AI Summary & Key Takeaways
AI InsightThe findings reveal a thematic shift from biomechanical and imaging metrics toward minimally invasive interventions and digital health applications.
📌 3 Key Takeaways
Abstract
Background: This study aims to systematically map the thematic evolution of temporomandibular joint (TMJ) research from 2000 to 2025, identifying shifts in clinical and methodological priorities using transformer-based topic modeling. Material and Method: A total of 18,107 TMJ-related publications were retrieved from the Scopus database. Titles and abstracts were embedded using a biomedical transformer model. Dimensionality reduction was performed via UMAP, followed by clustering with HDBSCAN. Topics were labeled using class-based TF–IDF. Annual topic proportions were analyzed using ordinary least squares regression with false discovery rate correction. Results: Twenty-seven distinct topics were identified across six domains: biomechanics, clinical symptoms, imaging, pathology, surgery, and therapies. Eleven topics showed significant upward trends, notably arthrocentesis, ankylosis, and mouth-opening interventions. Six topics declined, including disc-focused MRI metrics and elastography. Ten topics remained stable. The findings reveal a thematic shift from biomechanical and imaging metrics toward minimally invasive interventions and digital health applications. Conclusion: Transformer-based topic modeling offers a panoramic view of TMJ research trends, highlighting a growing focus on patient-centered, conservative, and technology-assisted care. These insights provide a framework for future research prioritization and underscore the need for interdisciplinary approaches.