Interoceptive Sensibility Predicts Anxiety in Children on the Autism Spectrum

Background: Anxiety is a major co-occurring feature of autism spectrum disorders (ASD). Interoception refers to the sensation of the internal state of the body including afferent information from the viscera, respiratory and genitourinary systems (Cameron, 2001). The incidence of anxiety symptoms in this population has been associated with problems with interoceptive processing (Garfinkel et al., 2016). Despite the developmental nature of autism, and findings of declining interoceptive accuracy with age (Khalsa et al., 2009), this relationship has only previously been investigated in adults.

Objectives: Here, we assessed the relative contribution of interoceptive dimensions to anxiety symptoms in autistic and typical children. Mirroring Garfinkel et al.’s previous research our hypotheses were that 1) a diagnosis of ASD will be associated with impaired interoceptive accuracy (reduced performance on an objective test of interoception); 2) individuals with ASD will display enhanced interoceptive sensibility (their subjective belief about their interoceptive aptitude); 3) the discrepancy between interoceptive accuracy and interoceptive sensibility, operationalised as interoceptive trait prediction error (ITPE), will be predictive of anxiety symptoms. In addition to this, extrapolating from Khalsa et al., (2009) we expect 4) interoceptive accuracy to decline with age, peaking in late childhood or early adolescence.

Methods:  We assessed the interoceptive accuracy (using heartbeat detection tasks), interoceptive sensibility (subjective sensitivity to internal sensations on a self-report questionnaire), and state and trait anxiety of 20 children diagnosed with an ASD and 30 age-matched typically developing children, aged between 6 and 18 years.

Results: We found higher levels of anxiety in autistic children and adolescents than typically developing children on parent-report, but not self-report measures, and poorer interoceptive accuracy in autistic than typical children, replicating previous findings (Garfinkel et al., 2016). Intriguingly, we identified a significant negative relationship between age and interoceptive accuracy, whereby younger children had better interoceptive awareness in comparison with older adolescents, extending previous findings of a negative relationship (Khalsa, Rudrauf, & Tranel, 2009) into a younger sample. In contrast to findings in adults, we observed a negative relationship between ITPE and self-report anxiety, across the entire sample. In typical and autistic individuals, the best predictor of anxiety symptoms was interoceptive sensibility, the subjective report of interoceptive symptoms. The more interoceptive symptoms an individual reported, the higher their anxiety score.

Conclusions:  The findings presented here support the notion that emotional processing difficulties in autism arise in part through a compromised interoceptive channel (Quattrocki & Friston, 2014). Reduced accuracy in the detection and prediction of interoceptive signals could disrupt their use in forming emotional inferences, causing uncertainty and anxiety about the external and internal world. Finally, these data underscore the need for research examining the characteristics of anxiety and interoception in autism within a developmental framework.