Inertial sensor-based gait parameters reflect patient-reported fatigue in multiple sclerosis.

Area of Science:

• Neurology
• Biomedical Engineering
• Rehabilitation Science

Background:

• Multiple sclerosis (MS) is a neurological condition causing significant gait disorders and fatigue, a prevalent symptom affecting 80% of patients.
• Previous gait analysis in MS relied on stationary systems and short tests, limiting comprehensive assessment.
• Wearable inertial sensors offer a novel approach for analyzing gait during longer, continuous activities.

Purpose of the Study:

• To evaluate the association between fatigue and spatio-temporal gait parameters derived from wearable foot-worn sensors in MS patients.
• To develop a predictive model for estimating fatigue levels based on gait data.

Main Methods:

• Forty-nine MS patients (mean age 41.6 years) were equipped with bilateral inertial measurement units (IMUs).
• Spatio-temporal gait parameters were collected during a 6-minute walking test, with fatigue assessed using the Borg scale.
• Normalized gait parameters were transformed into principal components and used in a Random Forest regression model.

Main Results:

• Six principal components, explaining over 90% of data variance, were utilized in the predictive model.
• The Random Forest model achieved a mean absolute error of 1.38 points in predicting fatigue.
• Stride time, maximum toe clearance, heel strike angle, and stride length significantly contributed (67%) to fatigue prediction.

Conclusions:

• Fatigue levels in MS patients can be accurately predicted using spatio-temporal gait parameters from IMU-based systems.
• This technology can assist therapists in monitoring fatigue and evaluating treatment efficacy in clinical and home settings.
• IMU-based gait analysis offers a scalable solution for remote patient monitoring, reducing patient and therapist burden.