Mirroring emotional expressions of others in one’s own bodies can, as “embodied simulation”, facilitate their interpretation. Yet, individuals differ in their physiological responses to others’ emotions. For example, when encountering a smiling person, some people automatically mimic the smiles less than others. Additionally, feedback from one’s own body, such as the mimicked smile, may be integrated to a varying degree in the interpretation of others’ expressions.

In an online and a lab experiment (N = 99 and N = 100), we examined individual differences in the link between mimicry of facial emotional expressions and their interpretation in relation to autistic trait levels. Additionally, we explored whether the ability to accurately sense one’s internal signal (i.e., interoceptive accuracy) may be relevant within this context.

In line with previous studies, people with higher autistic trait levels had, compared to people with lower autistic trait levels, lower recognition rates for some emotional expressions. Those were, however, not explained by differences in facial mimicry or interoceptive accuracy. Thus, the strength of “embodied simulations” and the ability to correctly sense physiological changes may not determine how well emotions of others can be categorized. Nevertheless, they might still influence the intensity, which we perceive in others’ emotional expressions, as indicated by our exploratory analysis. Namely, when facial muscles were more activated, expressions were rated as more emotionally intense by people with higher versus lower trait interoceptive accuracy.

In contrast, the link between facial muscle activity and perceived emotional intensity was less pronounced in people with higher versus lower autistic trait levels. A potential explanation could be that, with higher autistic traits, changes in one’s own physiology may be less integrated in judging emotionality in others’ facial expressions. Alterations in interoceptive processing are known in autism and have been linked to difficulties in various domains. The results from our study suggest that altered interoceptive processing should also receive more attention in research aiming to understand social interaction difficulties in autism, and particularly when it comes to understanding others’ emotions. While different paths can be taken to make sense of a counterparts’ expression, the integration of “embodied simulations” may, at least sometimes, help.