Reproducibility of a new signal processing technique to assess joint sway during standing

详细信息    查看全文

• 作者：Fabian Marcel Rast^a ; ^{fabian.rast@zhaw.ch} ; Judith Verena Graser^a ; ^b ; André ; Meichtry^a ; Markus Josef Ernst^a ; Christoph Michael Bauer^a
• 关键词：Postural control ; Motion analysis ; Spine ; Reproducibility of results ; Signal processing
• 刊名：Journal of Biomechanics
• 出版年：2017
• 出版时间：25 January 2017
• 年：2017
• 卷：51
• 期：Complete
• 页码：133-136
• 全文大小：291 K
• 卷排序：51

文摘

Postural control strategies can be investigated by kinematic analysis of joint movements. However, current research is focussing mainly on the analysis of centre of pressure excursion and lacks consensus on how to assess joint movement during postural control tasks. This study introduces a new signal processing technique to comprehensively quantify joint sway during standing and evaluates its reproducibility. Fifteen patients with non-specific low back pain and ten asymptomatic participants performed three repetitions of a 60-second standing task on foam surface. This procedure was repeated on a second day. Lumbar spine movement was recorded using an inertial measurement system. The signal was temporally divided into six sections. Two outcome variables (mean absolute sway and sways per second) were calculated for each section. The reproducibility of single and averaged measurements was quantified with linear mixed-effects models and the generalizability theory. A single measurement of ten seconds duration revealed reliability coefficients of .75 for mean absolute sway and .76 for sways per second. Averaging a measurement of 40 seconds duration on two different days revealed reliability coefficients higher than .90 for both outcome variables. The outcome variables’ reliability compares favourably to previously published results using different signal processing techniques or centre of pressure excursion. The introduced signal processing technique with two outcome variables to quantify joint sway during standing proved to be a highly reliable method. Since different populations, tasks or measurement tools could influence reproducibility, further investigation in other settings is still necessary. Nevertheless, the presented method has been shown to be highly promising.