Repeatability of the Vibroarthrogram in the Temporomandibular Joints

Area of Science:

• Biomedical Engineering
• Biomechanics
• Medical Acoustics

Background:

• The repeatability of temporomandibular joint (TMJ) sound examination is currently inconclusive.
• Vibroarthrogram (VAG) analysis offers a potential method for objective TMJ sound assessment.

Purpose of the Study:

• To investigate the repeatability of specific features of the TMJ vibroarthrogram (VAG) using accelerometers.
• To compare the diagnostic performance of a k-nearest neighbors (KNN) classifier with a joint vibration analysis (JVA) decision tree for TMJ sound analysis.

Main Methods:

• VAG accelerometers were used to record TMJ sounds in 94 participants across two sessions.
• Participants performed 10 jaw open/close cycles per session, guided by a metronome.
• Intraclass correlation coefficient (ICC) was calculated for seven VAG signal features, and a KNN classifier was trained and evaluated.

Main Results:

• Reliability varied across VAG features, with 'integral below 300 Hz' showing excellent ICC, and 'peak amplitude' showing poor ICC.
• The KNN classifier achieved accuracy scores up to 0.81, significantly higher than the JVA decision tree (up to 0.60).

Conclusions:

• Specific VAG features demonstrate good to excellent repeatability for TMJ sound analysis.
• The KNN classifier shows promise for improving the diagnostic accuracy of TMJ sound-based assessments.
• This research could establish a new avenue for TMJ disorder diagnosis through advanced VAG feature analysis.