Explicit vs. applied theory of mind competence: A comparison of typically developing males, males with ASD, and males with ADHD

Highlights

• Children with attention deficit hyperactivity disorder (ADHD) behave as though they have Theory of Mind (ToM) deficits even though they can pass laboratory-type ToM tasks.

• By contrast, children with autism spectrum disorder (ASD) show ToM deficits on tasks designed to assess both conceptual and applied ToM competence.

• Children with ASD and children with ADHD do not share a common underlying Theory of Mind (ToM) deficit despite the fact that they often share many of the same behavioral problems.

• Social skills curricula that aim to explicitly teach social-cognitive principles may be appropriate for children with ASD but are highly questionable for children with ADHD.

Abstract

Using laboratory-type Theory of Mind (ToM) tasks (our measure of ‘explicit’ ToM competence) and a more ecologically-valid measure of ToM (our measure of ‘applied’ ToM competence), we found that for composite scores, typically developing (TD) males performed near ceiling levels on both indices and age-matched males with autism spectrum disorder (ASD) performed near floor levels on both indices. The scores for age-matched males with attention-deficit hyperactivity disorder (ADHD) showed a different pattern such that the ADHD group had high scores on the explicit measure and low scores on the applied measure. Subscale scores (early, basic, advanced ToM) for the two indices also revealed that (1) despite variable complexity, explicit ToM almost always distinguished the ASD group from the other two groups but never distinguished the ADHD and TD groups and (2) level of complexity was critical for distinguishing groups with regard to applied ToM. We suggest that although children with ADHD can calculate the content of traditional laboratory ToM tasks, this explicit ToM competence fails to be applied and expressed in real world demonstrations of ToM (especially when advanced ToM skills are assessed). By contrast, the ToM difficulties of children with ASD seem to be attributable to a deeper metarepresentational deficit. Our results have implications for practice and extend current models of social cognition in developmental disabilities by isolating variable aspects of competence that predict specific and testable models for future research.

Introduction

Children with developmental disabilities evidence social and behavioral dysfunction which, left untreated, can lead to peer rejection, social isolation, and psychological maladjustment (Hoza, 2007, Hutchins and Prelock, 2013). Attention-deficit hyperactivity disorder (ADHD) and autism spectrum disorder (ASD) are common and debilitating neurodevelopmental disorders with a chronic course. Although they are distinct diagnostic categories, these psychiatric disorders are frequently comorbid with a little over 28% of persons with ASD also meeting criteria for ADHD (Simonoff et al., 2008). An overlap of symptoms and impairments between the two disorders is also well-documented (Memari, Ziaee, Mifrazeli, & Kordi, 2012). For example, compared to their typically developing (TD) peers, children with ADHD and children with ASD demonstrate more inattention and over-activity (Dickerson, Calhoun, Mayes, & Molitoris, 2012), mood and behavior problems (Mayes & Calhoun, 2007), irritability, anger (Brereton, Tonge, & Einfled, 2006), anxiety, and depression (Matson and Cervantes, 2014, Stratis and Lecavalier, 2013), and have higher rates of executive dysfunction (Lawson, Papadakis, Higginson, Barnett, & Willis, 2014) and language delay (Miniscalco, Hagberg, Kadesjo, Westerlund, & Gillberg, 2007). Although there is agreement that children with ASD and children with ADHD have many similar comorbid problems, consensus is lacking as to whether, or to what extent, certain ‘Theory of Mind’ deficits typical of ASD are common in ADHD.

Theory of Mind (ToM) has been defined as “a body of conceptual knowledge that underlies access to both one’s own and others’ mental states” (Sodian, Hulsken, & Thoermer, 2003, p. 778). Originally used in a narrow sense (i.e., merely to describe performance on the false belief task), ToM has come to be construed as a broad, complex, and multifaceted construct. To illustrate, ToM includes (but is not limited to) the ability to engage in joint attention and pretense, the understanding of play pragmatics, empathy, intentionality, and the capacity to distinguish appearance from reality and the mental from the physical world. It involves affect recognition, first- and second-order thinking, visual perspective-taking, and the understanding that seeing leads to knowing. One with a mature ToM also comprehends the mind as an active interpreter and can make inferences and reason about the causes and consequences of one's own and others' thoughts and feelings. Indeed, ToM has been described as a construct that “refuses to be corralled” (Astington & Baird, 2005, p. 4) and it is often used interchangeably with terms like ‘social cognition’, ‘mind-reading’, ‘mentalizing’, and ‘perspective-taking’ (Hutchins, Prelock, & Bonazinga, 2012).

The widening scope of the term ToM is primarily attributable to the breadth and pervasiveness of the social-cognitive impairments that have been documented in ASD. Owing to nearly three decades of robust empirical findings revealing deficits in ToM in ASD (for a review see Wellman & Peterson, 2013), energy and enthusiasm for the notion of ToM as a core deficit has endured. Specific empirical evidence for this ‘Theory of Mind Hypothesis’ of autism (Baron-Cohen, 1995) first came from two landmark studies (Baron-Cohen, Leslie, & Frith, 1985; Wimmer & Perner, 1983) demonstrating that children with ASD had significant difficulties understanding that others could have a belief that contradicted reality (i.e., a false belief). The fact that individuals with ASD performed poorly on a variety of ToM tasks but succeeded on carefully designed control tasks suggested that ToM impairments were not the result of more general cognitive dysfunction. Although differences in the measurement of ToM are known to influence performance (e.g., van Buijsen, Hendriks, Ketelaars, & Verhoeven, 2011), research shows that persons with ASD generally underperform TD individuals on assessment of an extensive range of mental states (e.g., Sterck & Begeer, 2010).

Although it has some serious limitations (see Hutchins & Prelock, 2015), the ToM hypothesis makes intuitive sense and parsimoniously explains the social communication and social interaction deficits that are the defining features of ASD. These deficits include a limited range of communicative functions, less diverse and elaborate functional play, difficulty modulating the use of prosody and gesture to aid communication, lack of social responsiveness, and failure to establish or maintain eye contact, to name a few. Whatever specific explanation is provided for the social communication and social interaction difficulties in ASD, it seems clear to many researchers and practitioners that they are a result of some underlying cognitive process that has come to be broadly referred to as ToM.

Social cognition deficits are not strictly limited to people with ASD and there are a variety of circumstances and disorders that can hinder ToM development (Dyck, Ferguson, & Shochet, 2001). Some ADHD researchers have pointed to impairment in specific social cognitive domains (e.g., empathy; Demurie, De Coral, & Roeyers, 2011; emotion recognition, prosody perception; Uekermann et al., 2010) as a basis for social dysfunction in ADHD. Others, explaining social deficits more generally (i.e., not specific to ADHD), have adopted a procedural model (i.e., Crick & Dodge, 1994) postulating disruption in a series of social-information processing steps that children go through when faced with social situations (i.e., encoding of cues, interpretation of cues, clarification of goal states, accessing/generating responses, response decision, and behavioral enactment). Indeed, compared to TD children, children with (or at risk for) ADHD tend to encode social information less accurately and, in turn, have difficulty integrating social cues and formulating appropriate responses (Coy, Speltz, DeKlyen, & Jones, 2001; Milch-Reich, Campbell, Pelham, Connelly, & Geva, 1999). Still other researchers have implicated biased or distorted social attribution as a causal factor underlying the social dysfunction in ADHD. For example, Andrade et al. (2012) reported that children with ADHD not only detected fewer social cues but also attributed more negative and less positive intent to peers and generated fewer positive responses compared to TD children.

Of course, there are multiple pathways to social difficulties and research employing more traditional ToM tasks has highlighted the potential of executive dysfunction (not ToM impairment) as a root cause of the social deficits of ADHD. While three studies (Charman, Carroll, & Struge, 2001; Happé & Frith, 1996; Perner, Kain, & Barchfeld, 2002) have reported no impairment on advanced ToM tests in school-aged children with (or at risk for) ADHD, findings are mixed and at least two additional studies do report impaired performance on advanced ToM tasks (Buitelaar, van der Wees, Swaab-Barneveld, & van der Gaag, 1999; Papadopoulos, Panayiotou, Spanoudis, & Natsopoulos, 2005). In an attempt to reconcile the variable results, Sodian et al. (2003) noted that previous studies were “not motivated by analysis of possible interactions between conceptual content and inhibitory demands of the tasks” (p. 780). In their study, TD and ADHD children were no different on standard second-order tests but children with ADHD were delayed on a test of advanced understanding that required online representation of another's mental state. Sodian et al. (2003) concluded that inhibitory control is important for higher-order ToM but that the effects are more important for online mind-reading than for ToM reasoning in general.

Children with ADHD often experience social difficulties that are similar to those of children with ASD (Happé & Frith, 1996). Both groups may exhibit significant problems with peer functioning, demonstrate fewer neutral nonverbal behaviors (such as attending to other children or adults; Pelham & Bender, 1982) and show deficits in emotion understanding and social-cognitive aspects of interpersonal problem solving (Cook, Greenberg, & Kusche, 1994; Dodge & Price, 1994). Researchers have recently begun to consider whether, which, and to what degree the ToM impairments typical of ASD are common to ADHD. The research in this area, however, is sparse and mixed. To our knowledge, only eight studies have specifically compared the social cognition abilities of children with ASD and children with ADHD. A review of those studies is presented in Table 1.

Several findings are relevant in the studies reviewed as they highlight the likely competence deficits in individuals with ASD vs. the more typical performance deficits in individuals with ADHD. First, in six of the eight studies, individuals with ASD performed more poorly on ToM tasks than either individuals with ADHD or those who were TD (Baribeau et al., 2015, Buitelaar et al., 1999; Demopoulos, Hopkins, & Davis, 2013; Demurie et al., 2011, Gonzalez-Gadea et al., 2013; Yang, Zhou, Yao, Su, & McWhinnie, 2009) confirming what would be expected in our understanding of a core ToM deficit in persons with ASD. In the remaining two studies, however, some unique findings were reported. Downs and Smith (2004) found that children with ADHD combined with oppositional defiant disorder (ODD) exhibited more social-emotional and behavioral difficulties than TD children or children with ASD suggesting a specific ToM deficit in emotion understanding for children with ADHD-ODD. Buhler, Bachman, Goyert, Heinzel-Gutenbrunner, and Kamp-Becker (2011) found greater inhibitory control differences for children with ADHD than those with ASD, and ToM differences between the two groups were characteristic of younger vs. older children. Their findings suggested a more foundational and conceptual ToM deficit for children with ASD and a developing ToM deficit for older children with ADHD. In addition, ToM complexity appeared to be an influencing factor for ToM performance that differentiated persons with ASD from those with ADHD. Individuals with ASD often performed more poorly on first order ToM tasks (Buitelaar et al., 1999) and had more severe deficits (Demopoulos et al., 2013, Gonzalez-Gadea et al., 2013) than either those with ADHD or typical development, although the general patterns of ToM deficits in those with ASD and ADHD were similar. Finally, inhibition appeared to be tied to executive function and ToM performance. Individuals with ADHD appear to have greater challenges in inhibitory control and emotion recognition, which may influence ToM performance where individuals with ASD may have a more foundational ToM deficit (Buhler et al., 2011, Gonzalez-Gadea et al., 2013, Yang et al., 2009).

First, it is noteworthy that while much education and psychological research speaks to explicit and applied competencies, it typically does so by employing different terminology to highlight slightly different contrasts, which in turn, encourages somewhat different discussions. In lieu of ‘explicit’ might be reference to ‘basic’ or ‘conceptual knowledge’ (Hiebert & Lefevre, 1986), the ‘cognitive level’ of functioning (Morton, 2004), or the notion of ‘competence’ more broadly (Chomsky, 1959) whereas ‘skill’, ‘procedural knowledge’, the ‘behavioral level’ of function, or the notion of ‘performance’ can act as stand-ins for ‘applied’. The distinction between explicit and applied competence that we emphasize here is not synonymous with any of these distinctions, but it draws upon all of them. We employ the contrast between explicit and applied competence because it is precise, relatively circumscribed, and thus, has utility for describing two types of functioning relevant to our understanding of social communication impairments in developmental disorders. The mixing of terms ‘explicit’ and ‘applied’ (instead of ‘explicit vs. tacit’ or ‘basic vs. applied’) is also deliberate as the more traditional pairings span disciplines and carry theoretical baggage that might confuse the issues and frustrate attempts to link each competency to a reasonable operational definition. The explicit/applied distinction, however, is not atheoretical and its interest lies in the potential for these two competencies to converge or dissociate as described below. Finally, the distinction is practical as questions about how children fail to acquire ToM (and how they should be taught) often turn on speculations about which type of competency has the most developmental significance, how support in one might buttress or otherwise affect the other, or what might be an optimal balance between the two.

We define explicit ToM competence as ToM knowledge that is conceptual, operational, and logical. It is also calculable within a limited experimental context and, as such, can be readily accessed and verbalized. By contrast, we characterize applied ToM competence as the ability to deploy ToM knowledge to successfully address ToM dilemmas as they are presented in real-world samples of behavior. Applied ToM competence is revealed in day-to-day performance and it is ostensibly affected by a variety of endogenous (e.g., executive functioning, motivation, sensitivity) and exogenous (e.g., physical setting) factors.

As noted above, explicit and applied ToM competencies may dissociate: being able to compute the logic of mental states is no guarantee that one can or will apply the principles more broadly. To illustrate, Bowler (1992) reported that although adults with Asperger syndrome could pass explicit ToM tasks, their responses to justification questions lacked reference to mental states. Bowler suggested that persons with Asperger syndrome may use a logical rather than mentalistic process to compute the correct solution. He reasoned that this might explain good performance on ToM tasks in persons with persistent and significant social impairments. More recently, Senju, Southgate, White, and Frith (2009) reconfirmed this notion by demonstrating that although individuals with Asperger syndrome showed an explicit understanding of desires and beliefs, they nevertheless failed tasks designed to assess the spontaneous attribution of mental states.

Just as children may be conceptually competent but fail to perform, they may also succeed in practice without having the knowledge that explicit ToM tasks are intended to assess. This potential is rarely entertained in the developmental disabilities literature (primarily because performance is viewed as a downstream consequence of conceptual knowledge) and it seems counterintuitive: how can someone be competent in an applied way (e.g., doing well on a mathematics exam) without mastering the relevant underlying concepts (e.g., not understanding mathematical operations) ? Yet, ToM is only sometimes necessary but never sufficient for explaining social competence (Astington, 2003). “Rather than deploying a theory of mind constantly, people rely on social scripts, behavioral cues, and narratives to guide their actions in social situations” (Hughes & Devine, 2015, p. 151). Indeed, Senju (2012) showed that whereas high functioning adults with ASD could pass explicit tests of false belief while failing tests of spontaneous false belief, TD infants showed the opposite pattern and passed spontaneous false belief tests well before they could articulate their understanding. Of course, the latter finding is not surprising in light of Vygotskian theories that emphasize children’s interactions with the world as shapers of children’s cognitive development.

In designing interventions to support appropriate social interaction and communication, researchers and practitioners are increasingly invoking ToM in an effort to understand which (in) competencies are the putative causes of social dysfunction. One problem is that ToM competencies can dissociate and a window onto one cannot be assumed to be a window onto the other. Another problem is that conceptual complexity appears to moderate performance on ToM tasks yet few studies have systematically isolated its effects. The purpose of this study was to examine, in depth, the ToM profiles of children with ASD and children with ADHD, contrast their patterns of competency, and compare them to TD children. Our examinations were exploratory given the previous mixed findings on ToM competence in ADHD and the fact that this was the first study to specifically contrast indices of explicit and applied competencies in ASD, ADHD, and TD groups across various levels of complexity.