Understanding Sarcasm in a Speaker's Remark: An fMRI Study in Children and Adolescents with ASD

Background: Social communication and language deficits are core to autism spectrum disorders (ASD). In particular, individuals with ASD show consistent impairment in processing pragmatic language when attention to multiple social cues (e.g. facial expression, tone of voice) is often needed to infer a speaker’s communicative intent. Differing activation patterns have been reported in individuals with ASD during pragmatic language tasks (Knaus et al., 2007; Tesink et al., 2009). In these studies, individuals with ASD showed stronger activity in primary language and secondary language association areas as well as more bilateral activation profiles in these regions than neurotypical controls. We also observed hyperactivity in right inferior frontal gyrus as well as in bilateral temporal regions when ASD children relied upon event knowledge or prosodic cues to ascertain whether a speaker’s comment was sincere or sarcastic, likely reflecting more effortful processing (Wang et al., 2006).

Objectives: Here we sought to build upon our prior work examining how facial affect and prosodic cues aid in the ability to infer a speaker’s communicative intent (Wang et al., 2007). specifically, we examined, both at behavioral and neural levels, whether children and adolescents with ASD differ from typically developing (TD) children and adolescents in their processing of sincere versus sarcastic remarks. 

Methods: While undergoing fMRI, a sample of ASD children and adolescents and matched TD controls viewed cartoon drawings of characters in different everyday situations while listening to short scenarios. Each vignette ended with either a sincere or sarcastic comment made by one of the characters. Eighteen scenarios were presented, each lasting fifteen seconds (interspersed with eight blocks of rest). As in the prior study we conducted using the same stimuli (Wang et al., 2007), participants decided whether or not a speaker meant what they said. However, unlike our prior study where we manipulated task instructions to direct the subjects’ attention to different social cues, all participants were given neutral instructions to simply pay attention. Here we sought to directly compare activity associated with processing sarcastic versus sincere remarks both within and between groups.

Results: Children and adolescents with ASD and their matched TD controls performed the task of determining whether a speaker’s remark was sincere or sarcastic equally well, showing longer response times for scenarios ending with sarcastic remarks. Both groups also showed significant activity in canonical language areas, as well as visual cortices for both types of scenarios (sincere versus sarcastic endings). However, only the TD group showed significantly greater activity in Broca’s Area for sarcastic compared to sincere remarks. Also, while the TD group showed strong left lateralized responses in both frontal and temporal language areas, the ASD group showed a more bilateral pattern of activity within language networks.

Conclusions: The present findings contribute to a growing body of evidence suggesting decreased left lateralization in individuals with ASD during language processing tasks.  At least for high-functioning individuals with ASD, the increased activity in right hemisphere homologues of canonical language areas in the left hemisphere may reflect compensatory mechanisms supporting normative behavioral performance.