The mobile-paradigm as measure of infants' sense of agency? Insights from babybot simulations

The `sense of agency' refers to the experiential state that one's actions cause events in the world. Developing a sense of agency allows infants to learn from interacting with the social and physical world in ways that would not be possible otherwise [1]. To date, few empirical studies seem to target this phenomenon in infancy directly. Notable exceptions are the work by Rochat and colleagues (e.g. [2]-[4]) and Watanabe and Taga [5], [6]. In these studies, researchers report an increased movement frequency for movements that cause effects in the world. For instance, Watanabe and Taga [5], [6] used a mobile-paradigm [7] to investigate whether infants learned a causal action-effect relation. In this paradigm, one of the infant's limbs is connected to a mobile above their crib by a ribbon. When the infant moves this limb, the mobile moves. This study replicated the effect that is typically found, namely that infants increase the movement frequency of the connected limb relative to baseline. The increased movement of the connected limb has been interpreted as evidence for a sense of agency in young infants [6]. However, it is not clear that the increase in movement frequency necessarily means that infants have built a representation of the cause-effect relation. Here, we used computer simulations to assess whether or not data patterns found in mobile-paradigm studies can be explained by a mechanism that assumes no internally represented cause-effect relations at all, and hence can be argued to be insufficient for a sense of agency [8].