Positively valenced outcomes increase outcome binding in human-robot interaction

Sense of agency (SoA) refers to the experience of being an agent who initiates actions and brings about their consequences. Prior work suggests that positive outcomes may produce a stronger sense of agency than negative outcomes and that predictability of outcome valence may modulate these differences. One approach used to investigate SoA is temporal binding. However, the measure can comprise two components, action binding and outcome binding, which appear to be dissociable and may depend on partly different cognitive mechanisms. In the present study, we examined whether positive outcomes strengthen binding relative to neutral outcomes and whether predictability affects the two components differently. Participants completed a Libet clock task in which a voluntary keypress caused the humanoid robot iCub to produce either a positive emotional vocalization or an acoustically similar neutral tone. iCub's facial expression indicated whether the valence of the upcoming sound was predictable or unpredictable. Positive vocalizations produced stronger outcome binding than neutral tones, whereas predictability did not reliably influence outcome binding or alter the effect of valence. Action binding was not significantly modulated by either valence or predictability. The two components were also not associated when measured across participants. These findings suggest that positive outcome valence can strengthen outcome binding relative to a neutral consequence, without a reliable change in action binding, and further support treating the two components of temporal binding separately.