Home - Imaging - New Software Tool Helps Water Molecules Sharpen the Picture
Research Brief

New Software Tool Helps Water Molecules Sharpen the Picture

SOLID improves diffusion MRI accuracy by salvaging data from outliers and noise.

Valerie Brown, Contributor
Thursday, July 26, 2018


Diffusion magnetic resonance imaging tracks the movement of water molecules in a given volume and is useful in studying brain connectivity, tissue microstructure and various central nervous system pathologies. But when the patient moves or the hardware shifts during a scan, some of the image slices could become partially corrupted, though they may still contain useful information. However, existing software tends to rely on model estimations to correct for geometric distortions and often spreads outlier data across image slices, distorting the data. The distorted data is then dropped from the scan, creating sampling gaps that makes model fitting more difficult.

Viljami Sairanen and colleagues at the University of Helsinki, Finland; University Medical Center Utrecht, Netherlands; and Cardiff University, Wales have developed a software tool to rescue the usable data from distorted slices. Called Slicewise OutLIer Detection (SOLID), the tool finds and flags these areas before the software attempts to correct their geometry. This increases model accuracy and speeds up the analysis considerably. SOLID is now available through the MATLAB graphical toolbox ExploreDTI.

To test SOLID’s accuracy, the researchers ran numerous simulations based on a partial single-individual data set. They also applied the tool to in vivo data from 54 neonatal patients in which the outliers were identified visually before the tool was applied. In both trials the tool performed as well as or better than previous approaches. The team suggests that, with further refinement, incorporation of SOLID into dMRI data analysis will be useful in large population studies. Their result was published in NeuroImage.