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: | Tong Wang |
| Institution: | Shanxi Medical University |
| Department: | Department of Health Statistics, School of Public |
| Country: | |
| Proposed Analysis: | we are interested in is to construct a joint model of multiple longitudinal variables based on functional data analysis to dynamically predict Alzheimer’s disease progression. Joint modeling (JM) of longitudinal and survival data is usually a popular framework in dynamic prediction and most JM approaches only investigated the association between a single longitudinal outcome and a time-to-event outcome. Including more longitudinal variables is expected to improve prediction accuracy. However, the shared random effect model and Bayesian framework would be computationally difficult or prohibitive when a large number of longitudinal variables are used. Multivariate functional principal component analysis is a useful method for the joint model of multiple longitudinal markers, but most of these methods require a limited number of continuous longitudinal variables. For high-dimensional sparse data such as image data, the longitudinal information of each brain area is not used. Therefore, we plan to extend the functional data analysis method to a more general joint model that can handle high-dimensional longitudinal variables and obtain a more accurate AD dynamic prediction model. We use the ADNI database as the training set and build the model, then expect to verify it in the AIBL database. |
| Additional Investigators | |
| Investigator's Name: | Juping Wang |
| Proposed Analysis: | Alzheimer’s Disease Neuroimaging Initiative (ADNI) database focus on the collection of longitudinal assessments, magnetic resonance imaging and positron emission tomography imaging measures, as well as other biomarkers from blood and cerebrospinal fluid.Of those covariates collected in the cohort, many are time-varying. We apply for this data to solve the following two problems. The first problem we are interested in is to construct a joint model of multiple longitudinal variables based on functional data analysis to dynamically predict Alzheimer’s disease progression. Joint modeling (JM) of longitudinal and survival data is usually a popular framework in dynamic prediction and most JM approaches only investigated the association between a single longitudinal outcome and a time-to-event outcome. Including more longitudinal variables is expected to improve prediction accuracy. However, the shared random effect model and Bayesian framework would be computationally difficult or prohibitive when a large number of longitudinal variables are used. Multivariate functional principal component analysis is a useful method for the joint model of multiple longitudinal markers, but most of these methods require a limited number of continuous longitudinal variables. For high-dimensional sparse data such as PET and MRI, the longitudinal information of each brain area is not used. Therefore, we plan to extend the functional data analysis method to a more general joint model that can handle high-dimensional longitudinal variables and obtain a more accurate AD dynamic prediction model. In addition，A central problem in neural mediation analysis is to identify which brain regions are positioned along the pathway between treatments and disease status. So we are also interested in identifying the brain areas intermediate between signal exposure or multiple exposures (including genetic information) and the development of brain diseases. Therefore, we plan to use the data to establish a high-dimensional functional mediation analysis method (including the mediation analysis method of survival outcome) to identify the brain area that plays a mediating role and quantify the mediation effect. |
| Investigator's Name: | Qian Gao |
| Proposed Analysis: | Alzheimer’s Disease Neuroimaging Initiative (ADNI) database focus on the collection of longitudinal assessments, magnetic resonance imaging and positron emission tomography imaging measures, as well as other biomarkers from blood and cerebrospinal fluid.Of those covariates collected in the cohort, many are time-varying. We apply for this data to solve the following two problems. The first problem we are interested in is to construct a joint model of multiple longitudinal variables based on functional data analysis to dynamically predict Alzheimer’s disease progression. Joint modeling (JM) of longitudinal and survival data is usually a popular framework in dynamic prediction and most JM approaches only investigated the association between a single longitudinal outcome and a time-to-event outcome. Including more longitudinal variables is expected to improve prediction accuracy. However, the shared random effect model and Bayesian framework would be computationally difficult or prohibitive when a large number of longitudinal variables are used. Multivariate functional principal component analysis is a useful method for the joint model of multiple longitudinal markers, but most of these methods require a limited number of continuous longitudinal variables. For high-dimensional sparse data such as PET and MRI, the longitudinal information of each brain area is not used. Therefore, we plan to extend the functional data analysis method to a more general joint model that can handle high-dimensional longitudinal variables and obtain a more accurate AD dynamic prediction model. In addition，A central problem in neural mediation analysis is to identify which brain regions are positioned along the pathway between treatments and disease status. So we are also interested in identifying the brain areas intermediate between signal exposure or multiple exposures (including genetic information) and the development of brain diseases. Therefore, we plan to use the data to establish a high-dimensional functional mediation analysis method (including the mediation analysis method of survival outcome) to identify the brain area that plays a mediating role and quantify the mediation effect. |
| Investigator's Name: | Shuting Chen |
| Proposed Analysis: | Alzheimer’s Disease Neuroimaging Initiative (ADNI) database focus on the collection of longitudinal assessments, magnetic resonance imaging and positron emission tomography imaging measures, as well as other biomarkers from blood and cerebrospinal fluid.Of those covariates collected in the cohort, many are time-varying. We apply for this data to solve the following two problems. The first problem we are interested in is to construct a joint model of multiple longitudinal variables based on functional data analysis to dynamically predict Alzheimer’s disease progression. Joint modeling (JM) of longitudinal and survival data is usually a popular framework in dynamic prediction and most JM approaches only investigated the association between a single longitudinal outcome and a time-to-event outcome. Including more longitudinal variables is expected to improve prediction accuracy. However, the shared random effect model and Bayesian framework would be computationally difficult or prohibitive when a large number of longitudinal variables are used. Multivariate functional principal component analysis is a useful method for the joint model of multiple longitudinal markers, but most of these methods require a limited number of continuous longitudinal variables. For high-dimensional sparse data such as PET and MRI, the longitudinal information of each brain area is not used. Therefore, we plan to extend the functional data analysis method to a more general joint model that can handle high-dimensional longitudinal variables and obtain a more accurate AD dynamic prediction model. In addition，A central problem in neural mediation analysis is to identify which brain regions are positioned along the pathway between treatments and disease status. So we are also interested in identifying the brain areas intermediate between signal exposure or multiple exposures (including genetic information) and the development of brain diseases. Therefore, we plan to use the data to establish a high-dimensional functional mediation analysis method (including the mediation analysis method of survival outcome) to identify the brain area that plays a mediating role and quantify the mediation effect. |
| Investigator's Name: | Lun Huang |
| Proposed Analysis: | Alzheimer’s Disease Neuroimaging Initiative (ADNI) database focus on the collection of longitudinal assessments, magnetic resonance imaging and positron emission tomography imaging measures, as well as other biomarkers from blood and cerebrospinal fluid.Of those covariates collected in the cohort, many are time-varying. We apply for this data to solve the following two problems. The first problem we are interested in is to construct a joint model of multiple longitudinal variables based on functional data analysis to dynamically predict Alzheimer’s disease progression. Joint modeling (JM) of longitudinal and survival data is usually a popular framework in dynamic prediction and most JM approaches only investigated the association between a single longitudinal outcome and a time-to-event outcome. Including more longitudinal variables is expected to improve prediction accuracy. However, the shared random effect model and Bayesian framework would be computationally difficult or prohibitive when a large number of longitudinal variables are used. Multivariate functional principal component analysis is a useful method for the joint model of multiple longitudinal markers, but most of these methods require a limited number of continuous longitudinal variables. For high-dimensional sparse data such as PET and MRI, the longitudinal information of each brain area is not used. Therefore, we plan to extend the functional data analysis method to a more general joint model that can handle high-dimensional longitudinal variables and obtain a more accurate AD dynamic prediction model. In addition，A central problem in neural mediation analysis is to identify which brain regions are positioned along the pathway between treatments and disease status. So we are also interested in identifying the brain areas intermediate between signal exposure or multiple exposures (including genetic information) and the development of brain diseases. Therefore, we plan to use the data to establish a high-dimensional functional mediation analysis method (including the mediation analysis method of survival outcome) to identify the brain area that plays a mediating role and quantify the mediation effect. |
| Investigator's Name: | Liuqing Peng |
| Proposed Analysis: | Alzheimer’s Disease Neuroimaging Initiative (ADNI) database focus on the collection of longitudinal assessments, magnetic resonance imaging and positron emission tomography imaging measures, as well as other biomarkers from blood and cerebrospinal fluid.Of those covariates collected in the cohort, many are time-varying. We apply for this data to solve the following two problems. The first problem we are interested in is to construct a joint model of multiple longitudinal variables based on functional data analysis to dynamically predict Alzheimer’s disease progression. Joint modeling (JM) of longitudinal and survival data is usually a popular framework in dynamic prediction and most JM approaches only investigated the association between a single longitudinal outcome and a time-to-event outcome. Including more longitudinal variables is expected to improve prediction accuracy. However, the shared random effect model and Bayesian framework would be computationally difficult or prohibitive when a large number of longitudinal variables are used. Multivariate functional principal component analysis is a useful method for the joint model of multiple longitudinal markers, but most of these methods require a limited number of continuous longitudinal variables. For high-dimensional sparse data such as PET and MRI, the longitudinal information of each brain area is not used. Therefore, we plan to extend the functional data analysis method to a more general joint model that can handle high-dimensional longitudinal variables and obtain a more accurate AD dynamic prediction model. In addition，A central problem in neural mediation analysis is to identify which brain regions are positioned along the pathway between treatments and disease status. So we are also interested in identifying the brain areas intermediate between signal exposure or multiple exposures (including genetic information) and the development of brain diseases. Therefore, we plan to use the data to establish a high-dimensional functional mediation analysis method (including the mediation analysis method of survival outcome) to identify the brain area that plays a mediating role and quantify the mediation effect. |
| Investigator's Name: | Limei Cao |
| Proposed Analysis: | Alzheimer’s Disease Neuroimaging Initiative (ADNI) database focus on the collection of longitudinal assessments, magnetic resonance imaging and positron emission tomography imaging measures, as well as other biomarkers from blood and cerebrospinal fluid.Of those covariates collected in the cohort, many are time-varying. We apply for this data to solve the following two problems. The first problem we are interested in is to construct a joint model of multiple longitudinal variables based on functional data analysis to dynamically predict Alzheimer’s disease progression. Joint modeling (JM) of longitudinal and survival data is usually a popular framework in dynamic prediction and most JM approaches only investigated the association between a single longitudinal outcome and a time-to-event outcome. Including more longitudinal variables is expected to improve prediction accuracy. However, the shared random effect model and Bayesian framework would be computationally difficult or prohibitive when a large number of longitudinal variables are used. Multivariate functional principal component analysis is a useful method for the joint model of multiple longitudinal markers, but most of these methods require a limited number of continuous longitudinal variables. For high-dimensional sparse data such as PET and MRI, the longitudinal information of each brain area is not used. Therefore, we plan to extend the functional data analysis method to a more general joint model that can handle high-dimensional longitudinal variables and obtain a more accurate AD dynamic prediction model. In addition，A central problem in neural mediation analysis is to identify which brain regions are positioned along the pathway between treatments and disease status. So we are also interested in identifying the brain areas intermediate between signal exposure or multiple exposures (including genetic information) and the development of brain diseases. Therefore, we plan to use the data to establish a high-dimensional functional mediation analysis method (including the mediation analysis method of survival outcome) to identify the brain area that plays a mediating role and quantify the mediation effect. |
