High Definition Fiber Tracking Laboratory

Diffusion MRI has arisen as the only non-invasive way to map white matter bundles and assess their structural integrity in the human brain. With fast imaging sequences, diffusion MRI, particular its high angular resolution variants, can be acquired on standard clinical scanners. This advancement has gained considerable interest because of its roles in mapping human connectome and potential for accessing neuropsychological disorders. There is a growing interest in large-scale analysis of diffusion MRI to explore its promising applications in biomedical research as an imaging biomarker of neuropathology.

Beyond DTI

Studies have used DTI, a modeling method for diffusion MRI (dMRI), to track brain pathways in the individual brain. It offers a sketch of the brain’s circuitry showing possible cortical connections. However, DTI cannot map a substantial amount of important crossing structures and sometimes even worse, create false connections due to tracking failure in the crossing regions. To address this limitation, the novel diffusion MRI conducted will acquire hundreds of sampling directions (DTI acquires only 30 or 60) at multiple diffusion sensitizations (DTI acquires only one). Exponential improvements have also been demonstrated by stronger gradient coils to achieve higher-spatial-resolution scans at 1.25 mm isotropic resolution (a typical DTI has only 2.5 mm). This advanced dMRI acquisition renders construction of a high-angular-resolution and high-spatial-resolution tractography possible.

Fiber Tracking with High Accuracy

We developed a high-accuracy fiber tracking method powered by generalized q-sampling imaging (Yeh et al., 2010) and its derived tracking method (Yeh et al., 2013). The method was released as an open-source tool to the public known as “DSI Studio” (http://dsi-studio.labsolver.org).

In 2015, the International Society for Magnetic Resonance in Medicine (ISMRM) held an open competition to examine the performance of different diffusion MRI fiber tracking pipelines. Our method achieved the highest valid connection score among 96 different approaches submitted by a total of 20 groups all over the world (www.tractometer.org/ismrm_2015_challenge/)(92.49%, ID:03)(Maier-Hein et al., 2017).

Our fiber tracking method is a long-term collaborative effort between diffusion MRI radiologists, psychologists, and neurosurgeons (Fernandez-Miranda et al., 2012; Yeh et al., 2013b; Yeh et al., 2010). The approach was previously optimized using capillary phantoms and validated by experienced neurosurgeons (Yeh et al., 2013b), leading to its outperformance over other competing algorithms.

Histology and Cadaver Validation

We conducted several studies to show that our tracking approach correlates well with histology (Gangolli et al., 2017; Modo et al., 2016) and cadaver microdissection in mapping several fiber pathways (Fernandez-Miranda et al., 2012; Fernandez-Miranda et al., 2014; Meola et al., 2016a; Meola et al., 2015; Meola et al., 2016b; Wang et al., 2016; Wang et al., 2013; Yoshino et al., 2016) and is a major improvement over DTI tractography (Abhinav et al., 2014a; Abhinav et al., 2014c; Abhinav et al., 2014d; Yeh et al., 2013b).

From a clinical perspective, our tracking method provides accurate structural connectivity studies in patients with intracerebral lesions (Abhinav et al., 2014b; Fernandez-Miranda et al., 2012; Yeh et al., 2013a) with other numerous applications in neurological and psychiatric disorders (see more than 200 journal publications in the last three years using DSI Studio at http://dsi-studio.labsolver.org/publications), allowing for qualitative and quantitative white matter damage assessment, aiding in understanding lesion patterns of white matter structural injury, and facilitating innovative neurological and psychiatric applications.


Fang-Cheng (Frank) Yeh


Department of Neurological Surgery

UPMC Presbyterian, Suite B-400

200 Lothrop Street

Pittsburgh, PA 15213