Quantitative gait analysis and prediction using artificial intelligence for patients with gait disorders

Area of Science:

• Biomedical Engineering
• Artificial Intelligence
• Clinical Biomechanics

Background:

• Quantitative Gait Analysis (QGA) provides objective gait performance metrics.
• Predicting gait quality progression is crucial for managing gait disorders.
• Current methods for gait progression prediction require improvement.

Purpose of the Study:

• To design an artificial intelligence (AI) system for predicting gait quality progression.
• To utilize kinematic data from QGA for gait prediction.
• To establish a novel AI-driven approach for gait analysis.

Main Methods:

• A gait database of 734 patients with gait disorders was used.
• Kinematic data from QGA was processed to generate Gait Profile Scores (GPS).
• Two AI approaches were developed: signal-based (raw gait cycles) and image-based (2D FFT gait cycles).

Main Results:

• Both signal-based and image-based AI approaches achieved an Area Under the Curve (AUC) above 0.72.
• The AI models demonstrated efficiency in predicting gait quality variations.
• This represents the first application of neural networks for gait prediction.

Conclusions:

• AI, particularly neural networks, can effectively predict gait quality progression using QGA data.
• The developed AI models offer a promising tool for monitoring and managing gait disorders.
• This research pioneers AI-driven gait prediction, enhancing clinical biomechanics applications.