Technology-enabled telerehabilitation for Parkinson's disease: a scoping review of digital rehabilitation systems, delivery architectures, and implementation challenges.

INTRODUCTION: Digital technologies are increasingly integrated into neurorehabilitation programs for Parkinson's Disease (PD), enabling remote delivery of therapy, continuous monitoring of motor performance, and adaptive feedback during rehabilitation training. Telerehabilitation systems incorporating wearable sensors, virtual reality platforms, mobile applications, and artificial intelligence (AI) have expanded rapidly in recent years. However, the evidence base remains fragmented across heterogeneous technological configurations, clinical contexts, and delivery models, limiting a comprehensive understanding of how digital rehabilitation systems are implemented in PD care. METHODS: This scoping review maps the current literature on telerehabilitation for PD with a focus on the technological architectures, care settings, and delivery models used in digital rehabilitation programs. Peer-reviewed studies published between 2020 and 2025 were identified through searches in PubMed with reference to Scopus and Web of Science. Eligible studies investigated remote rehabilitation interventions for PD using digital technologies such as wearable sensors, mobile health applications, virtual reality systems, and AI-supported monitoring tools. Evidence was analyzed across three domains: (i) technological components and digital rehabilitation systems, (ii) rehabilitation setting, and (iii) delivery model. RESULTS: Fifty-three studies met the inclusion criteria. Most interventions were home-based and implemented multi-component digital architectures combining teleconferencing platforms, wearable sensors, and mobile applications. Wearable sensing technologies were used in nearly half of the studies to quantify gait, balance, or tremor, while video platforms and mobile applications supported remote supervision and exercise delivery. Virtual reality systems and serious games were used to enhance engagement and taskspecific training, whereas AI techniques were increasingly integrated to support movement detection, monitoring, and adaptive feedback. Despite generally high usability and acceptability, substantial heterogeneity was observed in outcome measures, terminology, and safety reporting. Few studies explicitly described care pathways, delivery architectures, or long-term clinical outcomes. DISSCUSSION: Telerehabilitation for Parkinson's disease is evolving toward integrated digital rehabilitation ecosystems combining wearable sensing, software platforms, and AI-enabled monitoring. Although feasibility and patient acceptance are consistently reported, current evidence remains limited by heterogeneous reporting standards and insufficient integration between technological systems and clinical workflows. Future research should focus on standardized outcome frameworks, scalable hybrid care models, and the development of interoperable, explainable digital rehabilitation systems capable of supporting long-term neurorehabilitation in real-world settings.