Analysis of Gait Posture Characteristics of People with Sarcopenia under Dual-Task Based on MLP Neural Network Algorithm
Objective:Multilayer Perceptron(MLP)neural network algorithm was used to evalu-ate the gait posture of sarcopenia under the dual-task paradigm.Methods:A total of thirty subjects were selected from the sarcopenia group and the healthy older group.The two groups of subjects were asked to complete the gait test in three modes:single-task(ST),cognitive-motor dual-task(CMDT)and motor-motor dual-task(MMDT),respectively,and the Vicon infrared capture system collected the data.Two-way ANOVA was used to screen the gait and posture characteristics of people with sarcopenia,and the weight analysis was performed by MLP neural network model to measure each characteristic index.Results:1)Compared with the normal aged group,there were significant differences in 25 indexes such as stride speed,stride length,stride width,gait cycle,and gait variability index in the sarcopenia group under the same task(P＜0.05).2)Com-pared with ST walking,18 indices such as step frequency,gait cycle,support period,and percentage variation index of support period were significantly different in the sarcopenia group under the dual task(P＜0.05).3)A total of 14 gait posture indices had significant differences between groups and between single and dual tasks(P＜0.05),and the weight analysis of gait characteris-tics in sarcopenia was included.The results showed that the top five feature weights in the datas-et were stride length,stride frequency,right double support percentage,stride length variability,and left double support percentage.Conclusions:Individuals with sarcopenia exhibited higher gait variability and decreased postural control during dual-task walking compared to healthy older adults.Health managers of the elderly population can optimize the postural control adjust-ment strategy of the sarcopenia population and reduce the fall risk of the older population by paying attention to the gait and postural control characteristics of people with sarcopenia during dual-task walking.
NETL
NSTL
万方数据
