From association to intervention: Semantic trajectories and knowledge frontiers in epilepsy-gut microbiota research revealed by bibliometrics and NLP

Background: Research on the epilepsy-gut microbiota axis is accelerating, yet its conceptual evolution and translational signals remain difficult to track across disciplines. Here, we propose a semantic-bibliometric fusion framework that combines network-based science mapping with phrase- and topic-level natural language processing (NLP) to resolve both structural and linguistic dynamics in the field. Methods: Publications indexed in the Web of Science Core Collection from January 1997 to November 17, 2025 were retrieved and analyzed using CiteSpace, VOSviewer, Pajek, and Scimago Graphica, complemented by an NLP pipeline for topic modeling and temporal phrase tracking. In total, 516 records were included. Results: Annual output rose sharply after 2021 and peaked in 2024, with China and the United States leading contributions. Beyond conventional hotspots, semantic analyzes revealed a field-wide shift from early infection/safety-oriented discourse to mechanism- and intervention-driven themes dominated by the microbiota-gut-brain axis. Conclusion: The most prominent and expanding intervention clusters involved the ketogenic diet, short-chain fatty acids, probiotics, and fecal microbiota transplantation, while "Mendelian randomization" emerged as a nascent frontier signaling movement from correlation to causal inference. By integrating knowledge-network topology with discourse dynamics, this study delivers an up-to-date, multi-resolution map of epilepsy-microbiome research and highlights actionable directions for future work, including causal validation, multi-omics stratification, and trial-ready microbiome-informed interventions.