When planning brain tumor surgery, neurosurgeons use MRI images to assess the location of tissues affected by the tumor. A key question in this planning is the location and/or displacement of important neuronal connections. For visualization of neuronal connections surgeons rely on fiber tractography results derived from diffusion MRI acquisitions. Current fiber tractography methods however fail to individually detect fiber bundles crossing at angles less than 40. This fiber detection limitation hampers the reliability of fiber tractography in neurosurgery and neuroscience research. Indeed, in neurosurgery applications, tractography struggles with absent or limited visualization of the lateral corticospinal tract, temporal projections of the arcuate fasciculus and anterior optic radiations, nerve bundles essential for preservation of respectively motor, language and visual function. This proposal is dedicated to the development of novel fiber identification methods for diffusion MRI inspired by MR Fingerprinting. In fingerprinting approaches diffusion weighted signals are matched to a pre-computed signal library of potential fiber configurations. Our preliminary data show that fingerprinting-based fiber identification makes better use of the information available in the diffusion MRI measurement and hence outperforms current methods. This improved characterization of the diffusion signal will better inform tractography algorithms on the underlying tissue microstructure, increasing adherence of tractography results to the biological truth. The initial goals of the proposal are to establish the fiber identification performance of the proposed method using simulations, Human Connectome Protocol datasets and a biomimetic hollow fiber phantom in order to achieve smaller angular resolution and to validate the improved tractography results in an animal model. The progress made in these initial goals will be employed to retrospectively assess the impact of the proposed Fingerprinting-based fiber identification on fiber tractography in in vivo brain, the final goal of the proposal. Attainment of these goals will significantly further the adherence of tractography to tissue microstructure aiding in the understanding and visualizing of brain structure in both fundamental research and clinical applications.
We will develop tools to improve the visualization of the brain's neuronal connections for brain surgery and neuroscientific studies. Current methods fail to identify all connections crossing at a single point hence limiting the reliability of the visualization for neurosurgery and neuroscience research. The proposed work will improve our ability to identify crossing connections and increase the precision of the visualization in general.
