Diffusion tensor based global tractography of human brain fiber bundles

Abstract

Tractography is a non-invasive process for reconstruction, modelling and visualization of neural fibers in the white matter (WM) of the human brain. It has emerged as a major breakthrough for neuroscience research due to its usefulness in clinical applications. In this research, we have investigated deterministic and probabilistic tractography approaches. We have evaluated the performance of different approaches on fiber bundle tracking using diffusion tensor imaging (DTI) data having multiple gradient directions. Experimental results show that global tractography is best for reconstruction of kissing and crossing fibers compared to deterministic tractography. We have also shown that DTI acquisition with a higher number of gradient directions provides better tracking results given limitations of acquisition and computation time.