(B-71) Quantifying biomechanical property changes of the human cadaver brain using tagged magnetic resonance imaging

Introduction:: Computational models of brain injury play an important role in assessment of head-neck biomechanics and designing protective equipment to prevent traumatic brain injury (TBI). Experimental TBI data frequently uses cadaveric specimens. Because of the degradation of human brain tissue after death, the investigation of the evolution of the brain’s biomechanical properties post-mortem is required to better understand the time window at which cadavers can be viably tested. Previous techniques to characterize brain stiffness in vivo were based on direct contact measurements such as tissue indentation [1, 2]. Those techniques are invasive and require a partial removal of the skull in mostly animal models that may alter the properties. To better examine the cadaver head degradation effects on the brain biomechanical properties, this study analyzes image data obtained from tagged magnetic resonance imaging (MRI) performed on a human cadaver head under mild loading conditions. Unlike conventional MRI scans, tagged MRI uses magnetic tags to label specific regions of the brain and track their movements in real-time [3, 4]. By providing a non-invasive, high-resolution view of brain mechanics, tagged MRI has the potential to unlock new insights into the motion of the brain under load and improves our ability to understand TBI. Because the loading conditions are at low levels, the data can also be compared to in vivo measurements.


Materials and Methods:: A custom device was designed and built to support and constrain the head rotation motion, as used in previous in vivo experiments (Fig. 1a) [4]. A rotation of approximately 32 degrees about the inferior-superior axis toward the left shoulder was performed on a female adult cadaver head (91 years) with no known TBI history. Consistent mild accelerations of the head were operated inside the imaging coil of a clinical MR scanner (Siemens Aera 1.5T Scanner) at 4 scan days (Day02, Day03, Day30, and Day58) after death. The cadaver head was preserved in air-tight wrapping at 4°C between scan days and removed on scan days to come up to room temperature. A passive head rotation motion was initiated by releasing a latch followed by rapid deceleration caused by a padded stop. The MRI acquisition was similar to [4, 5], employing a 1-1 SPAMM tagging sequence (Fig. 1b-d). The angular position for each drop was measured by an optical sensor [4, 5]. Images were acquired with two orthogonal tag line directions in the axial plane and one tag line direction in the sagittal plane. Motion was tracked between image frames using harmonic phase with 3-dimensional finite element analysis [6]. Lagrangian strain tensors were computed from the displacement vectors. Volume fractions of the strain field in which the maximum principal stain (E1) exceeds 3% were determined for the whole brain and by regions (cortical gray matter (CGM), deep gray matter (DGM), and white matter (WM)), based on T1-weighted image and SLANT segmentation [7].


Results, Conclusions, and Discussions:: The net weight of the cadaver head was 4.67 kg. The volume of the whole brain increased from 1239 cm³ (Day02) to 1320 cm³ (Day58) (6.56% increase), while the volume of the WM decreased from 523 cm³ (Day02) to 465 cm³ (Day 58) (11.0% decrease) (Fig. 2). On each scan day, 444 rotations were executed, and 12 axial and 13 sagittal plane slices were acquired. The average peak angular acceleration for all rotations was 311.0±41.3 radians/s². The 95th percentiles of the E1 in the first 100 ms measured at Day02 were 0.043, 0.050, 0.029, and 0.042 for the whole brain, CGM, DGM, and WM, correspondingly, which were slightly greater than the measurements at Day03 (Fig. 3). However, these quantities decreased to 0.033, 0.036, 0.021, and 0.033, correspondingly at Day30, and 0.031, 0.033, 0.022, and 0.031, correspondingly at Day58 (Fig. 3). This indicates that the brain tissues became stiffer after one month. In addition, the volume fractions were calculated at the time of peak angular acceleration, which was also the time when the peak E1 of the time series was observed. The volume fraction of the whole brain was 16.9% for E1 >3% at Day02 while it decreased to 5.9% at Day58 (Fig. 4). The volume fraction for E1 >3% in CGM was higher than the fractions in WM and DGM at Day02 (20.2%, 3.6%, and 19.5%, correspondingly), and these volume fractions decreased to 7.9%, 0.8%, and 6.0% at Day58, correspondingly (Fig. 4). The E1 and its volume fractions observed in this cadaver head tested on Day02 and Day03 were comparable to healthy volunteers published in previous studies [4, 5]. In summary, we found that by two months after death, the whole brain volume has largely increased up to 6.56%, likely due to the accumulation of fluid in the CGM and DGM, while shrinkage of volume was observed in the cerebellar WM by 11.0%. The brain tissues also became increasingly stiff after 1- to 2-month cooling preservation. Therefore, it is important for researchers to utilize post-mortem brain tissue as soon as possible after death to minimize the material property changes.


Acknowledgements (Optional): : This work was partially supported by NINDS grant U01-NS112120, the Department of Defense in the Center for Neuroscience and Regenerative Medicine, 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, the Uniformed Services University of the Health Sciences or the Department of Defense.


References (Optional): : [1] Miller et al. 2000. J. Biomechanics 33(11): 1369-1376.

[2] Gefen et al. 2004. J. Biomechanics 37(9): 1339-1352.

[3] Bayly et al. 2005. J. Neurotrauma 22: 845-56.

[4] Knutsen et al. 2014. J. Biomechanics 47: 3475-81.

[5] Knutsen et al. 2020. Brain Multiphysics 100015.

[6] Gomez et al. 2018. IEEE Trans Biomed Eng, 66: 1456-1467.

[7] Huo et al. 2019. NeuroImage, 194: 105-119.