This book explores how integrated computation and communication can be achieved to support scalable and user-centric Extended Reality (XR) service provisions. Leveraging the digital twin technique as a key enabler, this book aims to introduce a novel data-centric AI framework to support XR from the perspective of communication and networking. Specifically, the authors present architectural designs and algorithmic solutions of data-centric AI that support the cross-layer and intelligent collection, processing, and analysis of XR user data, thereby enabling user-centric service provision. The book presents a digital twin-based framework that encompasses conceptual architecture, workflow, and operation functions to support diverse XR modules involving both communication and computation. In addition, the book explores the role of data-centric AI in XR resource provisioning, with a particular focus on how data-centric AI enhances both the quality and quantity of data available for AI model training and decision-making. Various learning paradigms, including supervised learning and reinforcement learning, are examined to demonstrate how AI enhances the efficiency and adaptability of resource management