Ontology-based knowledge representation in rehabilitation: a systematic literature review

Rehabilitation increasingly relies on integrating heterogeneous data sources, including clinical records, assessment protocols, and sensor-based monitoring systems. In this context, ontologies have emerged as a promising approach for representing domain knowledge, enabling semantic interoperability, and supporting intelligent decision-support systems. This paper presents a systematic literature review of ontologies in the rehabilitation domain, with the aim of analyzing their purposes, modelling approaches, and application contexts. Following the PRISMA review methodology, relevant studies were identified, screened, and analyzed to examine how ontologies are designed and used within rehabilitation research and practice. The results show that most rehabilitation ontologies focus on the conceptual modelling of limited portions of the rehabilitation domain. Many works employ ontologies to facilitate data integration, semantic interoperability, and reasoning-based decision support, particularly in digital rehabilitation platforms and assistive technologies. However, the review also highlights several limitations, including the fragmentation of ontology development efforts, limited reuse of existing models, non-adherence to FAIR principles, and the relatively static nature of many current models. The discussion identifies key challenges and opportunities for advancing ontology-based approaches in rehabilitation. In particular, future research should explore integrating data-driven technologies and machine learning (ML) methods with ontological modelling. Such integration could support automated ontology population, ontology evolution, and the discovery of new knowledge from large rehabilitation datasets. Moreover, ontologies may play an important role in explainable artificial intelligence by providing semantic structures that enhance the transparency and interpretability of AI-driven rehabilitation systems.