Water diffusion in biological tissues is highly sensitive to the structural organization and geometry of tissue substrates. Consequently, diffusion MRI is rapidly becoming an important imaging probes for the investigation of tissue microstructure in the human brain. In the first part of this study, we optimized methods for diffusion tensor imaging of the brain. We also developed novel tractography methods for estimating white matter pathways in the brain. Using these tools, we can now segment many of the major white matter pathways in the human brain. We also found that diffusion tensor imaging was very useful for evaluating the relationship between brain tumors and functional white matter tracts. Finally, we demonstrated that the simple diffusion tensor model breaks down when multiple tissue components are mixed in a measurement region. In this continuation study, we will develop a framework of innovative and advanced tools for diffusion image analysis. Novel approaches for constraining and characterizing white matter tractography results will be implemented. A probabilistic white matter atlas of diffusion tensor properties in the healthy human brain will be developed. Algorithms to automatically segment most of the individual major white matter tracts will be implemented using the probabilistic atlas. Automated segmentation of specific white matter tracts will be extremely valuable for testing specific hypotheses relating disease or behavior to white matter anatomy. Further, we will develop advanced voxel-based diffusion image analysis tools that will use multivariate diffusion feature statistics to measure and characterize differences between groups. We will develop an efficient 3D diffusion-weighted method that will acquire full-brain diffusion images with truly isotropic spatial resolution without spatial distortion. Measurement methods for characterizing diffusion in regions with complex diffusion behavior will be optimized. Pilot studies will be performed to evaluate if these innovative diffusion imaging methods are promising for clinical applications.
Showing the most recent 10 out of 47 publications
