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: | Mina Gholami |
| Institution: | Tarbiat modares university |
| Department: | statistics |
| Country: | |
| Proposed Analysis: | Alzheimer's disease (AD) is an irreversible and chronic neurodegenerative disorder with progressive impairment of memory and other important mental functions. MRI scans provide detailed information about the internal anatomical structures and the morphology of brain tissues such as white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). They have been recognized as an important image biomarker for AD progression and have been widely studied to develop computer-aided diagnosis systems using pattern recognition methods (Hinrichs et al., 2009; Hinrichs et al., 2011; Hosseini-Asl et al., 2016; Jiao et al., 2017; Kloppel et al., 2008, epub; Liu et al., 2016; Liu et al., 2018; Yu et al., 2017). Instead of manually extracting features on the expert’s knowledge about the target domain, deep learning can discover the latent and discriminant representations of image data by incorporating the feature extraction into the task learning process. In addition, deep learning can construct multi-layer neural networks to transform image data to task outputs (e.g., disease/normal) while learning hierarchical feature representations from data. Longitudinal analysis of MRI images is important in modeling and measuring progression over time for a more accurate diagnosis of the disease. We present a classification framework based on a combination of convolutional and recurrent neural networks for longitudinal analysis of structural MR images at variable-length time series for AD diagnosis. Increasing accuracy, reducing network training time, and extracting features from images accurately are important issues on which the present research project is based. Criteria of accuracy and precision of competitive methods in simulation studies and analysis of real data are going to compare. |
| Additional Investigators | |
| Investigator's Name: | Taban Baghfalaki |
| Proposed Analysis: | Alzheimer's disease (AD) is an irreversible and chronic neurodegenerative disorder with progressive impairment of memory and other important mental functions. MRI scans provide detailed information about the internal anatomical structures and the morphology of brain tissues such as white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). They have been recognized as an important image biomarker for AD progression and have been widely studied to develop computer-aided diagnosis systems using pattern recognition methods (Hinrichs et al., 2009; Hinrichs et al., 2011; Hosseini-Asl et al., 2016; Jiao et al., 2017; Kloppel et al., 2008, epub; Liu et al., 2016; Liu et al., 2018; Yu et al., 2017). Instead of manually extracting features on the expert’s knowledge about the target domain, deep learning can discover the latent and discriminant representations of image data by incorporating the feature extraction into the task learning process. In addition, deep learning can construct multi-layer neural networks to transform image data to task outputs (e.g., disease/normal) while learning hierarchical feature representations from data. Longitudinal analysis of MRI images is important in modeling and measuring progression over time for a more accurate diagnosis of the disease. We present a classification framework based on a combination of convolutional and recurrent neural networks for longitudinal analysis of structural MR images at variable-length time series for AD diagnosis. Increasing accuracy, reducing network training time, and extracting features from images accurately are important issues on which the present research project is based. Criteria of accuracy and precision of competitive methods in simulation studies and analysis of real data are going to compare. |
