PROJECT SUMMARY Sensory and motor dysfunctions are potential biomarkers of preclinical Alzheimer’s disease (AD) as they may precede cognitive impairments. However, how the sensory and motor dysfunctions affect the brain functional connectivity (FC) between white matter (WM) and gray matter (GM) remains uninvestigated. This is because previous functional MRI analyses have overlooked WM due to weaker signal in WM. We recently demonstrated that FC between WM and GM (i.e., WM-GM FC) is reliably detectable and sensitive to AD, confirming its potential in capturing brain function and its changes. However, three major methodological obstacles hinder accurate evaluations and analyses of WM-GM FC at multiple scales. First, the unique topology of WM-GM FC network cannot be adequately modeled by a conventional graph model. Second, the most advanced WM spatial smoothing method, which requires diffusion MRI data for guidance, is complex and impractical for standalone fMRI data. Third, the elongated shape of WM functional architecture compromises traditional group-level voxel- scale statistical analysis. Therefore, our overall goal is to develop and apply a novel family of methods to investigate multi-scale alterations in the WM-GM functional connectome resulting from sensory and motor dysfunctions in preclinical AD. This goal will be achieved through two aims. Aim 1 is to develop innovative methods to characterize, detect and analyze WM-GM FC at multiple scales, including network-, region- and voxel- scales. Specifically, we will develop a bipartite-graph model to characterize WM-GM FC networks and quantify network properties, create a 4D atlas composed of diffusion-informed smoothing kernels for all WM voxels to simplify WM smoothing for any fMRI images, and develop a functional tract-based spatial statistics (fTBSS) method to optimize group-level voxel-scale analysis. Aim 2 is to apply the developed methods to investigate changes in WM-GM FC at multiple scales resulting from motor dysfunctions in preclinical AD. Exploiting our developed methods and existing databases, we will test three hypotheses. 1) specific WM- GM FC network metrics, especially within the somatomotor-related networks, may alter in preclinical AD subjects with MD relative to elderly controls and more altered WM-GM FC metrics may be associated with more severe MD; 2) the relationship between MD and WM-GM FC metrics may be moderated by other biological factors (e.g., sex, APOE ε4 status and brain atrophy). 3) the WM-GM FC metrics may act as mediators in the relationship between amyloid deposit and MD. The outcomes of this project will fill the gaps in our knowledge of how sensory and motor dysfunctions influence WM-GM functional connectome in preclinical AD and enrich the set of biomarkers for prediction of early AD, which will eventually enhance the well-being of both the aging population and their caregivers. The released code and atlases will benefit a broad community of investigators in...