Ultra-High Resolution, Large Strain Analysis of Ex Vivo Bovine Brain Tissue Using Magnetic Resonance Elastography

Introduction: : Traumatic brain injury (TBI) is a neuropathology that leads to alterations in brain function. It is caused by the application of large forces or rapid accelerations on the head (Menon, 2010). These injuries often cause long-term consequences, including an increased risk of neurodegenerative diseases and cognitive impairments (Bramlett, 2015). Although nearly 80% of cases are classified as mild, TBI remains a major contributor to approximately one in every three injury-related deaths (Abdelmalik, 2018). Large strains and strain rates are critical factors in TBI, as the deformation of brain tissue under the impact of a head injury generates internal strains that contribute to cellular damage and dysfunction. Quantifying the nonlinear mechanical behavior resulting from these large strains can provide deeper insights into the damage mechanisms associated with TBI and potential therapeutic targets. While magnetic resonance elastography (MRE) of the living human brain offers valuable insights into tissue mechanical properties and their relationship to disease, traditional in vivo MRE is limited to measuring linear mechanical properties within the small-strain regime. Though beneficial, characterizing the viscoelastic properties within the small-strain regime fails to account for the mechanical nonlinearity present during large strain impacts and injury. To address this limitation, we have adopted and modified an ultra-high resolution preclinical MRE experiment to induce high strains in ex vivo bovine brain tissue samples, enabling us to capture mechanical nonlinearity and enhancing our understanding of TBI-related tissue deformation.

Materials and

Methods: : Fresh bovine brains were acquired from a local butcher and stored at 4°C. Whole brain tissue was segmented into smaller sections to fit properly into a 50 mL conical tube and encapsulated in agarose gel (0.75 wt.%) to assist with signal and wave penetration. Prior to scanning, the samples were brought to room temperature to limit any temperature-related effects.

Data was collected on a preclinical Bruker Biospec 94/20 9.4T MRI scanner using a 40 mm transmit/receive volume coil (Figure 1). Tissue-agar phantoms were scanned using a custom EPI sequence with 0.5 mm isotropic resolution and mechanical waves generated from a piezoelectric actuator (APA150M-NM, Cedrat Technologies, Meylan, France). MRE requires a combination of motion encoding gradients (MEGs), which are incorporated into the pulse sequence to map displacement from the actuation to the phase of the MR signal, and applied mechanical actuation to obtain sufficient wave motion. Typically, MEGs are played at their highest allowable amplitude to maximize the signal-to-noise (SNR) ratio with only mild vibration required. However, here we increase the applied mechanical actuation while lowering the MEG amplitude to induce large strains into the brain tissue samples while maintaining equivalent data quality. A nonlinear inversion (McGarry, 2012) was used to output maps of the apparent mechanical properties, namely stiffness and damping ratio, from the resulting displacement fields. MRE displacement data was then further processed to determine the large strain formulation at each level of induced strain.

Results, Conclusions, and Discussions:: Our experimental protocol was validated by conducting tests across multiple vibration frequencies to determine the optimal scanning frequency for analysis. These results (Figure 2A) indicate that frequencies between 200 and 400 Hz are ideal as there is high SNR with sufficient wave propagation through the tissue. For analysis, we examined the brain tissue segment within the phantom as a whole and further subdivided into white and gray matter sections for independent evaluation. Maximum strain and average apparent stiffness values were quantified in each region. Figure 2B shows data from a representative phantom where a general strain softening effect is observed, as increasing the strain from 0.51% to 5.56% resulted in the apparent stiffness of the whole brain decreasing 19.6% (2.93 to 2.35 kPa). Comparison between gray and white matter, as depicted in Figure 3A, showed a 38.4% decrease in gray matter (2.60 to 1.58 kPa) and a decrease of 37.6% (4.17 to 2.60 kPa) in white matter. Figure 3B illustrates the effect of frequency, where data collected at lower frequencies typically achieved higher strain values and resulted in a higher percent stiffness change with a 37.8% decrease at 200 Hz compared to only an 8.74% decrease at 500 Hz.

Our results indicate success in achieving large strains using our modified preclinical MRE methods across multiple tissue-agar phantoms and match a similar nonlinear strain softening behavior previously identified in porcine white matter (Begonia, 2021) and bovine brain (Bilston, 2001) under shear stress. Further analysis and quantification of tissue nonlinearity, specifically differences between white and gray matter, will prove relevant to properly characterizing brain tissue nonlinearity. Though we are restricted by limitations in the maximum mechanical actuation able to be induced, future work aims to further manipulate the balance of applied and encoded motion to maximize the amount of induced strain. This experiment serves as proof-of-concept for inducing strain to capture tissue nonlinearity, subsequent investigations will ensure that proper mechanical testing guidelines for sample preparation are followed. Long-term applications of this study will aim to use the nonlinear mechanical property values from our MRE experiments to supplement computational TBI models.

Acknowledgements and/or References (Optional):: This project was supported by the NSF CMMI-2331296 and NIH U01-NS112120 awards.