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: | Cassandra Morrison |
| Institution: | McGill |
| Department: | Psychology |
| Country: | |
| Proposed Analysis: | MRI can be used to measure cortical thickness (a proxy for atrophy and neurodegeneration) and white matter hyperintensities (a proxy of cerebral small vessel disease). We know that cardiovascular risk factors (cholesterol levels, exercise, blood pressure) play a negative role in cognitive decline and dementia. We assume that the combination of MRI, lifestyle factors, and cognitive data will have heightened discriminatory power compared to the use of these factors individually. Using machine learning, I will combine cerebrovascular lifestyle risk factors, MRI features (e.g., cortical thickness, white matter hyperintensities), cognitive scores, and demographic information (sex, age, education, APOE status) to build predictive models. These models will predict if which people with subjective cognitive decline will actually decline to mild cognitive impairment and Alzheimer’s disease. The SCD model will detect which individuals with SCD are likely to decline cognitively, as well as the rate of decline We will ensure equal accuracy (sensitivity and specificity) for men and women. If there is sufficient data, we will also evaluate race. For each model, we will use a 10-fold cross-validation strategy, where 90% of the all the retrospective data available is used for training and model building, and the remaining 10% for testing at each of ten iterations through the data. |
| Additional Investigators | |
| Investigator's Name: | Louis Collins |
| Proposed Analysis: | Using machine learning, we will combine cerebrovascular lifestyle risk factors, MRI features (e.g., cortical thickness, white matter hyperintensities), cognitive scores, and demographic information (sex, age, education, APOE status) to build predictive models. These models will predict if which people with subjective cognitive decline will actually decline to mild cognitive impairment and Alzheimer’s disease. The SCD model will detect which individuals with SCD are likely to decline cognitively, as well as the rate of decline We will ensure equal accuracy (sensitivity and specificity) for men and women. If there is sufficient data, we will also evaluate race. For each model, we will use a 10-fold cross-validation strategy, where 90% of the all the retrospective data available is used for training and model building, and the remaining 10% for testing at each of ten iterations through the data. |
