Speaker
Description
According to the learning progress theory of curiosity, infants’ attention is guided by information in the environment. Yet, the role of social interaction and, in particular, the caregiver’s behavior in interaction with information provided by the environment remains unclear. The aim of this combined behavioral and computational modeling project was to investigate infants’ visual attention to information gain in an interactive context.
80 9-month-old infants (m=276 days; 39 girls) and their caregivers (68 mothers) took part in the experiment (preregistered at https://osf.io/gd6cb). The dyad sat at a table watching toys appear in one of the three boxes in front of them. Half of the caregivers were instructed not to interfere during the experiment while the other half was encouraged to interact with their infants as they would at home. For each trial, information gain (i.e., based on where the toy appears) was calculated as Kullback-Leibler divergence.
Based on previous research, we hypothesized that infants monitor and update the probability distribution of toy appearances in different locations. A computational Bayesian model was built to determine which factors play a role in infants’ looking behavior as a proxy for visual attention. It takes into account information gain, presence or absence of scaffolding from the caregiver's side, time passed from the start of the experiment and differing attention spans of infants, represented hierarchically in the model. All data are collected, and first results will be presented.
