Utilizing Facial Muscle Electromyography Signals in Maximum Intercuspation for Malocclusion Classification

Accurate classification of dental malocclusion is essential for timely and effective orthodontic intervention. This study introduces a novel, non-invasive diagnostic framework based on surface electromyography (sEMG) signals acquired from facial muscles during maximum intercuspation. EMG data were collected from ۳۲ male subjects aged ۲۲–۲۷, clinically categorized into three groups: ۱۷ with Class I (normal occlusion), ۶ with Class II, and ۹ with Class III malocclusion. Signals from the masseter and temporalis muscles were recorded using an MP۱۵۰ acquisition system at a sampling rate of ۱۰۰۰ Hz. A comprehensive set of ۳۴ temporal and spectral features was initially extracted, including frequency shift, fatigue time, and fatigue coefficient. Principal Component Analysis (PCA), supported by Singular Value Decomposition (SVD), was employed for dimensionality reduction. Eight principal features exhibiting statistically significant interclass differences (p < ۰.۰۵) were selected through ANOVA and post-hoc tests. Among them, the temporalis muscle frequency shift (p = ۰.۰۰۰۱۱) and fatigue coefficient (p = ۰.۰۰۰۰۰۲) emerged as key discriminators. A neural network classifier trained on these optimized features achieved a classification accuracy of ۹۴.۵۹%. The proposed method demonstrates substantial potential as a cost-effective and radiation-free alternative to conventional imaging-based malocclusion diagnosis, particularly in early-stage orthodontic screening.

کلیدواژه ها:

EMG signals ، Malocclusion ، Maximum Intercuspation ، Dental Classification ، Feature Extraction