The dynamic impact behavior of the human neurocranium

Realistic biomechanical models of the human head should accurately reflect the mechanical properties of all neurocranial bones. Previous studies predominantly focused on static testing setups, males, restricted age ranges and scarcely investigated the temporal area. This given study determined the biomechanical properties of 64 human neurocranial samples (age range of 3 weeks to 94 years) using testing velocities of 2.5, 3.0 and 3.5 m/s in a three-point bending setup. Maximum forces were higher with increasing testing velocities (p ≤ 0.031) but bending strengths only revealed insignificant increases (p ≥ 0.052). The maximum force positively correlated with the sample thickness (p ≤ 0.012 at 2.0 m/s and 3.0 m/s) and bending strength negatively correlated with both age (p ≤ 0.041) and sample thickness (p ≤ 0.036). All parameters were independent of sex (p ≥ 0.120) apart from a higher bending strength of females (p = 0.040) for the 3.5 -m/s group. All parameters were independent of the post mortem interval (p ≥ 0.061). This study provides novel insights into the dynamic mechanical properties of distinct neurocranial bones over an age range spanning almost one century. It is concluded that the former are age-, site- and thickness-dependent, whereas sex dependence needs further investigation.


Mechanical testing. A custom-built drop tower
was used for the dynamic mechanical tests in this study.
The setup consisted of a semi-circular high-strength aluminium indenter with a radius of 1.5 mm that was fixed to a weighted cross-head (total weight including indenter = 643 g), which was attached to two vertically oriented stainless steel rods using polytetrafluoroethylene bushings. The indenter was connected to a piezoelectric force transducer (Model 9021A, Kistler AG, Winterthur, Switzerland) with the signal being fed to a charge amplifier (Model 5015A, Kistler AG) and recorded with an oscilloscope (TDS 360, Tektronix, Beaverton, OR, USA) using a sampling rate of 200 MHz. The bone samples were placed on two support beams with a radius of 2 mm. The

Results
Maximum force was significantly higher at increasing testing velocities while bending strength increased on a statistically non-significant level only. The     Consistency analysis of tested groups. The individual testing velocity groups were statistically non-different with regards to age (p > 0.546), PMI (p > 0.371) and specimen thickness (p > 0.700). When only temporal bone was analyzed, the velocity groups were non-different regarding age (p > 0.614), PMI (p > 0.723) and specimen thickness (p > 0.355). The pooled temporal samples (4.34 mm, IQR = 2.11) were thinner compared to the pooled other samples (5.42 mm, IQR = 1.92, p = 0.046). The thickness of all samples significantly increased with age (r = 0.421, p < 0.001), which also applied to the temporal bones only (r = 0.425, p = 0.002), and when samples older than 18 years were evaluated exclusively (r = 0.271, p = 0.023). There was no significant difference of sample thicknesses between the two sexes (r = 0.009, p = 0.471).

Discussion
Human cadaveric material continues to be the preferred source for investigations determining the biomechanical behavior of the human neurocranium, thereby obtaining fundamental data for head injury criteria 23 . Only a few studies so far investigated the force tolerance limits of the human neurocranium under dynamic loading conditions 17,19,20,22 . This lack of study is partly related to tissue availability, and partly also to the experimental setup required to obtain meaningful and reproducible data. Fifty years ago, the dynamic tensile properties of 30 human neurocrania were determined on the frontal, temporal and parietal area of cadavers aged between 15 and 95 years using testing velocities between 0.0001 and 3.8 m/s 22 . No site dependency of the mechanical parameters was observed between the tested regions of the neurocranium, with only nine tested temporal specimens in total 22 . In contrast, the results of this given study revealed a significantly lower F max of the temporal bone compared to the frontal and parietal areas. The data in this given study was obtained utilizing a three-point bending setup, which is the most frequently used setup for the mechanical characterization of bones 1,12,[14][15][16]18,20 . Therefore, the tensile data obtained in the aforementioned study 22 cannot be directly compared to other recent studies. Another study tested frontal and parietal bone samples of eight human crania aged between 62 and 97 years using dynamic testing velocities between 0.5 and 2.5 m/s in a three-point bending setup comparable to the study presented here 20 . The observed averaged F max for the frontal, left parietal and right parietal bones using the highest testing velocity of 2.5 m/s were 1316 N, 1227 N and 1162 N 20 , respectively, which seems much higher compared to the here reported 716 N for all samples using an identical testing velocity. This may be explained by www.nature.com/scientificreports/ the following two facts: firstly, the here given study predominantly consisted of temporal samples, which were significantly thinner compared to the pooled frontal, parietal and occipital samples. This thickness difference is in line with previous studies on the thickness of the various regions of the neurocranium 16 . The quantity of diploë in temporal bone is low when compared to the other flat bones of the human neurocranium. It may even be nonexistent in particular areas 30 . The authors here hypothesize that diploë thickness appears to be of major interest for failure mechanisms of the human neurocranium as the results of the here given dynamic study and a previous quasi-static study 24 revealed the lowest F max values for temporal bone samples when compared to the major neurocranial flat bones. Secondly, the mean age of the cadavers of 48 years in the here given study was much lower compared to the 81 years of the aforementioned study 20 . A study that investigated the fracture loads of 94 cadavers in a static testing setup using a testing velocity of 0.0003 m/s reported mean F max values between 435 and 515 N for temporal bone 16 . The here presented temporal F max values were 638, 722 and 1136 N using testing velocities of 2.5, 3.0 and 3.5 m/s, respectively, being well in line with the former observations and showing an increasing trend with increasing velocities. Even though no statistical comparison can be provided between the two studies, the highest here used testing velocity of 3.5 m/s resulted in higher F max values for the temporal samples compared to the values provided in the aforementioned study using a static testing setup 16 . This can likely be explained by the time-dependent, viscoelastic and strain rate-dependent character of osseous tissue 10 . However, it has to be noted that this given study and the one by Torimitsu et al. 16 have technical differences including statistical considerations of the non-normally distributed F max values being compared with mean values 16 and the different distances of the lower supports with 13 and 50 mm. These impede the direct comparison of the data between the two studies. Also, temporal samples in this given study were thinner compared to the samples used in the referred static experiment with a median of 4.3 mm compared to a mean of 5.5 to 6.6 mm 16 . This study highlighted the strong correlations between F max and sample thickness, the here stated findings can be considered well in line with previous investigations after all, but provide a much broader age span and tissues were obtained from all major neurocranial bones. The age-dependent thickness increase of bone shown in this study might, even after adolescence, contribute to the increased failure loads in studies conducted on elderly people. This highlights that human head models have to be adjusted for an increasing thickness with increasing age as this has direct implications for the biomechanical behavior of the respective neurocranial bones. Supported by the here presented results, the thickness of the cranial sample seems the most predictive parameter to influence the applicable F max of the individual sample 24,31 , which further supports the necessity of modelling realistic thicknesses for all bone parts of the neurocranium in both physical and computational head models. A study applying dynamic three-point bending tests (testing velocities between 1.2 and 2.8 m/s) conducted on fetus and infant neurocrania revealed that bending modulus and ultimate stress were independent of the applied strain rate 19 . In contrast to these findings, dynamic studies on adult tissues showed that increasing loading rates led to significantly higher elastic moduli, maximum bending stresses, F max 20 and breaking strains 22 . The results of this study support the increase of F max with increased loading rates, although a statistically significant result was only observed between the lowest and the highest testing velocity. Similarly, in the study by Motherway et al. a significant increase in F max was only observed between testing velocities of 2.5 m/s vs. 1 and 0.5 m/s, respectively 20 . The high inter-individual scatter of mechanical parameters such as F max might require larger sample sizes than presented here and formerly 20 to detect minute differences in F max at lower testing velocities or much higher velocity differences between the groups. The testing velocities in this given study are in line with head impacts in contact sports, including elbow to head impact (1.7 to 4.6 m/s) or head to head impacts (1.5 to 3.0 m/s) in football 32 . It has to be acknowledged that a variety of head impacts, including falls from moving bikes (approximately 11 m/s) 33 or gunshot wounds, where the ammunition moves at a velocity of several hundred meters 34 are far beyond the here tested velocity range. However, the authors hypothesize that the velocity range of head impacts in vivo covers the entire range from almost static to the highest velocity gunshot impacts.
Age-related biomechanical data of human tissues are of increasing interest, as the population is ageing, and both males and females are known to suffer from hormone-induced changes in bone metabolism associated with reduced mechanical bone strength related to osteoporosis 35,36 . Contrary to femoral neck or vertebral fractures, which are associated with an age-related decrease in bone density 37 , the age-associated fracture characteristics of the neurocranium are less obvious. This might be due to the lower frequency of injuries in this region overall. Some studies concluded that the neurocranium does not show a particular age-related decrease of mechanical strength 23 , whereas others reported a decrease of fracture loads with age 16 . To the authors best knowledge this is the first study to investigate dynamic biomechanical properties of the human neurocranium over the expected human life span of more than nine decades. F max was unrelated to the age at death, which contrasts findings of another study that observed an age-dependent decrease of F max applying static testing velocities 16 . B strength decreased consistently with age in all testing velocities, which was present in both the overall group and the temporal group when considered independently. This may partly be explained by the increasing thickness of the neurocranium with age observed here, which is inversely proportional to the squared thickness of the bone sample 38 according to the B strength equation 12 . Failure stress, failure strain as well as energy absorption of human neurocranial samples remained constant with increasing age in a dynamic tensile setup 22 . The values F max , B strength and thickness were independent of sex in the here investigated samples except for the significantly higher B strength of females compared to males in the highest here tested velocity. Also, the F max of occipital and parietal samples were shown to be higher for males in a static testing setup 16 . However, both observations should be investigated in a larger sample size to confirm these sex-related differences as the general trend appears to lean towards sexindependence of the biomechanical properties of the human neurocranium. Noteworthy, mechanical properties of female neurocranial bone samples are scarcely investigated in general 17,20 , and studies so far frequently investigated an uneven ratio of males compared to females 1,16,22 . This uneven ratio might partially be explained by the overall higher number of males compared to females in forensic autopsies 39  www.nature.com/scientificreports/ tissues in all age ranges 13,[40][41][42][43] . This is in contrast to the oftentimes chemically altered anatomical dissections or pathologically altered clinical post mortems. The inclusion of sufficient female samples for statistical comparisons in biomechanical characterization studies as provided here are critical to acknowledge potential sex-dependent differences in both physical and computational human head models in the future. This may help to overcome the sex bias towards men in tissue mechanics and its respective application fields 44 . Lastly, the independence of the here investigated mechanical parameters from PMI within a median of 70 h indicates that human neurocranium samples have consistent failure characteristics within this given time frame, and, therefore, can be used for biomechanical purposes as well as grafts in transplant surgery when kept at a maximum temperature of 4 °C.
Limitations. The study is limited in sample size as the here given number of fresh tissues was the maximum that could be allocated to this project. The neurocranial samples that were used for the three-point bending tests in this study represent an inhomogeneous, anisotropic material with a varying cross-section along their lengths. Therefore, it is an oversimplification to assume a straight beam with a rectangular surface for B strength calculations. As the human neurocranial bone forms an anisotropic material 45,46 the here obtained mechanical parameters are only valid for the applied loading axis, which was chosen in order to reflect the extracranial-intracranial loading axis in vivo to the best possible extent.

Conclusions
The fracture loads of human neurocranial samples increase with increasing sample thickness and testing velocities between 2.5 and 3.5 m/s in a three-point bending setup. The B strength of neurocranial samples decreases depending on age at death and the thickness of the tested sample. The fracture loads of the human neurocranium are sexindependent, but sex-dependence of B strength warrants further investigation. The here investigated biomechanical parameters of human neurocranial bone remained constant over a median time frame of 70 h post mortem when being cooled at 4 °C.