UC Berkeley

A sensitive and accurate imaging technique capable of tracking the disease progression of Alzheimer’s Disease (AD) driven amnestic dementia would be beneficial. A currently available method for pathology detection in AD with high accuracy is Positron Emission Tomography (PET) imaging, despite certain limitations such as low spatial resolution, off-target error, and radiation exposure. Non-invasive magnetic resonance imaging (MRI) scanning with quantitative magnetic susceptibility measurements can be used as a complementary tool. The human brain can be magnetized by the strong magnetic field of an MRI scanner. The corresponding magnetic susceptibility can be computed from MRI signals and is found to be in the order of 10^(-8). In AD, the accumulation of amyloid plaques, tau neurofibrillary tangles, and iron are crucial signatures of the underlying neurodegenerative process. Amyloid and tau proteins are diamagnetic while biological iron is paramagnetic. The opposing magnetic susceptibility of these sources complicates the quantification of susceptibility changes in the brain. To date, quantitative susceptibility mapping (QSM) has widely been used in tracking deep gray matter iron accumulation in AD. The present work proposes that by compartmentalizing quantitative susceptibility into paramagnetic and diamagnetic components, more holistic information about AD pathogenesis can be acquired. Computational susceptibility-source separation techniques such as DECOMPOSE-QSM can be used to compartmentalize paramagnetic and diamagnetic susceptibilities on a voxel level. In this work, we demonstrated significant voxel-wise differences between AD patients and healthy controls using paramagnetic and diamagnetic susceptibility derived from DECOMPOSE-QSM. Particularly, diamagnetic component susceptibility (DCS) have shown promise in tracking protein accumulation in the grey matter (GM), demyelination in the white matter (WM), and relevant changes in the cerebrospinal fluid (CSF). Through regions of interest (ROI) based analysis, we observed meaningful associations between pathological measures vs paramagnetic and diamagnetic susceptibility in several GM regions typically affected by tau deposition in AD. Finally, we explored the relationship of susceptibility distribution parameters with clinical, pathological, and genetic markers of AD. It has been found that the separated paramagnetic and diamagnetic susceptibility can be used to track pathological neurodegeneration in different tissue types and regions of the brain. With the initial evidence, we believe the usage of compartmentalized susceptibility demonstrates substantive potential as an MRI-based technique for tracking AD-driven neurodegeneration.