Differences in Interoception and the Bodily Representation of Emotion in Autism and Typical Development

Background: Interoception is the detection of signals that relate to the physiological condition of the body, encompassing internal sensations such as those arising from the cardiac, respiratory and gastrointestinal systems, as well as external sensations such as temperature and pain (Ceuen et al., 2016; Craig, 2002; Sherrington, 1948). The perception of interoceptive signals has been considered of importance for understanding emotion since the seminal work of James and Lange (Lange & James, 1922), who suggested that emotion relies on the interpretation of physiological signals. It is now known that emotions are associated with discrete bodily maps of where they are experienced (Nummenmaaa et al., 2014), and these maps become more differentiated throughout child development (Hietanen et al., 2016). Difficulties in identifying and describing emotions are an integral part of autism (Griffin et al., 2015; Hill, Berthoz, & Frith, 2004).

Objectives: Here, we sought to investigate if there are differences in the bodily representation of emotion in children and adolescents with autism, compared to typical development and if so, do these differences relate to interoceptive processing.

Methods: Participants comprised 79 children and adolescents aged 6 to 19 years (mean=13.5, SD=3.0), of which 40 were reported to be typically developing, and 39 were reported to have an independent clinical diagnosis of ASD, verified by ADOS-2 administration. These groups were matched for chronological age and IQ. The Embody emotion coloring task (Nemmenmaa et al., 2014) was used to measure the bodily representation of seven emotions and interoceptive accuracy was gauged using the heartbeat tracking task (Schandry, 1981).

Results: We find significant differences in how emotions are embodied between typical development and autism (t(77)=2.425, p=0.018) and these group differences remain after controlling for age, gender and IQ. Further, interoceptive accuracy significantly explains more variance in the bodily representation of emotion, over and above group membership.

Conclusions: Autistic children and adolescents show significant differences in how they report the bodily sensations associated with emotions. These data suggest one possible cause of these differences is reduced interoceptive accuracy. Finally, the Embody emotion coloring task (Nemmenmaa et al., 2014) may offer a means of identifying difficulties in describing emotions that does not rely on the individual’s ability to recognize that they have a difficulty, as is the case with traditional questionnaire measures.