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: | Leonard Wee |
| Institution: | Maastricht University |
| Department: | Clinical Data Sceince |
| Country: | |
| Proposed Analysis: | Medical imaging is a wealth of data that could be mined using machine algorithms for image-based digital biomarkes of disease. As a cancer researcher, I am particularly interested in cognitive decline and how this might be exacerbated by different forms of chemo- and radiation therapy. The small-scale problem we are tackling at the moment is how to automatically calibrate image intensity in MR, uniformly resize images with minimal content loss and automatic segmentation to isolate only whole brain from images corrected for field, gradient and other sources of distortion. The ADNI dataset appears to be a highly relevant use case for us to develop on, leading to a process where vast amounts of incompatible brain imaging MR data might become more easily used for large scale deep-learning computer-vision studies. |
| Additional Investigators | |
| Investigator's Name: | Chaitanya Kulkarni |
| Proposed Analysis: | This research is a portion of a PhD project. Computer vision analysis across multiple datasets needs standardized images. Even corrected MR images are heterogeneous in regards to size and intensity scale and are are usually not annotated. The analysis plan is to obtain a rapid and small-footprint deep learning algorithm that resizes images and rescales intensity with minimal loss using a condition Generative Adversarial Network. In the process, we wish to also develop a semi-supervised algorithm to automatically segment only the whole brain and mask out the remaining information to enhance privacy when using large private MR collections for computer vision research. |
