Theory of Mind in Autism Spectrum conditions: a task-based neuroimaging approach

Difficulties in everyday social interactions and communication are one of the hallmarks of autism spectrum disorder (ASD; American Psychiatric Association, 2013). One of the leading hypotheses about the causes of these social communicative difficulties is a deficit in theory of mind (ToM) or mentalizing (Baron-Cohen et al., 1985). ToM is defined as the ability to represent mental states of ourselves and others. Inferring other people’s beliefs and thoughts is crucial for having a conversation with someone as well as interacting with other people more broadly. Most of the tasks that claim to target ToM have shown that they elicit activity in a core network of brain regions, which comprises the bilateral posterior temporoparietal junction (TPJ) and the medial prefrontal cortex (mPFC; Schurz et al., 2014).

Many tasks have been developed to assess ToM abilities. Not all of these tasks, however, target ToM specifically. For example, performance on the Reading the Mind in the Eyes task (Baron-Cohen, Wheelwright, Hill, et al., 2001) can be attributed to alexithymia, and not ToM (Oakley et al., 2016). Another potential drawback of most tasks is that they rely on explicit mentalizing, which means that participants are asked to verbally describe or categorize the beliefs or thoughts of others (among others Frith-Happé animations; Castelli et al., 2002; False Belief; Fletcher et al., 1995). This contrasts with real-life use of ToM, for instance in communication, where interlocutors most often represent the beliefs of others without being asked to report about them. Interestingly, autistic people perform on the same level as neurotypicals on explicit mentalizing tasks, but make less use of implicit sources of information on the beliefs of another person, such as eye gaze (Baez & Ibanez, 2014; Schilbach et al., 2006; Senju et al., 2009). For these reasons, an implicit ToM task would correspond more closely to mentalizing in everyday situations.

In addition to the reliance on explicit ToM, the reliance of many ToM tasks on complex language processing is another aspect that prevents a thorough investigation of differences in autistic and non-autistic people in ToM ability. Examples of such tasks are the Faux Pas task (Stone et al., 1998) and the Strange Stories task (Happé, 1994). For instance, in a study using the Strange Stories task, language ability was a better predictor of task performance than ToM in siblings of autistic children (Shaked et al., 2006). It is important to take language processing into account, because a large meta-analysis has recently shown that autistic people are impaired compared to neurotypical people in their structural language ability, e.g. verbal comprehension, verbal fluency and verbal learning (Velikonja et al., 2019). Studies also point to differences between the two groups in some aspects of pragmatic language ability (Angeleri et al., 2016; Rundblad & Annaz, 2010). Therefore, studies on autism and ToM need to be tightly controlled for language ability or use tasks that are non-verbal. Interestingly, a large multicenter study involving 205 autistic people using a nonverbal ToM task found no differences in brain activation between autistic and non-autistic participants (Moessnang et al., 2020). Several other studies that tested for ToM ability while controlling for language ability did not find differences between autistic and non-autistic people (Capps et al., 1998; Norbury, 2005; Tager-Flusberg & Sullivan, 1994).

Because of these drawbacks in existing tasks, evidence of autistic people having decreased mentalizing ability is not conclusive (Gernsbacher & Yergeau, 2019).
What is needed is a non-verbal, naturalistic, implicit ToM task that avoids the pitfalls summarized above and is therefore well-suited for comparing mentalizing between autistic and neurotypical populations. Recently, a ToM localizer task has been developed that meets these criteria. In the task, participants watch an animated movie (‘Partly Cloudy’ by Pixar Animation Studios) where characters interact with each other non-verbally (Jacoby et al., 2016). The task activates the core mentalizing network in neurotypical individuals and it has already been employed as a ToM localizer (Paunov et al., 2019).

Yet, this task has only been used in neurotypical populations. Therefore, the objective of the proposed study is to compare ToM processing between autistic and neurotypical individuals with the localizer as conceived by Jacoby et al. (2016).

日常社交互动与沟通的困难是自闭症谱系障碍（ASD，美国精神病学协会，2013年）的显著特征之一。关于这些社交沟通困难原因的主要假设之一是心智理论（ToM）或心理化能力的不足（Baron-Cohen等人，1985年）。心智理论被定义为表征自身和他人的心理状态的能力。推断他人的信念和思想对于与他人进行对话以及更广泛地与他人互动至关重要。大多数声称针对ToM的任务表明，它们会激发大脑核心区域网络的活动，该网络包括双侧后颞顶交界（TPJ）和内侧前额叶皮层（mPFC；Schurz等人，2014年）。

已开发出许多任务来评估ToM能力。然而，并非所有这些任务都专门针对ToM。例如，在阅读眼神中的思维（Baron-Cohen，Wheelwright，Hill等人，2001年）任务中的表现可以归因于情感迟钝，而不是ToM（Oakley等人，2016年）。大多数任务潜在的另一个缺点是它们依赖于显式的心理化，这意味着参与者被要求口头描述或归类他人的信念或思想（例如，Frith-Happé动画；Castelli等人，2002年；错误信念；Fletcher等人，1995年）。这与现实生活中心智理论的使用形成对比，例如在沟通中，对话者通常在未要求报告的情况下代表他人的信念。有趣的是，自闭症患者在显式心理化任务中的表现与神经典型个体相同，但他们在利用关于他人信念的隐含信息源（如眼神）方面做得较少（Baez & Ibanez，2014年；Schilbach等人，2006年；Senju等人，2009年）。因此，一个隐含的ToM任务将更符合日常情境中的心理化。

除了对显式ToM的依赖外，许多ToM任务对复杂语言处理的依赖也是阻止对自闭症和非自闭症人群在ToM能力上的差异进行彻底调查的一个方面。这类任务的例子有 faux pas任务（Stone等人，1998年）和奇异故事任务（Happé，1994年）。例如，在一项使用奇异故事任务的研究中，语言能力比ToM更能预测自闭症儿童兄弟姐妹的任务表现（Shaked等人，2006年）。考虑到语言处理的重要性，因为最近的一项大型荟萃分析表明，与神经典型人群相比，自闭症患者在结构语言能力方面存在损害，例如口头理解、口头流畅性和口头学习（Velikonja等人，2019年）。研究还指出，两组在实用语言能力的某些方面存在差异（Angeleri等人，2016年；Rundblad & Annaz，2010年）。因此，关于自闭症和ToM的研究需要对语言能力进行严格控制或使用非言语任务。有趣的是，一项涉及205名自闭症患者的多中心研究表明，在使用非言语ToM任务时，自闭症和非自闭症参与者在脑激活方面没有差异（Moessnang等人，2020年）。其他一些在控制语言能力的同时测试ToM能力的研究也没有发现自闭症和非自闭症人群之间的差异（Capps等人，1998年；Norbury，2005年；Tager-Flusberg & Sullivan，1994年）。

由于现有任务的这些缺点，自闭症人群心智化能力降低的证据并不具有结论性（Gernsbacher & Yergeau，2019年）。所需要的是一种非言语、自然主义、隐含的ToM任务，它避免了上述陷阱，因此非常适合比较自闭症和非神经典型人群的心理化。最近，一种符合这些标准的ToM定位任务已被开发出来。在任务中，参与者观看一部动画电影（《部分多云》由皮克斯动画工作室制作），其中角色以非言语方式相互交流（Jacoby等人，2016年）。该任务激活了神经典型个体的核心心理化网络，并且已经被用作ToM定位器（Paunov等人，2019年）。

然而，这项任务仅用于神经典型人群。因此，所提出研究的目的是比较自闭症和非神经典型个体在Jacoby等人（2016年）构想定位器上的ToM处理。