Feature Selection Of Postural Summary Statistic Scores Based On Principal Component Analysis In Parkinson’s Disease
Free (open access)
L. Rocchi, L. Chiari, A. Cappello & F. B. Horak
The present study addresses the challenge of identifying the features of the centre of pressure (CoP) trajectory that are most sensitive to postural performance, to progress in the process of transforming CoP data into useful information and to promote standardization in quantitative posturography. For this purpose, we singled out the features that characterize postural sway in subjects with Parkinson’s disease (PD) with and without levodopa (on and off states). The feature selection was performed using principal component analysis. Results suggest that CoP in subjects with PD can be primarily characterised by four parameters, in both off and on states, describing: the size of the path over the support surface; the principal sway direction; and the shape and bandwidth of the power spectral density plot. The similarity of the present results with a previous study that considered young healthy subjects, allows us to define more confidently the minimum set of measures to recommend for specific applications as optimal descriptor of CoP sway in quiet stance. Keywords: principal component analysis, posture, Parkinson’s disease. 1 Introduction Body posture is the output of complex interactions between central nervous system control mechanisms and the musculo-skeletal actuators acting against the support surface. Because of its complexity body posture is challenging to
principal component analysis, posture, Parkinson’s disease.