Head injuries are a major concern for athletes, coaches, and parents. One of the first steps in delineating the effects of head impacts occurring during soccer play should be to determine an accurate head acceleration measurement. In most studies, the head acceleration occurring during soccer play is estimated on the basis of laboratory recreation (3,14,22,24). To understand the risk, real-time head acceleration measurements are necessary during soccer play.
Head acceleration has been successfully measured during sporting events previously (2,5,12,20,21,25), but it has been a challenge for researchers to successfully do this during soccer games or scrimmages because of the lack of headgear (6,13,14,24). Attempts have been made to measure soccer head accelerations but only in a laboratory setting (13,14,24). Although these efforts provide a starting point, data may not represent actual on-field events.
These previous attempts to measure head acceleration in soccer have been hindered by instrumentation limitations (13,14,24). Naunheim et al. (14) used a football helmet with a triaxial accelerometer mounted on the helmet’s vertex to measure head accelerations of high school soccer players, but the acceleration measured was the helmet acceleration. It has been shown previously that measuring helmet acceleration without adjustment, as opposed to head acceleration, vastly overestimates head acceleration (11). To address these instrumentation issues, researchers have used instrumented headpieces (13,24), instrumented bite plates (13,24), and ear accelerometers (22). Although these instrumentation changes allowed for head acceleration measurement, the subjects were still tethered to a data collection system and hindered by wired instrumentation (14,22,24).
The majority of soccer head acceleration measurement has focused on heading events (13,14,22,24), but it has been shown that the majority of head injuries result from other mechanisms, specifically, collision with another player (1). It is expected that, when comparing on-field head accelerations for header versus nonheader events, the nonheader events will have higher accelerations because they are generally the impacts that cause injury. There has been limited research into nonheader soccer impact events (10,26). Two studies have been performed, both using mechanical head surrogates. Whereas Janda et al. (10) focused on mitigation of head injuries resulting from goal posts, Withnall et al. (26) measured head accelerations in simulated player collision events. Further investigation is required to evaluate the accuracy of these recreations.
Laboratory recreations have provided insight into heading impact events (13,14,22,24). These recreations provide valuable kinematic and muscle activation data but are only representative of the least severe cases for head acceleration. However, all of the necessary instrumentation changes the ability of the player to move freely and, therefore, alters the dynamic of the impact. Ultimately, a wireless acceleration measurement system that does not inhibit movement and provides no head protection is needed to determine the linear and angular head accelerations during actual soccer play. The Head Impact Telemetry System (HITS) (Simbex, LLC, Lebanon, NH), a wireless head acceleration measurement system, was recently implanted into a soccer headband that has had the padding removed (Fig. 1) and validated for the soccer head acceleration environment (9). Although the validation did not assess the effect of hair on the potential movement of the system in relationship to the head, this potential motion was expected to be minimal. This instrumentation development facilitated the ability to evaluate on-field head acceleration measurements during soccer play.
The exact contribution of linear versus angular acceleration for a given impact when heading the ball is related to several factors including how quickly the neck muscles are recruited and the overall intent of the redirection (13,19,23). There is much debate as to whether linear or angular acceleration should be investigated when studying mild traumatic brain injuries (mTBI) because both have been shown to predict injury (8,17). For the current study, each will be evaluated independently along with various other head injury criteria, which have established thresholds for mTBI. These include the Head Injury Criterion (HIC) and the Gadd Severity Index, which were developed primarily for automotive impact. These criteria take into account acceleration over a period (7,15) and are useful only to assess one individual impact. There are no suggested head acceleration limits for multiple impact events.
The purpose of the current study was to collect on-field data during soccer scrimmages. It is hypothesized that higher linear and angular head accelerations will be measured during nonheader events. This is expected because the nonheader events are generally the ones that result in injury.
A total of 24 girls’ youth soccer players in the U14 age group agreed to participate in the study. Some participants were involved in more than one of the scrimmages and were processed as a new player for each. Before any testing, approval from Wayne State University’s Human Investigation Committee was obtained. All players and their parents gave written consent before participation in the study.
Players were fit with the HITS headgear by the researchers; a snug fit was required to maintain accelerometer contact with the back of the head (Fig. 1). The system has six (±250g) single-axis linear accelerometers (Analog Devices, Inc., Norwood, MA) placed tangentially to the head to measure both linear and angular accelerations during an impact event (9). After proper headgear fit was obtained, players were then asked to participate in a scrimmage on a grass soccer field. Data were recorded at 1000 Hz for each of the six separate scrimmages. Some participants were involved in more than one of the scrimmages and were processed as a new player for each. Players wore the headgear for the duration of the scrimmages, which lasted 30 to 65 min. When any event exceeds 10g, the system triggered and recorded 8 ms before the event and 32 ms after the event. The events were then downloaded to the sideline computer for later analysis. If the system was out of range, up to 100 impacts could be stored within the headband. Games were videotaped for later analysis in determining what type of impact occurred at each downloaded time point.
After field data collection, data analysis was performed using a validated algorithm provided by Simbex (3,4). Each individual impact was analyzed using this algorithm. Information obtained included several parameters including HIC and the resultant linear and angular accelerations. These values were then compared with currently available injury tolerance values (27). Along with this information, the number of impacts, location of impact, and type of impact event (header or nonheader) were assessed. Nonheader events included were as follows: player collisions with other players, player falls, collisions with goalposts, and unintentional collisions with the ball. The type of impact was determined from video analysis and by using an observational record taken during the scrimmages. Videos were captured at 60 frames per second using a single standard camera at the sideline that was set to capture the entire field.
Descriptive statistics were calculated for each of the header locations and types of nonheader impacts for linear head acceleration, angular head acceleration, and HIC 15. In addition, comparisons were made between headers for each location for linear head acceleration, angular head acceleration, and HIC 15 to determine differences in severity by location. This was done using an ANOVA (P < 0.05). For this analysis, the left and right side impacts were combined into one category (side). All three criteria were found to be statistically significant in the ANOVA test (P < 0.05). The Games–Howell post hoc test was performed because of the small group sizes and the lack of homogeneity of variance between groups. Comparisons were also made between the header and nonheader groups to determine severity differences in linear head acceleration or angular head acceleration using individual t-tests (P < 0.05).
A total of 47 header impacts and 20 nonheader impacts were observed during the study. The front location of the head experienced more headers than the other locations (n = 17), followed by the top of the head (n = 9), the right side (n = 8), the left side (n = 7), and the back of the head (n = 6). The nonheader impacts included player collisions (n = 8), player falls (n = 6), unintentional ball-to-head impacts (n = 3), player contact with the ground (n = 2), and collision with the goal (n = 1). A detailed description of the type of event along with the resulting acceleration and injury criteria measurements can be found in Table 1.
The peak linear acceleration from a header impact ranged from 4.5g, occurring to the front of the head, to 62.9g, which occurred on the right side of the head (Fig. 2). The average linear head acceleration for header impacts was 11.9g ± 5.9g (back), 17.4g ± 8.4g (front), 19.5g ± 10.7g (top), 27.2g ± 14.4g (left side), and 28.1g ± 20.8g (right side). The maximum number of header impacts a single player experienced in a scrimmage was four with players 3 and 18 both having that total. When comparing linear head acceleration by header location, the only statistical differences seen were that the side headers had significantly greater linear head acceleration than the back headers.
The peak angular head acceleration from a header impact was 8869.1 rad·s−2 and occurred to the right side of the head (Fig. 3). The average angular head acceleration for header impacts was 723.2 ± 220.3 (back), 1657.5 ± 954.5 (front), 1851.8 ± 1061.7 (top), 2586.6 ± 1501.5 (left side), and 3003.4 ± 2823.9 rad·s−2 (right side). Angular head accelerations for the front and side headers were significantly greater than that for the back. No other significant differences were found when assessing angular acceleration for the header groups.
Nonheader impacts were considered by the type of impact as opposed to the impact location on the head. Player collisions, accounting for the majority of nonheader impacts, resulted in an average resultant linear head acceleration of 22.3g ± 15.0g with a maximum of 56.7g and an average angular head acceleration of 1805.7 ± 949.8 rad·s−2. Player falls averaged 14.4g ± 6.1g of linear acceleration with a maximum of 23.7g with an average of 1246.5 ± 1050.4 rad·s−2 of angular head acceleration. Linear head acceleration was minimal in the other types of nonheader impacts resulting in peak linear accelerations of 32.2g ± 8.8g for unintentional ball-to-head impacts and 27.1g from a collision with the goal (Fig. 4). Comparisons were made between the nonheader group and the header group, but no significant differences were found when looking at linear head acceleration, angular head acceleration, or HIC 15.
HIC values were generally minimal for both header and nonheader impacts. The highest HIC value (154.1) recorded during the study occurred during a heading impact to the right side of the head. The peak HIC values for a single header impact were 79.4 (left side), 35.6 (top), 24.0 (front), and 7.4 (back). HIC values resulted in no significant differences between any of the locations. For nonheader impacts, the highest HIC value (90.8) took place during a player collision. The remaining HIC values were much lower and were well under the concussion threshold. The peak HIC values were 25.4 (unintentional impact with ball), 16.5 (collision with goalpost), 13.7 (player fall), and 4.1 (player impact with the ground).
The majority of players participating had only one nonheader event occurring during their scrimmage, and many had no occurrences. Three players, however, had multiple instances. These were players 12, 13, and 29 (Table 1). Player 12 also had multiple header events during her scrimmage. The highest peak linear acceleration of 56.7g, to player 28, was found to occur from a player collision (Fig. 4) and had a corresponding angular acceleration of 2910.3 rad·s−2. This was the only impact that player 28 experienced that was high enough to trigger the HITS.
Player 22 was the only player to have a nonheader impact that exceeded the current recommended tolerance values. This collision with the goal resulted in the peak angular acceleration of 5179.5 rad·s−2 (Fig. 5). In addition to this nonheader impact, player 22 had four header impacts during her scrimmage. Although none of the header impacts exceeded tolerance values for linear head acceleration, one of the headers had an angular acceleration of 4509.1 rad·s−2.
The current study provides valuable data and insight into the head acceleration experienced by players during a variety of impact types that occur during soccer play. Although there have not been any previous studies recording on-field measurements of head acceleration during soccer play, there have been laboratory recreations of many of the measured events. Therefore, the current study will be compared with previous laboratory studies.
There were comparable head accelerations between this study and Shewchenko et al. (24). Shewchenko et al. (24) determined that the average peak linear acceleration for the heading scenarios did not exceed 19.8g, which was found to be very similar to the peak average linear acceleration of 20.4g from the current study. A peak average angular acceleration of 2.41 krad·s−2 was found for the laboratory study with 1.94 krad·s−2 for the current study. However, when compared with Naunheim et al. (13), the current study exceeded the laboratory analysis in measures of maximum linear and angular accelerations. In the laboratory study, Naunheim et al. (13) did not have linear or angular accelerations in excess of 20.3g and 1.46 krad·s−2, in contrast to a maximum linear acceleration of 62.9g and a maximum angular acceleration of 8.87 krad·s−2 determined in the on-field study. This is a 42.6g difference in the maximum linear accelerations seen between the two studies and a 7.41 krad·s−2 difference in the angular acceleration measurements.
In contrast to the comparison with the results of Naunheim et al. (13), head accelerations for heading events were found to be lower in the current study than in the laboratory evaluation of Self et al. (22). Self et al. used higher impact velocities (12.2 and 16.5 m·s−1) than previous laboratory studies and male players to assess head acceleration during heading events. An average peak head acceleration of 29.3g and 32.6g was found for simulated corner kick headers and for traditional straight headers, respectively, in comparison with the average peak acceleration of 20.4g in the current study. This difference would be expected because of the use of high velocities for all headers in the laboratory study as well as the use of adult male players. Angular head acceleration was not collected and cannot, therefore, be compared with the results of the current study.
Although the majority of previous laboratory studies focused on header events, Withnall et al. (26) conducted the only previous study that investigated similar nonheader head accelerations. In this study, Withnall et al. looked at specific types of player collisions in a laboratory setting. The current study did not delineate between specific types of player collisions, so direct comparisons are challenging. Withnall et al. determined that an average linear head acceleration of 21.3g occurred when a player’s head was impacted by another player’s elbow. This is quite similar to the average peak linear head acceleration of 22.3g seen in player collisions in the current study. However, the highest linear acceleration in the current study (56.7g) fell within the range of Withnall et al. of the average peak linear head acceleration for the two velocities tested for head-to-head impacts (35.1g–86.7g). This indicates that the overall player collision range from the current study matches up well with the previous laboratory data.
It was hypothesized that the head accelerations from nonheader events would be higher than those in header events, but that was not what was found when comparing the two. This hypothesis did not hold true for any of the measures (linear acceleration or angular acceleration) because none of these variables were statistically different when comparing header and nonheader impacts. Because of the on-field nature of these impacts, there was a large variance within groups, and with the limited number of impacts in these groups, statistical significance was not found.
Data for both header events and nonheader events were compared with mTBI head injury tolerance values proposed previously (6,16–18,20,27). Tolerance values have been created to evaluate the risk of concussion using National Football League (NFL) reconstructions (18,27), human volunteers (7,20), and scaled animal data (16). Using NFL injury incidents, Pellman et al. (18) and Zhang et al. (27) determined that linear accelerations of 66g, 82g, and 106g resulted in a 25%, 50%, and 80% risk of concussion, respectively. Using the same data set, angular head accelerations of 4600, 5900, and 7900 rad·s−2 resulted in a 25%, 50%, and 80% risk of concussion, respectively. Pellman et al. (18) also determined that a HIC value of 250 should be the limitation for risk of concussion in the NFL population. Human volunteer data have since suggested that the tolerance values developed using recreations may be conservative (7,20). Funk et al. (7) developed tolerance values using the HITS in collegiate football where 165g and 9000 rad·s−2 represented a 10% risk of concussion. Recently, Rowson and Duma (20) updated these curves of Funk et al. (7) using additional human volunteer data finding the average linear head acceleration in concussive cases of 105g.
None of the impacts, header or nonheader, exceeded the 66g threshold, which was the 25% risk of injury tolerance level (27) or the HIC value of 250 (18). Angular accelerations, however, did exceed the suggested limits. Two angular acceleration measurements for heading events (5298.3 and 8869.1 rad·s−2) and one for a goalpost collision (5179 rad·s−2) exceeded the 25% risk of injury threshold of 4600 rad·s−2 with one of those headers exceeding the 80% risk of head injury (27). Although there were angular accelerations that exceeded the recommended tolerance values (Fig. 6), none of the players who participated in the current study were diagnosed with a concussion. This could be due to the fact that they fell within the percentage of the population that would not be injured at the suggested tolerances, or it is possible that these tolerance levels are not representative of the types of impacts that occur during soccer heading. In addition, recent research has indicated that players may be able to compensate during the impact events that they are experiencing (2) and reduce their risk of injury. This is evidenced by the number of football impacts assessed, which exceeds the tolerance values where no injury occurred (2). It has also been suggested that the injury probabilities used are conservative (6) because they were developed using only injury data as opposed to a combination of both injury and noninjury data; however, this does not affect the results of the current study because neither injury nor injury risk due to soccer-related impacts were assessed.
Many of the players in the current study experienced multiple impacts. All of the players who experienced tolerance-exceeding impacts had other impacts as well (Fig. 6). On the basis of standard tolerance levels, it is challenging to determine whether the combination of all of those impacts in a single scrimmage increases the likelihood of injury. Further research is necessary to determine whether the level of multiple impacts has an effect on mTBI probability. In addition, further research is needed to determine whether symptoms are occurring after scrimmages where head accelerations are known.
The limited study population used in the current study did not allow for evaluation of sex, age, or skill with respect to head acceleration in nonheader events. In addition, the investigation of soccer scrimmages and not actual games most likely resulted in lower head accelerations than would be seen in games. Scrimmages were chosen because of the challenge of getting players to wear equipment that is not required during competition but are representative of on-field activities.
This study was also supported in part by the Anthony and Joyce Danielski Kales Scholarship.
The authors thank the National Organizing Committee on Standards for Athletic Equipment for funding this project.
The authors also thank Charlene Brain and Sarah Stojsih for their assistance during data collection and all of the athletes for their participation.
There were no conflicts of interest for either author.
The results of the previous study do not constitute endorsement by the American College of Sports Medicine.
1. Andersen TE, Arnason A, Engebretsen L, Bahr R. Mechanisms of head injuries in elite football. Br J Sports Med. 2004; 38 (6): 690–6.
2. Broglio SP, Schnebel B, Sosnoff JJ, et al.. Biomechanical properties of concussions in high school football. Med Sci Sports Exerc. 2010; 42 (11): 2064–71.
3. Chu JJ, Beckwith JG, Crisco JJ, Greenwald RM. A novel algorithm to measure linear and rotational head acceleration using single-axis accelerometers. In: Proceedings of the 5th World Congress of Biomechanics. Munich (Germany): 2006. p. S534.
4. Crisco JJ, Chu JJ, Greenwald RM. An algorithm for estimating acceleration magnitude and impact location using multiple nonorthogonal single-axis accelerometers. J Biomech Eng. 2004; 126 (6): 849–54.
5. Duma SM, Manoogian SJ, Bussone WR, et al.. Analysis of real-time head accelerations in collegiate football players. Clin J Sport Med. 2005; 15 (1): 3–8.
6. Funk JR, Cormier JM, Bain CE, Guzman H, Bonugli E, Manoogian SJ. Head and neck loading in everyday and vigorous activities. Ann Biomed Eng. 2011; 39 (2): 766–76.
7. Funk JR, Duma SM, Manoogian SJ, Rowson S. Biomechanical risk estimates for mild traumatic brain injury
. Annu Proc Assoc Adv Automot Med. 2007; 51: 343–61.
8. Gadd CW. Use of weighted-impulse criterion for estimating injury
hazard. In: Proceedings of the Tenth Stapp Car Crash Conference. New York (NY): Society of Automotive Engineers, Inc.; 1966. p. 164–74.
9. Hanlon E, Bir C. Validation of a wireless head acceleration measurement system for use in soccer play. J Appl Biomech. 2010; 26 (4): 424–31.
10. Janda DH, Bir C, Wild B, Olson S, Hensinger RN. Goal post injuries in soccer. A laboratory and field testing analysis of a preventive intervention. Am J Sports Med. 1995; 23 (3): 340–4.
11. Manoogian S, McNeely D, Duma S, Brolinson G, Greenwald R. Head acceleration is less than 10 percent of helmet acceleration in football impacts. Biomed Sci Instrum. 2006; 42: 383–8.
12. Mihalik JP, Blackburn JT, Greenwald RM, Cantu RC, Marshall SW, Guskiewicz KM. Collision
type and player anticipation affect head impact severity among youth ice hockey players. Pediatrics. 2010; 125 (6): e1394–401.
13. Naunheim RS, Bayly PV, Standeven J, Neubauer JS, Lewis LM, Genin GM. Linear and angular head accelerations during heading of a soccer ball. Med Sci Sports Exerc. 2003; 35 (8): 1406–12.
14. Naunheim RS, Standeven J, Richter C, Lewis LM. Comparison of impact data in hockey, football, and soccer. J Trauma. 2000; 48 (5): 938–41.
15. Newman JA, Shewchenko N, Welbourne E. A proposed new biomechanical head injury
assessment function—the maximum power index. Stapp Car Crash J. 2000; 44: 215–47.
16. Ommaya AK, Goldsmith W, Thibault L. Biomechanics and neuropathology of adult and paediatric head injury
. Br J Neurosurg. 2002; 16 (3): 220–42.
17. Ommaya AK, Hirsch AE. Tolerances for cerebral concussion
from head impact and whiplash in primates. J Biomech. 1971; 4 (1): 13–21.
18. Pellman EJ, Viano DC, Tucker AM, Casson IR, Waeckerle JF. Concussion
in professional football: reconstruction of game impacts and injuries. Neurosurgery. 2003; 53 (4): 799–812; discussion 812–4.
19. Rowson S, Brolinson G, Goforth M, Dietter D, Duma S. Linear and angular head acceleration measurements in collegiate football. J Biomech Eng. 2009; 131 (6): 061016.
20. Rowson S, Duma SM. Development of the STAR evaluation system for football helmets: integrating player head impact exposure and risk of concussion
. Ann Biomed Eng. 2011; 39 (8): 2130–40.
21. Rowson S, Goforth MW, Dietter D, Brolinson PG, Duma SM. Correlating cumulative sub-concussive head impacts in football with player performance—biomed 2009. Biomed Sci Instrum. 2009; 45: 113–8.
22. Self BP, Beck J, Schill D, et al.. Head accelerations during soccer heading
. In: Moritz EF and Haake S, eds. The Engineering of Sport 6. New York (NY): Springer; 2006. p. 81–6.
23. Shewchenko N, Withnall C, Keown M, Gittens R, Dvorak J. Heading in football. Part 1: development of biomechanical methods to investigate head response. Br J Sports Med. 2005; 39 (1 suppl): i10–25.
24. Shewchenko N, Withnall C, Keown M, Gittens R, Dvorak J. Heading in football. Part 2: biomechanics of ball heading and head response. Br J Sports Med. 2005; 39 (1 suppl): i26–32.
25. Stojsih S, Boitano M, Wilhelm M, Bir C. A prospective study of punch biomechanics and cognitive function for amateur boxers. Br J Sports Med. 2010; 44: 725–30.
26. Withnall C, Shewchenko N, Gittens R, Dvorak J. Biomechanical investigation of head impacts in football. Br J Sports Med. 2005; 39 (1 suppl): i49–57.
27. Zhang L, Yang KH, King AI. A proposed injury
threshold for mild traumatic brain injury
. J Biomech Eng. 2004; 126 (2): 226–36.