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: | Tanya Garcia |
| Institution: | Texas A&M University |
| Department: | Biostatistics |
| Country: | |
| Proposed Analysis: | I propose to develop a new statistical method to identify and estimate the effects of high-dimensional neuroimaging measures in relation to age of Mild Cognitive Impairment (MCI). The method will be applied to T1-weighted MRI images of the whole brain, and adjust for demographic and clinical information. I will model age at MCI using Cox regression. This adaptation easily incorporates time-varying, patient-specific features, high-dimensional (tensor) covariates, and censoring. All computations will be implemented in Matlab which has existing packages for Cox regression and tensor computations. I will estimate the effects of tensor covariates and time-varying vector covariates by maximizing an objective function defined by the partial log-likelihood of the data. The maximization will be done using a block relaxation algorithm: an approach that alternates between updating estimates for array components of the tensor coefficient and the vector coefficient. The algorithm has been successfully used in generalized linear models with tensor covariates and permits two computational advantages: At each step, (i) the algorithm simplifies the problem to a low-dimensional Cox model which can be solved using the coxph function in Matlab; (ii) the algorithm identifies regions of the brain linked to age-at-onset using standard variable selection techniques available in the glmnet package of Matlab. An alternative to the Cox model is the accelerated failure time model which also has routine functions in Matlab (i.e., aft function). |
| Additional Investigators |
