Simulation and Clinical Assessment of Blast-Induced Traumatic Brain Injury

 

 

Authors: Paul A. Taylor, John S. Ludwigsen,  Andrei A. Vakhtin, Corey C. Ford

 

Introduction

Recent combat statistics report that over 267,000 US soldiers deployed in Iraq and Afghanistan have sustained traumatic brain injury (TBI), with over 48,000 of those categorized at the moderate-to-severe level, 69% caused by blast exposure [1–3]. The principal source of these brain injuries was one or more encounters with the blast wave produced by a detonated improvised explosive device (IED). Primary blast injury is associated with direct exposure of the head and body to the blast wave without other blunt injury mechanisms [4]. The role of direct or primary blast exposure in the development of TBI is not well understood and has been the focus of our research.

Modeling and simulation-based investigations into the causal relationship between explosive blast and TBI have recently begun to appear in the literature [5–7]. In an earlier study [6], we identified the presence of blast-induced, early-time stress waves that occur within the brain well before the onset of any head motion.  These studies also revealed the need for a more complete head model to better define the important structures of the brain.

In response, we developed a high resolution, head-neck model and simulated the effects of blast direction on intracranial stress waves and the deposition of wave energy [8]. We also investigated a group of veterans with mild TBI (mTBI), whose injury scenarios were primarily limited to blast exposure [9]. Our goal was to establish correlations between simulation predictions of intracranial wave energy deposition and the brain injury observed in these subjects. The next step would be to construct a brain injury threshold criterion to define the limits of wave physics variables (e.g., stress or energy maxima) that would lead to brain injury. A modeling and simulation (M&S) approach, in conjunction with a brain injury threshold criterion, would facilitate investigations of the mechanisms of blast-induced brain injury and provide the means to assess helmet design effectiveness and strategies to protect against blast induced brain injury.

 

Modeling and Simulation

We constructed a virtual head-neck model based on the National Library of Medicine’s Visible Human dataset [10]. The model possesses anatomically correct distributions of bone, white and gray brain matter, falx and tentorium, cerebral spinal fluid (CSF), and muscle-scalp (see Figure 1). Anatomical details of the model are defined at a 1 mm resolution. Constituent material properties are defined for bone, white and gray matter, membranes, cerebral spinal fluid, and muscle-scalp. All constitutive model descriptions for our bio-materials have been reported in detail in a separate article [8]. We assigned a non-linear equation-of-state representation for dry air, specifically designed for shock wave simulations [11].

figure_1_sim_NTL.jpg

Figure 1. Head-neck model. Top row: front, rear, and left side views. Bottom row: coronal, axial, and mid-sagittal cuts displaying internal structure.

We use two simulation methods, each chosen for its ability to capture the relevant physics of the injury scenario under investigation. Blast simulations are performed using the shock wave physics code CTH [12], and for blunt impact, we employ the transient dynamics code PRESTO [13]. PRESTO was also used to validate the head-neck model by simulating the magnetic resonance tagging experiments of Sabet et al. [14] and Feng et al. [15].

Simulations of direct blast exposure to the unprotected head were performed to study the effects of both the blast magnitude and direction (Figure 2). Localized brain injury may correlate with one or more of three possible stress wave energy quantities: isotropic compressive energy (ICE), isotropic tensile energy (ITE), and deviatoric shear energy (DSE).

figure_2_sim_NTL.jpg

Figure 2. Stop-action plots of blast-generated pressure waves propagating through the head model from the front (left image), rear (center image), and lateral (right image) directions.

Brain tissue is infused with a significant amount of fluid and therefore is essentially incompressible; however, ITE could result in cavitation if local fluid pressure is reduced to partial vacuum levels.  Fluid could undergo a phase transformation from liquid to vapor initiating bubble formation. When these bubbles collapse they could generate micro-shock wavelets that are thought to cause tissue damage in the vicinity of the collapse [16–18]. DSE is thought to cause tissue damage as a result of axonal membrane tearing and cytoskeletal disruption [19,20]. Figure 3 displays maximum predicted levels of these energy terms for a 360 KPa blast wave directed at the head from the rear. Although ICE has, thus far, not been associated with brain injury, it is plotted in Figure 3 for completeness.

figure_3_sim_NTL.jpg

Figure 3. Simulation plots of maximum energy in the mid-sagittal (left column) and axial (right column) planes for a 360 KPa rear blast. Top row: maximum isotropic compressive energy (blue: 1 J/m3; red: 300
J/m3).Middle row: maximum isotropic tensile energy (blue: 1 J/m3; red: 200 J/m3). Bottom row:
maximum deviatoric shear energy (blue: 1 J/m3; red: 300 J/m3). Black denotes that the plot variable maximum limit has been exceeded.

A significant unexpected finding in the simulations was the prediction of independence of ITE and DSE deposition on blast direction [8]. Deposition of these two energy quantities occurred in the same regions of the brain with the same magnitudes regardless of the blast direction (front, rear, or side). This result suggests that it is not necessary to take blast direction into account, an important result for the designers of protective headgear.

 

Clinical Assessment of TBI

We conducted functional magnetic resonance image (fMRI) studies on 13 combat veterans, who were diagnosed with mTBI. These veterans were given a battery of 12 neuropsychological tests. Their averaged t-scores defined a Gaussian distribution with a mean of 44, 6 points below normal controls [9]. The mTBI subjects demonstrated statistically significant deficits in tasks measuring attention and processing speed.

We then applied independent component analysis (ICA) to resting state fMRI data, which has shown potential to be more sensitive to small individual differences than conventional fMRI analyses [21–23]. Using ICA, temporal correlations between multiple brain regions can be examined [21]. These techniques allow detection of brain networks associated with attention, vision, motion, hearing, and other functions.

We compared our mTBI group and a cohort of normal controls taken from an extensive dataset of normal controls presented by Calhoun et al. [24] and Erhardt et al. [25]. A large fMRI study was presented by Allen et al. [26], who identified 28 independent resting state networks in a large sample of over 600 normal subjects. These 28 independent components were categorized into 7 resting state network groups: sensory-motor, attentional, visual, frontal, auditory, basal ganglia, and default mode, all of which were identified in our mTBI subjects.

Three main aspects of the ICA components can be tested [27]. A time course spectra analysis allows examination of differences in the power of specific blood oxygenation level-dependent (BOLD) signal frequencies between groups. This has the potential to identify abnormalities in specific brain networks. Cross correlation between component time courses defines functional network connectivity (FNC) between specific brain regions that may be functionally disrupted by injury. Abnormal FNC has the potential to explain cognitive impairments observed in TBI subjects and suggests where investigations of specific white matter tracts connecting these regions should be focused using techniques such as diffusion tensor imaging (DTI).

The details of this approach are reported elsewhere [9] and briefly summarized here. ICA identified significant spatial map differences in the mTBI group’s frontal and visual networks. The mTBI group displayed higher activity in bilateral temporo-parietal junctions (visual network) and lower activity in the left inferior temporal lobe (frontal network) relative to controls (Figure 4). Time course spectra between mTBI and control groups were significantly different in the attentional, frontal, and default mode networks. Lastly, FNC in the mTBI group was impaired in 6 network pairs relative to controls. FNC differences were detected between attentional-sensorimotor, attentional-frontal, frontal-default mode, default mode-basal ganglia, and sensorimotor-sensorimotor network pairs.

figure_4_sim_NTL.jpg

Figure 4. Locations of hyperactivity (blue) and hypoactivity (red) as determined by fMRI analysis of our TBI subject cohort.

Although these studies involved a small subject group, our interpretation of the results is that healthy hyperactive regions of the brain may be working overtime to compensate for functionally disconnected or damaged regions. The two regions that appear to be involved include the ventral and dorsal streams of visual processing. This fits with neuropsychological results showing impairments in attention and processing speed in mTBI subjects.

 

Correlation of Simulation Prediction and Clinical Outcomes

A major goal in our work is to establish a correlation between simulation predictions of selected wave physics variables, and the presence of localized brain injury. We were able to qualitatively compare simulation predictions with clinical results to see what spatial overlap may exist between regions of predicted wave energy deposition and spatial maps of altered functional network activity.

figure_5_sim_NTL.jpg

 

Figure 5. Overlay of fMRI results from the TBI subject cohort on simulation predictions of maximum deviatoric shear energy. Hyperactive regions (magenta stripping) shown in images (a)-(c)
and hypoactive (red) region (identified by arrows) shown in images (d)-(f).


Figure 5 shows locations of hyper- and hypo-active regions, identified by our analysis of the mTBI group, overlaid on our prediction of blast-induced deviatoric shear energy deposition. Figures 5(a)-(c) show the hyperactive brain regions of our mTBI group residing in areas predicted to receive low levels of deviatoric shear energy. It is possible that hyperactive regions could be compensating for other regions to which they are connected. Conversely, Figures 5(d)-(f) show the hypoactive region to reside in a region of the brain that sustained elevated levels of deviatoric shear energy. This suggests that deviatoric shear energy deposition may be associated with local brain injury.

 

Conclusion

We have established a viable modeling and simulation capability with which to investigate traumatic brain injury from blast and blunt force impact. Our results predict that for blast-induced TBI, important metrics of intracranial wave motion may be independent of blast direction. This result should reduce the complexities of correlating simulation results with clinical measures of TBI and of creating new designs of head protection gear.

We have attempted to correlate simulation predictions of brain injury with clinical measures of blast-induced mTBI. Although our results are encouraging, we do not suggest that any definitive correlation exists at this point in time. Although this is a goal that we strive to achieve, there is much more work to perform before such an accomplishment can be realized.

 

Acknowledgements

The authors acknowledge the National Library of Medicine and the Visible Human Project as the source of the Visible Human Data Set used to construct the digital head-neck model employed in this research.

This work is funded through the U.S. Naval Health Research Center, Office of Naval Research, Mr. James Mackiewicz, project funding manager.

Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the United States Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.

 

References

1. Defense and Veterans Brain Injury Center. DoD Worldwide Numbers for TBI | DVBIC. Available from: http://www.dvbic.org/dod-worldwide-numbers-tbi

2. Fischer H. U.S. military casualty statistics: operation New Dawn, operation Iraqi Freedom, and operation Enduring Freedom. Congressional Research Service; 1991 pp. 1–8. Available from: https://opencrs.com/document/RS22452/

3. Warden D. Military TBI during the Iraq and Afghanistan wars. Journal of Head Trauma Rehabilitation 2006;21:398–402.

4. DePalma RG, Burris DG, Champion HR, Hodgson MJ. Blast injuries. New England Journal of Medicine 2005;352(13):1335–1342.

5. Moore DF, Radovitzky RA, Shupenko L, Klinoff A, Jaffee MS, Rosen JM. Blast physics and central nervous system injury. Future Neurology 2008 May;3(3):243–250.

6. Taylor PA, Ford CC. Simulation of blast-induced early-time intracranial wave physics leading to traumatic brain injury. Journal of Biomechanical Engineering 2009;131(6):061007.

7. Moore DF, Jérusalem A, Nyein M, Noels L, Jaffee MS, Radovitzky RA. Computational biology — modeling of primary blast effects on the central nervous system. NeuroImage 2009 August;47, Supplement 2:T10–T20.

8. Taylor PA, Ludwigsen JS, Vakhtin AA, Ford CC. Investigation of blast-induced traumatic brain injury. Brain Injury 2013;submitted for publication.

9. Vakhtin AA, Calhoun VD, Jung RE, Prestopnik JL, Taylor PA, Ford CC. Changes in intrinsic functional brain networks following blast-induced mild traumatic brain injury. Brain Injury 2013;submitted for publication.

10. Anon. The National Library of Medicine’s visible human project. 2007. Available from: http://www.nlm.nih.gov/research/visible/visible_human.html

11. Hertel ES, Kerley GI. CTH reference manual: The equation of state package. Albuquerque, NM: Sandia National Laboratories; 1998.

12. Hertel ES, Bell R, Elrick M, Farnsworth A, Kerley G, McGlaun J, Petney S, Silling S, Taylor P. CTH: a software family for multi-dimensional shock physics analysis. Proceedings of the 19th International Symposium on Shock Waves 1993;1:377–382.

13. SIERRA Solid Mechanics Team. Sierra/SolidMechanics VOTD user’s guide. Albuquerque, NM: Engineering Sciences Center, Sandia National Laboratories; 2011.

14. Sabet AA, Christoforou E, Zatlin B, Genin GM, Bayly PV. Deformation of the human brain induced by mild angular head acceleration. Journal of Biomechanics 2008;41:307–315.

15. Feng Y, Abney TM, Okamoto RJ, Pless RB, Genin GM, Bayly PV. Relative brain displacement and deformation during constrained mild frontal head impact. Journal of The Royal Society Interface 2010 May 26;7(53):1677–1688.

16. Lubock P, Goldsmith W. Experimental cavitation studies in a model head-neck system. Journal of Biomechanics 1980;13:1041–1052.

17. Brennen CE. Cavitation in biological and bioengineering contexts. In: Proceedings of the 5th International Symposium on Cavitation. Osaka, Japan; 2003.

18. Nakagawa A, Fujimura M, Kato K, Okuyama H, Hashimoto T, Takayama K, Tominaga T. Shock wave-induced brain injury in rat: Novel traumatic brain injury animal model. Acta Neurochirurgica Supplement 2008;102:421–424.

19. Zhang L, Yang KH, King AI. Comparison of brain responses between frontal and lateral impacts by finite element modeling. Journal of Neurotrauma 2001;18(1):21–30.

20. Zhang L, Yang KH, King AI. A proposed injury threshold for mild traumatic brain injury. Journal of Biomechanical Engineering 2004;126(2):226.

21. Bell AJ, Sejnowski TJ. An information-maximization approach to blind separation and blind deconvolution. Neural Computation 1995;7:1129–1159.

22. Biswal B, Yetkin F, Haughton V, Hyde J. Functional connectivity in the motor cortex of resting human brain using echo-planar MRI. Magnetic Resonance in Medicine 1995;34(4):537–541.

23. Koch W, Teipel S, Mueller S, Buerger K, Bokde ALW, Hamper H, Coates U, Reiser M, Meindl T. Effects of aging on default mode network activity in resting state fMRI: does the method of analysis matter? NeuroImage 2009;51(1):280–287.

24. Calhoun VD, Adali T, Pearlson GD, Pekar JJ. A method for making group inferences from functional MRI data using independent component analysis. Human Brain Mapping 2001;14(3):140–151.

25. Erhardt EB, Rachakonda S, Bedrick EJ, Allen EA, Adali T, Calhoun VD. Comparison of multi-subject ICA methods for analysis of fMRI data. Human Brain Mapping 2011;32(12):2075–2095.

26. Allen E, Erhardt EB, Damaraju E, Gruner W, Segall JM, Silva RF, Havlicek M, Rachakonda S, Fries J, Kalyanam R, et al. A baseline for the multivariate comparison of resting-state networks. Frontiers in Systems Neuroscience 2011;5(2):1–23.

27. Jafri MJ, Pearlson GD, Stevens M, Calhoun VD. A method for functional network connectivity among spatially independent resting-state components in schizophrenia. NeuroImage 2008 February 15;39(4):1666–1681.

 

Corresponding author, address: Sandia National Laboratories, MS 1160, P.O. Box 5800, Albuquerque, NM 87185-1160; tel: 505-844-1960, fax: 505-284-1373, pataylo@sandia.gov.

Editor’s note: The views and opinions expressed in the articles contained in the International Neuro-Trauma Letter are those of the authors and contributors alone and do not necessarily reflect the views, policy or position of the International Brain Injury Association or all members of the NTL Editorial Board. The NTL is provided solely as an informational resource and the inclusion of any particular article does not establish or imply IBIA’s endorsement of its contents.