Holistic Evaluation of Gait Development in Children by Principal Component Analysis

Abstract

要旨： 4 7 4 1 2 キーワード： 3 澤留 朗 2014 多田充徳 竹村 裕 河内まき子 持丸正明 ー 50 ー Abstract : Entire gait characteristics were quantized by principal component analysis (PCA) for eight children (longitudinal survey from the age of 4 years to 7) to reveal the age-related difference in children. Temporal sequence of lower-limb angle and joint moment were compressed into lower dimensional space by PCA those were then evaluated in terms of change in gait pattern and variability among different age group. As a result of PCA, both first and second principal component (PC1 and PC2) were revealed to represent the role ofankle joint. This change was reported as typical difference between children and adults in the previous research. We found out that the gait pattern continued to change linearly in the lower dimensional space, while its variability were converged into constant value with age.These results suggest that development of gait pattern and acquiring its repeatability were independent phenomena. Also, it is clear that PCA enabled us to evaluate change in entire gait pattern keeping as much information as possible that can not be achieved by kinematic index such as step length or cadence.

A comprehensive analysis of sit‐to‐stand movement in a living space involving principal component analysis

“…PCA synthesized a 505‐dimensional principal component vector (PCVs) $$$ \boldsymbol{W} $$$ from the input standardized data $$$ \boldsymbol{X} $$$, and principal component scores (PCSs) $$$ \boldsymbol{T} $$$ were calculated for the observed values on each principal component axis using Equation (). $$$$ \boldsymbol{T}=\boldsymbol{XW} $$$$ Observed values refer to each subject. Each principal component space has its input information, and in this study, the top k components with an explanation rate exceeding 80% were extracted 16 . PCA reconstructed the data using only specific PCVs.…”

“…Each principal component space has its input information, and in this study, the top k components with an explanation rate exceeding 80% were extracted. 16 PCA reconstructed the data using only specific PCVs. The standardized data b X reconstructed in the principal component space are calculated as in Equation (3.2) using specific PCVs W p and the corresponding PCSs…”

“…Sawadome et al. synthesized a new principal component space from all‐time series data of lower‐limb joints and moments during walking to investigate how gait changes with age 16 . However, there have been few cases in which PCA has been applied to STS time series data, especially in terms of focusing on the characteristics of aging.…”

“…14 Deluzio et al adopted PCA to compare the gait of 50 patients with end-stage knee osteoarthritis with that of 63 asymptomatic controls and analyzed the gait characteristics of the osteoarthritis patients. 15 16 However, there have been few cases in which PCA has been applied to STS time series data, especially in terms of focusing on the characteristics of aging. In this study, PCA is applied to STS analysis, and time series data of the entire movement are analyzed.…”