Exploring the possibilities with deep learning to compute shape measures of the brain's white matter connections

Motivation: Studies have shown the potential of tractography shape measures to provide insight into the brain's structural connections and their relationship to human cognition. However, existing shape computation methods can be highly time-consuming when dealing with large-scale tractography datasets. Goal(s): We investigate the possibility of deep learning to compute shape measures of the brain's white matter connections. Approach: We propose a novel framework that leverages a point cloud representation of tractography to compute shape measures. Results: TractShapeNet outperforms other point cloud-based models for shape computation. Results demonstrate that our approach enables faster and more efficient shape-measure computation than the conventional DSI-Studio. Impact: We investigate the possibility of deep learning to compute shape measures of the brain's white matter connections without intermediate steps to convert geometric tractography streamline data to an image data representation using a voxel grid with our novel framework, TractShapeNet.