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: | Yi-Ju Li |
| Institution: | Duke University Medical Center |
| Department: | Biostatistics and Bioinformatics |
| Country: | |
| Proposed Analysis: | The objective of our study is to investigate the shared risk factors between Alzheimer disease (AD) and Parkinson disease (PD). Specifically, we are interested in identifying biomarkers and genetic variants for the progression of AD and PD. For AD, we would like to take the full advantage of the longitudinal clinical and biological data collected in ADNI. For PD, we will apply data from PPMI (http://www.ppmi-info.org/). Our goal is to identify CSF, PET, and MRI biomarker predictors and genetic variants for different stages of cognitive decline progression. This will be different from conventional analysis using data from a single time point. To achieve this, in addition to the analysis of longitudinal data to create phenotypes for different disease states, we will take the advantage of multi-state model to identify biomarkers and genetic variants that may predict transitions from one state to the next state. For instance, the mutli-state model will allow us to differentiate risk factors and predictors between CN to MCI and MCI to dementia. We consider the outcomes of proposed analysis will contribute significantly to the development of the strategies toward early detection or personalized medicine. Aim 1: Identify biomarkers and create a risk score for the progression of AD and PD. Multi-state models (MSM) are used to describe the stochastic process of transition of individuals between a finite number of states. In medical research, the states might be based on clinical symptoms such as cognitive stages. MSMs have not been applied to AD or PD biomarker and genetic research, largely due to the lack of longitudinal data an investigator can access. Here, our plan is to identify predictors of progression between different stages of cognitive impairment and to create corresponding risk scores. The strength of the data available from ADNI and PPMI is that cognitive data are collected longitudinally, allowing tracking of the progression of cognitive decline over time. We will start by considering stages of cognitive decline in AD (CN, MCI, and AD) and cognitive categories in PD (CN, PD-MCI, and PD Dementia) to test the role of CSF, PET and MRI biomarkers in each transition. As the result of this analysis, we expect to create a risk score to predict the probability of progressing between stages over time. Aim 2: Implement multi-state model to analyze genome wide single nucleotide polymorphisms data As stated in Aim 1, MSM methodology possesses some advantages over the conventional analysis that test association between factors and each disease stage one at a time (e.g. AD vs. control, or MCI vs controls). As the genome wide SNP data are available in ADNI, we will implement multi-state models to identify genetic variant predictors of transitions between stages, in addition to biomarkers identified in Aim 1. That is, biomarkers will be included in the multi-state models as adjustment covariates. Ideally, we will perform the same analysis as described in Aim 1 in PPMI dataset if the genetic data is available, and look for overlapping and non-overlapping results to draw the conclusion between AD and PD. Alternatively, we will obtained the genetics data for PD from NCBI dbGaP for validation of the top variants. |
| Additional Investigators |
