Ground-truth effects in learning-based fiber orientation distribution estimation in neonatal brains
This work addresses the problem of accurate neonatal brain imaging for medical researchers, though it appears to be an incremental improvement in ground truth selection for an existing method.
The study investigated whether different ground truth reconstruction methods affect deep learning-based fiber orientation distribution estimation in neonatal brains, finding that single-shell three-tissue constrained spherical deconvolution (SS3T-CSD) produced more realistic fiber ratios and maintained robust performance across age groups compared to the commonly used multi-shell multi-tissue method.
Diffusion Magnetic Resonance Imaging (dMRI) is a non-invasive method for depicting brain microstructure in vivo. Fiber orientation distributions (FODs) are mathematical representations extensively used to map white matter fiber configurations. Recently, FOD estimation with deep neural networks has seen growing success, in particular, those of neonates estimated with fewer diffusion measurements. These methods are mostly trained on target FODs reconstructed with multi-shell multi-tissue constrained spherical deconvolution (MSMT-CSD), which might not be the ideal ground truth for developing brains. Here, we investigate this hypothesis by training a state-of-the-art model based on the U-Net architecture on both MSMT-CSD and single-shell three-tissue constrained spherical deconvolution (SS3T-CSD). Our results suggest that SS3T-CSD might be more suited for neonatal brains, given that the ratio between single and multiple fiber-estimated voxels with SS3T-CSD is more realistic compared to MSMT-CSD. Additionally, increasing the number of input gradient directions significantly improves performance with SS3T-CSD over MSMT-CSD. Finally, in an age domain-shift setting, SS3T-CSD maintains robust performance across age groups, indicating its potential for more accurate neonatal brain imaging.