Our brain is a complex network with multiple levels of organization in white matter (WM) and gray matter (GM). The axonal and dendritic organizations of local GM tissue form one of the structural bases of normal brain functions. In addition, recent histopathology evidence has consistently demonstrated alterations in GM architectures in several neurological diseases, e.g., the Alzheimer's disease, multiple sclerosis, autism and epilepsy. Capturing these microstructural changes using non-invasive imaging techniques is extremely important but has not been achieved. It is therefore critical to develop the next generation imaging technologies that can visualize complex microstructural organizations in human cortical and subcortical GM. Diffusion MRI is the main technique for non-invasive mapping of WM tracts and a key component of the human connectome project. However, its low spatial resolution and the distinct organization of GM microstructures have been the major barriers for this technique to be used for characterization of GM micro-architectures and their abnormalities. We propose to develop an ultra-high-resolution diffusion MRI technique for improved human GM microarchitecture characterization at a 0.5 x 0.5 x 0.5 mm3 or finer resolution. This represents a 64- fold reduction in individual pixel volume from the status quo (2 mm isotropic), a dramatic improvement in spatial resolution, and fits the requirement of RFA-EB-17-001. In this application, we first demonstrate that the current state-of-the-art clinical MRI systems meet the theoretical SNR limit necessary to support imaging at such a high resolution for at least the cortical GM region. We then demonstrate the feasibility of using selective excitation to significantly shorten the acquisition of three- dimensional (3D) high-resolution diffusion MRI data. This will be further aided by a fast 3D imaging sequence that combines compressed sensing and parallel imaging. Altogether, we expect to acquire ultra-high-resolution diffusion MRI data from local GM regions at approximately 0.5 mm isotropic resolution within 20 minutes and use it to map the structural organization of the human cortex and hippocampus (Aim 1). We will integrate the high-resolution local GM data with whole-brain diffusion MRI data at 2 mm resolution and generate a group- average atlas of the hippocampus and temporal lobe based on ultra-high-resolution diffusion MRI data from 20 healthy volunteers (Aim 2). The proposed approach promises to provide an important set of tools for neuroscientists and clinicians to investigate structural organizations and connectivity of human GM.
The human gray matter plays a central role in normal brain functions, but non-invasive imaging of human gray matter structure and connectivity has not been achieved mainly due to limited imaging resolutions. In this project, we propose to develop an ultra-high-resolution diffusion magnetic resonance imaging method based on three-dimensional fast imaging acquisition with selectively excitation, compressed sensing, and parallel imaging. The technique will achieve a 0.5 mm x 0.5 mm x 0.5 mm spatial resolution for imaging gray matter microstructures and connectivity, and we will use it to construct a group-average atlas of the hippocampus and temporal cortex.
