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: | Wieske de Swart |
| Institution: | Radboud University Nijmegen |
| Department: | Data Science |
| Country: | |
| Proposed Analysis: | For my Master thesis I will work on computational diagnosis of Alzheimer's Disease with deep learning using multiple sources of data and missing modalities. I would like to continue the work of Tao Zhou et. al on "Effective feature learning and fusion of multimodality data using stage-wise deep neural network for dementia diagnosis" (2019). In this study a deep neural network is trained on the MRI, PET and genetic basis data of the ADNI dataset for the diagnosis of Alzheimer's disease and mild incognitive impairment. Their network consists of three stages, where the first stage is associated with each individual data modality, the second with each pair of modalities and the third with all three modalities. This model uses region-of-interest based features from the MRI and PET images as input for the network. The use of these manual features might limit the amount of information that can be learned from these images. To use the full potential of deep learning I would like to try a network with convolutional layers that uses the images as input data. I will start with trying to reproduce the model of Tao Zhou et. al and will then try to improve this model by using the full images in combination with convolutional layers. Results of my analysis on the proposed modified architectures will provide the basis for a more in depth analysis of how to develop effective deep neural networks in the presence of missing modalities, in particular in the context of AD's diagnosis. |
| Additional Investigators |
