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: | Carlo Fabrizio |
Institution: | IRCCS Fondazione Santa Lucia |
Department: | Data Science Unit |
Country: | |
Proposed Analysis: | We intend to analyze ADNI neuroimaging data to validate a data-centric approach in developing a computer-aided diagnosis tool for AD. This application will follow Andrew Ng's new approach to AI modeling (https://landing.ai/data-centric-ai/). In particular we implement a “data boosting” procedure, selecting the best data to fill in the gaps in AI prediction accuracy, thus focusing on data management strategies instead of on model engineering practices. ADNI data will be used to train a Deep Learning model that will be subsequently validated on data from the Neuroimaging Lab of the IRCCS Fondazione Santa Lucia to test the efficiency of our data-boosting procedure. To use AI standardized best practices in healthcare imaging, we will work with the Medical Open Network in Artificial Intelligence framework (https://monai.io/). Furthermore, 3D attention maps will be generated on the trained model, allowing to highlight image areas which are relevant for classification purposes. Developing a deployment-ready, fast and accurate diagnostic tool for in vivo imaging would foster better and faster diagnosis, possibly helping Low-to-Middle-Income-Countries healthcare systems, often suffering from the lack of expert physicians. Our final objective is to build a clinical decision support system that will be deployed in a standardized and fully reproducible way. |
Additional Investigators | |
Investigator's Name: | Andrea Termine |
Proposed Analysis: | We intend to analyze ADNI neuroimaging data to validate a data-centric approach in developing a computer-aided diagnosis tool for AD. This application will follow Andrew Ng's new approach to AI modeling (https://landing.ai/data-centric-ai/). In particular we implement a “data boosting” procedure, selecting the best data to fill in the gaps in AI prediction accuracy, thus focusing on data management strategies instead of on model engineering practices. ADNI data will be used to train a Deep Learning model that will be subsequently validated on data from the Neuroimaging Lab of the IRCCS Fondazione Santa Lucia to test the efficiency of our data-boosting procedure. To use AI standardized best practices in healthcare imaging, we will work with the Medical Open Network in Artificial Intelligence framework (https://monai.io/). Furthermore, 3D attention maps will be generated on the trained model, allowing to highlight image areas which are relevant for classification purposes. Developing a deployment-ready, fast and accurate diagnostic tool for in vivo imaging would foster better and faster diagnosis, possibly helping Low-to-Middle-Income-Countries healthcare systems, often suffering from the lack of expert physicians. Our final objective is to build a clinical decision support system that will be deployed in a standardized and fully reproducible way. |