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: | Mason Kadem |
Institution: | McMaster University |
Department: | Biomedical Engineering |
Country: | |
Proposed Analysis: | a mathematical modelling approach based on the state of the art machine learning to identiy patients at the highest risk of dementia |
Additional Investigators | |
Investigator's Name: | Michael Noseworthy |
Proposed Analysis: | Differentiating mild cognitive impairment from preclinical Alzheimer’s disease (AD) that eventually progresses to AD dementia (ADD) is vital worldwide[1]. Diagnostic prediction models aide in early detection of AD prior to overt disease onset; yet the performance and explainability of these risk prediction models are limited. Furthermore, prior predictive models have included costly and invasive measures (e.g., cerebrospinal fluid (CSF) analysis of β-amyloid (Aβ42), P-tau, neurofilament light data, Aβ-positron emission tomography) which limits the availability of these biomarkers and their usage. Moreover, previous models relied on traditional statistics, which overlook complex interactions and nonlinear relationships among clinical features. Rapid, accurate, low-cost, easily accessible and early clinical evaluation of AD is critical for timely referrals to a memory clinic. To support healthcare decision-making and planning, and potentially reduce the burden of ADD, this research leverages the Alzheimer’s Disease Neuroimaging Initiative database and a mathematical modelling approach based on state-of-the-art explainable supervised machine learning to go beyond identifying 1) predictive markers of ADD and 2) patients at the highest risk of ADD. Specifically, a simple decision path with absolute thresholds for key clinical features is identified. This decision path provides an intuitive explanation of the key non-invasive features and decision process which are vital steps towards the clinical adoption of machine learning tools for ADD evaluation. |