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: | Roya Riahi |
| Institution: | Isfahan university of medical science |
| Department: | School of public health, Epidemiology and Biostati |
| Country: | |
| Proposed Analysis: | The main aim of our study is to investigate the true causal relationships among risk factors (lifestyle factors, cardiovascular factors and biomarkers, diabetes-related and other endocrine factors) and dementia and Alzheimer's. Estimating causal effects using observational data ideally requires the analyst to have a vast prior knowledge of the underlying causal structure and the domain of application. Indeed, unbiased and efficient estimation of causal effect in observational studies requires adjustment for confounding variables (variables are related to both the outcome and exposure). Investigators often bypass difficulties related to the identification and selection of confounders through the use of fully adjusted outcome models. Although this approach protects against bias, including covariates unrelated to the outcome may increase the variance of the estimated causal effect. Instead of using a fully adjusted model, model selection can be attempted. Some variable selection techniques aim to maximize the ability of a model to achieve a correct estimate of causal effect and are used to decrease variability of the estimated causal effect, but they ignore covariate associations with exposure and may not adjust for important confounders weakly associated with the outcome. We use a Bayesian causal effect algorithm based on Bayesian model averaging for estimating causal effects that simultaneously consider models for both outcome and exposure. Our purposed algorithm select true confounders from a set of candidate variables using both outcome and exposure model and estimate correct causal effect as weighted averages of estimates under different outcome models. |
| Additional Investigators | |
| Investigator's Name: | Sayed mohsen Hosseinin |
| Proposed Analysis: | The main aim of our study is to investigate the true causal relationships among risk factors (lifestyle factors, cardiovascular factors and biomarkers, diabetes-related and other endocrine factors) and dementia and Alzheimer's. Estimating causal effects using observational data ideally requires the analyst to have a vast prior knowledge of the underlying causal structure and the domain of application. Indeed, unbiased and efficient estimation of causal effect in observational studies requires adjustment for confounding variables (variables are related to both the outcome and exposure). Investigators often bypass difficulties related to the identification and selection of confounders through the use of fully adjusted outcome models. Although this approach protects against bias, including covariates unrelated to the outcome may increase the variance of the estimated causal effect. Instead of using a fully adjusted model, model selection can be attempted. Some variable selection techniques aim to maximize the ability of a model to achieve a correct estimate of causal effect and are used to decrease variability of the estimated causal effect, but they ignore covariate associations with exposure and may not adjust for important confounders weakly associated with the outcome. We use a Bayesian causal effect algorithm based on Bayesian model averaging for estimating causal effects that simultaneously consider models for both outcome and exposure. Our purposed algorithm select true confounders from a set of candidate variables using both outcome and exposure model and estimate correct causal effect as weighted averages of estimates under different outcome models. |
| Investigator's Name: | Sayed mohsen Hosseinin |
| Proposed Analysis: | The main aim of our study is to investigate the true causal relationships among risk factors (lifestyle factors, cardiovascular factors and biomarkers, diabetes-related and other endocrine factors) and dementia and Alzheimer's. Estimating causal effects using observational data ideally requires the analyst to have a vast prior knowledge of the underlying causal structure and the domain of application. Indeed, unbiased and efficient estimation of causal effect in observational studies requires adjustment for confounding variables (variables are related to both the outcome and exposure). Investigators often bypass difficulties related to the identification and selection of confounders through the use of fully adjusted outcome models. Although this approach protects against bias, including covariates unrelated to the outcome may increase the variance of the estimated causal effect. Instead of using a fully adjusted model, model selection can be attempted. Some variable selection techniques aim to maximize the ability of a model to achieve a correct estimate of causal effect and are used to decrease variability of the estimated causal effect, but they ignore covariate associations with exposure and may not adjust for important confounders weakly associated with the outcome. We use a Bayesian causal effect algorithm based on Bayesian model averaging for estimating causal effects that simultaneously consider models for both outcome and exposure. Our purposed algorithm select true confounders from a set of candidate variables using both outcome and exposure model and estimate correct causal effect as weighted averages of estimates under different outcome models. |
| Investigator's Name: | Mohammad Ali Mansournia |
| Proposed Analysis: | The main aim of our study is to investigate the true causal relationships among risk factors (lifestyle factors, cardiovascular factors and biomarkers, diabetes-related and other endocrine factors) and dementia and Alzheimer's. Estimating causal effects using observational data ideally requires the analyst to have a vast prior knowledge of the underlying causal structure and the domain of application. Indeed, unbiased and efficient estimation of causal effect in observational studies requires adjustment for confounding variables (variables are related to both the outcome and exposure). Investigators often bypass difficulties related to the identification and selection of confounders through the use of fully adjusted outcome models. Although this approach protects against bias, including covariates unrelated to the outcome may increase the variance of the estimated causal effect. Instead of using a fully adjusted model, model selection can be attempted. Some variable selection techniques aim to maximize the ability of a model to achieve a correct estimate of causal effect and are used to decrease variability of the estimated causal effect, but they ignore covariate associations with exposure and may not adjust for important confounders weakly associated with the outcome. We use a Bayesian causal effect algorithm based on Bayesian model averaging for estimating causal effects that simultaneously consider models for both outcome and exposure. Our purposed algorithm select true confounders from a set of candidate variables using both outcome and exposure model and estimate correct causal effect as weighted averages of estimates under different outcome models. |
