Quantifying the effect of the facemask on helmet performance

Original Article Quantifying the effect of the facemask on helmet performance Proc IMechE Part P: J Sports Engineering and Technology 2018, Vol. 232(2) 94 101 Ó IMechE 2017 Reprints and permissions: sagepub.co.uk/journalspermissions.nav DOI: 10.1177/1754337117707859 journals.sagepub.com/home/pip Bethany Rowson, Evan J Terrell and Steven Rowson Abstract Evaluating and improving helmet design play a crucial role in reducing sports-related concussions. Despite widespread use of facemasks by football and hockey players, no helmet standards currently exist to test helmets equipped with facemasks. The purpose of this study was to determine the effect that attached facial protection has on the head kinematics resulting from impacts to the helmet shell. Helmets were fit to a modified NOCSAE (National Operating Committee on Standards for Athletic Equipment) headform and subjected to blows from a pneumatic impactor. A total of 240 impact tests were performed to evaluate the effect of the facemask on four helmet models (two for football, two for hockey). For each helmet model, one sample was tested with a facemask and another without a facemask. Tests were conducted at two impact velocities (6, 9 m/s) and three impact locations (front, side, and rear boss) for a total of six impact conditions. Five trials were performed for each helmet sample at each condition. Two-factor analyses of variance were used to quantify effects on linear and rotational head acceleration and Severity Index due to impact location and facemask presence. Significant effects varied by helmet model and impact location and were more commonly associated with football helmets. Differences in facemask effects between sports are likely attributed to differences in facemask-shell attachment mechanisms, and differences in the structure of the facemask itself. The effects of the facemask on linear and rotational acceleration were small, approximately 5% for both football and hockey helmets. On average, peak accelerations were decreased with the addition of a facemask, but individual differences were mixed and varied by helmet type and impact location. These small differences would not greatly affect impact performance tests in the lab. The results of this study have direct applications toward helmet standard development. Keywords Concussion, biomechanics, head acceleration, hockey, football Date received: 12 April 2016; accepted: 2 April 2017 Introduction Of all organized sports, football and hockey result in the highest rates of concussion. 1 Recent studies suggest the possibility that repetitive concussive or even subconcussive impacts may lead to neurodegenerative processes and contribute to long-term cognitive problems. There are three primary intervention strategies to reduce the incidence of injury: modifying rules to reduce high-risk scenarios, training athletes to use proper technique to reduce high-risk head impacts, and improving protective equipment. For example, Daniel et al. 2 identified ways to reduce high-risk head impacts in youth football which led to rule changes in Pop Warner that eliminated high-risk drills and limited contact in practice. A follow-up study quantified the effect of this rule and saw nearly a 50% reduction in the number of head impacts. 3 Furthermore, Mihalik and colleagues 4,5 demonstrated the importance of training players to use proper technique in youth ice hockey to reduce concussion risk. While rule changes and player education focus on reducing the number of head impacts experienced by athletes, advanced protective equipment further reduces concussion risk when incidental head impacts occur. This study focuses on protective equipment, specifically related to laboratory tests used to evaluate helmet performance. No current helmet testing standard evaluates helmets equipped with facemasks, even though nearly all Department of Biomedical Engineering and Mechanics, Virginia Tech, Blacksburg, VA, USA Corresponding author: Bethany Rowson, Department of Biomedical Engineering and Mechanics, Virginia Tech, 440 Kelly Hall, 325 Stanger Street, Blacksburg, VA 24061, USA. Email: browson@vt.edu

Rowson et al. 95 Figure 1. All helmet models tested, from left to right: Reebok 11K, Bauer RE-AKT, Riddell Revolution Speed, and Schutt Vengeance VTD. Two samples of each model were tested, one without a facemask and one with a facemask. Each sample was tested five times at three locations and two speeds, for a total of 30 tests per sample. players in contact sports wear facemasks during play. The National Operating Committee on Standards for Athletic Equipment (NOCSAE) specifies a testing standard that evaluates whether or not football helmets are likely to prevent catastrophic head injury, such as skull fracture and intracranial hemorrhage. Similar standards for hockey helmets are specified by the Hockey Equipment Certification Council (HECC), the Canadian Standards Association (CSA), and the European Committee for Standardization (CEN). These standards have pass/fail criteria based on either linear head acceleration or Severity Index (SI). SI is a weighted function of linear acceleration over time with head injury tolerance limits determined from cadaver and animal injury data. 6,7 The Virginia Tech Helmet Ratings supplement these existing standards by providing comparative data to consumers for football and hockey helmets. 8,9 These testing protocols currently evaluate helmet performance without a facemask attached. To date, no data have been published in the literature describing the effect that facemasks have on overall helmet performance. In theory, attachment of a facemask can stiffen the structural response of the helmet shell and change the mass and center of gravity (CG) of the helmet, all of which can potentially affect the kinematic response of the head resulting from impact. It is imperative that test conditions used in the laboratory are representative of the impact conditions experienced by athletes during play so that laboratoryassessed performance is relevant to the field. The objective of this study was to investigate whether facemasks affect the performance of helmets during laboratory testing. The hypothesis that linear and rotational head accelerations resulting from helmet shell impacts differ between helmets with and without a facemask attached was tested. Such data are directly applicable toward current and future helmet testing standards. Methodology A total of 240 impact tests were performed to assess the effect of facemasks on helmet performance. Two hockey helmet models (Reebok 11K and Bauer RE- AKT) and two football helmet models (Riddell Revolution Speed and the Schutt Vengeance VTD) were evaluated (Figure 1). Two samples of each helmet model were tested, one with and one without an attached facemask. The size of all helmets was large. Two impact velocities were tested at three impact locations. Impact velocities of 6 and 9 m/s were selected to represent moderate to high-severity head impacts in hockey and football. The 6-m/s condition resulted in head accelerations consistent with the 95th percentile of on-field head impacts in both sports. 10,11 The 9-m/s condition resulted in head accelerations consistent with on-field concussive impacts. 12 Impact locations consisted of the front, side, and rear boss (Figure 2). These locations were selected to test shell impact locations with varying distances from the facemask interface. These locations were also representative of impacts experienced by players in both hockey and football. 13 16 Five repeated trials were performed at each test configuration based on a power analysis of samples needed to detect a 5-g effect size with a significance of 0.05 and power of 80% using mean values and variances from preliminary data (Table 1). A pneumatic impactor (Biokinetics, Ottawa, Canada) was used to emulate impacts experienced by athletes during play (Figure 3). The 14-kg ram was equipped with an impactor face designed to replicate the impacting characteristics of a typical helmet. 17 Impact velocity was measured for each test with a dualbeam light gate. A medium-sized NOCSAE headform was mounted on a Hybrid III 50th-percentile neck and was instrumented with a 5-degree-of-freedom sensor package (6DX Pro 2k-18k; DTS, Seal Beach, CA) consisting of three linear accelerometers and three angular rate sensors. 8 The base of the neck was mounted to a sliding mass (Biokinetics) representative of the effective mass of the torso during impact. 17,18 For each test, data were sampled at 20 khz (TDAS Pro; DTS). Acceleration data were filtered to channel frequency class (CFC) 180 and angular rate data were filtered to CFC 155 using a four-pole phaseless Butterworth low-pass filter. Angular rate data were

96 Proc IMechE Part P: J Sports Engineering and Technology 232(2) Figure 2. Three impact locations were tested to determine the effects of the presence of a facemask and impact location on head kinematics. Those locations were front (left), side (middle), and rear boss (right). These locations were selected to test shell impact locations with varying distances from the facemask interface. Table 1. Impact conditions for each helmet model tested. Facemask Location Speed (m/s) Number of impacts Yes Front 6 5 Side 6 5 Rear boss 6 5 No Front 6 5 Side 6 5 Rear boss 6 5 Total number of impacts per helmet model 60 A total of four helmet models (two for hockey and two for football) were tested, resulting in a total of 240 impact tests. differentiated to compute rotational accelerations of the headform. Acceleration data were transformed to the CG of the NOCSAE headform, as the sensor package was not located at the CG. Sensor data were transformed using a rotation matrix and rigid body kinematics equations. Peak linear and rotational head accelerations were determined, and the SI was calculated for each test (equation (1)) ð SI = at ðþ 2:5 dt ð1þ The peak accelerations and the SI values from repeated trials were averaged, and the difference between facemask conditions was determined for each helmet model at each impact configuration. Differences were reported as the average difference (6standard deviation). Two-factor analyses of variance (ANOVAs) were used to determine the significance of effects due to both the presence of the facemask and the location of impact, along with any interactions between the two factors (a = 0.05). ANOVAs were performed by helmet type and impact speed for linear and rotational Figure 3. Pneumatic linear impactor used to simulate head impacts in sports. The impacting ram strikes a helmeted NOCSAE headform mounted on a Hybrid III 50th-percentile neck. The head and neck are mounted on a slide table that represents the effective mass of the torso during impact. The slide table has 5 degrees of freedom to adjust the impact location. acceleration and SI, resulting in a total of 24 comparisons. Post hoc Tukey s HSD (Honestly Significant Difference) tests were performed to determine the source of any significant interactions. Results The average differences between peak linear and rotational accelerations and SI with and without a facemask varied by impact location and speed for all helmet models tested, with negative values indicating lower acceleration with a facemask (Figures 4 and 5). Location was a significant main effect (p 4 0.0278) across all helmet types and speeds for linear and rotational accelerations and SI, with the exception of SI for the speed football helmet at 9 m/s. However, facemask effects and interactions varied by helmet type, location, and speed.

Rowson et al. 97 Figure 4. Hockey helmet average differences in linear (top) and rotational (middle) acceleration and SI (bottom) for matched conditions with and without a facemask. Negative numbers indicate lower values with a facemask. Five repeated trials per impact condition without a facemask were averaged and subtracted from the average of the matched condition with a facemask. Average differences varied by impact location and speed, as well as helmet type. In some cases, values were lower with a facemask, while in others they were higher. Significant differences between facemask conditions are denoted with a red filled marker. For hockey helmets, the facemask was a significant main effect for linear acceleration with the RE-AKT model at 6 m/s (p \ 0.0001) (Table 2). For rotational acceleration, the facemask was a significant main effect with the 11K at 9 m/s (p = 0.0472) and the RE-AKT at 6 m/s (p = 0.0022). The facemask was a significant main effect for SI for the RE-AKT model at 6 m/s (p = 0.0025). Average differences in linear accelerations ranged from 27.5 g (65.0 g; RE-AKT, rear boss, 6 m/s) to 7.4 g (65.2 g; 11K, front, 9 m/s). None of these differences resulted in significant interactions between facemask presence and impact location. The average difference in linear acceleration across all impact conditions and both helmet models was 20.9 g. Average differences in rotational acceleration for hockey helmets ranged from 21105 rad/s 2 (6901 rad/s 2 ; 11K, side, 9 m/s) to 353 rad/s 2 (6496 rad/s 2 ; RE-AKT, side, 9 m/s). The only significant interaction was the 11K at 9 m/s (p = 0.0069), with a significant difference between facemask conditions at the side location (p = 0.0092). The average difference in rotational acceleration across all impact conditions and both helmet models was 2105 rad/s 2. Average differences in SI ranged from 241.1 (648.4; RE-AKT, rear boss, 9 m/s) to 11.3 (638.8; 11K, front, 9 m/s). None of these differences resulted in significant interactions between facemask presence and impact location. The average difference across all impact conditions and both helmet types was 26.8. For football helmets, the facemask was a significant main effect for linear acceleration with both the Speed (p \ 0.0001) and Vengeance (p = 0.0022) models at 9 m/s (Table 3). For rotational acceleration, the facemask was a significant main effect with the Speed (p = 0.0127) and Vengeance (p = 0.0241) at 6 m/s. The facemask was a significant main effect for SI for the Speed at 6 m/s (p = 0.0072) and the Vengeance at both 6 m/s (p \ 0.0001) and 9 m/s (p \ 0.0001). Average

98 Proc IMechE Part P: J Sports Engineering and Technology 232(2) Figure 5. Football helmet average differences in linear (top) and rotational (middle) acceleration and SI (bottom) for matched conditions with and without a facemask. Negative numbers indicate lower values with a facemask. Five repeated trials per impact condition without a facemask were averaged and subtracted from the average of the matched condition with a facemask. Average differences varied by impact location and speed, as well as helmet type. More significant differences were seen with football helmets than hockey helmets. Significant differences between facemask conditions are denoted with a red filled marker. differences in linear acceleration ranged from 211.5 g (68.6 g; Vengeance, rear boss, 9 m/s) to 10.3 g (62.8 g; Speed, front, 9 m/s). Both football helmet models had significant interactions between facemask presence and impact location for linear acceleration (p 4 0.0153), with significant differences between the facemask conditions for the Speed at the front, 6 m/s condition (p = 0.0061), the Speed at the front (p \ 0.0001) and side (p = 0.0121), 9 m/s conditions, and the Vengeance at the rear boss, 9 m/s condition (p = 0.0013). The average difference in linear acceleration across all impact conditions and both helmet models was 20.2 g. Average differences in rotational acceleration ranged from 2611 rad/s 2 (6332 rad/s 2 ; Speed, side, 6 m/s) to 307 rad/s 2 (6191 rad/s 2 ; Speed, rear boss, 6 m/s). There were significant interactions for the Speed at 6 m/s (p \ 0.0001) and the Vengeance at 6 m/s (p = 0.0108), with significant differences between facemask conditions for the Speed at the side, 6 m/s condition (p = 0.0004), and the Vengeance at the side, 6 m/s condition (p = 0.0049). The average difference in rotational acceleration across all impact conditions and both helmet models was 2190 rad/s 2. Average differences in SI ranged from 251.0 (68; Vengeance, front, 9 m/s) to 5.9 (612.4; Speed, front, 9 m/s). There were significant interactions for the Speed (p = 0.0054) and Vengeance (0.0065) at 6 m/s, with significant differences between facemask conditions for the Speed at the front, 6 m/s condition (p = 0.0049) and the Vengeance at the front (p = 0.0025) and rear boss (p = 0.0030), 6 m/s conditions. Discussion The objective of this study was to determine the effects of facemask presence on head kinematics at different

Rowson et al. 99 Table 2. Values of p for two-factor ANOVAs for hockey helmet comparisons. Parameter Helmet Speed Facemask Location FM 3 Loc Tukey s HSD Location p value Linear acceleration 11K 6 0.6146 0.0001 0.4109 11K 9 0.0818 0.0001 0.0552 RE-AKT 6 0.0001 0.0001 0.1114 RE-AKT 9 0.4385 0.0001 0.0502 Rotational acceleration 11K 6 0.8646 0.0001 0.5422 11K 9 0.0472 0.0001 0.0069 S 0.0092 RE-AKT 6 0.0022 0.0001 0.186 RE-AKT 9 0.1726 0.0001 0.5807 SI 11K 6 0.3593 0.0001 0.7054 11K 9 0.8761 0.0001 0.5358 RE-AKT 6 0.0025 0.0001 0.0763 RE-AKT 9 0.2707 0.0001 0.1157 SI: Severity Index; HSD: Honestly Significant Difference; S: side. p values are for the main effects of facemask and location, and the interaction between facemask and location (FM 3 Loc) for each helmet and speed. For comparisons with significant interactions, Tukey s HSD tests were performed to determine the locations that were significantly different for the facemask conditions. Significant differences (p \ 0.05) are shown in bold text. Table 3. Values of p for two-factor ANOVAs for football helmet comparisons. Parameter Helmet Speed Facemask Location FM 3 Loc Tukey s HSD Location p value Linear acceleration Speed 6 0.2089 0.0001 0.0004 F 0.0061 Speed 9 0.0001 0.0033 0.0014 FS 0.0001 0.0121 Vengeance 6 0.4202 0.0278 0.0153 Vengeance 9 0.0022 0.0001 0.0044 RB 0.0013 Rotational acceleration Speed 6 0.0127 0.0001 0.0001 S 0.0004 Speed 9 0.1381 0.0001 0.4759 Vengeance 6 0.0241 0.0001 0.0108 S 0.0049 Vengeance 9 0.1346 0.0001 0.3382 SI Speed 6 0.0072 0.0001 0.0054 F 0.0049 Speed 9 0.6111 0.1935 0.5582 Vengeance 6 0.0001 0.0001 0.0065 FRB 0.0025 0.003 Vengeance 9 0.0001 0.0001 0.2399 SI: Severity Index; HSD: Honestly Significant Difference; F: front; S: side; RB: rear boss. p values are for the main effects of facemask and location, and the interaction between facemask and location (FM 3 Loc) for each helmet and speed. For comparisons with significant interactions, Tukey s HSD tests were performed to determine the locations that were significantly different for the facemask conditions. Significant differences (p \ 0.05) are shown in bold text. impact locations and speeds. The results presented here suggest that facemasks do not greatly affect helmet performance in a laboratory setting. Although statistically significant differences were found between facemask conditions in hockey and football helmets, these differences were small and varied by helmet type, impact location, and speed. Impact locations closer to the facemask (front and side) were more often affected by facemask presence. In some cases the facemask increased acceleration, while in other cases it was decreased. On average, the addition of a facemask decreased both linear and rotational accelerations in hockey and football helmets. The average differences in linear and rotational accelerations across all impact conditions were approximately 65% for hockey and football helmets. Similar trends were observed for SI values. A number of clinical studies have compared concussion rates of hockey players wearing different types of face protection. 19 22 Two separate studies found no significant differences in concussion rates whether players wore no face protection, half, or full protection. 21,22 Another group found no significant difference in concussion rates between players wearing half or full facial protection, but found that players wearing half protection missed more playing time after injury than those wearing full protection. 19,20 These studies suggest that it is unlikely that the presence of a facemask or different types of face protection greatly alter helmet performance. In general, the facemask had a greater effect on football helmets than on hockey helmets. These effects are most likely due to different attachment mechanisms for football and hockey facemasks. For football helmets, the facemask is rigidly attached to the shell, limiting any relative motion between the two parts. Therefore, the addition of the facemask may add more to the

100 Proc IMechE Part P: J Sports Engineering and Technology 232(2) overall stiffness of the shell and have a greater influence on helmet performance. Hockey helmet facemasks, however, are hinged at the top and only have straps attached to the shell on the sides. This attachment allows for measurable relative motion during impact and accounts for the smaller effects seen with hockey helmets. Additionally, hockey helmets have a modular two-piece shell, which further contributes to relative motion between different parts of the helmet. Football facemasks also have a much larger gauge, weighing about twice that of a hockey facemask (520 g for football vs 250 g for hockey). The larger gauge may increase the overall stiffness in the system. There were several limitations in this study. The first is that only a single laboratory impacting system (linear impactor) was evaluated. There are a number of systems, including different styles of drop towers and pendulums that have previously been used for helmet evaluation. The effects of facemask use in these different systems may vary as they have different impacting surfaces and masses associated with them. Another limitation is the small sample of helmets tested. Although this study shows that there are small differences in head kinematics with facemask use, the results were mixed by helmet type and impact location and cannot be extrapolated to other helmet types. Furthermore, only the standard facemask sold with each helmet was tested, and it is possible that alternative facemasks would vary the effects. Conclusion The objective of this study was to determine whether the presence of a facemask on football and hockey helmets would have an effect on head kinematics during impact. The small differences in head kinematics seen in this study suggest that any changes in performance would not greatly affect laboratory standards testing. For helmet standards that use a pass/fail criterion, the use of a facemask would likely not affect whether or not a given helmet passes the standard. However, addition of a facemask would be more representative of onfield conditions. The differences seen were helmet and location dependent and can most likely be attributed to differences in attachment mechanisms and facemask structure. All impact locations in this study were to the helmet shell, but larger differences could be expected at locations directly interacting with the facemask. These results can be used to inform the development of future helmet standards. Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research reported in this publication was supported by the National Science Foundation REU Program (#EEC-1359073). References 1. Daneshvar DH, Nowinski CJ, McKee AC, et al. The epidemiology of sport-related concussion. Clin Sports Med 2011; 30: 1 17. 2. Daniel RW, Rowson S and Duma SM. Head impact exposure in youth football. Ann Biomed Eng 2012; 40: 976 981. 3. Cobb BR, Urban JE, Davenport EM, et al. Head impact exposure in youth football: elementary school ages 9 12 years and the effect of practice structure. Ann Biomed Eng 2013; 41: 2463 2473. 4. Guskiewicz KM and Mihalik JP. Biomechanics of sport concussion: quest for the elusive injury threshold. Exerc Sport Sci Rev 2011; 39: 4 11. 5. Mihalik JP, Blackburn JT, Greenwald RM, et al. Collision type and player anticipation affect head impact severity among youth ice hockey players. Pediatrics 2010; 125: e1394 e1401. 6. Gadd CW. Criteria for injury potential. In: Proceedings of the impact acceleration stress symposium, Brooks Air Force Base, 27 29 November 1961, pp.141 144. Washington, DC: National Academy of Sciences. 7. Gadd CW. Use of a weighted-impulse criterion for estimating injury hazard. SAE technical paper 660793, 1966. 8. Rowson B, Rowson S, Duma SM, et al. A methodology for assessing the biomechanical performance of hockey helmets. Ann Biomed Eng 2015; 43: 2429 2443. 9. Rowson S and 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: 2130 2140. 10. Crisco JJ, Wilcox BJ, Beckwith JG, et al. Head impact exposure in collegiate football players. J Biomech 2011; 44: 2673 2678. 11. Rowson B, Rowson S and Duma SM. Hockey STAR: a methodology for assessing the biomechanical performance of hockey helmets. Ann Biomed Eng 2015; 43: 2429 2443. 12. Rowson S and Duma SM. Brain injury prediction: assessing the combined probability of concussion using linear and rotational head acceleration. Ann Biomed Eng 2013; 41: 873 882. 13. 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 1340-1. 14. Viano DC, Withnall C and Halstead D. Impact performance of modern football helmets. Ann Biomed Eng 2012; 40: 160 174. 15. Post A, Oeur A, Hoshizaki B, et al. Examination of the relationship between peak linear and angular accelerations to brain deformation metrics in hockey helmet impacts. Comput Methods Biomech Biomed Engin 2013; 16: 511 519. 16. Wilcox BJ, Beckwith JG, Greenwald RM, et al. Head impact exposure in male and female collegiate ice hockey players. J Biomech 2014; 47: 109 114. 17. Pellman EJ, Viano DC, Withnall C, et al. Concussion in professional football: helmet testing to assess impact

Rowson et al. 101 performance part 11. Neurosurgery 2006; 58: 78 96; discussion 78 96. 18. Viano DC, Casson IR and Pellman EJ. Concussion in professional football: biomechanics of the struck player part 14. Neurosurgery 2007; 61: 313 327; discussion 327 328. 19. Benson BW, Mohtadi NG, Rose MS, et al. Head and neck injuries among ice hockey players wearing full face shields vs half face shields. JAMA 1999; 282: 2328 2332. 20. Benson BW, Rose MS and Meeuwisse WH. The impact of face shield use on concussions in ice hockey: a multivariate analysis. Br J Sports Med 2002; 36: 27 32. 21. Stevens ST, Lassonde M, de Beaumont L, et al. The effect of visors on head and facial injury in National Hockey League players. J Sci Med Sport 2006; 9: 238 242. 22. Stuart MJ, Smith AM, Malo-Ortiguera SA, et al. A comparison of facial protection and the incidence of head, neck, and facial injuries in Junior A hockey players. A function of individual playing time. Am J Sports Med 2002; 30: 39 44.