Abstract
The biomechanical responses of human head (translational head CG accelerations, rotational head accelerations, and HIC) under lateral impact to the parietal-temporal region were investigated in the current study. Free drop tests were conducted at impact velocities ranging from 2.44 to 7.70 m/s with a 40 durometer, a 90 durometer flat padding, and a 90 durometer cylinder. Specimens were isolated from PMHS subjects at the level of occipital condyles, and the intracranial substance was replaced with brain simulant (Sylgard 527). Three tri-axial accelerometers were instrumented at the anterior, posterior, and vertex of the specimen, and a pyramid nine accelerometer package (pNAP) was used at the contra-lateral site. Biomechanical responses were computed by transforming accelerations measured at each location to the head CG. The results indicated significant “hoop effect” from skull deformation. Translational head CG accelerations were accurately measured by transforming the pNAP, the vertex accelerations, or the average of anterior/posterior acceleration to the CG. The material stiffness and structural rigidity of the padding changed the biomechanical responses of the head with stiffer padding resulting in higher head accelerations. At the skull fracture, HIC values were more than 2–3x higher than the frontal skull fracture threshold (HIC=1000), emphasizing the differences between frontal and lateral impact. Rotational head accelerations up to 42.1 krad/s2 were observed before skull fracture, indicating possible severe brain injury without skull fracture in lateral head impact. These data will help to establish injury criteria and threshold in lateral impacts for improved automotive protection and help clinicians understand the biomechanics of lateral head impact from improved diagnosis.
INTRODUCTION
Head accelerations have been associated with head/brain injuries for more than five decades, for example, the widely adopted head injury criterion (HIC) is derived from resultant translational head center of gravity (CG) accelerations. Despite the fact that HIC is based on repeated drop tests to the forehead using post mortem human subjects (PMHS), it has been used to quantify all injuries related to the head (Mertz et al., 1996; Yoganandan and Pintar, 2000; Takhounts et al., 2003; Zhang et al., 2006).
As documented in a recent study of real-world cases, occupants can sustain severe brain injuries, such as diffuse axonal injury (DAI) from side-impact vehicle collisions. (Yoganandan et al., 2009). These real-world DAI cases were strongly associated with head contact, but less than 20% of these cases were associated with skull fracture. In previous studies, DAI has been produced in laboratory animals and physical models with pure rotational acceleration input and rotation-based brain injury criteria have been proposed (Gennarelli and Thibault, 1982; Margulies and Thibault, 1992; Ommaya et al., 2002). A recent computational study has also shown that rotational acceleration is a major contributor to brain strain in side impact (Zhang et al., 2006). Therefore, rotational head acceleration is an important bioengineering metric that needs to be documented in lateral head impact studies.
Other important factors that could affect the severity of sustained brain injuries include the stiffness and the structural shape of the contact surfaces. As a previous study indicated, almost all (>95%) of the DAI cases were associated with head contact to the interior surfaces of the vehicle (Yoganandan et al., 2009). Vehicular interior surfaces consist of a variety of materials with varying stiffness and structures in different shapes. The variations in contact surfaces can significantly change the impact loading and affect the severity of sustained head injury. A thorough understanding of how the properties of contact surfaces affect sustained brain injury during head impact can be used to mitigate trauma.
Consequently, the current research was designed to quantify biomechanical responses of the human head (translational head CG acceleration, rotational accelerations, and HIC) in lateral impacts with varying stiffness padding materials and shape.
METHODS
Unembalmed specimens were isolated from post mortem human subjects (PMHS) at the level of occipital condyles. The scalp was included in the preparation, and the intracranial contents were replaced with Sylgard Gel (Sylgard 527, Dow Corning, MI). Pretest X-ray and Computed Tomography (CT) images were obtained to confirm the integrity of the cranial vault. Free fall experiments were conducted to impact the left parietal-temporal region of the specimens (Fig. 1) with a flat surface padded with 40 (40D, 5 cm in thickness) or 90 durometer (90D, 5 cm in thickness) rubber plate, or a 90 durometer rubber cylinder (90D-Cyl, 5 cm in diameter). The mid-sagittal plane of the specimen was aligned at a 10 degree angle with respect to the horizontal plane such that the impact occurred to the left temprotoparietal region on each specimen.
Figure 1.
Schematic of lateral head impact experiment setup (left) and a side view of accelerometer locations (right)
Tests were conducted at increasing drop heights with impact velocities ranging from 2.44 to 7.70 m/s. Testing was concluded as soon as a skull fracture was identified biomechanically. The skull fracture was determined by reduced impact force with increasing drop height. Data from fractured tests were excluded from further analysis in this study. Three linear triaxial accelerometers were instrumented on the specimen at the vertex, posterior, and anterior regions, and a pyramid nine-accelerometer-package (pNAP) was instrumented on the contra-lateral side of impact (Yoganandan et al., 2006). Locations of the accelerometers were digitized using a FaroArm (FARO Technology Inc., Lake Mary, FL) with respect to the anatomical coordinate system with the origin at the CG. Acceleration-time histories were collected using a digital data acquisition system (TDAS Pro, DTS Technologies, Seal Beach, CA), and a common trigger system synchronized all the dynamic data. Data were collected at a sampling frequency of 12.5 kHz and filtered according to SAE Class 1000 specifications. Post-test CT images were obtained for all specimens to confirm skull fracture. Three-dimensional (3-D) rotational accelerations were computed from the pNAP using established pNAP data reduction procedures developed in our lab (Yoganandan et al., 2006). Three-dimensional translational accelerations at the CG of the head were computed by compensating rotational accelerations computed from the pNAP and transforming linear acceleration data to the CG (Yoganandan et al., 2006). A total of 53 non-fracture lateral head impact tests were conducted using ten PMHS specimens.
RESULTS
A summary of tests conducted on each specimen and skull fracture impact velocity are shown in Table 1.
Table 1.
Summary of test conditions
| Padding | PMHS ID | Number of tests | Fx (Y/N) | Fx Impact Velocity (m/s) |
|---|---|---|---|---|
| 40D | 1 | 7 | Y | 6.47 |
| 2 | 8 | Y | 8.08 | |
| 90D | 3 | 4 | Y | 5.47 |
| 4 | 5 | Y | 5.47 | |
| 5 | 6 | Y | 5.99 | |
| 6 | 3 | Y | 4.23 | |
| 90D-Cyl | 7 | 4 | Y | 4.89 |
| 8 | 6 | Y | 5.99 | |
| 9 | 6 | Y | 5.99 | |
| 10 | 4 | N | 4.89* |
The final test was conducted at this velocity. See results section for details.
In 40D padding tests, both specimens fractured at the last impact. Specimen 1 had left parietal fracture and left squamous temporal fracture at impact velocity of 6.47 m/s. Specimen 2 had right orbital roof fracture at impact velocity of 8.08 m/s.
In 90D padding tests, all four specimens fractured. Specimen 3 had non-displaced, non-depressed left frontoparietal fracture at an impact velocity of 4.23 m/s. Specimens 4 and 5 were fractured at an impact velocity of 5.47 m/s. Specimen 4 had a non-displaced, non-depressed left frontotemporal fracture and left zygomatic arch fracture. Specimen 5 had a comminuted and depressed left frontoparietotemporal fracture, which extended into the skull base. Specimen 6 had left frontoparietal fracture and squamous temporal fracture at impact velocity of 5.99 m/s.
In 90D cylinder tests, three specimens fractured. Tests on Specimen 10 were terminated due to a decrease in impact force at increased drop height at impact velocity of 4.89 m/s. However, there were no fractures diagnosed on the post-test CT images. Specimen 7 had a left frontoparietotemporal fracture at impact velocity of 4.89 m/s. Specimen 8 and 9 were fractured at impact velocity of 5.99 m/s; both of them had left frontoparietotemporal fracture. The fracture in Specimen 9 extended into the skull base.
Typical head CG translational accelerations computed from different measurement locations are shown in Figures 2–4.
Figure 2.
Computed head CG translational accelerations from different accelerometers at 2.44 m/s impact (90D)
Figure 4.
Comparison of computed head CG translational accelerations from the average of anterior/posterior accelerometers, the vertex, and the pNAP at 4.23 m/s impact (90D)
Comparisons of average resultant head CG translational accelerations, resultant rotational head accelerations, and HIC between 40D and 90D flat padding is shown in Figures 5–7.
Figure 5.
Comparison of average peak head CG translational accelerations with increasing impact velocity for 40D and 90D padding
Figure 7.
Comparison of average HIC with increasing impact velocity for 40D and 90D padding
Comparisons of average resultant head CG translational accelerations, resultant rotational head accelerations, and HIC between 90D and 90D-Cyl is shown in Figures 8–10.
Figure 8.
Comparison of average peak head CG translational accelerations with increasing impact velocity for 90D and 90D-Cyl
Figure 10.
Comparison of average HIC with increasing impact velocity for 90D and 90D-Cyl
The average fracture impact velocity was the highest (7.0 m/s) in 40D tests, followed by 90D-Cyl (5.64 m/s), and the lowest (5.29 m/s) in 90D tests. Skull fractures occurred after HIC values reached 2200, 2996, and 2353, and resultant rotational head accelerations were 20.1, 31.1 and 34.1 krad/s2 in 40D, 90D, 90D Cyl tests, respectively. At the same impact velocity before skull fracture, peak translational accelerations in the 90D tests were approximately 2x greater than in the 40D tests, and rotational acceleration were more than 2.5x.
DISCUSSION
In previous research, full-body PMHS in an upright seated position has been used and the head region was impacted with an impactor to study head dynamics (McIntosh et al., 1996). Another study has used free-fall techniques to drop a full body PMHS and impacted the head to the designated impact surface in order to study head impact dynamics (Got et al., 1978).
In the current study, instead of using full body PMHS, PMHS specimens were isolated at the level of occipital condyles to impact the lateral side using free fall impact technique. The specimens were totally unconstrained during the impacts. The method was justified based on the assumption that active and passive restriction from the neck did not significantly affect the biomechanics of the head during the short duration (a few milliseconds) of impact. Sometimes, post mortem storage leads to shrinkage of the brain tissues and results in air space in the cranium. The biomechanical effects of these post mortem changes are still unknown. In the present study, a brain simulant (Sylgard gel 527) was used to replace the intracranial contents of the specimen. The gel closely mimics the density and mechanical properties of human brain. It has been used as a physical substitute of brain in blunt head impact studies (Margulies et al., 1990; Brands et al., 1999; Bradshaw et al., 2001). The replacement resulted in direct skull-to-Sylgard gel contact, which better simulated the interior brain-skull contact boundary condition under the physiologic situation.
Linear accelerometers have been historically used to obtain translational head CG accelerations. Linear accelerometers have been placed at carefully selected locations on the skull in previous studies so that measurements taken on the skull can represent one or more components of the accelerations at the head CG as it is impossible to rigidly attach accelerometers at the CG of a human head. Such a crude method required pre-existing conditions and assumptions, such as plane of symmetry, minimal rotational head accelerations. Therefore, these methods can not be used for general head acceleration measurement.
Alternatively, the head CG accelerations can be obtained by taking acceleration measurements at an arbitrary location with known distance to head CG, and then transforming the accelerations to head CG by compensating with rotational accelerations. With modern digitalization techniques, locations of accelerometers can be obtained to the accuracy of millimeters.
To obtain rotational head accelerations, a nine accelerometer package (NAP) in a 3-2-2-2 configuration is needed to obtain stable measurements (Yoganandan et al., 2006). Early studies have used a three-arm design NAP. The three-arm NAP consisted of three diagonal (X-Y-Z) arms that join at the origin. Three of the nine linear accelerometers are mounted at the origin. The other six accelerometers are mounted on the end of the arms, with two accelerometers on each arm. The inherent open design of the three-arm NAP leaves it vulnerable to vibration. Studies have used such devices in PMHS tests with varying degrees of success. One of the studies suggested a low pass filter of 200 Hz in order to deal with the vibration of the instrument (McIntosh et al., 1996). In consideration of these technical difficulties, Yoganandan et al. (2006) proposed a new design of the NAP, in which a light weight pyramid shaped frame was used to house the nine accelerometers, and digitalization technique was used to obtain the locations of each accelerometer (Yoganandan et al., 2006). Tests using the pNAP on human head demonstrated less than 5% difference in translational head CG accelerations up to 272g and almost identical results on dummy tests compared to the built-in nine-accelerometer array head in Hybrid III. Hence, pNAP was chosen for acceleration measurement in the current study.
In theory, a pNAP is sufficient to obtain both rotational head accelerations and translational head CG accelerations if the head is a rigid body. However, the deformability of the human head may not be negligible under impact, especially at high impact velocities that produce skull fracture. Under these situations, measurement from a pNAP at one anatomical region may not be enough to represent the global dynamics of the head. Therefore, additional accelerometers (three tri-axial linear accelerometers) were used in addition to the pNAP to provide more acceleration data from different regions of the head. Linear accelerations were transformed to the head CG using the distance of the accelerometers to the head CG and the rotational accelerations computed from the pNAP assuming rotational accelerations were the same at all anatomical locations. This approach also provided a method of redundancy and cross-check to validate acceleration measurements from the tests.
At lower impact velocities (Fig. 2, v = 2.44 m/s), the translational head CG accelerations computed from different accelerometers agreed very well. Highest and lowest peak resultant accelerations (138g and 115g) differed within 10% of the average (125g).
The differences between head CG accelerations increased with increasing impact velocity. An obvious anterior-posterior high frequency oscillation can be clearly identified by plotting head CG accelerations from anterior and posterior accelerometers and their average on the same graph (Fig. 3, v = 4.23 m/s). This behavior, i.e., increased amplitude of high-frequency oscillation between anterior and posterior accelerometers, is consistent for all padding surfaces in the current study. This may be due to a “hoop effect” from large lateral skull deformation at high impact velocities, i.e., when the skull at the site of impact deforms inward, the skull at 90 degrees away from the impact deforms outward. At higher impact velocity (Fig. 3), the maximum difference between anterior and posterior head CG accelerations can be as much as 100%, e.g., at 0.88 ms, posterior CG acceleration was 120.7g and anterior CG acceleration was 244.0 g.
Figure 3.
Comparison of computed head CG translational accelerations from anterior and posterior accelerometers at 4.23 m/s impact (90D)
These oscillations between computed anterior and posterior translational head CG accelerations due to hoop effect are noticeable at the lower impact velocity, but their relative attribution to the total acceleration is very small (Fig. 2). With increasing impact speed, the oscillation became more obvious and significantly affected the accelerations measured at anterior and posterior of the head (Fig. 3).
Compared to the anterior and posterior tri-axial accelerometers, the vertex tri-axial accelerometer and the pNAP were less affected by the hoop effect due to their locations closer to the mid-coronal plane of the head. Average of the anterior and posterior head CG accelerations cancelled the hoop effect; results agreed very well with the head CG computed from the pNAP and tri-axial accelerometers at vertex (Fig. 4).
These results indicated skull deformation is not negligible, and the hoop effect may have a significant contribution to the anterior and posterior accelerations in lateral head impacts under high impact velocities. Translational head CG accelerations were more accurately measured by using pNAP, accelerations from the vertex, or the averaged accelerations from anterior and posterior accelerometers for lateral PMHS head impact tests. Consequently, further analysis and discussions in the current study were based on results obtained from the pNAP data.
Contact surface conditions, including material stiffness and surface curvatures, can change the impact response of the head. The effect of material stiffness can be studied by comparing results from 90D and 40D padding tests. Figures 5, 6, and 7 compare the average peak translational and rotational accelerations and HIC from 18 tests at five impact velocities for 90D padding with 15 tests at eight impact velocities for 40D padding. Linear trend lines are used to fit the experimental data. The linear equation and R2 values are shown in the figures by each data set. At the same impact velocity, the translational head CG acceleration, HIC, and rotational head acceleration for 90D padding were consistently higher than 40D padding, indicating elevated probability of severe injuries with stiffer padding material. The slopes of linear fit to the translational head CG acceleration, HIC, and rotational head acceleration for the 90D padding were 2.4x, 2.5x and 4.6x higher than the 40D padding, respectively. The consistent higher slope for 90D padding indicated a higher increment of impact severity with increased impact velocity for surfaces padded with stiffer materials. The even higher ratio of the slope (4.6x) in rotational head acceleration indicated a faster increase in rotational head accelerations when the head impacted a surface padded with a stiffer material. Rotation may become the dominant mechanism of injury at higher impact speed when impacted with a surface of stiff material.
Figure 6.
Comparison of average peak head rotational accelerations with increasing impact velocity for 40D and 90D padding
A comparison between the 90D and 90D-Cyl allowed the study of surface curvature effect. Both the translational head CG accelerations and HIC increased linearly with increasing impact velocity (Fig. 8, 9, 10). A linear regression fit the data very well with an R2 of 0.92 and 0.91 for translational head CG accelerations and HIC, respectively (Fig. 8, 10). In contrast, the rotational head accelerations did not demonstrate well-defined linear correlation to the impact velocity. A linear fit resulted in an R2 of only 0.61 (Fig. 9). This may due to the curvature of the cylindrical surface resulting in larger variation in point of contact on the head upon impact. As the contact locations changed with respect to the head CG, the momentum resulting from the contact force also varied, leading to the variation in rotational head acceleration.
Figure 9.
Comparison of average peak head rotational accelerations with increasing impact velocity for 90D and 90D-Cyl
Comparing the results from the 90D-Cyl to the 90D flat padding, the translational head CG accelerations, the HIC, and the head rotational accelerations for the 90D-Cyl were consistently lower than the 90D flat padding (Fig. 8, 9, 10). The slope of the 90D-Cyl was slightly lower than the 90D padding. The increased curvature of the cylinder may have reduced the overall structural stiffness and increased the stopping distance of the contact impact and, thus, resulted in lower translational head CG accelerations, HIC, and rotational head accelerations.
Average impact velocities at skull fracture were 7.27 m/s for the 40D padding, 5.29 m/s for the 90D padding, and 5.62 m/s for the 90D-Cylinder. Translational head CG accelerations, rotational head accelerations, and HIC at these fracture impact velocities according to the linear correlation determined from the current study are listed in table 2.
Table 2.
Biomechanical data at average skull fracture impact velocity from linear correlation
| Fracture Impact Velocity | Linear Fit | |||
|---|---|---|---|---|
| Trans. Acc. | Rot. Acc. | HIC | ||
| (m/s) | (g) | (krad/s2) | ||
| 40D | 7.27 | 263.1 | 13.6 | 2170 |
| 90D | 5.29 | 376.1 | 42.1 | 3392 |
| 90D-Cyl | 5.62 | 311.6 | 24.9 | 2154 |
McIntosh et al. (1996) positioned full body PMHS in an upright sitting position and subjected them to lateral impacts using an impactor with impact velocity ranging from 2.8 to 6.1 m/s. Results indicated the skull fractured at HIC values of 1,894 to 3,850 (McIntosh et al., 1996). Despite the difference in the method of applying the impact loading (free drop using isolated head in current study vs. impactor to upright full body PMHS), results from the current study agreed favorably with the study by McIntosh et al. (1996).
The HIC=1000 threshold has been associated with the probability of linear skull fractures in a frontal impact. Using this threshold, according to the linear relation from our study, a HIC=1000 would have indicated that a skull fracture will occur at 4.47 m/s for the 40D flat padding, 3.03 m/s for the 90D flat padding, and 3.74 m/s for the 90D-cylinder. In contrast, the fracture impact velocities in the current study were approximately 1.6x higher than those predicted using frontal skull fracture criteria (HIC=1000). This may due to the fact that the skull thickness in the parietal/temporal region is relatively thinner than in the frontal region. A thinner skull can sustain larger bending deformation before fracture, therefore, higher HIC values at fracture. Compared to frontal skull fracture criteria (HIC=1000), HIC in the current study was approximately 2x (40D padding and 90D cylinder) and 3.4x (90D padding) higher at the skull fracture.
Although, HIC is extensively discussed here in correlation with skull fracture, it should be emphasized it is not the authors’ intention to establish a HIC for lateral impact to the head, but rather to demonstrate the limitations of using HIC for side impact situations. HIC is derived only from translational head CG accelerations. It was originally correlated to linear skull fractures in frontal impact. Recent clinical and epidemiological studies have shown that, in vehicular side impact, severe brain injuries such as DAI may be associated with no skull fracture. Studies have demonstrated that, under these situations, rotational head acceleration is the major contributor to brain injury (Zhang et al., 2006).
Margulies and Thibault found that the threshold of DAI is about 9 krad/s2 (Margulies and Thibault, 1992). Ommaya et al. (2002) proposed a slightly higher threshold with 12.5 krad/s2 for mild DAI, 15.5 krad/s2 for moderate DAI, and 18 krad/s2 for severe DAI (Ommaya et al., 2002). Rotational head accelerations in the current study were approximately 13.6 krad/s2 for 40D padding, 24.9 krad/s2 for 90D Cylinder, and 42.1 krad/s2 for 90D padding. This suggests moderate to severe DAI may be caused from these lateral head impacts before skull fracture occurs. This conclusion agrees well with the field data reported in literature in which contact loading to the head in side motor vehicle crashes was associated with DAI but not necessarily accompanied with skull fracture (Yoganandan et al., 2009).
Hodgson and Thomas (Hodgson and Thomas, 1971; Hodgson and Thomas, 1973) conducted both full body and isolated head drops on 40D, 90D, glass, and rigid impact surfaces with embalmed PMHS. These early studies had limited acceleration data but served as the initial foundation for defining the response of the head. The larger dataset generated with superior instrumentation from the current study may offer an opportunity for updating/verifying the ISO 9790 side impact head response corridor.
Limitations of the current study include repeated impact tests on PMHS specimens. Prior non-fracture impacts could have potentially weakened the PMHS skull and led to lowered skull fracture impact velocity and head accelerations. However, this limitation may not affect the conclusions drawn from the current study, i.e., skull fracture occurs at higher HIC in lateral impact than frontal impact.
CONCLUSION
The biomechanical responses of the human head under lateral impacts to the parietal-temporal region were investigated in the current study. Translational head CG accelerations, rotational head acceleration, and HIC were quantified using accelerometers at multiple locations. HIC values for skull fracture in the current study were more than 2–3x the frontal skull fracture threshold (HIC=1000), emphasizing the biomechanical differences of the human head in different impact directions. Consequently, frontal criterion may not be applicable for lateral impact scenarios. Severe rotational head accelerations were found in these lateral head impacts. Rotational head accelerations were approximately 2–3x the suggested DAI thresholds before skull fracture. These results may explain severe brain injuries without skull fracture in motor vehicle side crashes reported in recent clinical and biomechanical literature (Yoganandan et al., 2009).
The study also investigated the effects of impacting surface material stiffness and structural rigidity on the biomechanical response of the human head. Stiffer material and more rigid structures resulted in increased injury metrics. The study also found that the “hoop effect” from skull deformation is significant and needs to be considered in lateral PMHS head impact studies.
Biomechanical data at different impact velocities and varying contact surfaces from this study are invaluable in the validation of finite element models in the lateral impact direction. Results from this study will help to establish skull fracture and rotational head injury tolerances in lateral impacts for improved automotive protection and help clinicians understand the lateral head impact biomechanics for improved diagnoses.
Acknowledgments
This research was supported in part by DTNH22-03-H-07147 and VA Medical Research.
REFERENCES
- Bradshaw DRS, Ivarsson J, Morfey CL, Viano DC. Simulation of acute subdural hematoma and diffuse axonal injury in coronal head impact. Journal of Biomechanics. 2001;34:85–94. doi: 10.1016/s0021-9290(00)00135-4. [DOI] [PubMed] [Google Scholar]
- Brands DWA, Bovendeerd PHM, Peters GWM, Wismans JSHM, Paas MHJW, van Bree JLMJ. Comparison of the Dynamic Behavior of Brain Tissue and Two Model Materials. 43rd Stapp Car Crash Conference; San Diego, California. 1999. pp. 350–358. [Google Scholar]
- Gennarelli TA, Thibault LE. Biomechanics of acute subdural hematoma. Journal of Trauma. 1982;22:680–686. doi: 10.1097/00005373-198208000-00005. [DOI] [PubMed] [Google Scholar]
- Got C, Patel A, Fayon A, Tarriere C, Walfisch G. Results of experimental head impacts on cadavers: the various data obtained and their relations to some measured physical paramenters. 22nd Stapp Car Crash Conf; Ann Arbor, MI. 1978. [Google Scholar]
- Hodgson VR, Thomas LE. Breaking strength of the human skull vs. impact surface curvature. Springfield, VA: US Department of Tranportation; 1971. [Google Scholar]
- Hodgson VR, Thomas LM. Breaking strength of the human skull vs impact surface curvature. Springfield, VA: National Technical Information Service; 1973. [Google Scholar]
- Margulies SS, Thibault LE. A proposed tolerance criterion for diffuse axonal injury in man. Journal of Biomechanics. 1992;25:917–923. doi: 10.1016/0021-9290(92)90231-o. [DOI] [PubMed] [Google Scholar]
- Margulies SS, Thibault LE, Gennarelli TA. Physical model simulations of brain injury in the primate. Journal of biomechanics. 1990;23:823–836. doi: 10.1016/0021-9290(90)90029-3. [DOI] [PubMed] [Google Scholar]
- McIntosh A, Sevensson N, Kallieris D, Mattern R, Krabbel G, Ikels K. Head impact tolerance in side impacts. 15th International Technology Conference on the Enhanced Safety of Vehicles; Melbourne, Australia. 1996. [Google Scholar]
- Mertz HJ, Prasad P, Nusholtz G. Head injury risk assessment for forehead impacts. SAE transactions. 1996;105:26–46. [Google Scholar]
- Ommaya AK, Goldsmith W, Thibault L. Biomechanics and neuropathology of adult and paediatric head injury. British Journal of Neurosurgery. 2002;16:220–242. doi: 10.1080/02688690220148824. [DOI] [PubMed] [Google Scholar]
- Takhounts EG, Eppinger RH, Campbell QJ, Tannous RE, Power ED. On the Development of the SIMon Finite Element Head Model. Stapp Car Crash Journal. 2003;47:1–18. doi: 10.4271/2003-22-0007. [DOI] [PubMed] [Google Scholar]
- Yoganandan N, Gennarelli TA, Zhang J, Pintar FA, Takhounts E, Ridella SA. Association of contact loading in diffuse axonal injuries from motor vehicle crashes. The Journal of Trauma. 2009;66:309–315. doi: 10.1097/TA.0b013e3181692104. [DOI] [PubMed] [Google Scholar]
- Yoganandan N, Pintar FA. Frontiers in Whiplash Trauma: Clinical and Biomechanical. The Netherlands: IOS Press; 2000. [Google Scholar]
- Yoganandan N, Zhang J, Pintar FA, King Liu Y. Lightweight low-profile nine-accelerometer package to obtain head angular accelerations in short-duration impacts. Journal of Biomechanics. 2006;39:1347–1354. doi: 10.1016/j.jbiomech.2005.03.016. [DOI] [PubMed] [Google Scholar]
- Zhang J, Yoganandan N, Pintar FA, Gennarelli TA. Brain Strains in Vehicle Impact Tests. Annu Proc Assoc Adv Automot Med. 2006;50:1–12. [PMC free article] [PubMed] [Google Scholar]