Quantifying Asymmetry Between Medial and Lateral Compartment Knee Loading Forces Using Acoustic Emissions

Area of Science:

• Biomechanics
• Biomedical Engineering
• Signal Processing

Background:

• Osteoarthritis commonly affects the knee, particularly the medial compartment, due to excessive loading.
• Understanding medial to lateral load distribution is crucial for managing knee osteoarthritis and preventing injuries.

Purpose of the Study:

• To develop and validate a novel method for quantifying knee joint load asymmetry in healthy individuals.
• To investigate the use of knee acoustical emissions and deep learning for assessing medial vs. lateral compartment loading.

Main Methods:

• Knee acoustical emissions were recorded during squat exercises simulating loading asymmetries.
• A deep neural network, specifically a convolutional autoencoder, was used to analyze time-frequency representations of the acoustic signals.
• Handcrafted audio features and automated features from the model were compared, along with multi-sensor fusion versus single-sensor approaches.

Main Results:

• A subject-independent classification model was developed with 83% accuracy in classifying medial and lateral joint load asymmetry.
• Wavelet coherence, a time-frequency correlation method, demonstrated the highest accuracy in the analysis.

Conclusions:

• Acoustic signals can effectively quantify the direction of medial to lateral load distribution in the knee.
• This method shows potential for wearable sensing technologies to monitor cartilage health and guide rehabilitation.