A New Method for Automatic Gait Event Detection: Analysis of Inertial Measurement Units in Healthy Subjects and Patients with Moderate to Severe Impairment

Cyril Voisard, Nicolas de L’Escalopier, Damien Ricard, Laurent Oudre. Journal of NeuroEngineering and Rehabilitation (2024) 21:104 https://doi.org/10.1186/s12984-024-01405-x

Background of the Study

Gait analysis is an important tool in medicine to assess the health status and disease progression in various patients. Inertial Measurement Units (IMUs) have been widely developed in clinical gait analysis due to their compact size, low cost, and ease of integration. However, although existing automatic Gait Event (GE) detection methods have achieved high efficiency in healthy subjects, challenges remain in patients with severely impaired gait.

Objective of the Study

This study aims to propose an improved GE detection method to extract data from IMU recordings, applicable to patients with severely impaired gait.

Subjects and Methods

The study recorded 10-meter gait IMU signals of 13 healthy subjects, 29 patients with multiple sclerosis, and 21 patients with post-stroke equinovarus foot. The research method first utilizes autocorrelation and pattern detection techniques to identify reference gait patterns and then applies multi-parameter dynamic time warping (mDTW) to label the pattern, thereby detecting all GEs in the signal.

Results

The GE detection F1 score for healthy subjects reached 100%, with a median absolute error of 8ms. For patients with multiple sclerosis and equinovarus foot, the F1 scores were 99.4% and 96.3%, respectively, with median absolute errors of 18ms and 26ms, respectively. The results indicate that the method in this study is consistent with existing techniques in healthy subjects and shows good accuracy in pathological patients.

Conclusion of the Study

This study provides an effective method for GE detection in IMU signals, even in the case of impaired gait. However, it should be evaluated within each patient group to ensure its reliability.

Highlights of the Study

The highlight of this study is the proposal of an effective IMU gait event detection method for both healthy subjects and pathological gait patients, particularly showing innovation and practical value in patients with moderate to severe gait impairment.