NA-MIC Project Weeks

Back to Projects List

DMRI population analysis library bridge

Key Investigators

• Ebrahim Ebrahim (Kitware, USA)
• Arthur Chakwizira (BWH, USA)

Project Description

We’ve been working on a python library, currently (badly) named abcdmicro, for diffusion MRI population analysis. Its goal is to make it easy to have the tools you need for processing population brain diffusion MRI in one convenient-to-set-up python package, with normally disparate processing steps getting linked together nicely. Slicer isn’t where one would typically do large population analysis, but it is an excellent for visualizing examples while putting together a pipeline, and it’s excellent for interacting with results. For this reason, I’d like to look into bridging abcdmicro with Slicer.

Objective

• Make it easy for someone who is building a dmri processing pipeline with abcdmicro to try out the steps of their pipeline in Slicer.

Approach and Plan

• Create vtk-mrml-based versions of some of the Resources in abcdmicro and conversion utilities that allow them to be created and used.

Progress and Next Steps

Added two example notebooks to abcdmicro. This helps introduce the functionality.

One notebook works up through NODDI estimation and tract segmentation for a particular subject:

The other demonstrates multimodal population template construction:

In a discussion with Arthur, here are some future directions we identified as valuable:

• Adding more pre-processing such as motion correction, gibbs ringing correction, etc.
• Incorporating strucural mri processing and handling co-registration
• Adding more microstructure models
• Support DICOM as an on-disk format, and consider having dicom-nifti conversion
• Adding (optional?) experimental parameters to DWI class (or a subclass?) to support things like time-dependent diffusion mri

Illustrations

Coming soon?

Background and References

• abcdmicro is the library we have currently been working on for which I’d like to experiment with bridging with Slicer.