# Infant Vocal Behavior During Contingent Vocal Imitation and Its Interruption as a Window Into the Emerging Sense of Agency

## TL;DR

This study explores how infants learn from vocal imitation by caregivers and how this might help them develop a sense of control over their actions.

## Contribution

A novel paradigm using an artificial agent to study vocal contingency learning in infants without a human interaction partner.

## Key findings

- Infants rapidly learned the contingency between their vocalizations and an imitative response.
- There was a descriptive trend of increased vocalizations when the contingency was discontinued, but it was not statistically significant.
- The paradigm proved viable for studying vocal contingency learning without human interaction.

## Abstract

Infants' emerging sense of agency is thought to be supported by caregivers' contingent responsiveness. However, it remains unclear which types of responses are most relevant to this process. Here, we examined the role of contingent vocal imitation, defined as the prompt repetition of an infant's vocalization by an interaction partner. To tease apart the contribution of contingent vocal imitation from other elements of social interactions, we developed a novel vocal contingency paradigm. First, we investigated whether 5‐ to 6‐month‐old infants could rapidly learn the contingency between their own vocalizing and a novel imitative response. Then, we examined whether infants tested this newly learned contingency when it was suddenly discontinued. Novel audio‐visual imitative responses were delivered and manipulated by an artificial agent. Infants' vocalizations were recorded while they experienced the novel contingency (connect phase) and its discontinuation (disconnect phase). Time‐course analyses indicated a significant linear increase in vocalization frequency over time in the connect phase, supporting the hypothesis that contingent vocal imitation enables rapid vocal contingency learning. Descriptively, data suggested a quadratic trend consistent with a vocal extinction burst during the disconnect phase. However, this trend did not reach statistical significance. Therefore, there was only partial support for the role of contingent vocal imitation in the emerging sense of agency (i.e., young infants quickly learned this contingency, but there was no evidence that they tested it upon discontinuation). Overall, our paradigm provided proof of concept that vocal contingency learning can be studied in the absence of a human interaction partner.