Inter- and intra-individual variability of ground reaction forces during sit-to-stand with principal component analysis

Variable reduction is an important issue in biomechanics, because the defin ition of a non-redundant set of variables necessary for a complete descript ion of a given motor act provides information about the motor strategy. A s ystematic tool for dealing with variable reduction problems is Principal Co mponent Analysis. In this paper, as an example of an application of this te chnique, the set of Ground Reaction Forces (GRFs) provided by a six-compone nt force plate, gained during standing up in a heterogeneous population of 82 normal individuals, was reduced to a set of fewer variables. Each subjec t was required to stand up from a chair five times at different, randomly s elf selected, speeds, obtaining a data set of 410 trials. Principal Compone nts (PCs) of GRFs were computed for each trial. On average, over the ensemb le of trials, first and second PCs (PC1 and PC2) explained together 90% of the data set variance, indicating that, during standing up, movement of the human body may be reliably described by using two PCs. Inter- and intra-in dividual repeatability of the first two PCs was investigated by examining t he correlation coefficient between PC waveforms obtained from the whole set of trials and within the set of trials performed by the same subject, resp ectively. While the PCI exhibited repeatable patterns, the second one, alth ough repeatable within the group of trials performed by the same subject, d isplayed marked inter-individual variability. Therefore, PCI was related to intrinsic aspects of the motor task and PC2 to intersubject features. (C) 1999 IPEM. Published by Elsevier Science Ltd. All rights reserved.