This is the first study to characterize social functioning in daily-life in a relatively large sample of adolescents and young adults with ASD and 22q11DS using EMA. Our main findings indicate that both participants with ASD and 22q11DS show similar social behaviors. In particular, they spent a comparable amount of time alone, more time with the people they live with but less time with familiar individuals (e.g., friends) than HC. Overall, participants with ASD and 22q11DS also reported a similar experience of aloneness (ExpA), with the exception that individuals with ASD reported feeling more excluded than both participants with 22q11DS and HC. By contrast, they reported markedly different social experiences (ExpSI), with individuals with ASD reporting worse ExpSI than both participants with 22q11DS and HC. The only similarity between the two clinical groups in terms of ExpSI was a higher desire to be alone when in company of other people compared to HC. Regarding the influence of context on affect, individuals with ASD reported less PA and more NA than the other two groups, regardless of the context. Finally, being in the company of other people had a beneficial impact on affect in HC, whereas this benefit was less clear in 22q11DS.
The present study challenges the commonly accepted idea that social withdrawal is a characteristic of neurodevelopmental disorders (4,15,16,24,25,58), as the three groups reported spending a similar amount of time alone (i.e. physical absence of other people or aloneness (59)). This is however in line with the findings of Hintzen and al. (45), who also used EMA, whereas the remaining studies used more classical approaches to measure social withdrawal, such as questionnaires and interviews. Moreover, the majority of studies used information reported by caregivers and not by the participants themselves. That being said, even if the present study highlights a comparable amount of time spent alone between the three groups, it also suggests that individuals with 22q11DS and ASD have a different involvement in the social world compared to their typically developing peers. Indeed, they both reported spending more time with the people they live with and less time in the company of familiar persons outside of the direct family circle, friends in particular. However, a similar amount of time spent with classmates and colleagues was reported between the groups. This is in line with previous studies reporting smaller social networks in individuals with ASD (13,60–63) and the fact that individuals with 22q11DS have been described to be more isolated from peers (4). These results also highlight the central role of the family environment in the lives of adolescents and young adults with neurodevelopmental disorders (e.g., 64). During development, adolescence typically represents a period of emancipation from the family circle (65), a process that appears to be challenged in both youth with 22q11DS and ASD. Several reasons could potentially explain this phenomenon.
First and in line with previous reports (e.g., 66), it suggests that individuals with 22q11DS and ASD have fewer opportunities of interactions, especially in less structured environments, which may prevent them from broadening their social network. Indeed, the fact that participants with 22q11DS and ASD reported spending a comparable amount of time with classmates or colleagues than their peers is probably explained by the fact that these interactions mostly take place in relatively structured contexts (e.g., school or work). These interactions might therefore be more predictable and more accessible than those one can have with friends. Of note, the majority of our sample was composed of people who were still attending school, which ensures at least a minimal number of social encounters through these structured contexts. This should be considered in light of the transition to adulthood, a period during which there might be less structured opportunities to be in contact with other people outside of the direct family circle. Indeed, several studies have highlighted a high percentage of unemployment among adults with ASD (67–69), especially younger adults (70), which reduces the opportunities of interactions in a structured environment. Of note, this phenomenon might be a shared characteristics of several clinical populations, including individuals with neurodevelopmental disorders and severe mental illness such as schizophrenia (71). This lack of opportunity to interact with peers, especially in less structured settings, may be related to the fact that social initiatives are more difficult for youth with ASD and 22q11DS. Incidentally, adolescents with ASD were found to rely more on parents to facilitate social relationships (14,25,72), which would mean that they spend time with family members in order to access time with friends. In the present study, the different types of company were computed to be mutually exclusive. For this reason, we could not identify if participants reported to be in the company of both a family member and a friend, which could support the idea of family as a way of accessing friends. Moreover, cognitive level was found to account for spontaneous initiations of interaction with peers (25). Particularly in the group of participants with 22q11DS, whose average intellectual functioning level is in the borderline range, this could contribute to explain the differences in the time spent with friends. Indeed, interactions with friends mostly take place in less structured contexts and therefore rely more on the ability to actively initiate an interaction and less on the ability to follow well defined “social scenarios”.
Secondly, it is possible that when HC refer to people as “friends”, individuals with ASD and 22q11DS refer to them as “classmates/colleagues”. Indeed, this distinction requires a deep and accurate understanding of the different types of relationships as well as a certain introspection to fully gasp the distinction between the different kind of relationships. Incidentally, it was found that children with ASD lack an intersubjective understanding of interactions with peers and social relationships when they were asked about the quality of their relationships and about their understanding and feelings of loneliness with self-report questionnaires about friendship and loneliness (73), which makes it hard to fully understand the complexity of social interactions. Interestingly, the number of friends reported by the caregivers were superior of the one reported by the children in the study of Bauminger & Kasari (73). This is in line with the precited hypothesis of a misunderstanding of the different degree of friendship and the derived terms (e.g., friend or classmate).
In summary, the present study shows that individuals with 22q11DS and ASD are not characterized by social withdrawal as such but that social interactions take place much more within the restricted family circle and in relatively structured environments. Since most of our participants were still attending school – which provides opportunities for structured social interactions – the results of the present study suggest that the transition after school and towards independent living should also be anticipated from a social perspective in order to avoid a decrease of the number of social contacts. A longitudinal follow-up of such a cohort would provide a unique opportunity to investigate how social behaviors evolve during this transition period from adolescence to adulthood.
Experience of aloneness (ExpA)
Contrary to our expectations, the three groups reported a similar subjective experience of aloneness. This is a major finding considering that, as previously said, social disinterest is typically considered to be a feature of both ASD and 22q11DS (e.g., 24,74). Taken together, the results of this study suggest that individuals with 22q11DS and ASD are not characterized by social withdrawal from a quantitative point of view (i.e. social behavior) but also do not report social disinterest from a qualitative point view (i.e. social experiences). Moreover, it is particularly interesting to note that the three groups did not differ on the item “I’d rather be with other people”, pointing toward a preserved, but of a relatively low intensity, motivation for interpersonal interactions, as pointed out by previous studies (25,75,76). It should be noted that the findings discussed in the context of the present study are based on group comparisons that may mask substantial interindividual variability. Future studies should aim to parse this heterogeneity in order to identify relevant subgroups of individuals characterized by distinct social profiles. Such an approach has recently been employed by Uljarevic et al. (77) who used a social functioning questionnaire to cluster individuals with ASD based on their social phenotype. However, given the limitations of classical measures of social functioning to assess social experiences in an ecologically valid way (34), future studies may use EMA to identify more relevant subgroups.
Despite an overall similar subjective experience of aloneness among the three groups, individuals with ASD reported higher levels of exclusion and isolation feelings than both participants with 22q11DS and HC when looking at individual items composing ExpA one by one, which points towards increased loneliness (i.e. a negative emotional experience (59)) in this population. These findings are in line with previous studies that found higher levels of loneliness in younger (14,25,e.g., 63,73,78) and older adolescents with ASD (e.g., 29,79,80), though none of them used EMA to measure this construct. As suggested by Maddox and al. (76), the subjective feeling of social isolation experienced by individuals with ASD could be explained by a lack of knowledge about how to form relationships. Besides, it was shown that adolescence is a transition phase during which individuals experience new relationships and the expectations towards these relationships evolve (81), inducing loneliness when there is a gap between the expectations and reality (59,82). Interestingly, one of the few studies that directly compared individuals with ASD and 22q11DS found higher levels of empathy, sense of humor and other complex social skills in 22q11DS than in idiopathic ASD (83), elements that could possibly play a role in preventing them to feel lonely. Indeed, if individuals with ASD have little access to these complex social skills, it could lead to a worse comprehension of social interactions, and therefore to feelings of rejection and a greater experience of loneliness. Of note, social anxiety was also found to be related to greater loneliness (78,84–88), and this comorbidity was reported twice as much in our sample of individuals with ASD compared to the 22q11DS group, playing a plausible role in explaining the distinct experience of aloneness between the two conditions. Moreover, social anxiety was described by Hintzen et al. (45) as the discrepancy between the fear of rejection and the desire for social interaction, a pattern that matches what is reported by individuals with ASD in the present study. Future studies should aim to better investigate loneliness in individuals with ASD and with 22q11DS by taking in account the potential impact of social anxiety.
Experience of social interactions (ExpSI)
Contrary to the experience of aloneness, participants with ASD reported a markedly different – and more negative – experience of social interactions than both individuals with 22q11DS and HC, providing information about how social interactions are experienced in daily-life and not only about the quantity of social interactions. This finding is not consistent with the previous report of Hintzen and al. (45), who observed that individuals with ASD mostly enjoyed the company of other people. This could be explained by the fact that our sample, being younger than the one of Hintzen and al. (mean age = 28.3), is characterized by less mature emotion regulation strategies. Indeed, the latter are known to improve with age (89), and in younger sample like ours, could contribute to the rather negative experience of social interactions. Additionally, emotion regulation difficulties have been shown to be inherent to ASD and contribute to the socioemotional and behavioral problems they experience (e.g., 90). The period of adolescence is also characterized by more conflictual social interactions (e.g., 91), and experiences of bullying and peer victimization appear to be particularly frequent among individuals with ASD at this age (e.g., 92). A longitudinal follow-up of this cohort of adolescents and young adults would allow to investigate whether changes in the experience of social interactions are occurring with increasing age.
Of particular interest, the present study highlights that the subjective experience of social interactions is markedly different between individuals with ASD and those with 22q11DS, with subjective reports of ExpSI in the 22q11DS group being similar to those of HC. Besides, if ASD has been described as a frequent comorbidity of 22q11DS (e.g., 93), other studies also highlighted differences between the two conditions in terms of social impairments (83,94,95). Moreover, previous findings suggests that individuals with both 22q11DS and ASD are characterized by social anhedonia (96–98), described as a diminished social interest and a lack of pleasure from social contact leading to withdrawal (99). However, the present study offers new insights regarding the subjective experience of social interactions that challenge this assumption. As already stated above, these discrepancies might arise from the fact that previous studies used parent-reported information collected in a laboratory setting, whereas the present study uses self-reported information collected in the daily-life of individuals. In our sample of participants with ASD, the profile of answers was more characteristic of social anxiety than of a diminished social interest. Indeed, they reported a negative experience of social interactions and a lower enjoyment of the company of others compared to both participants with 22q11DS and HC, but did not spend more time alone. The reports of participants with 22q11DS during social interactions were also not indicative of a diminished social interest, as they experienced social interactions rather positively and rated the pleasantness of their social company similarly to HC. When alone, they also reported wanting to be with other people to the same extant than HC, which is also suggestive of a preserved motivation to interact with others. Of particular interest, recent studies using EMA have also challenged the notion of social anhedonia in psychosis. In particular, adults with psychosis were found to spend less time with other people than HC but reported an intact hedonic experience of social interactions (71), a profile that is relatively consistent with the one observed in our sample of participants with 22q11DS. Of note, ExpSI was not influenced by the type of company in any of the three groups, suggesting that profile described above reflect how participants experienced their social interactions in general. However, it should be noted that the small number of occurrences during which participants were in company of unfamiliar people prevented us of examining ExpSI in this social context specifically. In line with previous studies (e.g., 45), it is likely that they would have resulted in a more negative ExpSI.
Interestingly, both participants with ASD and 22q11DS reported a greater preference for being alone when they were in the company of other people, regardless of the type of company. Incidentally, this was the only significant difference between individuals with 22q11DS and HC, who otherwise reported a similar positive appreciation of social interactions. This could possibly be explained by the cost of interacting: for individuals with ASD or 22q11DS, social interactions might require greater efforts than for HC, hence the higher desire to be alone, although social interactions were not as unpleasant for 22q11DS as they were for ASD. Of note, since participants with 22q11DS and ASD spent more time with the people they live with, it is also possible that they did not always choose to be interacting with them, contributing to this higher wish to be alone.
Altogether, the results regarding social behaviors and the subjective experience of social interactions suggest that individuals with 22q11DS and ASD could benefit from different therapeutic intervention targeting social impairments. Indeed, individuals with 22q11DS should probably be supported on how to initiate and maintain social interactions since they experience social interactions positively but experience difficulties in this domain. In ASD, interventions should rather focus on skills that might help to decrease the negative experiences of social interactions, such as increasing conversational skills. Support could also be offered on how to better discriminate and differentiate other people’s affective states during social interactions, which may increase interpersonal synchrony and contribute to experience social interactions in a more positive way. For example, it has been shown that a poor differentiation of other people’s emotions in individuals with ASD contributed to lower levels of empathy and to mimic the facial expression of others to a lesser extent (e.g., 100). Moreover, interventions targeting social anxiety have been proven to be benefic and even more effective when coupled with social skills training (e.g., 101).
Influence of context on affect
Individuals with ASD reported less PA and more NA than 22q11DS and HC, regardless of the context. Therefore, ASD participants reported a more negative emotional experience overall, whether they were alone or in company of other people. This is in line with several findings in autism that revealed intense negative and attenuated positive emotions (102,e.g., 103), although divergent findings have also been reported (104). The relatively high level of NA could partially be accounted by the psychiatric comorbidities present in our ASD sample: mood disorders were reported twice as much by participants with ASD than by participants with 22q11DS, the same being observed for social phobia and agoraphobia.
In the HC group, being in the company of other people was shown to have a beneficial impact on affective states. Indeed, higher levels of NA and lower levels of PA were observed when alone, the oppositive pattern being found when in the company of others. This is in line with a study using EMA that found more happiness and interest, as well as less sadness, pain and tiredness when individuals were engaged in social interactions as opposed to when they were not (e.g., 105). In individuals with 22q11DS, this benefit was less clear since the social context had a similar impact on NA than in HC but an opposite impact on PA. The fact that reduced NA were reported in both HC and 22q11DS when in company of others highlights the positive impact of being with others on NA. This is coherent with the positive ExpSI reported by both HC and 22q11DS, indicating that interactions with others were enjoyed and associated with lower NA. Of note, 22q11DS spent the majority of their time with people they live with, meaning that interactions took place in a structured and well-known environment that was probably reassuring and could have soothed NA. However, PA were higher when alone in 22q11DS. This could be explained by the fact that, spending a lot of time home, participants with 22q11DS didn’t necessarily always choose to be interacting with the people they live with, which could explain why they report greater PA when alone. Moreover, as already discussed, interactions probably required greater effort for participants with 22q11DS, which could contribute to making them feel more relaxed and joyful when alone.
Strengths, limitations and future directions
This is the first study comparing the social phenotype in daily-life of two neurodevelopmental conditions – ASD and 22q11DS. By using EMA, contextual information is taken into account and offers more granulated information that consider daily variations, therefore ensuring that participants’ answers reflect their actual environment accurately. Moreover, it shows feasibility of this method in neurodevelopmental disorders, reinforcing EMA literature that is still scare in this domain. Furthermore, the present study contributes to better distinguishing between two conditions often considered as characterized by the same social impairments, adding new information about social interactions functioning in daily-life, a domain still little studied.
However, results of the present study should be considered in light of several methodological limitations. First, EMA relies on participants subjective self-report. Although one member of the research team went over EMA items with all the participants, interpretation may still differ from one individual to another. Moreover, as the 22q11DS group had a significantly lower IQ than both ASD and HC, the level of comprehension could have been different between the groups. Of note, Wilson and al. (35) validated EMA feasibility in a population with moderate intellectual disability but this technique had rarely been used in 22q11DS population (39,40). This is why we took time to go through the protocol with each participant and to carefully read and explain each item, as well as closely monitoring them during the full EMA period. Being available for questions and technical issues experimented by participants also had an impact on study compliance, since only 5 participants were excluded from the original sample because of an insufficient number of answered beeps.
Secondly, heterogeneity within the 22q11DS and ASD groups should be considered. Indeed, various comorbidities and medications were present in both clinical groups, possibly having an impact on the results. However, comorbidities are more the rule than the exception in neurodevelopmental disorders (e.g., 106,107). Given the variety of comorbidities reported, it wasn’t possible to subdivide our groups accordingly but future studies should aim to further investigate this important question. Six participants with 22q11DS scored above the clinical cutoff on the SCQ, suggesting concerns for a potential diagnosis of ASD (the presence of an ASD diagnosis was not formally examined in this group). To examine the influence of these participants on the obtained results, all the analyses were conducted while excluding these participants and the results remaining unchanged. This suggests that the results obtained in the 22q11DS group are not explained by the presence of comorbid autistic traits in a subgroup of participants. Another limitation of the current study is that only 8% of the ASD group presented low intellectual functioning. For this reason, the obtained results cannot be extended to the low functioning part of the spectrum. However, contrary to previous research in ASD using EMA (35,42–44,46,47), not only high functioning individuals were recruited. Finally, alexithymia wasn’t examined in the present study. Given the high prevalence of alexithymia among individuals with ASD (108), it would have been useful to get a deeper comprehension on how this could influence their answers.