Gait analysis and prediction in Parkinson's disease using rhythmic auditory stimulation: A data-driven approach with machine learning models

IF 1.8 3区医学 Q4 ENGINEERING, BIOMEDICAL

Clinical Biomechanics Pub Date : 2026-02-01 Epub Date: 2025-12-02 DOI:10.1016/j.clinbiomech.2025.106725

Jingyuan Lin

{"title":"Gait analysis and prediction in Parkinson's disease using rhythmic auditory stimulation: A data-driven approach with machine learning models","authors":"Jingyuan Lin","doi":"10.1016/j.clinbiomech.2025.106725","DOIUrl":null,"url":null,"abstract":"<div><h3>Background</h3><div>Gait disturbance is a core feature of Parkinson's disease (PD), contributing to mobility loss and fall risk. Rhythmic auditory stimulation (RAS) can improve gait, yet response varies across patients. This study evaluated RAS effects on gait and developed a model to predict responders.</div></div><div><h3>Methods</h3><div>Three hundred PD patients were enrolled and assigned to RAS or control groups. Gait data were collected with a 3D motion capture system and force plates, and baseline assessments included UPDRS-III and MoCA score. Clinical and gait features (stride length, speed, swing time, double support) were analyzed. Paired analyses compared pre- and post-RAS changes. Machine learning models (Random Forest, XGBoost, SVM) were trained on demographic and baseline gait data to predict RAS responsiveness, defined as improvement in gait speed and double support time.</div></div><div><h3>Findings</h3><div>Among the 171 patients in the RAS group, 116 (67.8 %) were identified as responders based on improvements in gait speed and double support time.The RAS group showed significant improvements in gait speed (+0.14 m/s), stride length (+0.10 m), and reductions in double support (−0.026 s) and variability (−0.012), all <em>p</em> < 0.001. Among models, Random Forest classifier demonstrated the best balanced performance (AUC = 0.713, accuracy = 0.71, F1 = 0.64, recall = 0.87), outperforming SVM and XGBoost models. Feature importance highlighted UPDRS-III, stride length, MoCA score, and gait asymmetry as key predictors.</div></div><div><h3>Interpretation</h3><div>RAS significantly enhances gait performance in PD, though with individual variability. A data-driven machine learning framework enables reasonable prediction of responders, supporting personalized gait rehabilitation strategies in clinical practice.</div></div>","PeriodicalId":50992,"journal":{"name":"Clinical Biomechanics","volume":"132 ","pages":"Article 106725"},"PeriodicalIF":1.8000,"publicationDate":"2026-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Clinical Biomechanics","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0268003325002980","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2025/12/2 0:00:00","PubModel":"Epub","JCR":"Q4","JCRName":"ENGINEERING, BIOMEDICAL","Score":null,"Total":0}

引用次数: 0

Abstract

Background

Gait disturbance is a core feature of Parkinson's disease (PD), contributing to mobility loss and fall risk. Rhythmic auditory stimulation (RAS) can improve gait, yet response varies across patients. This study evaluated RAS effects on gait and developed a model to predict responders.

Methods

Three hundred PD patients were enrolled and assigned to RAS or control groups. Gait data were collected with a 3D motion capture system and force plates, and baseline assessments included UPDRS-III and MoCA score. Clinical and gait features (stride length, speed, swing time, double support) were analyzed. Paired analyses compared pre- and post-RAS changes. Machine learning models (Random Forest, XGBoost, SVM) were trained on demographic and baseline gait data to predict RAS responsiveness, defined as improvement in gait speed and double support time.

Findings

Among the 171 patients in the RAS group, 116 (67.8 %) were identified as responders based on improvements in gait speed and double support time.The RAS group showed significant improvements in gait speed (+0.14 m/s), stride length (+0.10 m), and reductions in double support (−0.026 s) and variability (−0.012), all p < 0.001. Among models, Random Forest classifier demonstrated the best balanced performance (AUC = 0.713, accuracy = 0.71, F1 = 0.64, recall = 0.87), outperforming SVM and XGBoost models. Feature importance highlighted UPDRS-III, stride length, MoCA score, and gait asymmetry as key predictors.

Interpretation

RAS significantly enhances gait performance in PD, though with individual variability. A data-driven machine learning framework enables reasonable prediction of responders, supporting personalized gait rehabilitation strategies in clinical practice.

查看原文本刊更多论文

节律性听觉刺激对帕金森病的步态分析和预测：基于机器学习模型的数据驱动方法

步态障碍是帕金森病（PD）的核心特征，导致行动能力丧失和跌倒风险。节律性听觉刺激（RAS）可以改善步态，但不同患者的反应不同。本研究评估了RAS对步态的影响，并建立了预测反应者的模型。方法将300例PD患者分为RAS组和对照组。步态数据通过3D运动捕捉系统和力板收集，基线评估包括UPDRS-III和MoCA评分。分析临床和步态特征（步幅、速度、摇摆时间、双支撑）。配对分析比较了ras前后的变化。机器学习模型（Random Forest, XGBoost， SVM）在人口统计学和基线步态数据上进行训练，以预测RAS反应性，定义为步态速度和双支撑时间的改善。在RAS组的171例患者中，根据步态速度和双支撑时间的改善，116例（67.8%）被确定为应答者。RAS组在步速（+0.14 m/s）、步长（+0.10 m）、双支撑（- 0.026 s）和变异性（- 0.012）方面均有显著改善，p均为0.001。其中，Random Forest分类器表现出最佳的平衡性能（AUC = 0.713，准确率= 0.71,F1 = 0.64，召回率= 0.87），优于SVM和XGBoost模型。特征重要性强调UPDRS-III，步幅长度，MoCA评分和步态不对称是关键预测因子。ras可显著提高PD患者的步态表现，但存在个体差异。数据驱动的机器学习框架能够合理预测响应者，支持临床实践中的个性化步态康复策略。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Clinical Biomechanics 医学-工程：生物医学

CiteScore

3.30

自引率

5.60%

发文量

189

审稿时长

12.3 weeks

期刊介绍： Clinical Biomechanics is an international multidisciplinary journal of biomechanics with a focus on medical and clinical applications of new knowledge in the field. The science of biomechanics helps explain the causes of cell, tissue, organ and body system disorders, and supports clinicians in the diagnosis, prognosis and evaluation of treatment methods and technologies. Clinical Biomechanics aims to strengthen the links between laboratory and clinic by publishing cutting-edge biomechanics research which helps to explain the causes of injury and disease, and which provides evidence contributing to improved clinical management. A rigorous peer review system is employed and every attempt is made to process and publish top-quality papers promptly. Clinical Biomechanics explores all facets of body system, organ, tissue and cell biomechanics, with an emphasis on medical and clinical applications of the basic science aspects. The role of basic science is therefore recognized in a medical or clinical context. The readership of the journal closely reflects its multi-disciplinary contents, being a balance of scientists, engineers and clinicians. The contents are in the form of research papers, brief reports, review papers and correspondence, whilst special interest issues and supplements are published from time to time. Disciplines covered include biomechanics and mechanobiology at all scales, bioengineering and use of tissue engineering and biomaterials for clinical applications, biophysics, as well as biomechanical aspects of medical robotics, ergonomics, physical and occupational therapeutics and rehabilitation.