Personalized Template-Based Step Detection From Inertial Measurement Units Signals in Multiple Sclerosis

Area of Science:

• Biomedical Engineering
• Neurology
• Gait Analysis

Background:

• Objective gait assessment is crucial for monitoring progressive multiple sclerosis (pMS).
• Inertial measurement units (IMUs) offer a practical solution for quantitative gait analysis in clinical settings.
• Existing automated step-detection algorithms struggle with the severely altered gait patterns observed in pMS.

Purpose of the Study:

• To develop and validate a novel, personalized template-based algorithm for automated step detection in pMS patients.
• To assess the algorithm's performance against a gold standard using synchronized IMUs and an electronic walkway.
• To evaluate the algorithm's ability to self-assess detection confidence and identify changes in gait over time.

Main Methods:

• A personalized template-based step-detection method (IITD and IGTD) was developed using IMU signals and gold-standard Initial/Final Contact times.
• The algorithm was tested on 22 individuals with pMS and 10 healthy subjects (HSs) across two 6-month follow-up visits.
• Performance metrics (precision, recall, F-measure) and a similarity index (SId) were computed to evaluate accuracy and self-assessment capabilities.

Main Results:

• The algorithm achieved high precision and recall (0.94-1.00 for IITD, 0.85-0.95 for IGTD) in pMS participants, outperforming previous methods.
• The similarity index (SId) correlated with detection accuracy and could reliably identify decreased performance (F-measure ≤ 0.95).
• The SId demonstrated the algorithm's ability to distinguish gait changes in individuals over a 6-month follow-up period.

Conclusions:

• The developed personalized step-detection algorithm demonstrates high performance for gait analysis in pMS, even with severely altered gait.
• The integrated SId provides a valuable self-evaluation metric, indicating the clinician's confidence in the automated step detection.
• The SId shows potential as a sensitive biomarker for tracking disease progression and changes in gait severity in pMS patients over time.