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: | Zehra Önen |
Institution: | Koç University |
Department: | Industrial Engineering and Operations Management |
Country: | |
Proposed Analysis: | There has been considerable research effort about modeling progression and treatment of AD, however the question of screening the population for AD has not been investigated as has been done in other diseases. Screening policies for various cancers such as breast cancer and prostate cancer have been extensively studied. Chronic diseases such as type II diabetes, infectious diseases such as HIV have also been modeled. This research focuses on developing optimal population screening policies for AD using Markov Decision Process (MDP) and Partially Observable MDP models. The states of the model include different stages of AD progression and treatment states corresponding to the severity of the illness. The screening tool we consider is the MMSE test score. The value function takes into consideration quality-adjusted life years and treatment costs. We have already built the elementary MDP version and implemented it using figures we obtain from published studies. However, these figures may be based on different studies and therefore lead to consistency issues. Moreover, we are not able to observe how transition probabilities change over time (i.e. with patient’s age) because we do not have access to underlying data. Our main objective by requesting ADNI data is to implement our models as consistently as possible and find optimal screening policies that realistically reflect current state of knowledge. Using ADNI data we aim to compute realistic transition probabilities between different states of the system. We also plan on incorporating clinical, imaging and/or genetic data to, see if these probabilities fit into any statistical distribution that will help us derive structural properties of the model. Also studying the effect of treatment for AD patients is of interest to use. Finally we wish to study sensitivity and specificity values of MMSE test using ADNI data and use these figures in our implementation. |
Additional Investigators | |
Investigator's Name: | Serpil Sayın |
Proposed Analysis: | There has been considerable research effort about modeling progression and treatment of AD, however the question of screening the population for AD has not been investigated as has been done in other diseases. Screening policies for various cancers such as breast cancer and prostate cancer have been extensively studied. Chronic diseases such as type II diabetes, infectious diseases such as HIV have also been modeled. This research focuses on developing optimal population screening policies for AD using Markov Decision Process (MDP) and Partially Observable MDP models. The states of the model include different stages of AD progression and treatment states corresponding to the severity of the illness. The screening tool we consider is the MMSE test score. The value function takes into consideration quality-adjusted life years and treatment costs. We have already built the elementary MDP version and implemented it using figures we obtain from published studies. However, these figures may be based on different studies and therefore lead to consistency issues. Moreover, we are not able to observe how transition probabilities change over time (i.e. with patient’s age) because we do not have access to underlying data. Our main objective by requesting ADNI data is to implement our models as consistently as possible and find optimal screening policies that realistically reflect current state of knowledge. Using ADNI data we aim to compute realistic transition probabilities between different states of the system. We also plan on incorporating clinical, imaging and/or genetic data to, see if these probabilities fit into any statistical distribution that will help us derive structural properties of the model. Also studying the effect of treatment for AD patients is of interest to use. Finally we wish to study sensitivity and specificity values of MMSE test using ADNI data and use these figures in our implementation. |