Diffusion Modelling and Analysis
Information about the diffusion of water molecules, as acquired in Diffusion MRI, can tell us a lot about the structure of the brain. This information can be used as a surrogate probe into the microstructure of brain tissue, particularly for axons that connect the neurons together in the brain. Modelling and analysis of diffusion imaging data allows us to measure changes in the microstructure of tissue and to infer things about disease processes as well as the direction of large axon bundles and thus the wiring of the brain.
Research in the area of Diffusion MRI Modelling and Analysis covers both aspects: voxel-wise analysis to look at microstructure, and global connectivity measures via tractography. We also investigate modelling techniques at a microscopic level, to relate imaging measures more directly to axonal fibre distributions and cellular properties. This modelling work is done in close collaboration with groups in Neuropathology (NDCN) and Microstructural imaging (NDCN, FMRIB Physics).
Structural connectivity, as measured using tractography, is a major theme in our research and has been extensively used for brain parcellation (e.g. identifying thalamic subdivisions), and is increasingly playing a role in multimodal connectivity research. As such we work closely with researchers in the Connectivity Modelling and Multimodal Analysis areas.
The methods generated by the group are available in FSL (via the FDT toolbox) such as TBSS (voxel-wise analysis of surrogates for microstructural changes), BEDPOST/PROBTRACKX (Bayesian probabilistic tractography and connectivity). In addition, much of this work feeds into the Human Connectome Project (HCP), the Developing Human Connectome Project (dHCP), the UK Biobank and their associated pipelines.
FDT is a toolbox within FSL that contains:
- Probabilistic Tractography (ProbtrackX)
- Bayesian Model Fitting (BEDPOST)
Other tools in FSL include:
- Analysis of local diffusion metrics (TBSS)
- Eddy-Current and Distortion Correction (EDDY and TOPUP)
Many of these tools are available in extremely fast GPU implementations.