Fuzzy Fibers: Uncertainty in dMRI Tractography - Scientific Visualization: Uncertainty, Multifield, Biomedical, and Scalable Visualization - page 76

coefficients. Since fibrous tissues such as nerve fiber bundles in the human brain con-
strain water molecules such that they diffuse more freely along fibers than orthogonal
to them, the apparent diffusivity depends on the direction of measurement, and allows
us to infer the main fiber direction.
Based on such data, tractography algorithms reconstruct the trajectories of major
nerve fiber bundles. The most classic variant is streamline tractography, in which
tracking starts at some seed point and proceeds in small steps along the inferred
direction. In its simplest form, this results in one space curve per seed point. It has
been observed that many of the resulting streamlines agree with known anatomy
[ 11 ]. Tractography is also supported by validation studies that have used software
simulations, physical and biological phantoms [ 25 ].
Tractography is currently the only technique for noninvasive reconstruction of
fiber bundles in the human brain. This has created much interest among neuroscien-
tists, who are looking for evidence of how connectivity between brain regions varies
between different groups of subjects [ 60 ], as well as neurosurgeons, who would like
to know the exact spatial extent of specific fiber bundles in individual patients.
However, drawing reliable inference from dMRI is challenging. Even though a
randomized controlled study has shown that using dMRI in cerebral glioma surgery
reduces postoperativemotor deficits and increases survival times [ 68 ], neurosurgeons
have observed that some methods for tractography underestimate the true size of
bundles [ 37 ] and they are still unsatisfied with the degree of reproducibility that is
achieved with current software packages [ 9 ].
In order to establish tractography as a reliable and widely accepted technique, it is
essential to gain a full understanding of its inherent sources of error and uncertainty.
It is the goal of this chapter to give an introduction to these problems, to present
existing approaches that have tried to mitigate or model them, and to outline some
areas where more work is still needed.
8.2 Noise and Artifacts
8.2.1 Strategies for Probabilistic Tractography
It is the goal of probabilistic tractography to estimate the variability in fiber bundle
reconstructions that is due tomeasurement noise. This is often referred to as precision
of the reconstructed bundle trajectory [ 33 ]. Due to additional types of error in data
acquisition and modeling, which will be covered later in this chapter, it is not the
same as accuracy (i.e., likelihood of a true anatomical connection) [ 35 ]. Current
approaches do not account for factors such as repositioning of the head or variations
in scanner hardware over time, which further affect repeatability in practice.
Rather than only inferring the most likely fiber direction, probabilistic approaches
derive a probability distribution of fiber directions from the data. The first generation
of probabilistic tractographymethods has done so by fitting the diffusion tensormodel