Virtual Reality Simulation of Prosthesis embodiment: A Multimodal Investigation Examining Gaze Behaviour and Sense of agency
Year: 2026
Type: preprint
Abstract: Suffering a limb amputation has a major impact on physical and psychological health, with prosthetic devices being a fundamental tool to mitigate it. Prosthesis embodiment (integrating a prosthetic device into the bodily self) is a fundamental process for accepting these devices, with the sense of agency (the experience of controlling one’s actions and, through them, the course of external events) as a key contributing factor. Current upper-limb myoelectric prostheses still present high abandonment rates, partly due to motor-control errors that may reduce the sense of agency associated with prosthesis movements. One possible solution is sharing motor autonomy between the user and a grasping intention-detection algorithm, with prosthetic movements combining user actions and inferred intentions. In this study, we investigated whether shared autonomy, simulated through automated grasp assistance by an idealised intention-detection system in immersive virtual reality, influences explicit (self-reported) and implicit (indirectly assessed) sense of agency, prosthesis embodiment, task performance, and gaze behaviour. Fifty-seven able-bodied participants completed an object repositioning task over an obstacle with prosthetic avatars that varied in motor autonomy and appearance. Shared autonomy increased self-reported agency and reduced failed attempts. However, neither appearance nor motor autonomy affected embodiment scores. Although explicit and implicit measures of agency were significantly correlated, they responded differently to experimental manipulations. Shared autonomy increased explicit but not implicit agency, while longer action–outcome delays were associated with reduced explicit agency but stronger intentional binding. revealing dissociations across conditions and action-outcome delays. Additionally, explicit agency was associated with reduced visual attention towards the prosthesis and manipulated object. Together, these results indicate that shared autonomy approaches can enhance control experience and improve task performance while highlighting distinctions between explicit and implicit agency measures during prosthesis control.
Source: PsyArXiv (OSF Preprints)
Language: English
Subfield: Human-Computer Interaction
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7223005462