Exploring Diffusion Weighted Imaging with dipy
by Matteo Visconti di Oleggio Castello
Sun 04 June 2017
I love Python and its ecosystem of scientific packages because it makes it very easy to experiment with new techniques. My undergraduate research assistant Manon and I have been recently playing with Diffusion Weighted Imaging (DWI) using
dipy. For those unfamiliar with it, DWI is a neuroimaging technique that allows to infer the most likely direction of fiber tracts in in-vivo brains by using computational modeling. I'm interested in using it for my research in human face processing. It also has a certain aesthetic appeal, because in the end you can make nice images like this one (each line is a reconstructed tract, and the colors indicate the direction in the 3D space):
I made a tutorial on using
dipy "start to finish" for a recent event we had in our department. You can find it on github. Take it as an exploration—I think I would optimize certain steps before using it in an experimental context, but it helped me understand what I'm dealing with much better than reading endless papers on the method.
Link to the tutorial: https://github.com/mvdoc/grr_dti_tutorial/blob/master/dti_tutorial.ipynb
If you have feedback on this post, please let me know! Send me a tweet or an email at firstname.lastname@example.org.