Deep Learning Captures More Accurate Diffusion Fiber Orientations Distributions than Constrained Spherical Deconvolution
This work addresses the need for more biologically relevant diffusion MRI models in neuroimaging, though it appears incremental as it builds on existing CSD methods.
The study tackled the problem of accurately estimating intra-voxel fiber orientation distributions from diffusion MRI by comparing deep learning with constrained spherical deconvolution (CSD), finding that deep learning captures more accurate distributions by exploiting additional information in the diffusion signal.
Confocal histology provides an opportunity to establish intra-voxel fiber orientation distributions that can be used to quantitatively assess the biological relevance of diffusion weighted MRI models, e.g., constrained spherical deconvolution (CSD). Here, we apply deep learning to investigate the potential of single shell diffusion weighted MRI to explain histologically observed fiber orientation distributions (FOD) and compare the derived deep learning model with a leading CSD approach. This study (1) demonstrates that there exists additional information in the diffusion signal that is not currently exploited by CSD, and (2) provides an illustrative data-driven model that makes use of this information.