Submillimeter resolution diffusion MRI of the medial temporal lobe in the earliest stages of Alzheimer's disease

Project Summary. Beginning in the medial temporal lobe, an interplay of neurofibrillary tangles and amyloid
plaques contributes to neurodegeneration in Alzheimer’s disease (AD). By the time cortical and hippocampal
volume loss are detected with structural imaging, there has been extensive neuronal loss. Since neuroprotective
therapies need to be administered prior to irreversible tissue loss, there is an unmet need for early detection of
cortical alterations in the medial temporal lobe. Given the importance of connectivity centered in the
hippocampus for long-term memory, early markers of microstructural degeneration and decreased connectivity
could serve as diagnostic biomarkers early in the disease course. Specifically, tau pathology starting in the
entorhinal cortex impacts the perforant pathway, the main fiber tract carrying input from the entorhinal cortex to
the hippocampal formation. While in vivo diffusion MRI demonstrated evidence of perforant pathway
degeneration in aging, detecting degeneration of the perforant pathway in vivo in AD has thus far remained
elusive due to the small size of this structure.
Recent diffusion MRI models such as NODDI have shown promise in studying microstructural degeneration
in AD, which may be detectable before frank atrophy. However, cortical thinning due to atrophy may confound
estimates of advanced diffusion metrics because of partial volume effects at low resolution. Observing
microstructural alterations in the medial temporal lobe cortices requires submillimeter resolution given that their
thickness can be as low as 1.5 mm. Unfortunately, high-resolution diffusion MRI in the medial temporal lobe is
hampered by technical and physiological barriers. The proximity of medial temporal lobe cortices to air-tissue
interfaces leads to large B0 inhomogeneity, severe signal pileup, and distortions in diffusion-weighted echo planar
imaging (EPI). A technical leap is needed to address the artifacts that plague EPI and provide high-resolution,
high-fidelity, and high-SNR diffusion MRI of the medial temporal lobe. We propose diffusion MRI technology that
will eliminate distortions, minimize blurring, and boost SNR to provide high-resolution, high-fidelity, and high-
SNR diffusion MRI of the medial temporal lobe in AD. We will develop a vendor-agnostic, distortion-free EPI
acquisition to eliminate distortions and signal pileups and provide sensitive measurements amenable to high-
resolution tractography and microstructural modeling. We will combine this with a motion-corrected volumetric
encoding strategy to boost SNR and motion robustness. We will scan 15 subjects each in groups of mild cognitive
impairment (MCI), AD, and controls using the developed technology, and test the hypothesis that perforant
pathway fiber density and cross section and microstructural metrics from NODDI will be significantly reduced in
the medial temporal lobe cortices of MCI and AD patients compared to cognitively healthy older adults. The
developed technology will offer a noninvasive imaging biomarker of cortical neurodegeneration early in the
disease course, rendering neuroprotective treatments viable before the onset of irreversible neuronal loss.

Grant Number: 5R21AG082377-02
NIH Institute/Center: NIH
Principal Investigator: Berkin Bilgic