Accéder directement au contenu Accéder directement à la navigation
Article dans une revue

Parameters Selection for Bivariate Multiscale Entropy Analysis of Postural Fluctuations in Fallers and Non-Fallers Older Adults

Sofiane Ramdani 1 Vincent Bonnet 1 Guillaume Tallon 2 Julien Lagarde Pierre Louis Bernard 2 Hubert Blain 3, 2 
1 IDH - Interactive Digital Humans
LIRMM - Laboratoire d'Informatique de Robotique et de Microélectronique de Montpellier
Abstract : Entropy measures are often used to quantify the regularity of postural sway time series. Recent methodological developments provided both multivariate and multiscale approaches allowing the extraction of complexity features from physiological signals; see "Dynamical complexity of human responses: A multivariate data-adaptive framework," in Bulletin of Polish Academy of Science and Technology, vol. 60, p. 433, 2012. The resulting entropy measures are good candidates for the analysis of bivariate postural sway signals exhibiting nonstationarity and multiscale properties. These methods are dependant on several input parameters such as embedding parameters. Using two data sets collected from institutionalized frail older adults, we numerically investigate the behavior of a recent multivariate and multiscale entropy estimator; see "Multivariate multiscale entropy: A tool for complexity analysis of multichannel data," Physics Review E, vol. 84, p. 061918, 2011. We propose criteria for the selection of the input parameters. Using these optimal parameters, we statistically compare the multivariate and multiscale entropy values of postural sway data of non-faller subjects to those of fallers. These two groups are discriminated by the resulting measures over multiple time scales. We also demonstrate that the typical parameter settings proposed in the literature lead to entropy measures that do not distinguish the two groups. This last result confirms the importance of the selection of appropriate input parameters.
Type de document :
Article dans une revue
Liste complète des métadonnées
Contributeur : Cécile Nowak Connectez-vous pour contacter le contributeur
Soumis le : lundi 8 octobre 2018 - 09:49:39
Dernière modification le : vendredi 5 août 2022 - 15:02:26




Sofiane Ramdani, Vincent Bonnet, Guillaume Tallon, Julien Lagarde, Pierre Louis Bernard, et al.. Parameters Selection for Bivariate Multiscale Entropy Analysis of Postural Fluctuations in Fallers and Non-Fallers Older Adults. IEEE Transactions on Neural Systems and Rehabilitation Engineering, IEEE Institute of Electrical and Electronics Engineers, 2016, 24 (8), pp.859-871. ⟨10.1109/TNSRE.2015.2473701⟩. ⟨hal-01889847⟩



Consultations de la notice