Diagnostic Accuracy of AI-Driven Wearable Sensors for Early Detection of Bruxism

Abstract

Background: Bruxism, characterized by repetitive jaw-muscle activity involving clenching or grinding, often goes undetected until significant dental or muscular complications arise. Conventional diagnostic approaches—including clinical examination and patient self-reports—frequently miss early or nocturnal episodes. With the emergence of artificial intelligence (AI) and wearable biosensors, continuous and objective monitoring may enhance early detection. Objective: To assess the diagnostic accuracy of AI-integrated wearable jaw-movement sensors compared with standardized clinical examination for early bruxism detection. Methods: A cross-sectional study of 120 adults aged 18–50 years was conducted in Lahore over four months. Participants underwent simultaneous bruxism evaluation using AI-driven wearable jaw-movement sensors and clinical assessment based on international diagnostic criteria. The supervised AI algorithm analyzed sensor-recorded jaw-movement patterns to distinguish bruxism events from normal motion. Diagnostic accuracy parameters including sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and ROC-AUC were computed using SPSS version 26. Results: The AI-enabled wearable device demonstrated a sensitivity of 96.6%, specificity of 89.3%, overall accuracy of 93.1%, PPV of 91.7%, and NPV of 95.2%. ROC analysis indicated excellent performance (AUC = 0.95). A strong correlation (r = 0.86, p < 0.001) was found between AI-detected and clinically diagnosed bruxism cases. Conclusion: AI-driven wearable jaw-movement sensors exhibit high diagnostic precision and strong agreement with clinical evaluation, supporting their utility as reliable, noninvasive tools for early bruxism detection and personalized dental care.