Background
The 22q11.2 deletion syndrome (22q11.2DS) results from a hemizygous deletion on chromosome 22q11.2 and is the most common microdeletion syndrome in humans, with an estimated prevalence of 1:4000 [
1,
2]. The syndrome is associated with a variety of clinical manifestations of highly variable expressions [
3] including a high prevalence of psychiatric disorders. Notably, schizophrenia spectrum disorders are seen in up to 41% of the clinically identified adult 22q11.2 deletion carriers [
4‐
9], and large cross-sectional case-control studies in schizophrenia have reported odds ratio estimates for the 22q11.2 deletion between 15 and 44 [
10‐
13]. Moreover, an epidemiological population study has shown that the risk of schizophrenia spectrum disorders among persons diagnosed with 22q11.2DS is approximately eight times the risk of the general population [
14].
The clinical presentation of schizophrenia in 22q11.2DS has considerable overlap with clinically defined idiopathic forms of schizophrenia [
7,
15] and as such, 22q11.2DS may represent a biological homogeneous “at risk group”, that can further the understanding of both the behavioral and biological mechanisms underlying development of schizophrenia in the general population. Also, since more 22q11.2 deletion carriers are being identified at a young age, they may provide insights into the early signs of schizophrenia pathology.
Social impairments are core features of schizophrenia [
16] and can be categorized into three multidimensional constructs, nested within each other. The most narrow construct can be defined by social cognitive deficits, which underlie the more widely concept of social skills that again contributes to the broad construct of social function [
17]. “Social cognition” refers to the mental processes involved in perceiving, attending to, remembering, thinking about, and making sense of the people in our social world [
18] and includes emotional processing and theory of mind (ToM) [
19]. “Social skills” refers to verbal and non-verbal behaviors with complex cognitive abilities necessary to engage in successful interpersonal interactions (e.g., initiating a conversation) [
20,
21]. Lastly, “social function” is operationalized in a variety of ways, encompassing performance across everyday domains (e.g., independent living, employment, interpersonal relationships, and recreation) [
22].
The reported social impairments in schizophrenia include social cognitive functions [
23‐
29], social skills [
30,
31], and social functioning [
31‐
33]. In particular, diminished everyday functioning and social skills are associated with negative rather than positive symptoms in schizophrenia [
31,
34]. Similar results have been reported for a group of children and adolescents (age 12–18 years) at clinical high risk for schizophrenia presenting with either prodromal or transient and recent-onset psychotic symptoms; a recent drop in daily functioning combined with schizotypal personality disorder or a first degree relative with a psychotic disorder [
35]. Moreover, the Dutch Genetic Risk and Outcome in Psychosis (GROUP) study has shown correlations between measures of social cognitive impairments (i.e., The Degraded Facial Affect Recognition task) and negative symptoms severity in both patients with non-affective psychotic disorders and subclinical negative symptoms in their healthy siblings. This finding may suggest that there is a common origin of the social cognitive functions and negative symptoms [
23].
Previous studies have reported that children and adolescents with 22q11.2DS have impairments in social functioning [
36,
37], social skills [
38], and social cognition, e.g., [
38‐
41]. Diminished ToM has previously been reported to associate with the severity of psychotic symptoms in adolescences with 22q11.2DS [
42] and, Chow and collaborators [
43] found more impairments of ToM among 22q11.2DS adults with schizophrenia compared to 22q11.2DS adults without psychosis. However, there is still limited knowledge on the inter-relationship between different layers of social competences in 22q11.2DS and how impairments in the individual layers of the social competences relate to symptoms of schizophrenia in young 22q11.2 deletion carriers.
The aims of this study were twofold; to characterize impairments across social functioning, skills, and cognition and their inter-relationship in 22q11.2 deletion carriers who are at risk for developing schizophrenia and to investigate the strengths of correlation between these different layers of social impairments and subclinical levels of negative and positive symptoms in a young 22q11.2DS cohort without schizophrenia. We expected that the 22q11.2DS group would show deficits at all three layers of social competence that the different measures of social competences were inter-correlated and that the level of social impairments associated with the subclinical levels of negative and positive symptoms.
Results
Demography
There was an overall nominally significant sex difference in mean age (
t(56) = 2.2,
p = 0.03). The 22q11.2DS group had mean IQ well below the general population mean and differed significantly from the mean IQ in the control subjects (
t(56) = 7.4,
p = 2.14 × 10
−10). A total of six individuals with 22q11.2DS had IQ below 70 but within the mild (IQ 50–69) to moderate range (IQ 35–49). The 22q11.2DS cohort presented with reduced everyday functioning as indicated by a mean ABAS
composite rating, which were more than two standard derivations below the mean of the control group. There was a significant group
F(1,51) = 62.3,
p = 2.14 × 10
−10, age
F(1, 51) = 20.9,
p = 3.16 × 10
−5, and sex
F(1, 51) = 9.8,
p = 0.003 effect for ABAS
composite ratings. Table
1 shows that there was consistency between the lifetime and current prevalence of all psychiatric disorders but anxiety and phobias. This is due to the presence of specific phobias (e.g. arachnophobia) in both cases and control individuals, which people rarely seek health care for.
Overall, the 22q11.2DS cohort had significantly elevated mean SIPS scores for negative, positive, and disorganized but not generalized symptoms, relative to control peers (Table
2). The majority of the individuals with 22q11.2DS (
n = 27) had negative symptoms scores of two standard derivations above the mean of the control group, and six of these individuals presented at least one attenuated psychotic symptom (i.e., SIPS score above three in one or more positive symptoms). Only three individuals with 22q11.2DS presented with elevated disorganized symptoms. A sensitivity analysis leaving out the six individuals with attenuated positive symptoms provided similar nominal significant between-group difference for negative (
W = 657,
p = 6.51 × 10
−10, positive (
W = 459,
p = 0.01), disorganized (
W = 566,
p = 6.35 × 10
−7) as well as generalized (
W = 355,
p = 0.60) symptom scores.
Table 2
Characteristics of negative and positive SIPS scores and social competences in 22q11.2DS and control peers
Psychopathology |
Positive symptomsb
| 3 | 0–12 | 0 | 0–9 | | 625 | 7.5 × 10−4
|
Negative symptomsb
| 6 | 1–17 | 0.38 | 0–2 | | 831 | 5.8 × 10−11
|
Disorganized symptomsb
| 1 | 0–9 | 0 | 0–2 | | 723.5 | 1.2 × 10−07
|
General symptomsb
| 0 | 0–7 | 0 | 0–4 | | 469.5 | 0.32 |
Social functioning |
ABASsocial
c
| 99.89 (18.6) | 256–606 | 126.1 (9.26) | 110–138 | −2.77 | 61.02 (1,50) | 2.9 × 10−10
|
Social skills |
SRStotal
b
| 64 | 15–122 | 15.17 | 3–44 | 5.00 | 678 | 6.2 × 10−9
|
Social cognition |
TASITtotal
b
| 38 | 16–56 | 52.83 | 37–58 | −3.83 | 44.5 | 8.0 × 10−9
|
ERTtotal
c
| 40.83 (12.09) | 16.67–64.44 | 64.52 (7.75) | 47.67–77.78 | −2.96 | 88.07 (1,53) | 7.39 × 10−13
|
Social competence, age, and sex
The 22q11.2DS cohort was significantly impaired on all measures of social function, skills, and cognition, relative to the control group (Table
2).
The 22q11.2DS group had mean ratings and scores of more than two standard derivations below the mean of the control individuals for both ABAS
social and ERT
total. There was a statistically significant main effect of group as determined by two-way ANCOVA for ABAS
social and ERT
total (Table
2) as well as a significant effect of age,
F(1, 51) = 10.8,
p = 0.002, and sex
F(1, 51) = 14.90,
p = 0.003 for ABAS
social ratings. There was also a significant main effect of age,
F(1, 53) = 10.58,
p = 0.002 but not sex
F(1,53) = 3.26,
p = 0.08 in ERT
total scores. There was no pairwise interaction between age, sex, and group for either ABAS
social or ERT
total (
p > 0.05). Post hoc analyses showed that relative to the control group, the 22q11.2DS group was significantly less accurate in detecting the ERT subdomains of feelings of sadness, fear, disgust, anger, and happiness but not surprise (Table
3).
Table 3
Impairments in subdomains of social cognitive functions in 22q11.2DS relative to controls
ERT subdomains |
Disgust | 44 | 6.77 x10−9
|
Sadness | 90.1 | 4.44 × 10−7
|
Anger | 96 | 5.95 × 10−7
|
Happiness | 189 | 5.0 × 10−4
|
Fear | 197 | 8.0 × 10−4
|
Surprise | 340 | 0.3 |
TASIT subscales |
Paradoxical sarcasm | 53 | 5.55 × 10−8
|
Simple sarcasm | 84.5 | 1.16 × 10−6
|
Sincere | 167.5 | 6.0 × 10−4
|
Wilcoxon rank-sum tests showed statistical significant differences in SRS
total and TASIT
total scores between the 22q11.2DS and control group (Table
2).
Post hoc analysis revealed that all three types of social interactions accessed using TASIT (i.e., sincere, simple sarcasm, and paradoxical sarcasm) were significantly impaired in the 22q11.2DS group compared to the control group (Table
3).
Spearman’s correlation analysis showed an age-related trend in SRStotal ratings (r
s
= −0.38, p = 0.053) and TASITtotal scores (r
s
= 0.35, p = 0.07) in the 22q11.2DS but not in the control group (SRStotal: r
s
= −0.14, p = 0.48; TASITtotal: r
s
= 0.21, p = 0.29). Using Wilcoxon rank-sum test, we found no sex differences in either SRStotal ratings among 22q11.2DS (W = 357, p = 0.83) or control individuals (W = 75, p = 0.63). Likewise, there were no sex differences in TASITtotal scores within the 22q11.2DS cohort (W = 73, p = 0.43) or the control group (W = 85.5, p = 0. 68).
Inter-relationship between the three layers of social competencies
Spearman’s correlation analysis showed significant correlations between the ABASsocial and SRStotal ratings in both the 22q11.2DS (r
s
= −0.78, p = 3.174 × 10−6) and the control group (r
s
= −0.53, p = 0.003). Nominally, significant pairwise correlations were also seen between ERTtotal and the TASITtotal score within the 22q11.2 deletion carriers (r
s
= 0.37, p = 0.053) but not within the control individuals (r
s
= 0.00, p = 0.99). There were no significant correlations between ABASsocial or SRStotal and TASITtotal or ERTtotal (r
s
< 0.3, p > 0.05).
Explorative analyses of social competencies, everyday functioning, and IQ
We used explorative pairwise correlation analyses to explore if the social impairments across the different layers were related to overall everyday functioning (ABAScomposite) and IQ. Spearman’s correlation analysis showed significant correlations between the ABAScomposite and SRStotal ratings in both the 22q11.2DS (r
s
= −0.79, p = 1.37 × 10−6) and the control group (r
s
= −0.48, p = 0.008). Pearson’s correlation analysis showed a strong inter-relationship between ABAScomposite and ABASsocial in both groups (22q11.2DS: r = 0.84, p value = 4.48 × 10−8; controls: r = 0.86, p = 1.78 × 10−9). There was a trend in the pairwise correlations between ABAScomposite ratings and the ERTtotal score among the control individuals (r = 0.35, p = 0.07) but not the 22q11.2 deletion carriers (r = 0.19, p = 0.34).
Explorative correlation analysis showed that ERTtotal (r = 0.43, p = 0.02) but not TASITtotal (r
s
= 0.16, p = 0.41), SRStotal (r
s
= −0.02, p = 0.92), or ABASsocial (r = −0.04, p = 0.86) correlated significantly with IQ in the 22q11.2DS group. Within the 22q11.2DS cohort, there were no significant differences in mean test score of the four independent measures of social competence when we stratified the group into those with IQ below and above 70, respectively (p > 0.05). Similar results were obtained for the control group in which IQ did not correlate with either SRStotal (r
s
= −0.15, p = 0.42), or ABASsocial (r = 0.02, p = 0.91); but there was a trend in the correlations between IQ and ERTtotal (r = 0.35, p = 0.06). In contrast to the findings in the 22q11.2DS group, there was a nominal significant correlation between IQ and TASITtotal among the control individuals (r
s
= 0.41, p = 0.03).
Associations between social competence and negative and positive symptoms in 22q11.2DS case-only analyses
Linear regression analyses revealed an overall significant effect of the age, sex, and IQ corrected models on negative SIPS symptoms, which individually included ABAS
social, SRS
total, and TASIT
total and a nominally significant effect of the model including the ERT
total score (Table
4) within the 22q11.2DS cohort. The analysis of variance showed that the full models that included ABAS
social, SRS
total, and TASIT
total explained a significant fraction of the variation in negative symptoms than the basics models alone (Table
4) whereas the model that included ERT
total did not. Table
4 shows that there was an independent significant effect of ABAS
social, SRS
total ratings, and TASIT
total but not age and sex on the 22q11.2DS cohort.
Table 4
Linear models for negative and positive symptoms: the effect of social function, skills, and cognition adjusted for age, sex, and IQ
Intercept | Estimate | 3.90 | 3.55 | 3.66 | 4.18 | 2.17 | 1.81 | 1.95 |
T | 4.90 | 4.34 | 4.22 | 4.17 | 1.73 | 1.37 | 1.45 |
Pr(>|t|) | 6.76 × 10−5
| 0.0003 | 0.0003 | 0.0004 | 0.1 | 0.18 | 0.16 |
Age | Estimate | 0.04 | 0.06 | 0.03 | −0.01 | 0.17 | 0.19 | 0.11 |
T | 0.88 | 1.25 | 0.71 | −0.20 | 2.20 | 2.23 | 1.43 |
Pr(>|t|) | 0.39 | 0.22 | 0.48 | 0.84 | 0.04 | 0.04 | 0.17 |
Male sex | Estimate | −0.35 | −0.05 | 0.02 | 0.002 | −1.18 | −0.82 | −0.54 |
T | −1.21 | −0.21 | 0.08 | 0.006 | −2.61 | −1.98 | −1.37 |
Pr(>|t|) | 0.24 | 0.84 | 0.94 | 0.99 | 0.02 | 0.06 | 0.18 |
IQ | Estimate | −0.02 | −0.03 | −0.02 | −0.02 | −0.03 | −0.03 | −0.02 |
T | −3.09 | −3.33 | −2.94 | −2.16 | −2.63 | −2.70 | −2.03 |
Pr(>|t|) | 0.005 | 0.003 | 0.007 | 0.04 | 0.02 | 0.01 | 0.05 |
Social measure | Estimate | −0.51 | 0.52 | −0.38 | −0.19 | −0.64 | 0.63 | −0,25 |
t | −3.67 | 3.78 | −2.86 | −1.23 | −2.93 | 2.89 | −1.18 |
Pr(>|t|) | 0.001 | 0.001 | 0.009 | 0.23 | 0.008 | 0.009 | 0.25 |
Full modela
| Rb
| 0.42 | 0.44 | 0.39 | 0.22 | 0.29 | 0.28 | 0.15 |
F (df) | 5.71 (4,22) | 5.93 (4,21) | 5.28 (4,23) | 2.92 (4,23) | 3.60 (4,22) | 3.47 (4,21) | 2.17 (4,23) |
P | 0.003 | 0.001 | 0.004 | 0.04 | 0.02 | 0.03 | 0.10 |
Basica vs. full modelb
| F | 13.44 | 14.31 | 8.18 | 1.51 | 8.59 | 8.38 | 1.38 |
p | 0.001 | 0.001 | 0.009 | 0.23 | 0.008 | 0.009 | 0.25 |
Moreover, there was a significant effect of IQ in the models that included ABASsocial, SRStotal, and TASITtotal.
An extended model that included both ABASsocial and TASITtotal (F = 5.82; p = 0.03), but not a model that included both ABASsocial and SRStotal (F = 2.31; p = 0.14), explained nominally 10% more (r
2 = 0.52) of the variation in negative SIPS symptoms than the model including ABASsocial alone (r
2 = 0.42).
Separate linear models including ABAS
social or SRS
total explained a nominal significant part of the variance in the positive symptoms seen among the 22q11.2 deletion carriers (Table
4). In these two models, ABAS
social and SRS
total ratings contributed significantly to the model, and there was a nominal effect of IQ, age and sex (Table
4). However, the results did not survive correction for multiple testing.
Discussion
To the best of our knowledge, this is the first study that examines multiple layers of social impairments in the same cohort of young 22q11.2 deletion carriers, the pairwise interrelationship between the different layers of social impairments, and their relationship with schizophrenia-related symptoms. We found impairments across all layers of social competences in this young non-psychotic 22q11.2DS cohort, and as anticipated, we saw that the level of impairment in each social layer was associated with the severity of predominately subclinical negative symptoms.
Social impairments within and across different layers
In line with previous studies of 22q11.2DS, we found strong deficits in social function [
36,
37,
68] and social skills [
69,
70]. With a mean total SRS score of 64, we saw the same magnitude of impairment’s for comparable age groups as that previously reported by Ho and collaborators (mean SRS total score of 66–72). In contrast to our expectations, the inter-relationship among the measures varied and only the measures of social function and social skills were highly correlated. Hence, our results suggest that deficits in other cognitive functions than emotional recognition and ToM are also important for the more global social skills and functional abilities. The lack of strong correlations between the four social test variables and everyday functioning and IQ, respectively, indicated that overall, the social impairments across the different layers of interest were not solely accounted for by the overall reduced everyday (ABAS
composite) or intellectual (IQ) functioning within our 22q11.2DS cohort.
Social impairments and negative and positive symptoms
Our data showed that 22q11.2 deletion carriers who had worse social functioning (i.e., those who had lower ratings on ABAS
social) and who had worse social skills (i.e., higher SRS
total ratings) had significantly more subclinical negative and positive symptoms. Notably, the ratings of social functioning and social skills emerged as those most associated to both negative and positive symptoms, which is consistent with findings in clinical high-risk subjects [
31]. Social functioning has previously been shown to deteriorate from childhood into early adolescence among 22q11.2 deletion carriers who later developed schizophrenia [
71]. In line with our results, others have also found that worse scores on parent-reported sociability, peer relations, and interests were significantly associated with higher levels of schizotypy symptoms in 22q11.2DS [
72]. A previous study has also reported that adolescents with 22q11.2DS and psychotic symptoms have more social withdrawal and less adaptive socialization skills in comparison with young 22q11.2DS individuals without psychotic symptoms [
73]. Taken together, these lines of evidence indicate that a drop in social functioning and social skills may predate the onset of psychosis and schizophrenia in 22q11.2DS.
ToM deficits have previously been described for a cohort of 22q11.2DS adults with schizophrenia in comparison with a 22q11.2DS cohort without schizophrenia [
43]. ToM impairments (mean TASIT
z-score of approximately −1.75 compared to control peers) have also been reported in a mixed cohort of 22q11.2DS adolescents where some presented with psychotic illness [
42]. Jalbrzikowski and collaborators showed that TASIT scores were correlated with both positive and negative SIPS symptoms levels. Similarly, we found pronounced deficits in ToM as well as a strong association between TASIT
total scores and level of subclinical negative but not in positive symptoms in our cohort of young 22q11.2DS individuals. We also saw that ToM deficits were apparent in all three TASIT interaction types, with the more difficult interaction type (i.e., paradoxical sarcasm) showing the most significant difference between 22q11.2 deletion carriers and control peers. Due to the low levels of positive symptoms in the cohort, our data did not meet the assumption of linearity between TASIT scores and positive symptom levels. The discrepancies between our data and the results previous reported by Jalbrzikowski et al. (2012) on positive symptoms may result from difference in schizophrenia prevalence between the two 22q11.2DS cohorts under investigation. In particular, the average levels of positive symptoms differ by threefold between the two 22q11.2DS studies.
In line with previous reports [
38‐
41], we found that the 22q11.2DS group had strong emotional recognition impairments with difficulties in perceiving emotions from facial expressions across the panel of negative emotions (i.e., anger, fear, and disgust), but that the ability to identify expressions of surprise was conserved. In contrast to these previous reports, our 22q11.2DS cohort showed reduced ability to recognize happy facial expressions. However, the CANTAB ERT test applies morphed images with fast cover up times to avoid ceiling effects. As a consequence, the CANTAB ERT test may be more sensitive to more subtle deficits compared to other emotion recognition tasks, and the inter-study discrepancies may relate to differences in the tools applied between studies.
A recent meta-analysis has shown moderate associations between impairments in emotional recognition, and the severity of negative symptoms in idiopathic schizophrenia [
74] and findings from the Dutch Genetic Risk and Outcome in Psychosis (GROUP) study suggest that social cognitive deficits and negative symptoms are etiologically related [
23]. Although we only observed a minor association between emotional recognition deficit and the severity of subclinical negative symptoms, in our non-psychotic 22q11.2DS group, we did observe a stronger association between the negative symptoms and the other social cognitive function accessed by TASIT. This could indicate that there is a shared origin of these phenomenon’s in 22q11.2DS. Moreover, we cannot exclude the possibility that a pathological progression of negative symptoms may result in stronger correlations between these phenomena if some of the 22q11.2 deletion carriers develop schizophrenia. This would provide stronger confirmation on the etiologically relatedness between the social cognitive deficits and the negative symptoms. Taken together, these lines of evidence suggest that 22q11.2DS may be a valuable genetic model for schizophrenia per se or for the subgroup of schizophrenia patients more highly affected by negative symptoms.
Similar to findings in idiopathic schizophrenia, e.g., [
74], the social impairments in our 22q11.2DS cohort were not accounting for all of the variation in symptom severity. Hence, our results suggest that deficits in other cognitive functions are contributing to the mechanisms driving the negative and particular positive symptomatology. Moreover, the extended models that included ABAS
social and SRS
total did not improve the predictability of the linear models compared to the model for ABAS
social alone. This suggests that the effect of social skills on the negative symptom levels were captured by the measure of social function (ABAS
social). In contrast, when the combined effect of both ABAS
social and TASIT
total were considered the model accounted for a significant greater proportion of the variation in negative symptoms compared to the model for ABAS
social alone. Together with the lack of strong inter-relationship between the ABAS
social rating and the TASIT
total score, this may suggest that the effect of ToM on the negative symptom levels is mediated by slightly different mechanisms that are not fully encompassed within the measure of social function (ABAS
social).
Clinical implications
Identification of the early signs of schizophrenia is mandatory for early intervention and prevention strategies in psychiatry. Emerging evidence from studies in idiopathic schizophrenia suggests that psychosocial interventions are effective in improving functionality [
21,
75‐
77] and that early intervention may improve treatment outcome and even reduces the disease transition rate [
78].
From our study, the association between social impairments and subclinical negative and positive symptom severity emphasizes the importance of clinical awareness of potential subclinical psychosis symptoms when caregiver or self-reporting suggests substantial social impairments. Based on these findings, we speculate if psychosocial interventions could be efficient in improving functioning and perhaps even reduce the severity of schizophrenia symptoms in 22q11.2DS.
Intervention models for targeted psychosocial training in 22q11.2DS have already been proposed [
79,
80] and cognitive enhancement therapy of tailored exercises, which include perspective taking and identification of non-verbal social clues, has been shown to improve social cognitive abilities in adolescents with 22q11.2DS [
80]. Also, the training scheme applied by Glaser et al [
79] which targets focusing on the eyes, emotion recognition and understanding, and working memory has shown short-term improvements in these functions. However, the long-term effect on psychopathology of these interventions is still to be evaluated. Due to the lack of strong correlations between the social cognitive abilities and the more applied social skills and functioning, seen in our study, we speculate if psychosocial training should target both the cognitive domain and the applied skills in order to improve the overall functioning or mitigate symptom severity in 22q11.2DS individuals. Hence, longitudinal studies that can improve our understanding of the underlying mechanisms of social functioning and the relevance of social impairments as early markers of schizophrenia, as well as potential targets for psychosocial training in 22q11.2DS are of the essence.
Limitations
Our study showed that social impairments were more pronounced than negative and positive symptoms in our 22q11.2DS cohort where none of the participants meet the clinical diagnostic criteria within the schizophrenia spectrum. However, the cross-sectional study design is not suited to address the direction of the association between the social impairments and the subclinical symptoms (i.e., if social impairments develop before the onset of negative and positive symptoms or vice versa?). Despite this limitation, we only observed relatively low levels of current symptoms among the recruited carriers of the 22q11.2 deletion and the associations between social impairments and negative symptoms were strong and in the same direction as shown in the previous reports [
42,
71]. However, the robustness of our results regarding positive symptoms may be diminished due to sample size limitations. Likewise, the uneven sex ratio and the skewed between-sex and age distributions are likely sample size dependent, and although we saw the same magnitude of social skill deficits as previously reported for 22q11.2DS [
69,
70], we did not observe the usually reported superior female SRS ratings [
81] in neither 22q11.2DS nor controls. Our results may also be influenced by the application of one test block only in the ERT test; nonetheless, overall, our findings of diminished emotional recognition deficits were in accordance with results reported by Jalbrzikowski et al. [
42] and more recently in a group of children and adolescents with 22q11.2DS [
41].
The measures of social competences, which we have applied, were not evaluated relative to a negative control measure (i.e., a cognitive function unrelated to the symptomatology) and therefore we cannot fully specify their impact on the negative and positive symptoms. The area of social functioning is complex and can be conceptualized and measured as a very important area of adaptive functioning and as autistic social and communicative deficits. Although the reliability and validity of the Danish translation of the ABAS-II has not been assessed, ample reliability and validity data are included in the ABAS-II manual [
59]. Moreover, the average rating score of the 22q11.2DS cohort was more than two standard derivations below the mean of the control group indicating that everyday functioning was indeed impaired in the cases.
This study has evaluated two central aspects of social cognitive functions (ToM and emotion recognition). However, these functions may only reflect part of the multi-construct domains that they belong to. Other tasks will probably tap into overlapping or different mentalizing capacities that are equally important for performance in the social world and be implicated in negative and positive symptomatology.
Clinical case–control studies are often affected by ascertainment bias (i.e., cases being ascertained from subspecialty hospital departments and/or control individuals not being representative of the background population). However, the case group enrolled in the current study was not recruited from subspecialty hospital units. Moreover, based on the clinical profile of lifetime illnesses obtained from the Danish healthcare registries, we have previously shown that the case group is representative of the entire cohort of 22q11.2 deletion carriers known in Denmark [
44]. In contrast, our control participants are probably more altruistic (i.e., willing to invest time on something that is not to the benefit of themselves) and may have stronger mental resources than the true community control would have. However, we believe that this is a general limitation of case–control studies as such.
Acknowledgements
We are grateful to the participants and families who invested their time in our study and the Danish National 22q11DS Association for their active support to our work. We highly appreciate the efforts of Irene Cecilie Jensen, Lisbeth Juhl Mikkelsen, Morten Møller Kristensen, and Catherina Juel for their dedicated assistance; Gerda Demant Olesen and Lisbeth Nymark Jørgensen for lab support; Thomas Hempel Sparsøe for genetic assistance; and Rasmus Torp and Niels Feldsted Thorsen for their database support as well as Dorte Helenius for the fruitful statistical discussions, and Shantel Weinsheimer for the critical review of the manuscript. We would also like to express our gratitude to the staff involved in the Danish Blood Donor Study, Capital Region Blood bank, Glostrup and the staff at the Danish Neonatal Screening Biobank.