DeepTract: A Probabilistic Deep Learning Framework for White Matter Fiber Tractography

Itay Benou, Tammy Riklin-Raviv

Abstract

We present DeepTract, a deep-learning framework for estimating white matter fibers orientation and streamline tractography. We adopt a data-driven approach for fiber reconstruction from diffusion weighted images (DWI), which does not assume a specific diffusion model. We use a recurrent neural network for mapping sequences of DWI values into probabilistic fiber orientation distributions. Based on these estimations, our model facilitates both deterministic and probabilistic streamline tractography. We quantitatively evaluate our method using the Tractometer tool, demonstrating competitive performance with state-of-the art classical and machine learning based tractography algorithms. We further present qualitative results of bundle-specific probabilistic tractography obtained using our method. The code is publicly available at: https://github.com/itaybenou/DeepTract.git.

Introduction

Tractography is the process of using diffusion MRI (dMRI) to visualize the physical connections—white matter fibers—in the living brain. This is crucial for studying neural connectivity and diagnosing neurological disorders. Traditionally, this process relies on 'model-based' approaches that make mathematical assumptions about how water molecules diffuse in tissue. However, these models often struggle with 'crossing fibers' (where multiple pathways overlap in one voxel) and require high-quality data that isn't always available.

Recent Deep Learning (DL) attempts have tried to solve this, but most were 'deterministic,' predicting only a single direction at each step. This ignores the inherent uncertainty in the data and prevents 'probabilistic tractography,' which is essential for capturing the full range of possible neural connections.

Methods

DeepTract shifts the paradigm from regression to sequential classification. Instead of looking at a single point in isolation, DeepTract uses a Gated Recurrent Unit (GRU)—a type of Recurrent Neural Network (RNN)—to process a sequence of diffusion-weighted image (DWI) values along a path. This allows the model to use the 'history' of the fiber to inform its next step. By feeding the model sequences of DWI values rather than raw 3D coordinates, the model becomes invariant to spatial transformations, making it more generalizable.

The DeepTract architecture for training (a) and tracking (b) processes.

The authors represent the sphere of possible directions as 724 discrete points (classes). The model outputs a probability distribution over these classes, plus an extra 'End-of-Fiber' (EoF) class to know when to stop. To improve training, DeepTract uses a Gaussian kernel to smooth labels across the sphere, teaching the model that predicting a direction 5 degrees off is better than predicting one 90 degrees off. Furthermore, the paper introduces 'Conditional fiber Orientation Distribution Functions.' Because the RNN remembers the previous steps, the predicted direction at point B is conditioned on the path taken from point A, mimicking the physical continuity of biological fibers.

Key Results

DeepTract was evaluated using the Tractometer tool and compared against state-of-the-art classical and DL methods using the ISMRM tractography challenge DWI phantom dataset. It was the only method in the comparison to successfully detect all 25 gold-standard fiber bundles.

When supervised by ground truth, it achieved the highest Overlap (69.3%) and F1 score (70.1), significantly outperforming other deep learning models like those by Poulin et al. or Wegmayr et al. It also recorded the lowest rates of invalid connections (IC) and non-connections (NC), likely due to the entropy-based termination criteria which prevents 'stray' streamlines when uncertainty is high.

Qualitative comparison of whole-brain tractography between DeepTract, MITK, and Ground Truth.

Limitations and Future Work

While the results on the ISMRM phantom are impressive, phantom data is cleaner than real human brain scans. The authors acknowledge that larger, real-world dMRI datasets are needed to further validate the model. Additionally, the precision is currently limited by the 724 discrete directions (approx. 3.5° resolution). Future work will focus on increasing this resolution and reducing the dependency on high-quality supervision data to ensure the model does not inherit biases from its training sources.

Summary updated

Abstract

We present DeepTract, a deep-learning framework for estimating white matter fibers orientation and streamline tractography. We adopt a data-driven approach for fiber reconstruction from diffusion weighted images (DWI), which does not assume a specific diffusion model. We use a recurrent neural network for mapping sequences of DWI values into probabilistic fiber orientation distributions. Based on these estimations, our model facilitates both deterministic and probabilistic streamline tractography. We quantitatively evaluate our method using the Tractometer tool, demonstrating competitive performance with state-of-the art classical and machine learning based tractography algorithms. We further present qualitative results of bundle-specific probabilistic tractography obtained using our method. The code is publicly available at: https://github.com/itaybenou/DeepTract.git.
BibTeX
@inproceedings{benou2019deeptract,
  author = {Itay Benou and Tammy Riklin-Raviv},
  title = {{DeepTract: A Probabilistic Deep Learning Framework for White Matter Fiber Tractography}},
  booktitle = {International Conference on Medical Image Computing and Computer-Assisted Intervention},
  year = {2019},
  url = {https://openreview.net/forum?id=W-gcDAYOCk},
}
Figure