Rxivist logo

Learn to Track: Deep Learning for Tractography

By Philippe Poulin, Marc-Alexandre Côté, Jean-Christophe Houde, L Petit, Peter F. Neher, Klaus H. Maier-Hein, Hugo Larochelle, Maxime Descoteaux

Posted 06 Jun 2017
bioRxiv DOI: 10.1101/146688

We show that deep learning techniques can be applied successfully to fiber tractography. Specifically, we use feed-forward and recurrent neural networks to learn the generation process of streamlines directly from diffusion-weighted imaging (DWI) data. Furthermore, we empirically study the behavior of the proposed models on a realistic white matter phantom with known ground truth. We show that their performance is competitive to that of commonly used techniques, even when the models are used on DWI data unseen at training time. We also show that our models are able to recover high spatial coverage of the ground truth white matter pathways while better controlling the number of false connections. In fact, our experiments suggest that exploiting past information within a streamline's trajectory during tracking helps predict the following direction.

Download data

  • Downloaded 2,333 times
  • Download rankings, all-time:
    • Site-wide: 2,935 out of 92,253
    • In neuroscience: 403 out of 16,420
  • Year to date:
    • Site-wide: 7,479 out of 92,253
  • Since beginning of last month:
    • Site-wide: 5,220 out of 92,253

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide


Sign up for the Rxivist weekly newsletter! (Click here for more details.)