Neural fiber prediction with deep learning

Neurosurgery is a highly demanding field of medicine. Its subject is the central nervous system, which governs all critical functions of the human body. In recent years, artificial intelligence has assisted neurosurgeons' work through decision support systems that help better plan and perform safe surgeries. This paper introduces a novel method for analyzing diffusion data from Magnetic Resonance Imaging. It uses a deep learning to produce maps of white matter bundles. This information is suitable for multiple applications, including the computation of whole-brain maps of fiber crossing points. Neurosurgeons and radiologists can use such knowledge for pre-operative planning and intra-operative navigation.