It is well recognized that many everyday activities involve a risk of traumatic brain injury (TBI) such as bike riding, roller blading, or playing contact sports. An accidental yet serious blow to the head can cause brain injury due to extremely rapid head motions that translate into damaging deformation of brain tissue.1–6 Injury thresholds based on the biomechanics of head motion have been extensively characterized and used as standards to determine the effectiveness of head protection measures.7–13 However, in both the general press and medical literature there has been confusion over how head motions are linked to TBI.
Gravitational force (G-force), a directionless quantity of linear acceleration, is often inappropriately reported as the sole risk factor for brain injury. For instance, a series of medical case reports have described a potential causal relationship between intracranial hemorrhage and riding “high G-force” roller coasters.14–18 A recent review in a medical journal concluded that “emergency physicians should consider amusement park rides a possible cause of unexplained neurologic events,” related to “dangerously high G-forces.”19 Accordingly, news stories that quote physicians frequently attribute high G-forces induced by roller coasters to causing TBI in some riders.20–23
Misperceptions of the relationship between high risk activities, G-forces and TBI have had a surprisingly broad impact in our society. Without regard to years of scientific head injury research, news reports and anecdotal medical case reports have inspired legal suits for substantial monetary damages on behalf of roller coaster riders. As such, legislative acts have proposed to limit the level of G-forces on amusement park rides at the state level21,24,25 and regulation of the amusement park industry at the federal level.20,21,24,26 It is important to note, however, there is no scientific evidence demonstrating that G-forces induced by amusement park rides pose any risk of TBI. To the contrary, 2 independent scientific panels and an engineering consulting firm failed to find a connection between brain injury and roller coaster riding.27–29 In addition, by using a simple mathematical model, our group previously calculated that the peak head accelerations during roller coaster rides were far below standardized thresholds for injury.30
The source of confusion appears to be a fundamental misunderstanding of how G-forces play a role in the biomechanics of TBI.31–33 A long history of brain injury and motor vehicle safety research has shown that a peak G-force measurement alone is a very poor measure to determine the probability of injury to the brain.1,4–8,11,12,34,35 Rather, all the kinematic parameters of head motion must be considered. The direction (linear and rotational in 3 dimensions), duration and magnitude of the motion are all important parameters to accurately determine if TBI thresholds have been exceeded.
To address the general misunderstanding of the role of high G-forces and TBI in typical daily activities, we examined real-time 3-dimensional head motions of volunteers during roller coaster rides, low speed car crashes, and strikes with a pillow. The collective data were interpreted using established head kinematic parameters and compared with known thresholds of brain injury.
Selection of Participants
Four volunteers were selected ranging in age and weight to assess individual differences. Subject 1 was a 27-year-old man, 165 lbs, 70 inches tall; subject 2 was an 11-year-old male, 100 lbs, 57 inches tall; subject 3 was a 13-year-old male, 86 lbs, 55 inches tall; and subject 4 was a 24-year-old woman, 150 lbs, 62 inches tall. Volunteers were recruited by Six Flags Great Adventure to ride 3 roller coasters, participate in a pillow fight, and a 5 mph car bumper hit. A Six Flags review board of professional engineers familiar with amusement ride test protocols approved the experimental procedures and participant consent. Each volunteer was asked to sign an informed consent form. In the case of the minors, the consent forms were signed by the minors' guardians, who were present during the testing.
Study Design and Setting
Instrumented volunteers took part in 5 intense activities while the 3-dimensional kinematics of their heads was measured. Experiments took place on 3 distinctively different roller coasters at a Six Flags Amusement Park: (1) an inverted (track is overhead) looping coaster with a top speed of 55 miles per hour (mph), (2) a linear induction motor launch coaster that rapidly accelerates the riders to 65 mph and has several intense loops, and (3) a tall and fast noninverting coaster with significant drops and a top speed of 85 mph. Each volunteer rode in the middle of the train under normal operating conditions. Two additional tests were performed to gather relative data from 2 common experiences that involve rapid accelerations to the head: (4) a pillow fight and (5) a common car bumper hit of approximately 5 mph into a barrier.
Methods of Measurement
An instrument biteplate was built for each subject that could be placed in the mouth and held in place by biting, shown in Figure 1. Three linear accelerometers (Crossbow, model CXL25M3, San Jose, CA) and 3 angular rate sensors (Murata, model ENC-03J, Tenjin Nagaokakyo-shi, Kyoto, Japan) were mounted onto the biteplate in an orthogonal fashion. The x-axis lies in the anterior-posterior direction, y-axis in the lateral direction, and z-axis in the axial direction. Accelerometer and rate sensor data was collected on a TDAS Pro Acquisition Unit (DTS Inc., Seal Beach, CA) at a 10,000 Hz sampling rate with a 3000 Hz antialiasing filter. A total of 6 channels of data were collected for the entire duration of the test. All DTS data acquisition hardware meets the requirements of SAE J211 and is certified to the National Highway and Traffic Safety Administration, the Federal Aviation Administration, and ISO 6487 standards.
To remove measurement noise, linear accelerations were filtered with a 1650Hz low pass filter as set forth in SAE J211 for CFC 1000 data. Angular velocities were filtered with a low pass filter of 600 Hz.36 Amusement rides produce rigid body accelerations in the frequency range up to 1.5 Hz, therefore angular velocity measurements were also filtered with a high pass filter of 1.5 Hz to remove DC offsets in the data per SAE J211 and J1727. Filtering was performed only on raw data prior to further calculations.
Angular accelerations of the head were calculated by differentiating the angular velocity measurements (3 point centered difference) and linear velocities were obtained by integrating the linear acceleration measurements (Euler's Method) using Matlab Software (Mathworks, Inc., Natick, MA). The directional velocities and accelerations were then combined into resultant vectors and peak accelerations and velocities were identified.
The maximum accelerations and velocities were compared with reported tolerance levels for concussion,37–39 subdural hematoma,40,41 and DAI.8,35 In addition, we related our results to other reported measurements of head kinematics that occur during daily living activities,42 heading a soccer ball, football hits, hockey contact,13,43–46 boxing,47 and an 18 mph rear end car crash test48 (Kleinberger et al, Johns Hopkins University Applied Physics Laboratory, personal communication, February 11, 2005).
Head Injury Assessment Functions
The results in this study and comparative published data were used to calculate, where possible, mathematical predictors of head injury developed for use in automobile safety testing.11 The head injury criterion (HIC), a federally mandated motor vehicle safety standard, was calculated by the expression:
Here, the conservative HIC15 was used where the time interval (t2 – t1) was 15 milliseconds. Thresholds for the HIC15 range from a maximum value of 390 in car crash safety standards49 to a value of 151 for a mild injury.13
Since the HIC only evaluates linear accelerations of the head, the head impact power (HIP),12 a function of both linear and rotational accelerations and velocities in 3 dimensions, was also evaluated by the expression:
Calculation of the Probability of Concussion From Football Data
The occurrence of concussion as a result of football tackles was used to approximate the probability of sustaining a concussion in other recreational activities.46 In this study, the diagnosis of concussion in National Football League players strongly correlated to the value of the HIC15. A probability curve of concussion versus HIC was constructed using the Consistent Threshold method, a nonparametric method for ranking censored data.50,51
A time history of the 3-dimensional kinematics of the adult and child head was acquired during 3 roller coaster rides, a pillow fight and a 5 mph (2.2 m/s) car bumper hit. Rotational velocities and linear accelerations were measured during each activity, from which, rotational accelerations and linear velocities were calculated. Directional x, y, and z components were combined into resultant vectors and the peak values were identified (Table 1). In addition, head kinematic test data was collected for an 18 mph (8.1 m/s) car crash simulation, using test dummies from the Johns Hopkins University Applied Physics Laboratory (Kleinberger et al, personal communication, February 11, 2005). The results were compared with published kinematic data from peer-reviewed studies on nonpenetrating brain injury and to head motions measured in contact sports.
The 18 mph (8.1 m/s) car crash simulation resulted in the highest measurements of linear acceleration (29.3 × G), linear velocity (4.0 m/s), and rotational velocity (17.1 rad/s) of the head (Table 1). The highest level of rotational acceleration (2054 rad/s2) was measured during the pillow fight. Interestingly, the pillow fight generated peak head accelerations and velocities greater than the 3 roller coaster rides. Despite the difference in the 3 roller coaster rides (ie, speed, turns, loops), they lead to similar head motions. It is important to note that variations in head motions were small between the roller coaster rides, pillow fight and 5 mph (2.2 m/s) car bumper hit.
It is well accepted that individual peak values of acceleration or velocity of the head alone are not adequate to predict the risk of a brain injury.1,4–8,11,12,34,35 The time interval over which they occur must also be analyzed. Two well-established head injury assessment functions, the HIC and HIP, were calculated to evaluate the head motion as a function of time (Table 1). The 18 mph (8.1 m/s) car crash lead to the largest values of HIC15 = 28.1 and HIP = 3.41. In contrast, the accelerations and velocities associated with the roller coaster rides and pillow fight occur over shorter time frames and therefore much smaller HIC15 and HIP values. This highlights the importance of both the magnitude of the head motions and the time frame over which they occurred. For instance, the male adult in the pillow fight experienced the same linear acceleration as the male child 1 on roller coaster 2 (10.5 and 10.2 × G, respectively); however, the HIC15 values are different (1.3 and 4.1, respectively). The transient component of these head motions can also be seen on inspection of individual recordings shown in Figure 2. When compared with Federal Automobile Safety standards, we find that the HIC15 values are 2 to 3 orders of magnitude below the minimally accepted values of 390 for infants and 700 for adults.49
Test data from studies on nonpenetrating brain injury were collected from peer-reviewed literature to compare the potential risk of sustaining an injury from the intense activities of this study (Table 2). All data analyzed in this study are substantially below the lowest reported kinematic parameter that resulted in a measurable level of brain injury including: diffuse axonal injury,8 coma and concussion,37,38 tearing of bridging veins,41 and concussion from football tackles.13,46 Importantly, these thresholds are not limited to 1 measured parameter. They were based on all reported measurements. Included in Table 2 are data from head motions in several common activities that did not lead to injury. The data in this study are consistent with peak values of rotational accelerations and velocities seen in soccer heading, and linear accelerations similar to plopping in a chair, yet far below what a boxer or a football player experience.
Recently published data on the occurrence of concussion in football provides a reference database to evaluate the probability of sustaining a concussion in similar intense activities.46 The results demonstrated that the HIC15 correlated closely with the diagnosis of concussion. We used this data to construct a nonparametric probability curve for concussion, Figure 3A. This curve indicates that the probability of sustaining a concussion is zero up to a HIC15 value of 77, corresponding to the first diagnosed concussion as a result of a football tackle. The roller coaster, pillow fight and 5 mph (2.2 m/s) car crash were more than 19 times below this threshold. Even the 18 mph (8.1 m/s) car crash was 3 times below this threshold. Comparing the calculated HIC15 values of this study and from other contact sports in the literature, Figure 3B illustrates that only boxing and football tackles fall within the probability of suffering a concussion.
Real time 3-dimensional motions of the heads of volunteers were measured during rides on 3 different roller coasters, strikes with a pillow, and low speed car crash simulations. It was found that the peak head accelerations and velocities from these intense events were all comparable and fell far below the established biomechanical thresholds for TBI.
When predicting TBI thresholds, all the parameters of head motion must be considered. Specifically, injury to the brain is dependent upon (1) the direction of head motion, (2) the magnitude of velocity and acceleration, and (3) the time frame over which it occurs.7,11,12,30 For instance, linear motions result in focal injuries such as skull fracture, cerebral contusions, and hematomas at the site of impact to the head. Rotational motions, on the other hand, cause extensive deformation of the brain and vasculature, inflicting diffuse axonal injury throughout the white matter. Overt damage, such as tissue tears in the white matter and intraparenchymal hemorrhage, only occur from deformations caused by exceptionally high levels of rotational acceleration, over very short time periods.6,40 The results of this study show that the linear and rotational head accelerations produced by riding a roller coaster are similar to the well-tolerated head motions experienced during a pillow fight or heading a soccer ball (Table 2 and Figure 3).
The perception that riding roller coasters presents a risk of TBI is currently not supported by epidemiological or scientific data. Rather, this misperception appears to stem from a fundamental misunderstanding of the role of G-forces in TBI linked with a handful of case reports of patients suffering brain bleeding around the time of riding a roller coaster.14,15,17,18,52–59 The human body can withstand very large G-forces when they occur over very short time periods. For instance, a sneeze generates linear accelerations of up to 10 G s, but occurs over only 0.002 seconds with no ill effect.42 Likewise, a boxer can withstand approximately 3 milliseconds punches of 100 Gs or more with no overt signs of injury.47 In contrast, a fighter pilot will lose consciousness at 5 to 9 Gs if these forces are sustained for over 40 seconds. In this case, loss of consciousness results from restriction of blood flow rather than mechanical injury to the brain.60 These collective studies illustrate that scalar measurement of linear acceleration (G-force) alone is a poor measure to assess risk to injury. Indeed, this limitation leads to the development of the HIC15 and HIP by automobile safety researchers to integrate the linear and head rotational motions over their respective durations.11
A probability curve for mild levels of TBI was created to relate the degree of head motions during a roller coaster ride to those that can occur in contact sports (Figure 3A). This curve statistically correlates the injury risk of concussion injury in football players with the HIC15 value caused by a football tackle. The probability of concussion in football was compared with the HIC15 values from roller coaster rides in this study and from published reports of other intense activities,13,43–47,61–67 as shown in Figure 3B. The HIC15 values that occur during a roller coaster ride fall below the probability of injury in all contact sports using current data from the literature, demonstrating a lower risk of TBI than from playing a sport.
It is important to note that it is currently difficult to establish biomechanical thresholds that will cause mild levels of TBI. In particular, established HICs consider the entire kinematics of head motion, however, most peer reviewed experimental studies do not report all these parameters. In addition, it can be difficult to diagnose brain injury and injury severity clinically. The diagnosis of concussion, for example, is very controversial.68 Accordingly, there is a lack of published experimental and clinical data to directly correlate head motion to minor injuries to the brain.
Another important limitation in assessing the cause of brain injury is the unknown presence of pre-existing conditions that could augment a person's susceptibility to injury. For example, a pre-existing brain aneurysm might rupture during or near the time of a roller coaster ride, as has been demonstrated in at least 1 fatality case.69,70 However, rupture of the aneurysm could occur from many factors other than head accelerations, such as hypertension due to excitement. Even if these individuals had preexisting conditions, such as cerebral vascular malformations, it is unknown whether hemorrhage is more likely to be induced by the level of head motions during roller coaster rides as opposed to other daily activities. The current study does not address this possibility.
Our current empirical data supports 2 scientific panels' opinions27,28 as well as previous results from a computational model.30 Specifically, head motions during roller coaster riding fall within the range of normal activities and are far below thresholds of TBI in normal individuals.
The authors thank Michael Kleinberger, Jack Roberts, and Andrew Merkel of the Johns Hopkins University Applied Physics Laboratory for their contribution of the 18 mph car crash simulation data.
1. Holbourn AHS. The Mechanics of Brain Injuries. Br Med Bull
2. Holbourne AHS. Mechanics of head injuries. Lancet
3. Gennarelli TA. Head injury in man and experimental animals: clinical aspects. Acta Neurochir Suppl (Wien)
4. Gennarelli TA. Mechanisms of brain injury. J Emerg Med
. 1993;11(suppl 1):5–11.
5. Meaney DF, Margulies SS, Smith DH. Diffuse axonal injury. J Neurosurg
6. Smith DH, Meaney DF. Axonal damage in traumatic brain injury. Neuroscientist
7. Hardy WN, Khalil TB, King AI. Literature review of head injury biomechanics. Int J Impact Eng
8. Margulies SS, Thibault LE. A proposed tolerance criterion for diffuse axonal injury in man. J Biomech
9. Meaney DF, Smith DH, Ross DT, et al. Diffuse axonal injury in the miniature pig: biomechanical development and injury threshold. Crashworthiness and Occupant Protection in Transportation Systems. Paper presented at: ASME Winter Annual Meeting, New Orleans, La; 1993.
10. Meaney DF, Thibault LE, Gennarelli TA. Rotational brain injury tolerance criteria as a function of vehicle crash parameters. In: Proceedings of the International IRCOBI Conference on the Biomechanics of Impacts; 1994; Barcelona, Spain.
11. Newman JA. Head injury criteria in automotive testing. SAE paper No. 801317, Society of Automotive Engineers, Washington, DC; 1980.
12. Newman JA, Shewchenko N, Welbourne E. A proposed new biomechanical head injury assessment function–the maximum power index. Stapp Car Crash J
13. Zhang L, Yang KH, King AI. A proposed injury threshold for mild traumatic brain injury. J Biomech Eng
14. Nencini P, Basile AM, Sarti C, et al. Cerebral hemorrhage following a roller coaster ride. JAMA
15. Bo-Abbas Y, Bolton CF. Roller-coaster headache. N Engl J Med
16. Snyder RW, Sridharan ST, Pagnanelli DM. Subdural hematoma following roller coaster ride while anticoagulated. Am J Med
17. Fukutake T, Mine S, Yamakami I, et al. Roller coaster headache and subdural hematoma. Neurology
18. Yamakami I, Mine S, Yamaura A, et al. Chronic subdural haematoma after riding a roller coaster. J Clin Neurosci
19. Braksiek RJ, Roberts DJ. Amusement park injuries and deaths. Ann Emerg Med
20. Kornblut AE. Bigger, faster coasters tied to head injuries: findings in a new study alarming, doctors say. Boston Globe.
May 6, 2000.
21. Yoshino K. Rides are blamed for brain injuries. Los Angeles Times.
June 8, 2002.
22. Yoshino K. Study finds brain risk for riders of coasters. Chicago Tribune.
February 10, 2002.
23. Luna N. Study cites dangers of thrill rides to brain. Orange County Register.
February 5, 2002.
24. Redfearn S. The trill is. deadly? Washington Post.
May 21, 2002.
25. Writer S. New rules for coasters–Recent deaths, injuries get CAL-OSHA involved. Daily News Los Angeles.
May 8, 2002.
26. Himmelberg M. Mind-blowing amusement? Safety: faster, wilder coasters stir debate on whether they can harm brain. Orange County Register.
June 22, 2002.
27. BIAUSA. Blue Ribbon Panel Review of the Correlation Between Brain Injury and Roller Coaster Rides.
Alexandria, VA: Brain Injury Association of America; 2003.
29. Exponent Failure Analysis Associates. Investigation of Amusement Park and Roller Coaster Injury Likelyhood and Severity.
Alexandria, VA: Failure Analysis Associates; 2002.
30. Smith DH, Meaney DF. Roller coasters, g-forces, and brain trauma: on the wrong track? J Neurotrauma
31. McCarthy J. Roller coaster forces get stronger, faster. Florida Today.
November 27, 2005.
32. Schwartz M. Theme park safety: faster can mean more dangerous. Press-Enterprise.
June 16, 2005.
33. Powers S. Thrills, chills on mission: space. Orlando Sentinel.
June 25, 2006.
34. Gennarelli TA, Ommaya AK, Thibault LE. Comparison of translational and rotational head motions in experimental cerebral concussion. In: Proceedings of the 15th Stapp Car Crash Conference; 1971; Warrendale, Pa. SAE p-39, 797–803.
35. Meaney DF, Smith DH, Shreiber DI, et al. Biomechanical analysis of experimental diffuse axonal injury. J Neurotrauma
36. Martin PG, Crandall JR, Pilkey WD, et al. Measurement techniques for angular velocity and acceleration in an impact environment. SAE paper No. 970575, Society of Automotive Engineers, Warrendale, Pa; 1997.
37. Ommaya AK, Hirsch AE. Tolerances for cerebral concussion from head impact and whiplash in primates. J Biomech
38. Ommaya AK, Yarnell P, Hirsch AE, et al. Scaling of experimental data on cerebral concussion in sub-human primates to concussion threshold for man. In: Proceedings 11th Stapp Car Crash Conference; 1967; Warrendale, Pa. Paper 670906.
39. Ewing CL, Thomas DJ, Lustick L, et al. The Effect of the Initial Position of the Head and Neck on the Dynamic Response of the Human Head and Neck to -Gx Impact Acceleration. In: Proceeding of the 19th Stapp Car Crash Conference; 1975; Warrendale, PA. Paper 751157.
40. Gennarelli TA, Thibault LE. Biomechanics of acute subdural hematoma. J Trauma
41. Lowenhielm P. Strain tolerance of the vv. cerebri sup. (bridiging veins) calculated from head-on collision tests with cadavers. Z Rechtsmed
42. Allen ME, Weir-Jones I, Motiuk DR, et al. Acceleration perturbations of daily living: a comparison to ‘whiplash.' Spine
43. McCrory PR. Brain injury and heading in soccer. BMJ
44. Naunheim RS, Bayly PV, Standeven J, et al. Linear and angular head accelerations during heading of a soccer ball. Med Sci Sports Exerc
45. Naunheim RS, Standeven J, Richter C, et al. Comparison of impact data in hockey, football, and soccer. J Trauma
46. Pellman EJ, Viano DC, Tucker AM, et al. Concussion in professional football: reconstruction of game impacts and injuries. Neurosurgery.
2003;799–812; discussion 812–794.
47. Pincemaille Y, Trosseille X, Mack P, et al. Some new data related to human tolerance obtained from volunteer boxers. In: Proceedings 33rd Stapp Car Crash Conference; 1989; Washington, DC. 177–190.
48. Kleinberger M, Liming V, Merkle A, et al. The role of seatback and head restraint design parameters on rear impact occupant dynamics. Paper presented at: 18th International Technical Conference on the Enhanced Safety of Vehicles; 2003; Nagoya, Japan.
49. Code of Federal Regulations. Title 49–Transportation. Chapter 5–National Highway Traffic Safety Administration, Department of Transportation, Part 571.208.
50. Di Domenico L, Nusholtz G. Comparison of parametric and non-parametric methods for determining injury risk. Paper 2003-01-1362. Proc Soc Automot Eng
51. Nusholtz G, Mosier R. Consistent threshold estimate for doubly censored biomechancial data. SAE paper No. 01-0714: 1179–1191, Society of Automotive Engineers, Detroit, Mich; 1999.
52. Biousse V, Chabriat H, Amarenco P, et al. Roller-coaster-induced vertebral artery dissection. Lancet
53. Blacker DJ, Wijdicks EF. A ripping roller coaster ride. Neurology
54. Burneo JG, Shatz R, Papamitsakis NI, et al. Neuroimages: amusement park stroke. Neurology
55. Kettaneh A, Biousse V, Bousson V, et al. Roller-coaster syringomyelia. Neurology
56. Lascelles K, Hewes D, Ganesan V. An unexpected consequence of a roller coaster ride. J Neurol Neurosurg Psychiatry
57. McBeath JG, Nanda A. Roller coaster migraine: an underreported injury? Headache
58. Pelletier AR, Gilchrist J. Roller coaster related fatalities, United States, 1994–2004. Inj Prev
59. Stahlfeld KR, Roozrokh HC. Traumatic bilateral ECCA injury in a roller coaster enthusiast. Ann Vasc Surg
60. Whinnery JE, Whinnery AM. Acceleration-induced loss of consciousness: a review of 500 episodes. Arch Neurol
61. Pellman EJ, Lovell MR, Viano DC, et al. Concussion in professional football: neuropsychological testing–part 6. Neurosurgery
. 2004;55:1290–1303; discussion 1303–1295.
62. Pellman EJ, Powell JW, Viano DC, et al. Concussion in professional football: epidemiological features of game injuries and review of the literature–part 3. Neurosurgery
. 2004;54:81–94; discussion 94–86.
63. Pellman EJ, Viano DC, Casson IR, et al. Concussion in professional football: players returning to the same game–part 7. Neurosurgery
. 2005;56:79–90; discussion 90–72.
64. Pellman EJ, Viano DC, Casson IR, et al. Concussion in professional football: injuries involving 7 or more days out–part 5. Neurosurgery
65. Pellman EJ, Viano DC, Casson IR, et al. Concussion in professional football: repeat injuries–part 4. Neurosurgery
. 2004;55:860–873; discussion 873–866.
66. Pellman EJ, Viano DC, Tucker AM, et al. Concussion in professional football: location and direction of helmet impacts-part 2. Neurosurgery
. 2003;53:1328–1340; discussion 1340–1321.
67. Viano DC, Pellman EJ. Concussion in professional football: biomechanics of the striking player–part 8. Neurosurgery
. 2005;56:266–280; discussion 266–280.
68. Parkinson D. Evaluating cerebral concussion. Surg Neurol
69. Herubin D. Rides may have to post aneurysm signs. Orange County Register.
October 2, 2001.
70. Edwards H. Ride “last straw” for woman–coaster contributed to death, report finds. Daily News of Los Angeles.
July 28, 2001.