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: | Hamid Jazayeriy |
| Institution: | Babol Noshirvani University of Technology |
| Department: | Computer Engineering |
| Country: | |
| Proposed Analysis: | Improvements of health care technologies lead to an increase on the life expectancy and longevity index. One of the diseases related to the aging is Alzheimer's Disease (AD). In recent decades, due to increasing aging people the probability of Alzheimer's disease has been increased. Although no precise cure has been discovered to stop Alzheimer's disease yet, the available drugs only slow down the progress of the disease. Moreover, the expenditures of this disease are increasing that make Alzheimer's disease the most expensive chronic disease. The complicated mechanism of Alzheimer's disease has challenged the early diagnosis of this disease. By discovering of the stages of Alzheimer's disease, experts can diagnose the disease progresses. One of the biomarkers in Alzheimer's diagnosis is neuroimaging. MRI and PET neuroimaging are two common methods in Alzheimer's clinical diagnosis, as well fusion of Multi-Modalities can provide supplementary information in Alzheimer's diagnosis. Therefore, automatic multi-class diagnosis of Alzheimer's based on the fusion of MRI and PET images is the main goal of this research. The steps of solving the problem in this study are as follows: (1) preprocessing of images (2) fused images based on deep learning (3) feature extraction and learning based on attention model of deep learning Algorithm (4) classification of disease progression stages (AD, eMCI, lMCI and NC) (5) deep neural network interpretability. Although many studies have been presented for the automatic classification of Alzheimer's disease, the improvement of each of the above steps may improve the automatic classification and interpretability. Considering the successful performance of the deep networks in Alzheimer's diagnosis, in this proposal, deep learning will be used along with models based on spatial and channel attention. First, the MRI and PET images are fused by the deep learning, and then the spatial and channel attention model is used to develop more important features and suppress poor features. Then, for feature extraction deep networks are used to learn more features. Also, considering that the performance of deep networks in the diagnosis of Alzheimer's disease is not clearly understandable for the end user, interpretable methods will be used to explain the performance of the network in the diagnosis of Alzheimer's disease in a visual way to build trust in experts. The data used in this research are from well-known global databases. The proposed algorithm will be evaluated by common criteria and compared with previous recent studies. |
| Additional Investigators | |
| Investigator's Name: | Somayeh Mirzai |
| Proposed Analysis: | Improvements of health care technologies lead to an increase on the life expectancy and longevity index. One of the diseases related to the aging is Alzheimer's Disease (AD). In recent decades, due to increasing aging people the probability of Alzheimer's disease has been increased. Although no precise cure has been discovered to stop Alzheimer's disease yet, the available drugs only slow down the progress of the disease. Moreover, the expenditures of this disease are increasing that make Alzheimer's disease the most expensive chronic disease. The complicated mechanism of Alzheimer's disease has challenged the early diagnosis of this disease. By discovering of the stages of Alzheimer's disease, experts can diagnose the disease progresses. One of the biomarkers in Alzheimer's diagnosis is neuroimaging. MRI and PET neuroimaging are two common methods in Alzheimer's clinical diagnosis, as well fusion of Multi-Modalities can provide supplementary information in Alzheimer's diagnosis. Therefore, automatic multi-class diagnosis of Alzheimer's based on the fusion of MRI and PET images is the main goal of this research. The steps of solving the problem in this study are as follows: (1) preprocessing of images (2) fused images based on deep learning (3) feature extraction and learning based on attention model of deep learning Algorithm (4) classification of disease progression stages (AD, eMCI, lMCI and NC) (5) deep neural network interpretability. Although many studies have been presented for the automatic classification of Alzheimer's disease, the improvement of each of the above steps may improve the automatic classification and interpretability. Considering the successful performance of the deep networks in Alzheimer's diagnosis, in this proposal, deep learning will be used along with models based on spatial and channel attention. First, the MRI and PET images are fused by the deep learning, and then the spatial and channel attention model is used to develop more important features and suppress poor features. Then, for feature extraction deep networks are used to learn more features. Also, considering that the performance of deep networks in the diagnosis of Alzheimer's disease is not clearly understandable for the end user, interpretable methods will be used to explain the performance of the network in the diagnosis of Alzheimer's disease in a visual way to build trust in experts. The data used in this research are from well-known global databases. The proposed algorithm will be evaluated by common criteria and compared with previous recent studies. |
