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: | Mohammad Ashourzadeh |
| Institution: | Allameh Tabataba'i |
| Department: | Computer Science |
| Country: | |
| Proposed Analysis: | Title: Prediction of Alzheimer's disease using temporal fusion transformer Introduction: Alzheimer's disease is one of the psychological diseases of aging that is associated with changes in brain structure and decline in cognitive functions. Since predicting Alzheimer's disease is very important in order to prevent it and provide timely treatment, many researchers are trying to develop methods to predict this disease. One of the new and effective methods for predicting Alzheimer's disease is the use of neural networks with transformer architecture. Using attention modules, these networks can well discover complex patterns that exist in temporal data. In this research, we intend to predict Alzheimer's disease by using temporal fusion transformer architecture. suggested method: We will use existing medical data to predict Alzheimer's disease. First, the data related to the cognitive assessment indicators of the patients are collected in two different time periods, that is, before the onset of the disease and after the onset of the disease. Then, by using temporal fusion transformer, the complex patterns that exist in time data are identified and predicted for Alzheimer's disease. Conclusion: By using transformer networks and specifically temporal fusion transformer, we can achieve high accuracy for predicting Alzheimer's disease. This method can be used as an important tool in early diagnosis and treatment of this disease. This project can play an important role in improving the health of society. |
| Additional Investigators | |
| Investigator's Name: | Farzam Matinfar |
| Proposed Analysis: | Transformer networks are tools used in many natural language processing and machine vision applications. But in recent years, these networks have also been used in time series processing. In particular, in time series forecasting problems, Transformers are recognized as one of the best and most effective methods. The use of Transformer networks in time series forecasting is due to their features such as easy training and high speed, the ability to understand and understand more diverse models from data, and high capabilities to pay attention to data sequences. One of the methods of using transformers in predicting time series is the use of temporal fusion transformer. In this method, first, the time series data is divided into several parts. Then the temporal fusion transformer model is used to combine these parts as input so that accurate prediction of time series data can be made in the future. The use of temporal fusion transformer in time series forecasting improves the accuracy and efficiency of forecasted models, and in cases such as predicting temporal diseases such as Alzheimer's and diabetes, significantly improves the results obtained. |
