As longevity increases, Alzheimer's disease (AD) is becoming more prevalent, with 5.4 million Americans living with AD in 2012, costing society more than $100 billion per year. In this study we will develop high resolution in vivo anatomical magnetic resonance imaging (MRI) methods for evaluating microstructures of the medial temporal lobe (MTL), where neurofibrillary tangles first appear in early AD and beta-amyloid deposits accumulate, leading eventually to dementia and memory loss. The proposed imaging techniques may contribute to early detection and improved understanding of AD. Reliable early-disease-stage imaging biomarkers are critically important in trials evaluating targeted interventions like anti-beta-amyloid vaccines. Although our methods apply to whole brain imaging, we will optimize and validate using anatomical targets in the MTL. This is a technically challenging area for MRI due to its location relatively deep in the brain and near to areas of differing magnetic susceptibility. Specifically, we aim to image structures within the perirhinal cortex (specifically layers II - III and cortical thickness), entorhinal cortex (verrucae that contin cell rich islands in layer II) and hippocampus (subfields and molecular layer). From ex vivo imaging, we know the resolutions required to see these structures.
We aim to achieve in vivo resolutions of 3003 um3 and 1502x1200 um3. We will integrate well-developed methods with experimental approaches to enable robust, efficient, high resolution T1-weighted imaging. Receive coil design, field strength and acquisition time contributes substantially to SNR (therefore resolution). We will use the largest-N arrays, 3 T and 7 T fields, and scan times <= 3 hours. To ensure a narrow point spread function (PSF) despite subject motion during long scans, we will integrate real-time navigator-based motion, frequency drift and first order shim tracking into a low distortion MEMPRAGE with inner-loop GRAPPA (for better contrast and PSF). Data will be streamed to an external storage device for offline processing. Images will be reconstructed with optimally combined channels, optimal GRAPPA and complex averaging. Prior information may be added from lower-resolution images to decrease scan time. Ten ex vivo MTL samples (5 healthy, 5 AD) will be imaged at 1003 um3 so that all structures are visible and confirmed with histology. Volumes will be down-sampled to determine required resolution for in vivo volumes in 10 healthy volunteers and 20 older participants (10 healthy and 10 with mild cognitive impairment). An expert neuroanatomist and trained research assistant, blinded to acquisition parameters, will identify structures in all in vivo and ex vivo volumes. Image sharpness will be evaluated theoretically and empirically. In vivo motion corrected T2 imaging will be evaluated at 5003 um3 and 2502x2000 um3. With the resulting technology and knowledge, we will propose (1) a practical 3 T protocol for AD clinical trials, compatible with the mMR (MR-PET) system, (2) more intensive MR protocols for AD research at 3 T and 7 T and (3) a protocol for extremely high resolution tissue imaging, with a list of visible structures for each.

Public Health Relevance

One of the first areas to be affected by Alzheimer's disease (AD) is entorhinal cortex, with its microscopic layers differentially affected, followed by damage to the hippocampal formation and eventually the whole brain. We will develop extremely high isotropic resolution whole brain MR imaging methods, focusing on the structures of the medial temporal lobe, for researching and diagnosing AD and evaluating treatments.

National Institute of Health (NIH)
Exploratory/Developmental Grants (R21)
Project #
Application #
Study Section
Biomedical Imaging Technology Study Section (BMIT)
Program Officer
Hsiao, John
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Massachusetts General Hospital
United States
Zip Code
Tisdall, M Dylan; Reuter, Martin; Qureshi, Abid et al. (2016) Prospective motion correction with volumetric navigators (vNavs) reduces the bias and variance in brain morphometry induced by subject motion. Neuroimage 127:11-22
Reuter, Martin; Tisdall, M Dylan; Qureshi, Abid et al. (2015) Head motion during MRI acquisition reduces gray matter volume and thickness estimates. Neuroimage 107:107-15