Forensic Engineering Expert Witness Blog

Helmet-to-helmet impacts in football: Why are struck players concussed more often than striking players?

Posted on Tue, Jun 05, 2018 @ 08:28 AM

Helmet-to-helmet impacts in football cause more concussions in players getting hit than in players doing the hitting. While the occurrence of blindside hits may seem like a reasonable explanation for this pattern, the effects of player awareness and neck muscle tension on concussion are likely too small to fully account for this pattern.

Another possible explanation is that a striking player tends to use the front of their helmet to strike a wide range of sites, like the side and rear, on their opponent’s helmet. These different impact locations may be important for two reasons. 

First, we know from prior research that the brain is more vulnerable to rapid head twists (“no” motions of the head) than to rapid head nods (“yes” motions of the head). This difference means head impacts that cause twisting motions can lead to more brain injuries than head impacts that cause nodding motions. 

Second, helmeted heads aren’t spherical and identical like billiard balls, so the striking and struck players’ helmets and heads can experience different types of motions during the same impact. Some impact locations produce more linear motion while others produce more rotational motion. 

To test if these two phenomena could explain why struck players are concussed more than striking players, we ran two computer models of the skull and brain using head motions taken from laboratory football helmet impacts to 12 different locations on the helmet. Our results showed that for the same initial impact energy, the brain experienced lower strains for impacts to the forehead and crown of the helmet than for impacts to other regions of the helmet. Less brain strain means less potential for brain injury. 

These findings suggest that striking players are choosing impact locations that minimize their own risk of brain injury. Struck players, on the other hand, often don’t have this choice and thus see impacts over a wider range of impact locations. As a result, struck players’ brains can see higher levels of strain than striking players’ brains. Across all impacts, these higher strains will manifest as more concussions in players receiving the hit than in players delivering the hit.

concussion blog figure

Figure: Laboratory impact locations (top row) and heat maps of the peak brain strain in three orthogonal slices through the brain (bottom three rows) for two impacts with the same input energy, i.e., the same impactor speed.  The forehead impact represents a common location for a striking player’s head impact and the jaw pad impact represents a location more commonly associated with a struck player’s head impact. The heat maps, which are generated using the SIMon brain model, show higher brain strains from the twisting motion that develops in the jaw pad impact.

For more information, see: Elkin BS, Gabler LF, Panzer MB, Siegmund GP (2018). Brain tissue strains vary with head impact location: A possible explanation for increased concussion risk in struck versus striking football players. Clinical Biomechanics. doi: 10.1016/j.clinbiomech.2018.03.021.

Tags: Helmets, concussion

HOT SHOCK: Illuminating Evidence of Headlamp Use

Posted on Wed, Feb 21, 2018 @ 06:15 AM

Whether a vehicle’s lights were on or off can affect liability for a nighttime car crash. Fortunately, there is often physical evidence of whether a vehicle’s lights were on or off at the time of a collision.

An incandescent lightbulb emits visible light when an electric current is passed through the finely coiled wire filament suspended inside its glass bulb. The dense tungsten-based filament softens at hot incandescent temperatures and can be stretched by the filament’s own inertia when the bulb is abruptly accelerated during a crash. This stretching phenomenon is called “hot shock”[1] and its presence indicates that a bulb was on at the time of impact. If instead the bulb were off at impact, the filament would be cold (less soft and deformable), which can lead to a broken filament rather than a stretched filament, or no damage at all to the filament.

The figure below shows digital microscope images of a bulb with hot shock compared to an undamaged exemplar bulb. This examination was done non-destructively by carefully looking at the  filament through the glass envelope of the bulb. The stretching of the filament and uneven coil spacing show that this filament was energized at the time of impact. In contrast, the undamaged bulb has uniform coil spacing and no evidence of stretching.

Hot shock1.jpgFigure 1: Digital microscope images of a bulb with hot shock (Bulb A, Inset A) compared to an undamaged exemplar bulb (B, Inset B). The white arrows show regions of filament stretching in the hot shocked bulb.

Some caution and experience are needed to properly interpret bulb filaments following a crash. Hot shock can be confused with “age sag”, where the filaments of old bulbs can droop downward and stretch after prolonged use. Also, the lack of hot shock does not necessarily mean that a bulb was not on at impact: it may simply mean that the acceleration was not large enough to deform the filament.

This type of forensic lightbulb analysis can be performed on headlights, brakelights and other incandescent bulbs with filaments. Some modern lighting systems that include light emitting diodes (LEDs) do not have heavy coiled filaments and therefore do not yield the same kind of forensic information as incandescent bulbs. Lightbulb evidence is also very fragile and should be captured early by examining and/or carefully extracting the bulbs soon after the crash. Attempts to energize a car’s electrical system to test other vehicle or safety systems can spoil lightbulb filaments and therefore should not be done until the lightbulb evidence has been preserved.

[1] JS Baker, LB Fricke, KS Baker, TL Aycock, Lamp Examination for ON or OFF in Vehicle Collisions, Northwestern University Center for Public Safety,  (2003)

Tags: illuminating evidence, lightbulbs, Hot shock

Bicycle helmets stand the test of time

Posted on Thu, Jul 13, 2017 @ 08:00 AM

Helmet performance.jpg

Helmets withstand the test of time. Using 770 donated bicycle helmets, Alyssa DeMarco, Dennis Chimich, and Gunter Siegmund found little or no evidence of helmet age-related deterioration in impact performance. Within the limitations of the sample and test conditions, these findings do not justify replacing a damage-free helmet every 2-10 years as recommended by some helmet manufacturers. A full-text, view-only version of the study “Age has a Minimal Effect on the Impact Performance of Field-Used Bicycle Helmets” is available here: Thank you to our collaborators: Craig Good at Collision Analysis and Jeff Bakal at the University of Alberta.

Tags: Helmets

Concussion and Whiplash Injury

Posted on Thu, Mar 16, 2017 @ 08:00 AM

Can a concussion be caused by a low-speed rear-end crash? The prevailing science suggests the answer is no, but new research* has identified some cases where the answer could be yes.

Most people link whiplash injuries to low-speed rear-end crashes and concussions to direct head impacts. These links are sufficiently pervasive that the type of impact—rear-end crash versus a direct head impact—may influence a clinical diagnosis for what can be a similar pattern of symptoms. From a biomechanical perspective, rear-end crashes generate head impacts with the head restraint and direct head impacts generate forces in the neck, which mean that mechanisms for both whiplash injury and concussion often co-exist.

To answer the question of whether a concussion can occur in a low-speed rear-end crash, we compared how a computer model of the brain responds during sport-related head impacts to how it responds during head restraint impacts that occur in rear-end crashes.

We used head impact data from football and rear-end crash tests as input to the brain model and then examined the resulting brain strain, i.e., how much the brain stretches and distorts. We then compared these results to data from concussed NFL players whose head impacts were reconstructed from game video, and in particular to two players whose impacts were to the back of the head, the region of the head that strikes the head restraint. We found that football impacts caused higher brain strains than most of the rear-end crash tests we examined, but there were some instances where both types of impacts produced similar amounts of strain in the brain. In one crash involving a 15 km/h speed change and a low head restraint, the dummy’s head wrapped onto the top of the head restraint and caused similar brain strains to head impacts that were more severe than the two responsible for concussions in NFL players.

Image_Concussion_2017_brain strain.jpgFigure - Two images of the peak strains on the brain’s surface as calculated by the SIMon brain model. The left image shows the largest brain strains produced by a rear-end collision, which consisted of a 15 km/h (9 mph) vehicle speed change with a low head restraint. The right image shows the largest brain strains produced during a low rear football impact, which simulated a helmeted head impact with a closing speed of 33 km/h (21 mph). The simulated football impact was more severe than two low rear impacts that have caused concussions in NFL players.

These findings show that brain strains associated with some concussions can occur during rear-end crashes in some special cases. They also show that most exposures to rear-end crashes probably don’t generate the amount of brain strain associated with concussion. For clinicians, our findings mean that they should not rule out a concussion just because their patient experienced a rear-end car crash, particularly if the engineering analysis indicates a higher speed change combined with a poorly positioned head restraint.

* Elkin BS, Elliott JM, Siegmund GP (2016). Whiplash injury or concussion? A possible explanation for concussion symptoms in some individuals following a rear-end collision. Journal of Orthopaedic & Sports Physical Therapy 46(10):874-85.

Tags: conscussion, whiplash, brain strain

Distracted Driving: How it Affects Liability

Posted on Fri, Jan 22, 2016 @ 11:00 AM

Distraction has always affected drivers, but the addition of interactive devices like cell phones and navigation systems has raised both the opportunity for distracted driving and awareness of its perils. Although official US statistics show that only 17% of injury-causing collisions involved distraction,1 a yearlong study monitoring drivers during regular vehicle use showed that inattention (which includes distraction) was present in the three seconds before 78% of all crashes2. Based on these data, distracted driving is a big problem. 

Proving that a driver was distracted before a collision is difficult, but understanding how distraction affects driver performance can help determine whether a driver’s ability to avoid a collision was impaired by distraction.

Effects of Cognitive Distraction

Cognitive distraction, which is anything that takes a driver’s mind off the road and away from the primary task of driving, affects a driver’s visual behaviour and response time. Drivers who are cognitively distracted spend more time focused on the road ahead and less time scanning the periphery3-6. This narrowed focus is commonly called cognitive tunnel vision and can lead to missed or late detection of hazards outside of the forward field of view.

Cognitive distraction also increases response time7-8. A recent driving simulator study9 showed that response times vary with the type and location of a hazard, but in all cases cognitively distracted drivers had longer response times than undistracted drivers. Response times increased by as much as 0.4 seconds, which adds two extra car lengths to the stopping distance at 80 km/h or 50 mph.

Was a Driver Distracted?

Establishing liability based on distraction typically means proving that the driver was distracted (e.g., on their cell phone) and then showing the distraction affected the likelihood or severity of the collision. Witness evidence or phone records can establish the driver was engaged in a distracting task. Proving the distraction affected the collision is more nuanced and requires an investigator to perform a series of analytical steps. For instance, to establish that the distraction affected a driver’s ability to avoid a collision, the following steps might be needed:

  • Use the physical evidence from the collision to estimate the driver’s response time,
  • Compare the driver’s response time to the typical response times for drivers facing a similar situation, and
  • Assess whether a typical response time would have avoided the collision.

In some cases, a slow response time alone may be sufficient to show that the driver was inattentive or distracted before the collision, but this generally requires a response time significantly longer than the range for the general population.

Distractions reduce driver performance and can be partially or completely responsible for a collision and the resulting damages. Ask your collision reconstruction engineer or human factors expert if distraction played a role in your next case, and in the meantime, put away your cell phone and minimize your other distractions while driving.

  1. National Highway Traffic Safety Administration (2013). Distracted Driving 2011, DOT HS 811 737. NHTSA’s National Center for Statistics and Analysis: Washington, DC.
  2. Dingus TA et al (2006). The 100-Car Naturalistic Driving Study, Phase II – Results of the 100-Car Field Experiment, DOT HS 810 593. Virginia Tech Transportation Institute: Blacksburg, VA.
  3. Hammel KR, Fisher DL, and Pradhan AK (2002). Verbal and Spatial Loading Effects on Eye Movements in Driving Simulators: A Comparison to Real World Driving. Proceedings of the Human Factors and Ergonomics Society Annual Meeting, vol. 46, no. 26, pp. 2174–2178.
  4. Harbluk JL, Noy YI, Trbovich PL, and Eizenman M (2007). An on-road assessment of cognitive distraction: Impacts on drivers’ visual behavior and braking performance. Accident Analysis and Prevention, vol. 39, pp. 372–379.
  5. Recarte MA and Nunes LM (2000). Effects of verbal and spatial-imagery tasks on eye fixations while driving. Journal of Experimental Psychology: Applied, vol. 6, no. 1, pp. 31–43.
  6. Reimer B (2009). Impact of Cognitive Task Complexity on Drivers’ Visual Tunneling. Transportation Research Record: Journal of the Transportation Research Board, vol. 2138, pp. 13–19.
  7. Caird, J. K., Willness, C. R., Steel, P., & Scialfa, C. (2008). A meta-analysis of the effects of cell phones on driver performance. Accident Analysis & Prevention, 40(4), 1282–1293.
  8. Horrey, W. J., & Wickens, C. D. (2006). Examining the Impact of Cell Phone Conversations on Driving Using Meta-Analytic Techniques. Human Factors: The Journal of the Human Factors and Ergonomics Society, 48(1), 196–205.
  9. D’Addario P (2014). Perception-response time to emergency roadway hazards and the effect of cognitive distraction, Master of Applied Science. University of Toronto. 

Tags: Human Factors, distracted driving

Safely Towing a Trailer

Posted on Thu, Jun 11, 2015 @ 10:30 AM

Trailers are an important part of many family vacations, but they can also spell tragedy if they are not properly equipped, properly set up and properly towed.

Tow vehicles handle, brake and corner differently when towing a trailer. Stopping distances can be twice as long when towing a trailer, and during cornering the trailer’s wheels track inside of the tow vehicle’s wheels. Thus wide turns are needed to avoid the trailer striking the inside curb. Drivers, particularly those with little or no trailering experience, should drive slowly while they learn how the combined tow vehicle and trailer handle and stop.

As vehicle speed increases, the stability, stopping performance and the ability to swerve or maneuver are reduced. Avoid exceeding the speed limit and slow down before cornering and descending a hill. Trailer stability is reduced in turns because the trailer pushes the rear of the tow vehicle towards the outside of a turn – a phenomenon that can lead to jackknifing.

Trailer sway is a side-to-side oscillation of the trailer that can lead to a control loss and possibly a crash or rollover. Trailer sway can be induced by a sudden steering maneuver, lane change, or even a wind gust. Trailer sway will often diminish after a few oscillations, but high speeds and improper trailer weight distribution can cause the sway to increase with each oscillation until the vehicle loses control.


The likelihood of trailer sway increases if too little of the trailer’s weight rests on the hitch. The hitch load should generally be between 10% and 15% of the total trailer’s weight (including cargo). To achieve this hitch load, some trailer manufacturers recommend the 60/40 rule: put 60% of the trailer’s total weight forward of the trailer axles and 40% behind the axles.

When a trailer starts to sway, the best response is to hold the steering wheel straight, let off the gas and apply the trailer’s brakes. Applying only the trailer’s brakes counters the sway and will help realign the trailer with the tow vehicle. To prevent further sway, drive slower and redistribute the load if necessary.

Safe trailering begins with the proper equipment, proper setup and proper driving. Consult your owner’s manuals for instructions on how to properly equip, set up and drive your tow vehicle and trailer.

Tags: Failure Analysis, Trailers

Can A Priori Risk Establish Injury Causation?

Posted on Thu, Mar 26, 2015 @ 07:09 AM

The risk of sustaining an injury during a specific kind of crash can be calculated from real world accident data. We simply take the number of injured individuals following a crash and divide this by the number of individuals exposed to this kind of crash. The resulting number, expressed as a percentage, is the a priori risk of that injury for that kind of crash. Here, “a priori risk simply means the risk of being injured before the crash occurs.

Once a crash occurs, however, the a priori risk of injury becomes less meaningful to the exposed individual, who is either injured following the crash or not injured following the crash.  This diagnosis of an injury is made by a medical professional.

In our work, we are routinely asked to establish or refute the causal link between a diagnosed injury and an event alleged to have caused that injury. This question may be asked despite a medical diagnosis for various reasons, including: i) time has elapsed between the incident and when the injury was reported, ii) there is disagreement regarding the diagnosis, or iii) there may be other plausible explanations for the injury.

To assess this causal link, we routinely calculate the force applied to the injured tissue and then compare this force to that shown to cause the injury during controlled experiments. When the applied forces cannot be reliably calculated or the forces associated with the injury are not known, we sometimes turn to a priori risk data. These risk data, however, must be used carefully.

A priori risk data can tell us that a specific injury is possible in a specific kind of crash. This is useful because it confirms that other individuals exposed to a similar crash have been diagnosed with the injury, though these data alone fall short of demonstrating that the injury was probably caused by the crash.

A priori risk values are sometimes used to conclude incorrectly that a diagnosed injury is not causally linked to a specific crash. The argument goes like this: Since only 6% of individuals in a study sustained this specific injury in this specific kind of crash, it is unlikely that the injured party sustained his/her injury in this crash. This logic is flawed because it applies data from many individuals (the study population) to a single individual. Just as the a priori risk generated by the entire study population cannot be applied to a single subject within the study, the a priori risk cannot be applied to a single person outside the study.

In summary, estimates of the a priori risk of injury can help establish that an injury is possible in specific kinds of crashes; however, they cannot be used—in isolation—to conclude that an injury is probably or probably not related to a specific kind of crash.

Tags: Injury Biomechanics

When Should I Replace My Helmet?

Posted on Thu, Mar 12, 2015 @ 11:00 AM

It’s common knowledge that a helmet should be replaced after a significant impact. Exposure to extreme heat or chemicals, like gasoline, ammonia and solvents, can also degrade a helmet. But aside from impact, heat or chemical damage, does a helmet need to be replaced just because it is old? Do the impact attenuating properties of the helmet diminish with time? Put more simply, do helmets have an expiry date?

Owners’ manuals for motorcycle, bicycle and snow-sport helmets often suggest that helmets should be replaced every 2 to 10 years. The scientific basis for these recommendations is unclear, and a cynical consumer might view the short end of this range as self-serving for helmet manufacturers.

Few published studies have evaluated the effect of age on helmet impact performance. Previous tests we did showed that unused 10-year-old motorcycle helmets attenuated impacts as well as new helmets (see this link). But showing that one old helmet model performs like an array of newer helmets does not answer the more general question of whether helmet impact performance degrades with age.

To address this bigger question, we acquired and tested hundreds of undamaged used bicycle helmets of various makes, models and ages. Our preliminary findings, presented at the World Congress of Biomechanics in July 2014, showed that helmet age does not have a significant effect on the impact performance of bicycle helmets. Based on these data, age-related degradation in helmet impact performance is not a compelling reason to replace your old helmet.

As with all scientific studies, our study has limitations. Helmet performance relies on more than just the energy absorbing properties of the helmet’s foam. The straps, buckles, and other helmet features also affect helmet performance, and we did not study how age affects these components. We also tested only helmets without visible damage or significant wear and tear, and therefore our results do not mean that any old helmet will provide the same protection as a new helmet. Nevertheless, our results show that the impact properties of bicycle helmet foam are not significantly affected by age.

Helmet_headinjury_braininjury helmet_closeup

Figure 1 – Exterior and interior of a damaged helmet



Figure 2 – Impact performance (head acceleration) does not change significantly with the year of helmet manufacture.

Tags: helmet safety research, Injury Biomechanics, Helmets, head injury

Guardrail End Treatments

Posted on Tue, Dec 09, 2014 @ 02:01 PM

Roadside guardrails make our roads safer, but the ends of these guardrails need protection, or end treatments, to prevent them from becoming roadside hazards of their own. Recent court cases involving some end treatments have focused media and public attention on the potential hazards associated with these special features of guardrails.

The ends of guardrails are designed to absorb the energy of vehicles that crash head-on into them. There are many different designs to achieve this goal, but two “extruder-type” end treatments have received most of the recent attention[i] (Figure 1). These two extruders are mounted onto the end of common W-shaped steel guardrails and are then pushed along the guardrail by the vehicle during a crash. The moving extruder flattens the guardrail and bends it back on itself, a process that absorbs energy and slows down the vehicle. The posts supporting the guardrail also break away, absorbing additional energy.


                                Figure 1: Extruder type guide rail end treatment

Extruder end treatments are designed for primarily frontal vehicle impacts, and therefore are most commonly installed on straight roadways. Extruder end treatments can fail to operate properly under some conditions: if they are improperly installed, if their design is defective, or if vehicles strike them too obliquely. These failures can cause three outcomes:

  • The extruder head bends back with the steel beam and allows the vehicle to get behind the guardrail before slowing the vehicle;
  • The extruder head drops to the ground and the vehicle may then ride up and over the extruder and guardrail; or,
  • The extruder head “locks” on the steel beam and penetrates (spears) the vehicle.

All guardrail end treatments must pass a series of crash tests before being approved for roadside installation. Early tests conformed to a standard issued in 1981[ii], but most currently installed end treatments meet a revised, more comprehensive standard issued in 1992[iii]. Since 2009 end treatments are required to meet a third standard[iv] that is similar to the second standard. The crash tests in these standards attempt to capture typical impact conditions, and include passenger cars and larger pickups striking the end treatments at speeds up to 60 mph (100 km/h) at a maximum angle of 25 degrees. The test vehicles are driven along a straight path and only the front of the vehicle strikes the end treatment. In practical terms, this means that there are real-world crash conditions (e.g., angled or side impacts) that fall outside of the standard test conditions and may result in failures of the end treatments.

Cases involving the potential failure of a guardrail end treatment need to be properly investigated. Key steps in the investigation process include i) examining the vehicle, guardrail and end treatment, ii) determining if the installed device was approved, iii) assessing the appropriateness of the end treatment for the specific location, and iv) documenting how the end treatment was installed and maintained. All of these factors then need to be interpreted in the context of the specific vehicle and collision dynamics. A roadside safety engineer can help evaluate how well a guardrail end treatment functioned and whether a properly functioning device could have affected the collision outcome.


[i] Trinity ET-Plus and its predecessor the ET-2000 are the extruders involved in the recent court cases

[ii] National Cooperative Highway Research Program: NCHRP-230 “Recommended Procedures for the Safety Performance Evaluation of Highway Appurtenances”, TRB, 1981

[iii] NCHRP-350 “Recommended Procedures for the Safety Performance Evaluation of Highway Features”, TRB, 1993

[iv] “Manual for Assessing Safety Hardware (MASH)”, American Association of State Highway Officials (AASHTO), 2009

Tags: accident reconstruction

Three Factors for Assessing a Driver's Emergency Response

Posted on Wed, Nov 19, 2014 @ 03:47 PM

A driver who fails to avoid a roadway hazard, such as a pedestrian, cyclist or other car, could have his pre-crash responses closely scrutinized if the crash is investigated.  Driver responses are generally scrutinized in three ways: 1) response speed, 2) response choice, and 3) response magnitude.

Response Speed

The time it takes a driver to respond to a hazard is called the perception/response time. This time begins the instant a hazard becomes visible and ends when the driver begins to steer or brake. It captures everything from figuring out what the hazard is doing, deciding how to respond to the hazard, and then contracting the muscles needed to either steer or brake the vehicle.

Differences in individual drivers, hazards and environmental conditions affect the length of perception/response times.  By adjusting for these differences, a driver’s actual perception/response time can be compared to those measured in detailed studies of driver behaviour in similar situations. This comparison can then establish if a specific driver’s response was normal or unusually delayed.

Response Choice

When faced with a complex hazard, a driver may not know if braking, steering or doing both is the best way to avoid a crash.  For instance, a driver might brake and steer to the left to avoid a child who runs onto the road from their right.  This would be the right choice if the child sees the car and stops, but could be the wrong choice if the child continues running (see figure). In the latter case, not steering at all would have avoided the collision.

This sort of after-the-fact analysis of a driver’s choice is affected by hindsight bias1, which means that, with the benefit of hindsight, the analyst can clearly see the best choice because he knows what the child actually did.  At the time the driver had to choose a response, the driver did not know what the child would do.

For this reason, some legal jurisdictions acknowledge that drivers facing the “agony of collision” should not be held to a standard of perfection, but rather to a standard of what a “reasonable” driver would do in similar circumstances.  Thus, information about how typical drivers respond to a situation can help judges and juries see that a specific driver’s response was reasonable, despite the unfortunate outcome.


Figure 1: With hindsight, we can see that steering left away from the incoming child, contributed to the collision, whereas not steering would have avoided an impact.

Response Magnitude

Most drivers do not apply their vehicle’s brakes maximally in an emergency. A study of drivers who encountered a child dummy suddenly crossing their path found that 86% of drivers had a two-stage braking response: they quickly applied their brakes to an initial submaximal level and then paused to assess the situation before braking harder2. Drivers who used two-stage braking were 17% more likely to strike the dummy than drivers who braked hard immediately. Here again an analysis that assumes a driver braked hard immediately might show the collision could be avoided when the driver’s less perfect braking behaviour was the same as a typical driver.


Response speed, choice and magnitude need to be considered carefully in the context of a driver’s specific collision conditions to properly assess whether their response was typical of most drivers or atypical and possibly responsible for the collision.

[1] Dilich M, Kopernik D, Goebelbecker J (2006)  Hindsight judgment of driver fault in traffic accident analysis – Misusing the science of accident reconstruction.  Transportation Research Record: Journal of the Transportation Research Board, No. 1980, 1-7.

[2] Prynne K & Martin P (1995).  Braking Behaviour in Emergencies, 950969.  Society of Automotive Engineers.


Tags: accident reconstruction, Human Factors