PROJECT SUMMARY
Advanced Bayesian statistical methods for the analysis of functional magnetic resonance imaging (fMRI),
developed in parent grant R01EB027119, produce accurate measures of functional brain organization in
individual subjects. As a result, they are uniquely suited for advancing the development of fMRI-based brain
biomarkers for Alzheimer’s Disease (AD) and Mild Cognitive Impairment (MCI). Accurate biomarkers are
needed for early diagnosis and for identification of clinical trial participants who are likely to exhibit sufficient
decline across the trial to adequately test the intervention. Functional MRI-based biomarkers may serve as an
inexpensive, non-invasive, and widely-available first-line screening measure before positron emission
tomography (PET) imaging is used. The Alzheimer’s Disease Neuroimaging Initiative (ADNI) was launched in
2004 to develop and validate biomarkers for AD clinical trials. A principal goal of ADNI-3, the latest iteration of
ADNI, is to promote the development of diagnostic models and precision medicine approaches to identify
patients for therapeutic clinical trials, including the use of resting-state fMRI (rs-fMRI). In this supplement
project, we will apply the methods developed in the parent grant to extract features related to the topological
functional organization of the brain, functional connectivity, and texture of functional networks. These features
will be used to develop diagnostic and prognostic models to predict current disease status in ADNI-3.
Additionally, we will utilize the overlap between the ADNI-2 and ADNI-3 cohorts to investigate conversion to
MCI or AD. We will validate these models using a large independent holdout set to ensure rigor and
generalizability in our results. This project leverages an existing long-term collaborative environment between
Indiana University, Bloomington, and the University of Utah.