Evaluating the performance of poroelastic and poroviscoelastic models in intrinsic MR elastography

Motivation: Previous work showed that viscoelastic property reconstruction results in data-model mismatch and non-unique solutions in Intrinsic Magnetic Resonance Elastography (iMRE) and suggested that a poroelastic model may address these challenges. Goal(s): To evaluate and compare poroelastic and poroviscoelastic brain models in iMRE for the brain, and to assess the influence of two different driving frequencies. Approach: Poro- and poroviscous properties were reconstructed using the same displacement data set, and the performance was evaluated through spatial consistency, repeatability and left-right symmetry. Results: Both models performed well, with the poroviscoelastic model being the most consistent. The second harmonic proved valuable for multifrequency reconstruction. Impact: These results directly contribute to the improvement of iMRE by addressing previous limitations and by demonstrating the feasibility of poro(visco)elastic reconstruction, potentially enabling novel ways of studying cerebral microvasculature and (mechanical) neurovascular coupling in the brain.