MRI-Based Characterization of the Mechanical Properties of the Biofidelic ANGUS Phantom with Enhanced Skull-Brain Formulations

Introduction: : Traumatic brain injury (TBI) affects an estimated 69 million people annually and is a leading cause of disability and death. Although mild TBIs comprise 75–90% of cases, they can still lead to long-term cognitive, emotional, and neurological impairments. Understanding brain mechanics is critical in studying the forces and stresses the brain experiences during traumatic events. Since direct investigation of TBI mechanisms in humans is ethically and practically unfeasible, phantom models provide a safe and controlled alternative. These models enable targeted exploration of injury mechanisms by replicating key anatomical features and material properties. However, a major challenge in phantom development is accurately reproducing the mechanical properties of the human brain, including stiffness and damping ratio. To address this, our team has developed the Anthropomorphic Neurologic Gyrencephalic Unified Standard (ANGUS) phantom—an idealized gyrencephalic brain model composed of polyacrylamide (PAA) hydrogel [1]. In this study, we employ tagged magnetic resonance imaging (tMRI) and magnetic resonance elastography (MRE) to characterize the deformation response and mechanical properties of the refined ANGUS phantom and compare them to previously collected in vivo human brain data [1,2]. This follow-up investigation highlights improvements in phantom composition, including the integration of a falx and a subarachnoid space model, evaluating its biofidelity under mild impulsive loading using tMRI and dynamic mode decomposition (DMD) analysis [3], as well as mechanical properties estimated by MRE. Together, these techniques provide complementary insights for advancing the biofidelity of brain phantoms and improving the study of brain injury mechanics.

Materials and

Methods: : The ANGUS phantom consists of a 3D-printed skull with clear acrylic plates on the superior and inferior sides and is filled with 10 wt%/vol PAA hydrogel replicating cortical gyri and ventricles (Fig. 1a, 1b). The current iteration also includes a polylactic acid (PLA) falx extending into the interhemispheric fissure and a 2 wt%/vol PAA gel to simulate the combined cerebrospinal fluid (CSF), meningeal attachments, and vascular attachments within the subarachnoid space. tMRI tracks brain tissue deformation by noninvasively applying spatial tag lines to MRI images. In this study, tMRI was conducted on a 1.5T Siemens Magnetom Sola Fit scanner using two custom-built devices that produced controlled neck rotations (~32 degrees along the inferior-superior axis; Fig. 1c) and neck extensions (~7 degrees along the medial-lateral axis; Fig. 1d), with deceleration achieved through a latch release and padded stop [2]. Image acquisition employed a SPAMM tagging sequence with harmonic phase (HARP) analysis to calculate maximum principal strains (MPS) within the hydrogel (Fig. 1f(i-iv), 1g). DMD analysis was applied to tMRI strain tensors to investigate the natural frequencies and deformation damping ratios [3]. MRE was performed on a 3T Siemens Magnetom Prisma scanner (Fig. 1e). Displacement data were acquired using single-shot echoplanar imaging at 30, 50, and 70 Hz (Fig. 1f(v, vi)) and processed through a nonlinear inversion algorithm [4] to estimate mechanical properties (Fig. 1h). Measurements were compared to human data from ten healthy adults and to a previous phantom version that used an isotonic solution to simulate CSF.

Results, Conclusions, and Discussions:: In the tMRI experiments, the phantom exhibited average peak angular accelerations of 330.6 ± 4.5 radians/s² (rotation) and 369.7 ± 1.7 radians/s² (extension), significantly higher than the 183.0 ± 47.2 and 194.8 ± 85.8 radians/s² measured in human scans, respectively. To ensure consistent comparisons, human MPS values were adjusted based on previously established linear relationships between MPS and squared angular velocity using the same rotation and extension devices. This resulted in average peak 95th percentile MPS (rotation: 0.0077 ± 0.003; extension: 0.036 ± 0.006) (Fig. 2a, 2b). The current ANGUS phantom exhibited a more biofidelic dynamic response post-impact compared to the previous version. The ANGUS phantom exhibited peak strain within 70 ms, which then dissipated within 150 ms, closely following the corridor of adjusted human MPS data, likely due to the incorporation of the PLA falx and 2 wt%/vol PAA gel, which better replicate the skull-brain interface. The median natural frequencies derived from tMRI using DMD for the ANGUS phantom were 13.87 ± 0.52 Hz and 21.70 ± 0.66 Hz for rotational and extension motions, respectively, and were significantly higher than those observed in in vivo human data (Fig. 2c). The corresponding median deformation damping ratios were 0.21 ± 0.02 and 0.12 ± 0.03, aligning with the human range for rotational motion but falling below it for extension motion (Fig. 2d). Based on the MRE analysis, the ANGUS phantom showed increasing trends in median storage modulus, loss modulus, and shear stiffness from 30 Hz to 70 Hz, except for viscoelastic damping ratio (Fig. 3a–3d). Storage modulus and shear stiffness closely matched the human range and demonstrated considerable improvement over the previous phantom version, while loss modulus and viscoelastic damping ratio remained lower than human values across all frequencies, consistent with earlier findings (Fig. 3e–3h). Overall, the ANGUS phantom, designed to replicate key brain properties and anatomy, showed improved biofidelity under impulsive rotational loading. tMRI with DMD analysis and MRE enabled direct comparison to in vivo biomechanical measurements. Additional refinements, such as adjusting material properties or adding anisotropic features, may further improve its realism.

Acknowledgements and/or References (Optional)::

References:
[1] A. K. Knutsen et al., “Characterization of material properties and deformation in the ANGUS phantom during mild head impacts using MRI,” J. Mech. Behav. Biomed. Mater., vol. 138, p. 105586, 2023.
[2] A. K. Knutsen et al., “In vivo estimates of axonal stretch and 3D brain deformation during mild head impact,” Brain Multiphysics, vol. 1, p. 100015, 2020.
[3] J. D. Escarcega et al., “Natural oscillatory modes of 3D deformation of the human brain in vivo.,” J. Biomech., vol. 119, p. 110259, Apr. 2021.
[4] M. D. J. McGarry et al., “Multiresolution MR elastography using nonlinear inversion.,” Med. Phys., vol. 39, no. 10, pp. 6388–6396, Oct. 2012.

Acknowledgements:
This work was partially supported by NINDS grant U01-NS112120, the Department of Defense in the Military Traumatic Brain Injury Initiative, the Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., and the intramural research program of the NIH. The authors have no conflicts of interest to disclose. The opinions and assertions expressed herein are those of the author(s) and do not reflect the official policy or position of the Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., the Uniformed Services University of the Health Sciences, or the Department of Defense.