There are many active research projects accessing and applying shared ADNI data. Use the search above to find specific research focuses on the active ADNI investigations. This information is requested annually as a requirement for data access.
| Principal Investigator | |
| Principal Investigator's Name: | Federico Ferrari |
| Institution: | Duke University |
| Department: | Statistical Science |
| Country: | |
| Proposed Analysis: | It is very common to collect high-dimensional data and time-to-event outcomes in medical studies. In such cases, Cox proportional hazards models are commonly used with variable selection or regularization to deal with the issue of high dimensionality. Such approaches do not usually account for interactions among predictors, which would greatly increase dimensionality. Cases in which predictors can be moderately to highly correlated and heredity assumptions commonly used in high-dimensional interaction detection problems are unwarranted are especially of interest in this study. The primary goal of this study is to address these cases by proposing a factor analysis modeling approach in which latent factors underlying the predictors are included in a quadratic proportional hazards regression model. We design an efficient Markov Chain Monte Carlo (MCMC) algorithm for routine implementation as well as provide expressions for the induced coefficients in the space of covariates. We further show how our results are applicable to not only the Cox proportional hazards model but also to any generalized linear model with logarithmic link, namely Poisson and Negative-Binomial regression. |
| Additional Investigators |
