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: | Sheng XU |
| Institution: | The Hong Kong Polytechnic University |
| Department: | Department of Applied Mathematics |
| Country: | |
| Proposed Analysis: | To whom it may concern, Our research team led by Dr. Catherine Chunling Liu (Department of Applied Mathematics, The Hong Kong Polytechnic University, email: macliu@polyu.edu.hk) is doing a project related to the variable selection problems in functional regression with heteroscedasticity. We propose to analyze the clinical cognitive assessment outcomes in ANDI, for example, the mini-mental state examination (MMSE) score in the neurological examination, associated with the functional covariates, the longitudinal MRI measurements. We aim to select the significant regions of interest (ROIs) that affect the mean and variance of the cognitive assessment outcomes through a multiple functional regression model with heteroscedasticity. It is evident that MMSE might have non-normal and heavy-tailed distributions (Ma el al. 2018). Therefore, some ROIs might affect the variance of MMSE but hard to be identified. We plan to apply our proposed method to identify the significant functional predictors that affect the mean and the variance of MMSE simultaneously and separately. The following is the technical analysis procedure. The demographical and genetic data of each subject will also be used for eliminating the demographical and genetic effects on the cognitive assessment outcomes. A more robust loss function with concave penalties will be considered for functional regression so that the entire distribution of the outcomes could be characterized and the significant functional predictors, the signals in the MRI data, are able to be identified more accurately. Functional data analysis techniques have been applied for analyzing the data in ADNI study. (See Zhu ei al. 2014, Ma etal. 2018, among others.) However, to the best of our acknowledge, there are no methods that take the heteroscedasticity issues into consideration and are able to select the significant ROIs that affect the variance of the clinical cognitive outcomes. We believe that our proposed analysis for the data will discover more information and provide statistical evidence for the Alzheimer's disease studies. Reference: Ma, H., Li, T., Zhu, H., & Zhu, Z. (2018). Quantile regression for functional partially linear model in ultra-high dimensions. Computational Statistics & Data Analysis. Zhu, H., Khondker, Z., Lu, Z., & Ibrahim, J. G. (2014). Bayesian generalized low rank regression models for neuroimaging phenotypes and genetic markers. Journal of the American Statistical Association, 109(507), 977-990. |
| Additional Investigators | |
| Investigator's Name: | Catherine Chunling Liu |
| Proposed Analysis: | To whom it may concern, My research team is doing a project related to the variable selection problems in functional regression with heteroscedasticity. We propose to analyze the clinical cognitive assessment outcomes in ANDI, for example, the mini-mental state examination (MMSE) score in the neurological examination, associated with the functional covariates, the longitudinal MRI measurements. We aim to select the significant regions of interest (ROIs) that affect the mean and variance of the cognitive assessment outcomes through a multiple functional regression model with heteroscedasticity. It is evident that MMSE might have non-normal and heavy-tailed distributions (Ma el al. 2018). Therefore, some ROIs might affect the variance of MMSE but hard to be identified. We plan to apply our proposed method to identify the significant functional predictors that affect the mean and the variance of MMSE simultaneously and separately. The following is the technical analysis procedure. The demographical and genetic data of each subject will also be used for eliminating the demographical and genetic effects on the cognitive assessment outcomes. A more robust loss function with concave penalties will be considered for functional regression so that the entire distribution of the outcomes could be characterized and the significant functional predictors, the signals in the MRI data, are able to be identified more accurately. Functional data analysis techniques have been applied for analyzing the data in ADNI study. (See Zhu ei al. 2014, Ma etal. 2018, among others.) However, to the best of our acknowledge, there are no methods that take the heteroscedasticity issues into consideration and are able to select the significant ROIs that affect the variance of the clinical cognitive outcomes. We believe that our proposed analysis for the data will discover more information and provide statistical evidence for the Alzheimer's disease studies. Reference: Ma, H., Li, T., Zhu, H., & Zhu, Z. (2018). Quantile regression for functional partially linear model in ultra-high dimensions. Computational Statistics & Data Analysis. Zhu, H., Khondker, Z., Lu, Z., & Ibrahim, J. G. (2014). Bayesian generalized low rank regression models for neuroimaging phenotypes and genetic markers. Journal of the American Statistical Association, 109(507), 977-990. |
| Investigator's Name: | Catherine Chunling Liu |
| Proposed Analysis: | To whom it may concern, My research team is doing a project related to the variable selection problems in functional regression with heteroscedasticity. We propose to analyze the clinical cognitive assessment outcomes in ANDI, for example, the mini-mental state examination (MMSE) score in the neurological examination, associated with the functional covariates, the longitudinal MRI measurements. We aim to select the significant regions of interest (ROIs) that affect the mean and variance of the cognitive assessment outcomes through a multiple functional regression model with heteroscedasticity. It is evident that MMSE might have non-normal and heavy-tailed distributions (Ma el al. 2018). Therefore, some ROIs might affect the variance of MMSE but hard to be identified. We plan to apply our proposed method to identify the significant functional predictors that affect the mean and the variance of MMSE simultaneously and separately. The following is the technical analysis procedure. The demographical and genetic data of each subject will also be used for eliminating the demographical and genetic effects on the cognitive assessment outcomes. A more robust loss function with concave penalties will be considered for functional regression so that the entire distribution of the outcomes could be characterized and the significant functional predictors, the signals in the MRI data, are able to be identified more accurately. Functional data analysis techniques have been applied for analyzing the data in ADNI study. (See Zhu ei al. 2014, Ma etal. 2018, among others.) However, to the best of our acknowledge, there are no methods that take the heteroscedasticity issues into consideration and are able to select the significant ROIs that affect the variance of the clinical cognitive outcomes. We believe that our proposed analysis for the data will discover more information and provide statistical evidence for the Alzheimer's disease studies. Reference: Ma, H., Li, T., Zhu, H., & Zhu, Z. (2018). Quantile regression for functional partially linear model in ultra-high dimensions. Computational Statistics & Data Analysis. Zhu, H., Khondker, Z., Lu, Z., & Ibrahim, J. G. (2014). Bayesian generalized low rank regression models for neuroimaging phenotypes and genetic markers. Journal of the American Statistical Association, 109(507), 977-990. |
