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: | Junhyoun Sung Junhyoun Sung |
Institution: | University of Washington |
Department: | Department of Biostatistics |
Country: | |
Proposed Analysis: | In this study, we explore the heterogeneity of atrophy patterns in AD using a data-driven framework which yields multiple distinct localized components of gray matter atrophy. We are interested in extracting biologically meaningful components and measuring their association. For this purpose, we quantify subject-level loss coefficients with non-negative least squares and regard them as re ection of how the corresponding components contribute to atrophy loss. We show the ability of those coefficients to distinguish diagnosis groups. Considering the normality assumption may not hold, we assess their correlation of Alzheimer's disease using Gaussian copula graphical models (GCGMs) rather than ordinary Gaussian graphical models (GGMs). Finally, we use a hierarchical agglomerative cluster analysis to nd potential subtypes for each AD subject. |
Additional Investigators | |
Investigator's Name: | Kwun Chuen Gary Chan |
Proposed Analysis: | In this study, we explore the heterogeneity of atrophy patterns in AD using a data-driven framework which yields multiple distinct localized components of gray matter atrophy. We are interested in extracting biologically meaningful components and measuring their association. For this purpose, we quantify subject-level loss coefficients with non-negative least squares and regard them as reection of how the corresponding components contribute to atrophy loss. We show the ability of those coefficients to distinguish diagnosis groups. Considering the normality assumption may not hold, we assess their correlation of Alzheimer's disease using Gaussian copula graphical models (GCGMs) rather than ordinary Gaussian graphical models (GGMs). Finally, we use a hierarchical agglomerative cluster analysis to nd potential subtypes for each AD subject. |