Automated feature assessment in instrumented gait analysis

Area of Science:

• Biomechanics
• Medical Technology
• Rehabilitation Engineering

Background:

• Gait analysis is crucial for assessing motor function and treatment efficacy.
• Current methods often rely on subjective clinical expertise and manual parameter selection.
• Objective and automated gait assessment is needed to improve diagnostic accuracy and treatment monitoring.

Purpose of the Study:

• To present a modular framework for automated gait pattern assessment.
• To validate the framework using Botulinum Toxin A (BTX-A) treatment data for spastic equinus foot in cerebral palsy (CP) patients.
• To introduce a norm distance measure for quantifying gait deviations.

Main Methods:

• Developed a modular framework for automated gait data selection, parameter calculation, and evaluation.
• Applied mathematical methods to a dataset of 3670 gait parameters.
• Classified 42 diplegic CP patients against a reference group and analyzed pre- and post-treatment data.

Main Results:

• Identified highly relevant gait parameters that align with previous subjective findings.
• Demonstrated the framework's ability to differentiate between patient groups and assess treatment effects.
• Introduced a norm distance measure for quantifying deviations from normal gait patterns.

Conclusions:

• The proposed automated framework offers a robust and objective method for gait analysis.
• The norm distance measure provides a valuable tool for evaluating gait deviations and treatment outcomes.
• This approach has broad applicability beyond the specific clinical case studied.