Biomechanics research at the Bioengineering Center of Wayne State University began in 1939. The first project was an interdisciplinary effort to study head injury mechanisms. This particular computer model was developed over a period of over 10 years, utilizing biomechanical knowledge and modern-day high computational speeds to predict the stresses and deformations of the brain during a blunt impact. Its purpose was to find how and where the brain is injured by an impact to the head. The significance of the project is its capability to accurately predict the extent and location of the injury in the brain. It not only assists the physician in identifying the damaged area but also enables the designer of automobiles and helmets to improve protection against brain injury.

By way of background, head injury is the cause of 50% of all automotive-related fatalities and is also a common injury among bicyclists and athletes. Brain injury is never totally reversible, causing great disruptions to the lives of the victims as well as those of his/her family members. However, animal experiments that can help us understand the mechanisms of brain injury are not only costly to perform but are also a difficult public relations problem. Thus, a comprehensive computer model can drastically reduce the number of animal experiments needed to understand the mechanisms of injury. In fact, such a model can simulate a large variety of impact situations, some of which are hard to replicate experimentally. Our computer model of brain impact is by no means unique. Many have been developed in the past but this is the first one which can predict with reasonable accuracy the location and severity of injury when the model results are compared with animal injury data obtained by another research laboratory. It is also capable of predicting the pressures developed in the brain as measured in a cadaver. Additionally, it can explain why an impact to the rear of the head is more likely to cause some of the veins at the top of the brain to rupture and to produce a life-threatening injury called a subdural hematoma. The Discovery Channel recently aired a program on brain injury entitled: Death by Impact. The program starts with a story about a young woman who was killed when she was rearended by a fast moving vehicle. Her head did not appear to have hit anything but the damage to her car was severe. The TV crew that filmed the program came to Wayne State to find out about our model and our experimental program. During one of the interviews, we told them that the young woman must have died from a subdural hematoma. The TV crew was very surprised that we came up with the cause of death based solely on the information regarding the crash. This was, of course, not a lucky guess. It had been predicted by our model.

The major features of the model that make it a unique and the most up-to-date model available are the simulation of the gray and white matter of the brain as separate materials and the introduction of bridging veins at the top of the brain to study the mechanism of subdural hematomas. The model is also anatomically accurate. In addition to simulating the cerebral hemispheres of the brain and the cerebellum, it also simulates the scalp, skull, dura, cerebral spinal fluid, the membranes dividing the hemispheres and separating the hemispheres from the cerebellum. It also has a brain stem and a rudimentary face.

The model has many potential uses. It will improve automotive design for the protection of head injury. Even with the airbag, there are still many sources of head injury within the vehicle, including the windshield supports and the roof rail. It can be used to simulate a variety of impacts to different parts of the head to ensure that the protection is not limited to a few scenarios that can be simulated experimentally.

The present US Federal Motor Vehicle Safety Standard (FMVSS) 208 for head injury does not encompass the effects of head rotation on the brain. Our comprehensive model can be used to update the current standard to render it more compatible with current knowledge on head injury tolerance. Any change in US standards will have a worldwide effect because the US has been given a leadership role in mandating safety in cars.

Another potential application is in the design of helmets that are to be used for head protection. At present the plastic liner used in most of these helmets is of uniform thickness all the way around the head. This is because there is currently no logical way of determining where a thicker liner is needed. The model can be used design the liner for maximum protection without an undue increase in weight.

Clinically, the model can be used to pinpoint the location of injury within the brain if the approximate impact conditions are known. Although this can be accomplished by imaging techniques, such magnetic resonance imaging (MRI), head injured patients are difficult to control and may not remain still long enough for the completion of the scan. Identification of the part of the brain that is injured will enable the physician to direct treatment, including drugs, to that location, if and when these drugs become available.

This project is designed to help everyone in this modern age. Head protection is needed in many forms of mechanized travel as well as in many sporting activities. The intelligent design of heat protective surfaces as well as of helmets requires the use of tool which can predict the severity of brain injury for a given impact. At present, a rudimentary tolerance level, called the Head Injury Criterion or HIC, is the only measure used to assess injury risk. It is based on the Wayne State Tolerance Curve which was published almost 40 years ago and which is theoretically only valid for head impacts against rigid surfaces. Moreover, this criterion cannot account for brain injury caused by rotation of the head. Since the head almost always undergoes rotation in any impact, a better injury criterion is badly needed. A computer-based model which can be used to duplicate any conceivable impact to the head is a prime candidate for a new injury criterion.

On the highway, the model will benefit not only automotive occupants but also pedestrians, motorcylists and bicyclists. In sports, many athletes will benefit from this model. Those who use helmets will experience a decrease in the level of brain damage. For example, in American football, minor traumatic brain injury is of serious concern because these players sustain repeated head impacts while wearing a helmet. Perhaps, a helmet design that is scientifically based can reduce this risk for both amateur and professional football players. Other athletes who can benefit from better protection are hockey players, baseball players, boxers, skiers, ice skaters and riders of horses. It also has a potential application in the intelligent design of head gear for protection against intentional injury. This includes helmets for the foot soldier to those used by law enforcement officers in a riot control situation.

One of the major advantages of using a computer model to design for head protection is cost reduction. It minimizes trial and error design and can basically lead to a final design on the first or second try. The extent of protection is also improved dramatically because the model looks at all parts of the brain and not just a single parameter or a single area of the brain.

Clinically, the identification of the site of injury within the brain without subjecting the patient to an imaging study, such as an MRI, has its advantages. At present, the neurosurgeon or neurologist is unable to tell the site of injury based on knowledge of the site of impact. The model can narrow down the injured region for the application of site specific-drugs to treat the brain injury. Effective and rapid treatment is important because delays can result in serious secondary injuries.

To accomplish these tasks, a high performance computer is essential. The model is fairly complex and it can days to make a single simulation on a workstation. Supercomputers will reduce the time needed by at least 10-fold, rendering it feasible to do try and error studies using the computer.

The beneficiaries include the entire human race. People will be able to avoid head injuries or sustain a much lower level of injury when head protective devices are based on our computer model. We plan to make continuing improvements to the model to enhance its capabilities and the accuracy of its predictions.

The model used an existing software program which solves complex problems in stress analysis using the finite element method. The method of solution was adapted to solve problems involving impact on computers with large or small memories. We did not write new software but came up with a unique data set to simulate head impact. The data set contains a geometric mesh which describes all of the components of the head, including the scalp, skull, inner brain membranes, the hemispheres of the brain, the cerebellum, brain stem, ventricles, cerebrospinal fluid and bridging veins. All of these components were made up of 38,000 elements and 29,000 nodes (left). Another computer program was used to aid in the generation of this mesh although much of the mesh had to be generated manually to take care of the complex geometry of the human brain. The most difficult part was to obtain the boundaries of the gray and white matter. There is much interdigitation of the white matter into the outer gray matter and the large number of elements was due largely to the fact that the boundaries could not be accurately represented without making the elements very small. In the right figures, the green areas are the gray matter while the yellow areas are the white matter. The ventricles are shown in red. The data set also contains the material properties of the various tissues simulated.

The results are depicted graphically in terms of contours which show levels of stress at an instant of time during the impact. The shear stress contours in the brain at 1.7 milliseconds after initiation of impact are shown here. They show that the nerves connecting the left and right halves of the brain in the corpus callosum are subjected to high shear (red zones) during a frontal impact and that the effects of the injury to this area are a loss of the ability to process information and to respond normally to external stimuli. When there is a high angular acceleration associated with an impact, the bridging veins at the top of the brain can rupture causing a subdural hematoma or blood clot. This is a dangerous condition because the clot exerts large pressures on the brain and if they are not quickly relieved can deprive the brain of its blood supply, causing brain death. The stretch in some of the veins can reach 200% as shown in the top right.

Because of the large size of the model, a fast computer is needed to solve the problem in a reasonable amount of time. On a J-90 Cray computer with 4 parallel processors, the computer processor unit (CPU) time was 32 hours for a 15 millisecond simuulation. It could take days on a SPARC 10 Sun Workstation. According to automotive safety engineers, this range of CPU time on a Cray is still not good enough because in automotive safety testing, a large number of head impacts need to be analyzed to determine compliance with a safety standard. At present, the HIC takes only a fraction of a second to compute in comparison with the time needed to compute the stress or strain distribution within the brain. Thus, there is a need for at least a 10-fold increase in computational power to render the model useful for the assessment of safety designs for head protection. There is however every reason to expect that the reduction in computation time can be achieved long before the Government gets around to changing the safety standards.

The original features of the model include the modeling of the gray and white matter of the brain as separate materials and the inclusion of the bridging veins in the model. Additionally, the model is extremely detailed in terms of the modeling of the complex anatomy of the brain. All of this made possible because of the availability of high performance computers which can provide turn-around times in the order of hours instead of days. It is the state-of-the-art model for brain injury and has for the first time predicted sites of high strain that correlated with animal injury data. The project grew out of a need to determine sites of brain injury from a blunt impact of known magnitude and direction. Equivalent information could be obtained from animals but the number of tests that could be done is limited and the geometry of most animal brains is different than that of the human. Some forms of brain injury cannot be elicited from cadaveric brains. In view of this need in view of the long history of head injury research at Wayne State, it was decided that a comprehensive model should be developed as an adjunct to the cadaveric and animal testing that has gone on and that is still on-going. The modeling effort started about 10 years ago rather modestly, in the form of simple two-dimensional models. The master plan for the current phase of head injury research also includes the use of state-of-the-art high-speed x-ray equipment to measure brain motion in cadavers and rats. This is a unique piece of equipment which has a bi-axial x-ray source coupled to two high-speed digital cameras that yields clear images of extremely small targets, using computer technology. Additionally, computer enhancement of the images was used to determine the motion of these targets more accurately. The experimental data are now being used to validate and improve the computer model.

This project has met its goals and has the promise of exceeding them once it is fully validated. In its current form, it is operational and is already widely known. We have had inquiries about its availability from France, the United Kingdom, Canada, Japan and the US. The two Ph.D. students who developed this model were hired by the automotive industry as soon as they became available to them. The model was licensed by the University to a consulting firm in Canada and an earlier version of the model was licensed to the software company that developed the finite element code used to run our model This model is being used by the French Government for military helmet development. Other industries interested in our model include the aerospace industry, the automotive industry, professional football organizations and the military.

Although it is not possible to estimate the number of people who will benefit from the use of this model, there is no question that it will contribute to a reduction in brain injury in a large variety head impact situations.

In terms of future plans, we would like to improve the model further by adding more features to it. These features include more blood vessels, more accurate material properties for the tissues of the head, including the brain and improvements to the simulation of the cerebrospinal fluid. Eventually, we hope to simulate the response of individual neurons and axons for a better understanding of how they might be injured.

Several difficulties were encountered during the development of the model. There is a lack of knowledge in the area of material properties of brain tissue which are essential to the development of an accurate model. Much of the available data were obtained at low speeds and are not directly usable in a dynamic simulation. Many attempts were made to narrow down the range of values before a good correlation with experimental data was obtained. Then, it was discovered that the model could not predict the sites of injury as seen in a pig's brain, based on studies done at a collaborating university. To overcome this difficulty, the innovative idea of modeling gray and white matter as different materials was introduced into the model. This resulted in an accurate prediction of the sites of injury, using a two-dimensional model of the pig's brain. There were no difficulties in getting the model accepted by the biomechanics community, as evidenced by the widespread interest in the model and requests for its availability. Funding was provided by the National Center for Injury Prevention and Control, Centers for Disease Control and Prevention (CDC) in Atlanta, GA. We are now in the eighth year of the grant.