Functional Data Analyses of Gait Data Measured Using In-Shoe Sensors

Area of Science:

• Biomechanics
• Locomotion analysis
• Medical sensor technology

Background:

• Ground reaction force (GRF) measurement is crucial for gait analysis, biomechanics, and pathology detection.
• Gold-standard GRF measurement requires expensive, specialized in-lab equipment.
• In-shoe sensors present a more affordable alternative for continuous GRF monitoring.

Purpose of the Study:

• To explore the properties of continuous in-shoe sensor recordings using functional data analysis.
• To assess the potential of in-shoe sensors as a surrogate for in-lab GRF measurements.
• To analyze phase and amplitude variabilities in in-shoe sensor data.

Main Methods:

• Functional data analysis approach applied to in-shoe sensor data.
• Curve registration technique used to separate phase and amplitude variabilities.
• Analysis of phase shift correlations across multiple in-shoe sensors.
• Comparison of registered in-shoe sensor data with gold-standard GRF measurements.

Main Results:

• In-shoe sensor data exhibit both phase and amplitude variabilities.
• Curve registration successfully separated these variability sources.
• Phase shifts across sensors within a stance show correlated patterns.
• Exploration of associations between in-shoe sensor data and GRF measurements.

Conclusions:

• In-shoe sensors demonstrate potential as a cost-effective surrogate for in-lab GRF measurement.
• Functional data analysis provides a robust framework for analyzing in-shoe sensor data.
• Understanding phase and amplitude variability is key to validating in-shoe sensor technology for gait studies.