User Profiling and Modeling in Extended Reality: A Scoping Review

Extended reality (XR) systems increasingly seek to adapt to individual users by leveraging behavioral and contextual data captured during immersive interaction. Central to this capability is the construction of user profiles or user models, which represent user characteristics such as identity, states, or preferences to guide system behavior. Despite growing interest in XR personalization, research on how user profiles and models are constructed and applied remains fragmented across disciplines and application domains. This paper presents a scoping review of 161 studies on user profiling and modeling in XR. Our analysis shows that most XR user profiles are derived from behavioral interaction traces, particularly head, hand, and eye movements, and are primarily used for user identification, state inference, and adaptive system responses. We further observe that modeling methods are closely tied to the computational tasks they support, with learning-based models dominating behavioral inference and rule-based approaches often used for descriptive attributes. The findings suggest that XR user profiling and modeling is evolving from isolated recognition tasks toward behavior-aware infrastructures that support adaptive immersive systems.