Objective: The aim of this study was to quantify the motor symptoms of Parkinson’s disease (PD), including freezing of gait, using a motion capture device.
Background: The MDS-UPDRS III is used to evaluate the motor symptoms of PD, including freezing of gait, which is the most troublesome symptom when walking, however, the scores obtained by different evaluators may differ and the process is time-consuming. To overcome this problem, there is a need to establish an automatic measurement method for motor symptoms.
Method: We studied 23 patients with PD (13 males, 10 females; 60 ± 9.40 years old) and asked them to perform a task of straight-line walking and turning during off and on periods. The movements were recorded by using a motion capture system. From the coordinate data, we calculated various features such as Euler angles of joints, angular velocities, and the volume of space occupied by the joints that moved. A random forest model was created with each feature as an explanatory variable and the MDS-UPDRS III score as the objective variable, and cross-validation was performed using k-fold cross-validation (k = 5). Furthermore, the importance of the feature for prediction was evaluated using Gini importance. An algorithm was designed to detect the first step, and the time required to take the first step was defined as the score for freezing gait.
Results: The predictions of the MDS-UPDRS III score by the constructed model showed a linear relationship with the actual values. The model showed that the feature related to the joint coordinates of the part close to the trunk were important. Freezing gait measured by the proposed model was improved during on period, and was correlated with the freezing gait score of MDS-UPDRS III.
Conclusion: We have successfully established the basis for a system that can predict motor scores for PD, including freezing of gait, with only a few minutes of gate video recording.
To cite this abstract in AMA style:
W. Sako, M. Osawa, T. Iguchi, T. Hara, H. Haginiwa, S. Haji, N. Hattori. Automatic quantification of motor symptoms including freezing of gait in Parkinson’s disease through motion capture [abstract]. Mov Disord. 2025; 40 (suppl 1). https://www.mdsabstracts.org/abstract/automatic-quantification-of-motor-symptoms-including-freezing-of-gait-in-parkinsons-disease-through-motion-capture/. Accessed October 5, 2025.« Back to 2025 International Congress
MDS Abstracts - https://www.mdsabstracts.org/abstract/automatic-quantification-of-motor-symptoms-including-freezing-of-gait-in-parkinsons-disease-through-motion-capture/