ABSTRACT
Objective:To analyzed the characteristics of the cognitive domains of the montreal cognitive assessment(MOCA)scale in cognitively impaired or normal elderly people from the perspective of structural equation modeling(SEM).Methods:There were 335 old individuals in this study, including 166 cognitively normal individuals, 130 patients with mild cognitive impairment(MCI)and 39 individuals with dementia.The average age of the individuals was 81.5±9.0.Data on the Montreal cognitive assessment(MoCA, Beijing version)were gathered.Following exploratory factor analysis and selection of latent and manifest variables, a structural equation model was established.After assessment of data from the domains of the scale in the elderly, participants were divided into a normal group and a patient group, which formed the basis of the model.Results:The KMO value of the data calculated by exploratory factor analysis was 0.762.The dimensions measured by the scale were divided into four main latent variables: memory, visuospatial execution, language ability and attention.The CMIN value of the overall model was 44.039 and the P value was 0.168.The parameters of the overall model and individual dimensions all indicated a good fit.The model showed that visuospatial execution had the largest impact on cognitive function, with a path coefficient of 0.742, and language ability had the least impact, with a path coefficient of only 0.091.As a latent variable, attention had path coefficients of 0.372 and 0.236 for memory, 1.663 and 1.102 for visuospatial execution, and 1.090 and 0.798 for language ability, respectively, for the two groups, with clear statistically significant differences between the groups(all P<0.05). Conclusions:SEM can improve researchers' overall understanding of the impacts of the individual components of the scale and their use and interpretation of the scale.
ABSTRACT
Objective To improve the image quality of the electrical impedance tomography (EIT) by introducing the prior information into the regularization matrix.Methods The linear combination of the conductivity was established by background conductivity of dynamic variation,the covariance matrix was used here to remove the correlation between the background conductivity,and this prior information was introduced to construct the regularization matrix.Resnlts Compared with the traditional regularization matrix,the one involving in the prior information on the dynamic background gained more stable and better images.Conclusion Trials prove the efficacy of the regularization matrix on EIT imaging in 1 respiratory cycles (or heart beat),and following related researches may find theoretical references and support for feasibility.