Quantifying standing posture during multi-joint movements Clint Hansen a,c *, Qin Wei b , Jiann-Shing Shieh b , Nasser Rezzoug c , Philippe Gorce c and Brice Isableu a a University Paris-Sud. URCIAMS – Motor Control & Perception team, EA 4532, Orsay F-91405, France; b Department of Mechanical Engineering, Yuan Ze University, 135 Yuan-Tung Road, Chung-Li, Taoyuan, Taiwan; c University du Sud – Toulon Var, HandiBio, EA 4322, 83957 La Garde Cedex, France Keywords: center of pressure; multi scale entropy; multi-joint movement; upper-limb 1. Introduction The postural sway is a common measurement when evaluating the standing posture stability of a person. In everyday life, we interact with the environment and use our arms for gestures or to perform specific tasks. Since every person tries to prevent slipping or falling, movements of our upper limbs are usually performed in a controlled way. Changing the speed can cause a destabilisation of our posture, and postural adaptations are needed to avoid loss of our bipedal equilibrium. New methods such as the multi- scale entropy algorithm (Costa et al. 2005) have recently been developed to measure and quantify the intrinsic complexity of an individual’s posture while standing. The purpose of this study was to examine the influence of cyclic internal– external rotation movements of the upper limbs, performed at different speeds, on postural control mechanisms. Further- more, this study examines the different impacts on postural stability due to the dominance of our limbs. 2. Methods Fifteen subjects (22 ^ 3 years) voluntarily participated in the experiment after signing a statement of informed consent as required by the Helsinki declaration and the EA 4532 local Ethics Committee. Subjects’ limb dominance was determined using a 10-item version of the Edinburgh Handedness Inventory. During separate trials, the subjects performed a shoulder internal–external rotation of their dominant and non-dominant arm from upward (about 308 behind the vertical) to downward (weakly below horizontal) with a shoulder elevated at 908 in the frontal plane and the elbow flexed at 908. Two angular velocities were chosen, S (0.1 Hz < 0.3 rad s 21 ) versus F (2.0 Hz <6.3 rad s 21 ), and two sensory conditions, visuo-kinesthetic (VK) with eyes open versus kinesthetic (K) with eyes closed. The other arm was aligned with the body. In the K condition, the participants were allowed to control their movements visually and in the VK condtion, the movements were performed with eyes closed. Participants stood upright on a force plate (AMTI OR6001200) to measure postural sway (1000Hz). Each subject performed the task three times in each of the four conditions for a total of 24 trials per arm. The position of the feet was marked with chalk to ensure that the subjects always had the same feet position. A V8i VICON eight camera (Mcam2) motion capture system was used to record trunk and arm movements at a rate of 250 Hz (Vicon motion systems Inc., Oxford, UK). Multi-scale entropy (MSE) algorithm was used to calculate the complexity and the interaction between the postural sway (x and z) and the time series of the shoulder-to-elbow axis in x, y and z directions in concordance with the recommen- dations of the International Society of Biomechanics. The MSE algorithm converts the original time series into coarse-grained time series corresponding to a scale factor. Depending on the number of scales, sample entropy is calculated for every coarse-grained time series. If the signal becomes more predictable, the entropy value will diminish, whereas adding noise to the signal (velocity) will lead to an amplification of the entropy value. The integral of the MSE curve is the complexity index (CI), which allows easy comparison between subjects and groups of subjects. The higher the CI the higher is the stability of a person (Jiang 2012). 3. Results and discussion 3.1 Variability of shoulder-to-elbow axis displacements A multivariate repeated measures analysis of variance (MANOVA) combining two arm velocities (S vs. F), * two dominance profiles (D vs. ND) and two visual conditions (K vs. VK) with the variability of endpoint displacements of the shoulder-to-elbow axis (x, y , z directions) as the dependent variables showed significant: main effects for dominance (D vs. ND) [F(1, 14) ¼ 20.379, p ¼ 0.00049], ISSN 1025-5842 print/ISSN 1476-8259 online q 2012 Taylor & Francis http://dx.doi.org/10.1080/10255842.2012.713662 http://www.tandfonline.com *Corresponding author. Email: clint.hansen@u-psud.fr Computer Methods in Biomechanics and Biomedical Engineering Vol. 15, No. S1, September 2012, 256–258