Pelykh Olena, Klein Anke-Maria, Bötzel Kai, Kosutzka Zuzana, Ilmberger Josef
Department of Physical Medicine and Rehabilitation, University Clinics Munich, Marchioninistr. 15, Munich MA 81377, Germany; Department of Neurology, University Clinics Munich, Marchioninistr. 15, Munich MA 81377, Germany.
Department of Physical Medicine and Rehabilitation, University Clinics Munich, Marchioninistr. 15, Munich MA 81377, Germany.
Gait Posture. 2015 Sep;42(3):246-50. doi: 10.1016/j.gaitpost.2014.09.021. Epub 2014 Oct 2.
It has been suggested that dynamical measures such as sample entropy may be more appropriate than conventional measures when analyzing time series data such as postural sway. We evaluated conventional and dynamical measures of postural sway in Parkinson disease (PD) patients with and without freezing episodes.
COP (center of pressure) data were recorded during quiet standing with eyes open, eyes closed and while performing a dual task. Data for 16 patients with freezing of gait, 17 patients with no history of freezing and 24 healthy subjects were analyzed. The amount of postural sway was quantified using conventional measures, whereas for the characterization of the temporal structure of the COP data the normalized sway path and sample entropy was calculated.
Mean radius was higher and sample entropy was lower in patients with freezing symptoms as compared to healthy subjects in all three conditions. Dual-tasking significantly increased sway path length in patients with freezing, while normalized sway path did not change over conditions in this group.
Our findings show that postural sway is characterized by a combination of large radius, short normalized sway path and high regularity of the COP only in patients with freezing. This pattern becomes most prominent in a dual-task paradigm. This may explain higher occurrence of gait freezing in dual task situations with subsequent higher risk of falls. Results suggested that dynamic measures may add valuable information for characterizing postural stability in PD patients.
有人提出，在分析诸如姿势摆动等时间序列数据时，样本熵等动态测量方法可能比传统测量方法更合适。我们评估了有和没有冻结发作的帕金森病（PD）患者姿势摆动的传统测量方法和动态测量方法。
在睁眼、闭眼安静站立以及执行双重任务期间记录压力中心（COP）数据。分析了16例步态冻结患者、17例无冻结病史患者和24名健康受试者的数据。使用传统测量方法对姿势摆动量进行量化，而对于COP数据的时间结构特征，计算归一化摆动路径和样本熵。
在所有三种情况下，与健康受试者相比，有冻结症状的患者平均半径更高，样本熵更低。双重任务显著增加了有冻结症状患者的摆动路径长度，而该组患者的归一化摆动路径在不同条件下没有变化。
我们的研究结果表明，只有在有冻结症状的患者中，姿势摆动的特征是半径大、归一化摆动路径短以及COP的高规律性相结合。这种模式在双重任务范式中最为突出。这可能解释了在双重任务情况下步态冻结的发生率较高以及随后跌倒风险较高的原因。结果表明，动态测量方法可能为表征PD患者的姿势稳定性增加有价值的信息。
