The overall objective of the Biostatistical Core is to work closely with all 5 projects and the animal core to provide efficient data management, biostatistical design and data analyses, as well as aid in presentations of scientific data. To accomplish this goal, the Biostatistical Core will provide rigorous methodological and biostatistical support to all projects;and will conduct independent research and analyses related to the program projects.
As specific aims the Core will provide: (1) efficient data management;(2) routine statistical support for all 5 projects, including assistance with the design of studies, advice on database design and content, summarizing, plotting, analyzing data, interpretation of results, and collaboration in manuscript preparation for publication and/or presentation;and 3) development of new statistical methodology to address and support the statistical design and analyses for all projects. Statistical analysis for the Projects in this grant often involves data modeling, comparison among groups, and parameter estimations. The Biostatistical Core will combine data from different projects and cores of this program project to perform both univariate and multivariate statistical analyses. The Core will also correlate and share data for different projects on the same animals. Special emphasis will be placed on investigating correlations among longitudinal measures (such as motor function in animals of different ages) and endpoint values (such as dopamine levels and morphological correlates) from the same animal. For situations where the conventional parametric or semi-parametric theory is appropriate, the Core investigators will apply ANOVA or the linear mixed models to fit the data with univariate outcome. For multiple dependent outcomes, the Core investigators will use parametric global statistical tests, obtain optimal estimate of parameters of interest and the corresponding confidence intervals. When the parametric assumptions are in doubt, the Core investigators will apply nonparametric global statistical tests for both univarite and multivariate outcomes and repeated measures. For situations when no suitable statistical models in the literature can be applied, the Biostatistical Core will develop suitable statistical models and necessary statistical inference theory for all projects.
Showing the most recent 10 out of 30 publications
