2024 American Academy of Neurology Abstract Website

Ming-Yang Ho¹, Ming-Che Kuo², Ciao-Sin Chen³, Ruey-Meei Wu², Ching-Chi Chuang⁴, Chi-Sheng Shih⁵, Yufeng Tseng¹
¹Department of Computer Science and Information Engineering, National Taiwan University, ²Department of Neurology, National Taiwan University Hospital, ³Graduate Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, ⁴NTU IoX Research Center, National Taiwan University, ⁵Graduate Institute of Networking and Multimedia, National Taiwan University

Objective:

We present PathoOpenGait, a cloud-based platform for comprehensive gait analysis.

Background:

Gait assessment is crucial in neurodegenerative diseases such as Parkinson's and multiple system atrophy, yet current techniques are neither affordable nor efficient.

Design/Methods:

PathoOpenGait utilizes 2D and 3D data from a binocular 3D camera for monitoring and analyzing gait parameters. Our algorithms, including a semi-supervised learning-boosted neural network model for turn time estimation and deterministic algorithms to estimate gait parameters, were rigorously validated on annotated gait records collected from two medical centers. We subsequently developed a cloud-based platform to integrate our algorithms into clinical practice.

Results:

Our algorithms demonstrated high precision and consistency. Besides, the semi-supervised learning strategy was helpful for model adaptation across data collected from different medical centers. All six estimated gait parameters demonstrated "good" and "excellent" clinical reliability as shown by their ICC_3,1 values: stride length (ICC_3,1=0.97), stride width (ICC_3,1=0.88), stride time (ICC_3,1=0.93), velocity (ICC_3,1=0.98), cadence (ICC_3,1=0.88), and turn time (ICC_3,1=0.79). We further demonstrate PathoOpenGait's applicability in clinical settings by analyzing gait trials from Parkinson's patients and healthy controls. The analysis of gait trials from 122 Parkinson's disease (PD) patients and 124 healthy controls highlighted significant differences in gait parameters between the two groups. As PD progressed, notable changes in stride length (p<0.001), step width (p<0.001), and turn time (p<0.001) were observed. Moreover, gait parameters were found to correlate with factors like MMSE scores (p<0.001) and disease duration (p<0.05).

Conclusions:

PathoOpenGait is the first open-source, cloud-based system for gait analysis, providing a user-friendly tool for continuous patient care and monitoring. It offers a cost-effective and accessible solution for both clinicians and patients, revolutionizing the field of gait assessment. PathoOpenGait is available at https://pathoopengait.cmdm.tw. The source code is available on GitHub at https://github.com/Kaminyou/PathoOpenGait.