What eyes tell about beliefs? Pupil dilation reflects on-line belief updating in adults and 5-year-olds

• In Event: Poster Session 11
  In Poster Session: PS 11 Topic - Social Cognition

Abstract

Social interactions heavily rely on our ability to follow what our partners think and know about the world. A critical question regarding our ability to keep track of others’ mental states is how the process of belief monitoring and updating unfolds over time: do humans track others’ beliefs in a prospective manner or do they do so only post hoc (or ‘on demand’), when the context requires to predict or explain others’ behavior? To date, this issue has been overlooked in the literature, with most of the studies using end-of-scenario measurements to assess mind reading abilities. Here, to shed light on the temporal dynamics of mind reading, we employed eye tracking to assess online belief update processes by measuring correlates of mental effort in a modified false-belief task relying on referential communication (Kiraly et al., 2018). More specifically, we used pupil dilation to investigate whether the changes in pupil size reflect real-time updating of another agent’s belief. We tested 5-year-old children and adults.

Across three eye-tracking experiments (N = 36, per group), we introduced a situation where the observer discovers post hoc that the sunglasses worn by the protagonist when the locations of two objects were swapped were opaque (Figure 1). Therefore, if participants initially assumed the sunglasses to be transparent, and, thus, attributed a true belief about the location of the objects to the protagonist, now they could to update her true belief into a false belief. We observed an increase in pupil dilation time-locked to when participants received the information about the glasses in the belief-update condition compared to the no-update condition. Additionally, participants’ behavioral responses at the end of the scenario confirmed that participants had updated their representation of the protagonist’s belief. Importantly, these effects were not observed in a control experiment, which did not include the belief-induction phase.

Our results suggest that, from early on we update others’ beliefs online, just about a second after new information relevant to others’ mental states comes in (0.5 s for adults, 1.5 s for children). These belief-updating computations are indexed by an increase in pupil dilation, when an attributed belief has to be revised, both in adults and children. In addition, our findings reveal that the time-course of spontaneous online belief updating can be successfully assessed using eye-tracking technology. Such methodological advancements provide means for future research to better understand how belief tracking emerges in ontogeny and phylogeny.

Authors

• Barbara Pomiechowska, Central European University
  Presenting Author

• Erno Teglas, Central European University
  Non-Presenting Author

• Ágnes M. Kovács, Central European University
  Non-Presenting Author

• Ildiko Kiraly, ELTE Eotvos Lorand University Budapest
  Non-Presenting Author