Body sway and global equilibrium condition of the elderly in quiet standing posture by using competitive neural networks

Objective: The balance deficit (equilibrium condition) is one of the leading sources of falls of the elderly. Finding out the quiet standing point of equilibrium of body sway in older adults by using the competitive neural network for accessing body steadiness is presented in this study. Methods: The sample population consists of elderly aging from 60 to 80 years old quiet standing on a force measuring platform. To assess body steadiness, the values of the center of pressure (COP) are obtained by stabilometry. The center of pressure position and displacement are acquired over a time interval, COP(t), composing a body sway area. The competitive neural network use the COP(t) data as input signal for determining the Global Center of Pressure (G-COP) concerning the quiet standing point of equilibrium. Results: The competitive neural network based statokinesigram analysis is able to achieve the G-COP regardless the sort of dynamical body sway. Results demonstrate that the competitive neural network determines the point of equilibrium for patients with uniform, conservative, homogeneous body balance as well as for those patients presenting non-uniform, non-conservative, heterogeneous body sway in quiet standing posture. The proposed approach can be used to compute the G-COP both in off-line (when the entire COP data set is available) as well as in real-time (meanwhile COP data are being acquired); not requiring the entire COP data set be available. Conclusion: The competitive neural network comes to be a feasible alternative to compute the global center of pressure and, thus, to contribute to the postural steadiness and equilibrium condition analysis of the elderly. Such an approach is also able to be extended to any other aging group as well as in the presence of distinct pathologies that alter body control. Further, the proposed competitive neural network based statokinesigram analysis can be employed to work together with current techniques employed for steadiness analysis. (C) 2018 Elsevier B.V. All rights reserved.