Me in action: Distinct processing levels shape the relationship between sense of agency and sense of ownership

Competing neurocognitive models offer different accounts of how the sense of agency (the experience of controlling one's actions) and the sense of ownership (the experience that one's body belongs to oneself) relate within bodily self-awareness. The additive model proposes that agency entails ownership, the independence model treats the two experiences as functionally distinct, and the interactive model suggests reciprocal influences between them. However, empirical tests comparing these predictions within a single coherent experimental paradigm remain limited. Here, we developed an immersive virtual reality paradigm to orthogonally manipulate agency- and ownership-related cues within a goal-directed action task. Fifty-six participants pressed a (virtual) button to switch on a lightbulb. Agency was manipulated by comparing voluntary self-generated and externally induced passive movements and action-outcome delays, whereas ownership was manipulated through the anatomical plausibility of the avatar's hands. Direct and indirect agency and ownership indices were collected. Direct explicit judgements revealed both domain-specific and reciprocal influences: agency was modulated by action-related cues and by anatomical plausibility, whereas ownership was modulated by anatomical plausibility and by action-related cues, consistent with an interactive account. Indirect measures showed a distinct pattern: intentional binding was influenced by action-outcome contingency and was not affected by ownership manipulations, whereas proprioceptive drift varied as a function of anatomical plausibility. We propose a hierarchical two-stage account that integrates previous cognitive models across distinct processing levels, in which agency and ownership arise from partly independent sensorimotor processes at an implicit level but dynamically interact at the level of metacognitive, explicit self-attribution.