Developmental Changes in Neural Responses to an Interoception Task: Implications for Autism Symptoms.

Background: Given the developmental nature of autism spectrum disorder (ASD), it is important to understand how core features of ASD, like aberrant sensory reactivity, change across the lifespan. While most research has focused on external sensory reactivity, we also receive copious sensory information from within our own bodies, which is invaluable to monitoring body state and interpreting emotions. These internally-generated sensations, known as interoception, may be altered in ASD. Neural networks for interoception include insula, cingulate, and somatomotor cortex.
Objectives: We asked whether neural responses to an interoception task (heartbeat counting, compared to a visual counting control task) differed in ASD relative to a typically developing comparison (TC) group.
Methods: Participants included 46 individuals with ASD (age 8 to 54, mean= 20.84±10.11 years) and 55 individuals in a TC group (age 8 to 53, mean= 19.85±11.14 years). Data were processed using FSL, with age added as a covariate in the model given the wide age-range. Group maps investigated responses in the interoceptive task>visual task. A cluster threshold of z=2.3 and an FWE-corrected p-value=0.05 was used and then randomise was applied for non-parametric permutation testing (5000 iterations).
Results: The ASD group showed unilateral insula and secondary somatosensory (S2) responses, and bilateral response in the visual cortex and cerebellum, while the TC group showed bilateral insula and S2 responses. In this model, there was a positive age association with response in the bilateral insula and S2. Given this significant age finding, we then examined group results in children and adults separately. In children (age 8-17), both groups had significant clusters in the visual cortex, while the TC group also had a small cluster in S2. Additionally, there was a significant positive effect of age in the entire interoceptive network (insula, S2, caudate, putamen, thalamus). In adults, the ASD group had a much more widespread significant clusters in interoceptive regions compared to the TC group. There were no significant age findings in adults. Furthermore, percent signal change in the bilateral subdivisions of the insula all showed an inverted curvilinear relationship to age; neural responses in insula increased with age until they peaked in early adulthood and declined slightly in later adulthood. In adults, percent signal change in the insula was also associated with higher scores on the Social Responsiveness Scale.
Conclusions: This work highlights a potential developmental trajectory in interoception processing that could be relevant for ASD. Interestingly, once neural networks for interoception are refined in adulthood, there is still a wider distribution in ASD, suggesting more heterogenous interoceptive processing. Consistently greater recruitment of visual regions in the ASD relative to the TC group suggests the ASD group may rely more heavily on visual strategies to perform the interoception task. These data suggest that aberrant neural response to sensory cues in ASD is not limited to external cues, but also extends to interoception. Future work will need to understand how and when this developmental interoceptive trajectory may be altered in ASD and what impact it might have on behaviors in ASD.