Integration of steady-state diffusion MRI with Neural Posterior Estimation (NPE) for post-mortem investigations

Post-mortem diffusion MRI plays a key role in investigative pipelines to characterise tissue microstructure, with long scan times facilitating the acquisition of datasets with improved spatial/angular resolution and reduced artefacts versus in vivo. Diffusion-weighted steady-state free precession (DW-SSFP) has emerged as a powerful technique for post-mortem imaging, achieving high SNR-efficiency and strong diffusion weighting in the challenging imaging environment of fixed tissue. However, the sophisticated signal forming mechanisms of DW-SSFP limit the characterisation of tissue properties with conventional parameter estimation routines. Here, I investigate the integration of DW-SSFP with neural posterior estimation (NPE), a parameter inference technique leveraging concepts from Bayesian statistics and machine learning to directly estimate P({theta} | S) (i.e. the posterior distribution of parameters {theta} given signal S). A key challenge is that diffusion attenuation in DW-SSFP is dependent on tissue relaxation properties (T1/T2) and transmit inhomogeneity (B1), which must be estimated experimentally and incorporated into the NPE network for accurate modelling. Using a Tensor representation and NPE to estimate P({theta} | S, T1, T2, B1) (i.e. conditioning on S and known T1/T2/B1), evaluations using synthetic DW-SSP data give excellent agreement to simulation ground truth even in the presence of non-Gaussian noise in low-SNR regimes. Evaluations using experimental DW-SSFP data (whole human post-mortem brain) give excellent agreement to non-linear least-squares estimates, with NPE providing 1000s of posterior samples in a matched evaluation time. Taken together, findings provide a framework to perform rapid parameter estimation with DW-SSFP, and an intuitive approach to incorporate conditional dependencies with NPE.