Introduction
Autism spectrum disorder (ASD) is a heterogeneous neurodevelopmental syndrome, which is characterized by persistent deficits in social communication and social interaction across multiple contexts and restricted, repetitive patterns of behaviour, interest or activities (American Psychiatric Association
2013). Individuals with ASD show a large heterogeneity in phenotype, severity and type and frequency of symptoms. In addition, symptoms can change within one person over the course of development (Hill and Frith
2003; Levy et al.
2009; Wing
1997). However, as initially stated by Kanner (
1943), and still relevant today: ‘The outstanding, “pathognomonic”, fundamental disorder is the children’s inability to relate themselves in the ordinary way to people and situations from the beginning of life’ (p.242). This statement is supported by recent research showing that of the triad of symptoms, social impairments are the most stable symptom not only in children, but also in adolescents, adults and elderly and across all developmental levels (Shattuck et al.
2007). As such, they are considered more central and persistent than the other core symptoms of autism (James et al.
2006; Seltzer et al.
2004; Shattuck et al.
2007). These social impairments typically manifest themselves in abnormalities in reciprocal interactions and difficulties in the expression and recognition of emotions (Bauminger
2002).
Although the social difficulties are independent of age and developmental level, their impact may disproportionally increase during adolescence (Nicpon et al.
2010; Pellicano
2010). Compared with children, adolescents are more sociable and form more complex relationship with their peers. The formation of social networks contributes to their identity development and affects their self-esteem. Adolescents are also more sensitive to acceptance and rejection by their peers (Blakemore
2008; Steinberg and Morris
2001). That is, during this developmental period, opinions and evaluations of peers become increasingly salient and many adolescents with ASD begin to notice how they differ from their peers (Burnett et al.
2009; Crone and Dahl
2012; Steinberg
2005). This is especially noticeable in high-functioning adolescents with ASD as they, more than their low-functioning counterparts, seek and initiate social interaction with peers (Bauminger et al.
2003; Hauck et al.
1995). The effects hereof may be found in the development of anxiety or mood disorders and feelings of social loneliness (the perceived lack of social involvement with peers) in high-functioning adolescents with ASD (Bauminger
2002; White et al.
2009; White and Roberson-Nay
2009). Thus, despite the general assumption of social aloofness in ASD, these reports of loneliness and the relatively high degree to which high-functioning adolescents with ASD initiate social interactions with peers (Bauminger et al.
2003) indicate that they do want to take part in more satisfying social relationships. Despite the different impact social impairments have in the different developmental stages, numerous studies have been performed on the underlying mechanisms of social impairments in younger children with ASD or in adolescents with ASD with low to average intelligence levels, but only limited research has been done on social-cognitive function in high-functioning adolescents with ASD with high-average to above-average intellectual functioning.
For social interaction to be successful, it is necessary to understand another persons’ emotions, intentions, believes and knowledge (David et al.
2010). This information is needed to predict another person’s behaviour and adjust one’s own behaviour accordingly. The ability to impute and understand the mental states of others, and recognize that these states may differ from your own is called ‘mentalizing’ or having a theory of mind (ToM) (Baron-Cohen et al.
1985; Premack and Woodruff
1978; Vollm et al.
2006). The ability to mentalize depends on a range of both lower level mechanisms, such as face and emotion processing, gaze direction and the detection of animacy, as well as higher executive function mechanisms such as attention and working memory (see Stone and Gerrans
2006 for a review). Of these, particularly the human face and its emotional expressions play an important role in mentalizing, as they constitute an important source of information about a person’s inner state. Developmental trajectories of both mentalizing and emotion recognition abilities reach near-adult levels around age eleven, with a further refinement of these skills during adolescence and across the adult lifespan (Baron-Cohen et al.
1985; see Brune and Brune-Cohrs
2006 for a review; Perner and Wimmer
1985; Rump et al.
2009).
Although mentalizing problems in individuals with ASD have been a major topic in autism research since the early 1980s, there is still little information available on the mentalizing abilities in high-functioning adolescents with ASD. To the best of our knowledge, only two studies investigated these abilities in a group of high-functioning adolescents with either Asperger syndrome or autism (Kaland et al.
2007,
2008). Participants in these studies were 21 early-to-late adolescents with ASD and 20 typically developing matched controls. In their first study, Kaland and colleagues (Kaland et al.
2007) used an adjusted and contextually more complex version of the Strange Stories Task (Happé
1994), whereby the participants had to make mental-state inferences versus physical state inferences in a story context. Adolescents with ASD made more errors compared with controls and had slower reaction times especially on the mental-state interference task. The authors concluded that the adolescents with ASD had more problems compared with controls in making inferences about mental states. The second study was seemingly conducted with the same participant group (Kaland et al.
2008). In this study, the performance of the participants on three different advanced theory of mind tasks, that is, The Reading the Mind in the Eyes task (Baron-Cohen et al.
2001), the Strange Stories Test (Happé
1994) and the Stories of Everyday Life (Kaland et al.
2002), were compared. The results showed that the adolescents with ASD performed worse than the typically developing controls on all three mentalizing tasks. In both studies, the authors made no direct link between the performance of the adolescents with ASD on the mentalizing tasks and their more spontaneous mentalizing abilities as for example seen on the Autism Diagnostic Observation Schedule (ADOS), a standard diagnostic measure of ASD (Lord et al.
2010), or between the tasks and the adolescents’ social functioning outside the laboratory setting.
Also important for social interaction in daily life is the ability to perceive and label expressions. However, emotion perception has also only been marginally investigated in high-functioning adolescents with ASD (Rump et al.
2009; Tracy et al.
2011), with contradicting results. Some studies showed that high-functioning adolescents with ASD were able to recognize (complex) emotional facial expressions just as fast and accurate as typically developing teens (Rump et al.
2009; Tracy et al.
2011; Jones et al.
2011), whereas other studies showed that they were less accurate than their typically developing counterparts (Kuusikko et al.
2009; Smith et al.
2010; Mazza et al.
2014; Brosnan et al.
2015; Greimel et al.
2014; Wallace et al.
2011). Only the result of one study (Wallace et al.
2011) was related to the behavioural problems these adolescents with ASD have in everyday life. They found that diminished perceptual sensitivity to sad faces was positively related to the number of social communication symptoms (as measured with the ADOS) and adaptive functioning (as measured with the Adaptive Behaviour Assessment System-II) in a group of 42 high-functioning (IQ > 80) adolescents with ASD when compared with a normal control group.
In all, little attention has been paid to social cognition in high-functioning autism, especially in adolescence, a life phase in which social skills are crucial for developing and maintaining successful relationships with peers. Both ToM and emotion recognition are important aspects of social cognition. However, while there is abundant evidence for a ToM deficit in ASD, the findings on emotion recognition are mixed. Recently, Happé and Conway (
2016) even argued that although ToM deficits are a hallmark of ASD, deficits in emotion recognition are not. Furthermore, the relation between social-cognitive processes such as mentalizing and emotion perception and everyday social functioning is also unclear and especially poorly studied in high-functioning adolescents with ASD.
In the current study, we address these shortcomings by examining both mentalizing and emotion recognition abilities of high-functioning adolescents with ASD and matched typically developing controls. We aimed to relate their results with more spontaneous mentalizing skills as seen on the ADOS and their social functioning as measured with self-reports about friendship and feelings of acceptance. By doing so, we aim to get more insight into the mentalizing abilities of these high-functioning teens with ASD and find out if the skills we measure in the laboratory setting relate to the skills they are able to use in real life. Based on the existing literature, we assume to find impaired mentalizing abilities in high-functioning adolescents with ASD compared to the control group. Due to the scarcity and contradictory evidence of previous reports, emotion recognition abilities and the relation between the adolescents’ social abilities in- and outside the structured setting of the laboratory were assessed in an exploratory way.
Discussion
The first aim in this study was to compare mentalizing and emotion recognition abilities in high-functioning adolescents with ASD and a matched typically developing control group. We wanted to compare the groups on both their spontaneous social abilities and their task-related social abilities. Our results confirm the expected social impairments in high-functioning adolescents with ASD in daily life situations; however, we could not confirm these impairments during task performance.
In line with other studies, friendship characteristics of high-functioning adolescents with ASD differ from those of their normal developing peers (Bauminger et al.
2008a,
b; Kuo et al.
2013). According to their parents, the adolescents with ASD in our study have fewer real friends and they meet their friends less often outside school compared with the control group. In general, they report that their children are less able to get along with others. Considering the three dimensions of friendship in adolescence (i.e. companionship, intimacy-trust, closeness-affection), and the fact that individuals with ASD by definition have peer difficulties (Bauminger et al.
2008b; Mazurek and Kanne
2010), we expected to find these results.
This perspective of friendship characteristics as reported by their parents, however, is not in concordance with how the adolescents with ASD describe these friendships themselves. Although they do report being less able to get along with others, their reports on social acceptance, the number of close friendships and the frequency of contact with their friends were comparable with controls. These results may be explained by differences in the definition of friendship that parents and high-functioning adolescents with ASD use. For example, Kuo and colleagues (2013) suggested that adolescents with ASD may consider peers who have similar interests to be a friend, whereas their parents may consider someone a friend only then if there is a relationship with mutual interactions and emotional interchanges. Another explanation may be that parents underestimate the adolescents’ friendships because they assume that because their child has ASD, (s)he per definition would have fewer (closer) friends (Kuo et al.
2013). Considering that the adolescents with ASD in our study went to a special education school, their parents may also have less insight in the quality of the friendships, as out-of-school contact with friends mainly occurs via social media, or because adolescents in general may be inclined to a lesser extent to share their social life with their parents. However, scores on the item ‘Insight in social relationships’ from the ADOS show that more than 70% of the adolescents with ASD in our study have no or limited insight into the nature of social relationships and/or his/her own role in it. In combination with the fact that all adolescents with ASD in our study scored on or above the ASD cut-off level on ‘Reciprocal social interaction problems’ of the ADOS, it is more likely that these adolescents with ASD use a different definition of friendship. Despite these differences in friendship definition of high-functioning adolescents with ASD and the problems that they have in their everyday life with forming and maintaining social relationships (Scheeren et al.
2013), these problems were not confirmed on task level. Even on an advanced ToM test (Director Task) and on the recognition of more complex emotions such as ‘surprise’ of the ERT, where the participant has to make inferences about someone else’s thoughts (Loveland et al.
1997), we do not find significant differences between adolescents with ASD and the control group.
Although most often not investigated in high-functioning adolescents with ASD, this discrepancy between social interaction problems in everyday life and the absence of deficits on neuropsychological tests of social cognition has been repeatedly reported in the ASD literature (Dahlgren and Trillingsgaard
1996; Rump et al.
2009; Russell and Hill
2001; Scheeren et al.
2013; Senju et al.
2009; Tracy et al.
2011). This apparent discrepancy could be explained by the use of compensation techniques that high-functioning individuals with ASD are often able to apply in more structured settings. By using explicit cognitive or verbally mediated reasoning techniques, in contrast to more automatic social information processing that normal developing individuals use, high-functioning people with ASD are able to pass complex emotion recognition and mentalizing tasks (Harms et al.
2010; Loveland et al.
1997; Senju et al.
2009). Also, the demands of social interaction in daily life are not as explicitly defined as, for example, task instructions of social-cognitive tests. Therefore, mental-state attribution in high-functioning individuals with ASD may also be prompted by explicit task structure and instructions (Scheeren et al.
2013; Senju et al.
2009). A forced-choice paradigm, as in the ERT, may facilitate the recognition of facial emotions, especially if the adolescents with ASD are formally trained to identify emotions using such labels. This kind of training is often a standard component of intervention programs and may have influenced the test results of the ASD group in the current study.
Another additional explanation might be that high-functioning adolescents may not spontaneously be looking for, or detecting and using information from multiple simultaneous and dynamic sources in social situations, even though they might be able to comprehend them (Grossman et al.
2000; Jones et al.
2011). They may avoid making eye contact, fixate on the mouth region (Neumann et al.
2006) or miss a more general tendency to seek and initiate in social interactions. This may explain why high-functioning adolescents with ASD do not spontaneously attribute mental states, although they perform well on social functioning tasks. This is also consistent with the finding that training of mental-state attributions does not necessarily improve social adaptation in ASD (Begeer et al.
2011; Loveland et al.
1997; Ozonoff and Miller
1995; Senju
2013), although some reports show promising results (Bauminger
2002; Stichter et al.
2010) of task-related transfer to various non-trained social situations.
Both explanations, that is, the use of compensation techniques either or not in combination with the lesser tendency to search for and use social information in more natural environments, may also explain why we do not find a relationship between our tasks and daily functioning levels; they may not measure the same cognitive or psychological construct. A limitation of the current study is the relatively small sample size, which may have limited our power for detecting possible significant differences between our groups.
With respect to future recommendations, more research in larger samples is needed to investigate whether social behaviour and social cognition in high-functioning adolescents with ASD are in fact two sides of the same coin. A further challenge is to understand how exactly the aforementioned compensation in ASD takes place. That is, compensation in ASD has not received much attention so far, despite many open questions such as whether or not apparently successful compensation comes at a cost, and how compensation is modulated by the environment (see also a recent review by Livingston and Happé
2017). Moreover, it is still unclear which developing individuals with ASD will be able to acquire successful compensation in later life, as the heterogeneity in clinical presentation, course and potentially underlying neurocognitive mechanisms is large (cf. Happé and Conway
2016). Finally, the ecological validity of experimental social-cognitive paradigms, which are by definition highly structured, especially in a research setting, remains to be studied in more detail.