Reduction, classification and ranking of motion analysis data: an application to osteoarthritic and normal knee function data

Abstract

There are certain major obstacles to using motion analysis as an aid to clinical decision making. These include: the difficulty in comprehending large amounts of both corroborating and conflicting information; the subjectivity of data interpretation; the need for visualization; and the quantitative comparison of temporal waveform data. This paper seeks to overcome these obstacles by applying a hybrid approach to the analysis of motion analysis data using principal component analysis (PCA), the Dempster-Shafer (DS) theory of evidence and simplex plots. Specifically, the approach is used to characterise the differences between osteoarthritic (OA) and normal (NL) knee function data and to produce a hierarchy of those variables that are most discriminatory in the classification process. Comparisons of the results obtained with the hybrid approach are made with results from artificial neural network analyses.

Feature selection using a principal component analysis of the kinematics of the pivot shift phenomenon

“…Landry et al (2007) andO'Connor andBottum (2009) applied PCA to the biomechanics of cutting movements and successfully identified gender differences. Jones et al (2008) utilized a hybrid method that included PCA to characterize the differences between osteoarthritic and normal knee function. Lee et al (2009) showed that PCA allowed for the quantification of the effect of external loads on lower body kinematics.…”

Influence of wedges on lower limbs’ kinematics and net joint moments during healthy elderly gait using principal component analysis

“…It means that a variable only with values above these loadings, have a biomechanical interpretation in that portion of the curve. (Jones et al, 2008). For more information about PCA and its application previous studies are recommended (Jones et al, 2008;Sadeghi et al, 2002b).…”

“…The mean waveform of the three valid trials of each participant in each condition was used, and the PC score values (internal product from Matrix Z PC1, PC2 and PC3 to matrix A) for each subject in each condition were retained for analysis (PC1, PC2 and PC3 scores for the six waveforms, totalizing 18 PC score values per subject with each wedge). In the last phase, load vectors were calculated by normalizing the PC models (matrix Z) between À1 and 1 according to Jones et al (2008). After normalization, a threshold of ±0.71 was adopted to consider a load vector from one variable as relevant, and then to attribute a meaning for this PC (Knapp & Comrey, 1973).…”

“…In the last two decades the interpretation of gait data was improved by different methods of multivariate analysis (Deluzio & Astephen, 2007;Jones, Holt, & Beynon, 2008;Muniz & Nadal, 2009;Olney, Griffin, & McBride, 1998;Sadeghi et al, 2002a). Principal component analysis (PCA) is a powerful method used to reduce redundant information and it allows the comparison of the complete waveform, explaining much of the variance in the data with relatively few factors, or Principal Components (PCs) (Sadeghi et al, 2002a).…”

A new training algorithm using artificial neural networks to classify gender-specific dynamic gait patterns

“…Human gait is a subject of study in biomechanics, clinical analysis, robotics and biometrics (Yoo et al 2005;Jones et al 2008). As gait is assumed to be extremely individual and is stated to be impossible to repeat the same movement identically, every individual might be identified and classified by their own movement pattern (Hatze 1986;Schöllhorn and Bauer 1998;Schöllhorn 1999;Moustakidis et al 2008).…”