ASD (M = 72.1) and ADHD (M = 71.4), and borderline-clinical for ANX (M = 66.7). In addition, parents’ endorsements of children with ASD were in the clinical range on all subscales. ADHD group ratings were similar, with the exception that Participation scores were borderline-clinical. The ANX group’s PEERS-Q subscale scores were less elevated, falling in the borderline-clinical range, apart from normal ratings for Social Rules, and clinical level ratings for Relationships.

Social cognition: Child-direct assessment

Age-standardised PEERS Total, Composite and subtest scores are detailed in Table 4, and raw scores are presented in S1 Table.

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Comparisons of PEERS performances between TDC and clinical groups–scaled and composite scores.

https://doi.org/10.1371/journal.pone.0291929.t004

Supplementary subtests

For Say What, only the ADHD group performed significantly worse than TDC (p<0.01) overall, with performances characterized by more errors and slower completion times. The ASD group were less proficient in recognizing emotions in conversation. LISAS scores, integrating speed and accuracy, were significantly lower than TDC for all clinical groups. For Mind Read, scaled scores were similar across all four groups, and raw scores found no differences between TDC and ANX groups except for completion time for Level 2 (10.5 seconds, p = 0.022). The ASD group had slower completion times than TDC at all three conditions of the task, making more errors at Condition 1. Children with ADHD made more errors at Condition 1 and 3 and were slower on Condition 2. All clinical groups performed worse than TDC on Social Intent, with the ADHD group recording the greatest differences, followed by the ASD group (both p<0.001). Raw scores indicated more hostile, aggressive and negative (all p<0.01) responses than TDC for both ADHD and ASD groups. For the ANX group, subtest scores were lower than TDC, but only the Social Information Processing Score result reached significance (p<0.001).

Parent ratings of social competence and behavior

Parent ratings of children’s social competence distinguished clinical and TDC groups. Ratings of social skills (on SSIS) [13], while elevated, were not in the clinical range. In contrast, on PEERS-Q, which provides a fine-grained description of social competence [49], total scores were in the clinical range for all clinical groups. Further, for ASD and ADHD groups, clinical-level problems were present across all subscales, other than borderline-clinical participation for the ADHD group. For the ANX group, ratings were generally borderline-clinical with clinical problems for Relationships and normal ratings for Social Rules.

Social cognitive skills

Child direct assessment provided additional detail regarding social cognitive abilities and was able to distinguish social profiles across TDC and clinical groups. Consistent with expectations and parent ratings, the ASD profile was characterized by significant challenges in primary social processing, including recognizing, processing and labelling emotions across modalities (verbal, gesture, visual), as well as processing and interpreting simple social scenes. For complex social processing, moral maturity, emotional congruence, theory of mind, intent attribution and social information processing were all impaired compared to normative data (S1 Table). These poor results were exacerbated by cognitive difficulties, notably slowed information processing.

The ADHD profile included reduced cognition and primary social processing underpinned by specific weaknesses in cognitive efficiency and self-regulation, rather than primary social cognitive deficits. Children with ADHD recorded relatively faster response times and more errors when identifying emotions and interpreting social aspects of real-life situations, suggesting they struggled to manage speed-accuracy trade-off effectively (S1 Table). On complex intent attribution tasks, children with ADHD were more likely to interpret others’ actions as hostile or negative, rather than positive or neutral. Such responses may reflect prior negative social experiences, with ADHD associated with elevated risk of bullying by peers [32].

Children with ANX demonstrated relatively few difficulties in social cognition, in contrast to NDD groups and consistent with previous research [60]. Primary social processing difficulties were mild and specific to emotional labels, including facial and gestural representation of emotions. Consistent with the ADHD group, an intent attribution bias was identified, with this group also more likely to interpret others’ actions as hostile or negative, rather than positive or neutral. This is in line with findings that identified that those with social anxiety have a tendency to focus more on negative emotions in comparison to positive emotions [61]. Together, these findings suggest weaknesses, or hesitancies, in interpreting basic emotions, but intact abilities for complex social decision-making.

These distinct, disorder-specific social profiles provide meaningful information, superior to that of parent ratings, regarding the nature of social problems across clinical groups. This more nuanced characterization contributes to our understanding of whether impaired social competence is a ‘primary’ deficit, as for the ASD group, or reflects secondary consequences, as with ADHD, where social problems were characterized by reduced response inhibition and impulsivity and difficulties balancing the speed-accuracy trade-offs. Our findings are similar those from adult research that suggests performance on social cognitive tasks varies more subtly in those with ADHD compared to ASD [62]. In contrast to ADHD and ASD, and despite reports of day-to-day social difficulties, children with ANX had overall intact social competence, apart from slowed response times on some social information processing tasks, raising the possibility that social problems in these children may be secondary to anxiety and associated lack of confidence or inefficient decision making. Better understanding of the bases of everyday social behaviors will facilitate the implementation of intervention which best targets primary impairments.

Despite including a large sample and multiple measurement approaches, this study had some limitations. By excluding children attending special needs schools, it is likely that children with more severe ASD, ADHD and anxiety were not recruited. Thus, our findings likely reflect a relatively high-functioning subset of these children, thus restricting our ability to identify diagnosis-specific social profiles for more severe forms of these conditions. We were limited by the reliance on clinical diagnoses of ANX, ASD, ADHD, and a lack of dimensional data regarding children’s symptoms, and were unable to explore relationships between social profiles and diagnostic subtypes.