一种基于惯性传感器的帕金森病冻结步态识别方法研究

摘要

Objective Study on the method of measuring Parkinson's disease(PD)freezing of gait(FOG)using inertial sensor.Methods In this work,a frequency domain analysis method combined with the gait cycle from time domain analysis was employed to identify the FOG signal among the signal of 10-meter-round trip and the double task round trip.Results The databases of control group,PD patients and PD with FOG patients were employed in this work to verify the method.The experiment results show that this FOG labeling method can effectively and correctly identify FOG patients with a sensitivity of 100% and a specificity of 88.9% and identify FOG segments with a sensitivity of 84.7%.Conclusion The inertial sensor can detect the PD motor function accurately, and the improved time-frequency domain labeling method can acquire more accurate FOG recognition results than the existing method(P <0.05).The FOG accurate early recognition is of great significance for follow-up treatment.%目的 研究以惯性传感器测量数据识别帕金森病(PD)冻结步态(FOG)的方法及其准确性.方法 用惯性传感器对正常对照者、PD患者及PD合并冻结步态患者,进行10 m往返和双重任务往返行走的运动功能检测;采用改进的结合时域步态参数的频域分析方法进行测试数据分析.结果 惯性传感器测试数据经改进的方法分析,显著提升了识别的敏感性(P<0.05);针对实验的敏感性为100%,特异性88. 9%;针对冻结步态片段的敏感性为84.7%.结论 惯性传感器能对PD冻结步态精准检测,改良分析方法对其的识别更加有效.PD冻结步态的早期、精准识别,对临床治疗有重要指导意义.