Machine-learning-based head impact subtyping based on the spectral densities of the measurable head kinematics.

BACKGROUND: Traumatic brain injury can be caused by head impacts, but many brain injury risk estimation models are not equally accurate across the variety of impacts that patients may undergo, and the characteristics of different types of impacts are not well studied. We investigated the spectral characteristics of different head impact types with kinematics classification. METHODS: Data were analyzed from 3262 head impacts from lab reconstruction, American football, mixed martial arts, and publicly available car crash data. A random forest classifier with spectral densities of linear acceleration and angular velocity was built to classify head impact types (e.g., football, car crash, mixed martial arts). To test the classifier robustness, another 271 lab-reconstructed impacts were obtained from 5 other instrumented mouthguards. Finally, with the classifier, type-specific, nearest-neighbor regression models were built for brain strain. RESULTS: The classifier reached a median accuracy of 96% over 1000 random partitions of training and test sets. The most important features in the classification included both low- and high-frequency features, both linear acceleration features and angular velocity features. Different head impact types had different distributions of spectral densities in low- and high-frequency ranges (e.g., the spectral densities of mixed martial arts impacts were higher in the high-frequency range than in the low-frequency range). The type-specific regression showed a generally higher R2 value than baseline models without classification. CONCLUSION: The machine-learning-based classifier enables a better understanding of the impact kinematics spectral density in different sports, and it can be applied to evaluate the quality of impact-simulation systems and on-field data augmentation.

Predictive Factors of Kinematics in Traumatic Brain Injury from Head Impacts Based on Statistical Interpretation.

Brain tissue deformation resulting from head impacts is primarily caused by rotation and can lead to traumatic brain injury. To quantify brain injury risk based on measurements of kinematics on the head, finite element (FE) models and various brain injury criteria based on different factors of these kinematics have been developed, but the contribution of different kinematic factors has not been comprehensively analyzed across different types of head impacts in a data-driven manner. To better design brain injury criteria, the predictive power of rotational kinematics factors, which are different in (1) the derivative order (angular velocity, angular acceleration, angular jerk), (2) the direction and (3) the power (e.g., square-rooted, squared, cubic) of the angular velocity, were analyzed based on different datasets including laboratory impacts, American football, mixed martial arts (MMA), NHTSA automobile crashworthiness tests and NASCAR crash events. Ordinary least squares regressions were built from kinematics factors to the 95% maximum principal strain (MPS95), and we compared zero-order correlation coefficients, structure coefficients, commonality analysis, and dominance analysis. The angular acceleration, the magnitude and the first power factors showed the highest predictive power for the majority of impacts including laboratory impacts, American football impacts, with few exceptions (angular velocity for MMA and NASCAR impacts). The predictive power of rotational kinematics about three directions (x: posterior-to-anterior, y: left-to-right, z: superior-to-inferior) of kinematics varied with different sports and types of head impacts.

The relationship between brain injury criteria and brain strain across different types of head impacts can be different.

Multiple brain injury criteria (BIC) are developed to quickly quantify brain injury risks after head impacts. These BIC originated from different head impact types (e.g. sports and car crashes) are widely used in risk evaluation. However, the accuracy of using the BIC on brain injury risk estimation across head impact types has not been evaluated. Physiologically, brain strain is often considered the key parameter of brain injury. To evaluate the BIC's risk estimation accuracy across five datasets comprising different head impact types, linear regression was used to model 95% maximum principal strain, 95% maximum principal strain at the corpus callosum and cumulative strain damage (15%) on 18 BIC. The results show significantly different relationships between BIC and brain strain across datasets, indicating the same BIC value may suggest different brain strain across head impact types. The accuracy of brain strain regression is generally decreasing if the BIC regression models are fitted on a dataset with a different type of head impact rather than on the dataset with the same type. Given this finding, this study raises concerns for applying BIC to estimate the brain injury risks for head impacts different from the head impacts on which the BIC was developed.

Instantaneous Brain Strain Estimation for Automotive Head Impacts via Deep Learning.

Efficient brain strain estimation is critical for routine application of a head injury model. Lately, a convolutional neural network (CNN) has been successfully developed to estimate spatially detailed brain strains instantly and accurately in contact sports. Here, we extend its application to automotive head impacts, where impact profiles are typically more complex with longer durations. Head impact kinematics (N=458) from two public databases were used to generate augmented impacts (N=2694). They were simulated using the anisotropic Worcester Head Injury Model (WHIM) V1.0, which provided baseline elementwise peak maximum principal strain (MPS). For each augmented impact, rotational velocity (vrot) and the corresponding rotational acceleration (arot) profiles were concatenated as static images to serve as CNN input. Three training strategies were evaluated: 1) "baseline", using random initial weights; 2) "transfer learning", using weight transfer from a previous CNN model trained on head impacts drawn from contact sports; and 3) "combined training", combining previous training data from contact sports (N=5661) for training. The combined training achieved the best performances. For peak MPS, the CNN achieved a coefficient of determination (R2) of 0.932 and root mean squared error (RMSE) of 0.031 for the real-world testing dataset. It also achieved a success rate of 60.5% and 94.8% for elementwise MPS, where the linear regression slope, k, and correlation coefficient, r, between estimated and simulated MPS did not deviate from 1.0 (when identical) by more than 0.1 and 0.2, respectively. Cumulative strain damage measure (CSDM) from the CNN estimation was also highly accurate compared to those from direct simulation across a range of thresholds (R2 of 0.899-0.943 with RMSE of 0.054-0.069). Finally, the CNN achieved an average k and r of 0.98±0.12 and 0.90±0.07, respectively, for six reconstructed car crash impacts drawn from two other sources independent of the training dataset. Importantly, the CNN is able to efficiently estimate elementwise MPS with sufficient accuracy while conventional kinematic injury metrics cannot. Therefore, the CNN has the potential to supersede current kinematic injury metrics that can only approximate a global peak MPS or CSDM. The CNN technique developed here may offer enhanced utility in the design and development of head protective countermeasures, including in the automotive industry. This is the first study aimed at instantly estimating spatially detailed brain strains for automotive head impacts, which employs >8.8 thousand impact simulations generated from ~1.5 years of nonstop computations on a high-performance computing platform.

Validation of a finite element model with six-year-old child anatomical characteristics as specified in Euro NCAP Pedestrian Human Model Certification (TB024).

Accident statistics show that more than 80% of car-to-pedestrian collisions (CPC) occur when pedestrians cross the road. It is very important to establish a finite element model with natural walking posture to study the kinematics and injury mechanism of pedestrians. In this study, a finite element model of six-year-old child pedestrian is developed with detailed anatomical characteristics and posture parameters as specified in Euro NCAP Pedestrian Human Model Certification (TB024). The numerical human body model is validated in total twelve simulations in which the pedestrian is impacted against four generic vehicle models at speeds 30, 40, 50 km/h prescribed in TB024. The Head Impact Time (HIT), Contact Force and the Trajectories of HC, T12 and AC of all twelve simulations are compared with the reference corridors provided by Technical Bulletin 024. The results indicate that the numerical human body model of a six-year-old child can be used to demonstrate the suitability of the sensing system for the range of pedestrian sizes; the timing of system deployment, and the bonnet deflection due to body loading. Furthermore, the model could be a good tool for further research on pedestrian injury mechanism and the development of pedestrian protection devices.

Multi-Directional Dynamic Model for Traumatic Brain Injury Detection.

Given the worldwide adverse impact of traumatic brain injury (TBI) on the human population, its diagnosis and prediction are of utmost importance. Historically, many studies have focused on associating head kinematics to brain injury risk. Recently, there has been a push toward using computationally expensive finite element (FE) models of the brain to create tissue deformation metrics of brain injury. Here, we develop a new brain injury metric, the brain angle metric (BAM), based on the dynamics of a 3 degree-of-freedom lumped parameter brain model. The brain model is built based on the measured natural frequencies of an FE brain model simulated with live human impact data. We show that it can be used to rapidly estimate peak brain strains experienced during head rotational accelerations that cause mild TBI. In our data set, the simplified model correlates with peak principal FE strain (R2 = 0.82). Further, coronal and axial brain model displacement correlated with fiber-oriented peak strain in the corpus callosum (R2 = 0.77). Our proposed injury metric BAM uses the maximum angle predicted by our brain model and is compared against a number of existing rotational and translational kinematic injury metrics on a data set of head kinematics from 27 clinically diagnosed injuries and 887 non-injuries. We found that BAM performed comparably to peak angular acceleration, translational acceleration, and angular velocity in classifying injury and non-injury events. Metrics that separated time traces into their directional components had improved model deviance compare with those that combined components into a single time trace magnitude. Our brain model can be used in future work to rapidly approximate the peak strain resulting from mild to moderate head impacts and to quickly assess brain injury risk.

Effects of the Variation in Brain Tissue Mechanical Properties on the Intracranial Response of a 6-Year-Old Child.

Brain tissue mechanical properties are of importance to investigate child head injury using finite element (FE) method. However, these properties used in child head FE model normally vary in a large range in published literatures because of the insufficient child cadaver experiments. In this work, a head FE model with detailed anatomical structures is developed from the computed tomography (CT) data of a 6-year-old healthy child head. The effects of brain tissue mechanical properties on traumatic brain response are also analyzed by reconstruction of a head impact on engine hood according to Euro-NCAP testing regulation using FE method. The result showed that the variations of brain tissue mechanical parameters in linear viscoelastic constitutive model had different influences on the intracranial response. Furthermore, the opposite trend was obtained in the predicted shear stress and shear strain of brain tissues caused by the variations of mentioned parameters.

Development and validation of age-dependent FE human models of a mid-sized male thorax.

The increasing number of people over 65 years old (YO) is an important research topic in the area of impact biomechanics, and finite element (FE) modeling can provide valuable support for related research. There were three objectives of this study: (1) Estimation of the representative age of the previously-documented Ford Human Body Model (FHBM) -- an FE model which approximates the geometry and mass of a mid-sized male, (2) Development of FE models representing two additional ages, and (3) Validation of the resulting three models to the extent possible with respect to available physical tests. Specifically, the geometry of the model was compared to published data relating rib angles to age, and the mechanical properties of different simulated tissues were compared to a number of published aging functions. The FHBM was determined to represent a 53-59 YO mid-sized male. The aforementioned aging functions were used to develop FE models representing two additional ages: 35 and 75 YO. The rib model was validated against human rib specimens and whole rib tests, under different loading conditions, with and without modeled fracture. In addition, the resulting three age-dependent models were validated by simulating cadaveric tests of blunt and sled impacts. The responses of the models, in general, were within the cadaveric response corridors. When compared to peak responses from individual cadavers similar in size and age to the age-dependent models, some responses were within one standard deviation of the test data. All the other responses, but one, were within two standard deviations.

Impact response and biomechanical analysis of the knee-thigh-hip complex in frontal impacts with a full human body finite element model.

Changes in vehicle safety design technology and the increasing use of seat-belts and airbag restraint systems have gradually changed the relative proportion of lower extremity injuries. These changes in real world injuries have renewed interest and the need of further investigation into occupant injury mechanisms and biomechanical impact responses of the knee-thigh-hip complex during frontal impacts. This study uses a detailed finite element model of the human body to simulate occupant knee impacts experienced in frontal crashes. The human body model includes detailed anatomical features of the head, neck, shoulder, chest, thoracic and lumbar spine, abdomen, pelvis, and lower and upper extremities. The material properties used in the model for each anatomic part of the human body were obtained from test data reported in the literature. The human body model used in the current study has been previously validated in frontal and side impacts. It was further validated with cadaver knee-thigh-hip impact tests in the current study. The effects of impactor configuration and flexion angle of the knee on biomechanical impact responses of the knee-thigh-hip complex were studied using the validated human body finite element model. This study showed that the knee flexion angle and the impact direction and shape of the impactors affected the injury outcomes of the knee-thigh-hip complex significantly. The 60 degrees flexed knee impact showed the least impact force, knee pressure, femoral von Mises stress, and pelvic von Mises stress but largest relative displacements of the Posterior Cruciate Ligament (PCL) and Anterior Cruciate Ligament (ACL). The 90 degrees flexed knee impact resulted in a higher impact force, knee pressure, femoral von Mises stress, and pelvic von Mises stress; but smaller PCL and ACL displacements. Stress distributions of the patella, femur, and pelvis were also given for all the simulated conditions.

Analysis and evaluation of the biofidelity of the human body finite element model in lateral impact simulations according to ISO-TR9790 procedures.

The biofidelity of the Ford Motor Company human body finite element (FE) model in side impact simulations was analyzed and evaluated following the procedures outlined in ISO technical report TR9790. This FE model, representing a 50th percentile adult male, was used to simulate the biomechanical impact tests described in ISO-TR9790. These laboratory tests were considered as suitable for assessing the lateral impact biofidelity of the head, neck, shoulder, thorax, abdomen, and pelvis of crash test dummies, subcomponent test devices, and math models that are used to represent a 50th percentile adult male. The simulated impact responses of the head, neck, shoulder, thorax, abdomen, and pelvis of the FE model were compared with the PMHS (Post Mortem Human Subject) data upon which the response requirements for side impact surrogates was based. An overall biofidelity rating of the human body FE model was determined using the ISO-TR9790 rating method. The resulting rating for the human body FE model was 8.5 on a 0 to 10 scale with 8.6-10 being excellent biofidelity. In addition, in order to explore whether there is a dependency of the impact responses of the FE model on different analysis codes, three commercially available analysis codes, namely, LS-DYNA, Pamcrash, and Radioss were used to run the human body FE model. Effects of these codes on biofidelity when compared with ISO-TR9790 data are discussed. Model robustness and numerical issues arising with three different code simulations are also discussed.

Biomechanical Analysis of Human Abdominal Impact Responses and Injuries through Finite Element Simulations of a Full Human Body Model.

Human abdominal response and injury in blunt impacts was investigated through finite element simulations of cadaver tests using a full human body model of an average-sized adult male. The model was validated at various impact speeds by comparing model responses with available experimental cadaver test data in pendulum side impacts and frontal rigid bar impacts from various sources. Results of various abdominal impact simulations are presented in this paper. Model-predicted abdominal dynamic responses such as force-time and force-deflection characteristics, and injury severities, measured by organ pressures, for the simulated impact conditions are presented. Quantitative results such as impact forces, abdominal deflections, internal organ stresses have shown that the abdomen responded differently to left and right side impacts, especially in low speed impact. Results also indicated that the model exhibited speed sensitive response characteristics and the compressibility of the abdomen significantly influenced the overall impact response in the simulated impact conditions. This study demonstrates that the development of a validated finite element human body model can be useful for abdominal injury assessment. Internal organ injuries, which are difficult to detect in experimental studies with human cadavers due to the difficulty of instrumentation, may be more easily identified with a validated finite element model through stress-strain analysis.

Prediction and analysis of human thoracic impact responses and injuries in cadaver impacts using a full human body finite element model.

Human thoracic dynamic responses and injuries associated with frontal impact, side impact, and belt loading were investigated and predicted using a complete human body finite element model for an average adult male. The human body model was developed to study the impact biomechanics of a vehicular occupant. Its geometry was based on the Visible Human Project (National Library of Medicine) and the topographies from human body anatomical texts. The data was then scaled to an average adult male according to available biomechanical data from the literature. The model includes details of the head, neck, ribcage, abdomen, thoracic and lumbar spine, internal organs of the chest and abdomen, pelvis, and the upper and lower extremities. The present study is focused on the dynamic response and injuries of the thorax. The model was validated at various impact speeds by comparing predicted responses with available experimental cadaver data in frontal and side pendulum impacts, as well as belt loading. Model responses were compared with similar individual cadaver tests instead of using cadaver corridors because the large differences between the upper and lower bounds of the corridors may confound the model validation. The validated model was then used to study thorax dynamic responses and injuries in various simulated impact conditions. Parameters that could induce injuries such as force, deflection, and stress were computed from model simulations and were compared with previously proposed thoracic injury criteria to assess injury potential for the thorax. It has been shown that the model exhibited speed sensitive impact characteristics, and the compressibility of the internal organs significantly influenced the overall impact response in the simulated impact conditions. This study demonstrates that the development of a validated FE human body model could be useful for injury assessment in various cadaveric impacts reported in the literature. Internal organ injuries, which are difficult to detect in experimental studies with human cadavers, can be more easily identified with a validated finite element model through stress-strain analysis, especially in conjunction with experimental studies.