Concussions are a major concern in contact sports, and research is currently underway to explore preventative measures to reduce the risk of exposure (2,10,13,18). Most concussion prevention measures are focused on policy and equipment, including changes to rules or equipment with less focus on the individual player (9). One area of research that is gaining attention at the player level is the role the neck muscles, which may play in damping the acceleration experienced by the head (5,8,15,24,29). This is important because the primary cause of most concussions are the linear and rotational accelerations of the head resulting from impact (19).
Alsalaheen et al. (1) suggested that there are 2 important but unrelated constructs that influence the role the neck muscles in mitigating acceleration forces of the head and the neck: the musculoskeletal attribute of girth and strength, and the neuromuscular attributes of electromyographic (EMG) amplitude and latency. Biomechanical studies have shown that people with a stiffer and stronger neck experience less acceleration from impacts to the head (7,8,15,29). Eckner et al. (8) demonstrated that neck strength in each cardinal plane of motion is inversely proportional to the magnitude of head acceleration experienced from a low-level but sudden force applied to the head in that plane of motion. Jin et al. (15) used finite element modeling to find an inverse relationship between contraction velocity of the neck muscles and accelerations measured at the head; the faster the muscles respond to perturbation, the lesser the acceleration experienced by the head. Collins et al. (5) captured strength and anthropomorphic data in 6,662 high-school athletes and followed them for a standard sports season, during which 179 were diagnosed with a concussion. Using multivariate logistic regression, it was found that only composite (4-plane) neck strength remained as a significant predictor of concussion diagnosis. In a critical appraisal and synthesis of the relevant literature, Gilchrist et al. (11) found enough evidence to endorse training of the neck muscles through all planes of available motion as a potential strategy for concussion risk management. They further suggested that such training should move from single-plane isotonic resistance training focused on strength or hypertrophy, to dynamic plyometric-type training programs focused more on optimizing neck muscle reactivity. They theorized that the ability to rapidly dissipate anticipated or unanticipated impact energy could be improved through more efficient and shorter latency excitation-contraction coupling. In other words, the shorter amount of time it takes the neck to respond to a blow, the more efficiently it can absorb that force. The analogy of viewing the neck muscles as damping springs appears appropriate here, where more rapid recruitment of motor units of the neck muscle complex (including trapezius, sternocleidomastoid, scalenes, splenius capitis, and the deep segmental muscles) should, in theory, be more resilient to sudden load, reducing the energy transfer to the head and the brain. A focus on dynamic plyometric training of the neck muscles is also supported by Hrysomallis (14), following a systematic review of the relevant literature. Plyometric training has shown some potential for injury prevention in other body regions (23). Collectively, the current body of evidence, although not large, appears to support rigorous evaluation of more dynamic approaches to neck muscle training for risk management in any injury mechanism involving transfer of energy between the head and the trunk.
Toward this, a new neck training protocol focused on dynamic, multiplanar plyometric training has been designed. Akin to a “hula hoop” for the neck, the protocol leverages self-generated centripetal force using the neck to keep a weighted arm rotating about the top of a custom-fitted helmet. In theory, the use of self-generated force should present less risk of overexertion and injury than that created by an external machine, although several questions need to be answered before a fully powered clinical trial can be appropriately undertaken. These include recruitment and retention of participants, estimates of effect size following the training, and collection of any adverse events. As such, the purpose of this pilot study was to explore the feasibility of a planned future trial to explore the effectiveness of training on the novel neuromuscular strengthening device shown in Figure 1. This pilot study involved a 7-week training protocol in a sample drawn from a highly trained athletic population of university football players.
Experimental Approach to the Problem
This pilot study was designed as a quasiexperimental clinical trial with an intervention group and a control group, with a planned subanalysis of the change in performance metrics on the neuromuscular training device in the intervention group. All consenting participants continued to engage in team-mandated off-season training sessions. The intervention group also participated in a 7-week neuromuscular neck-training program (exposure variable) involving 2 sessions per week of approximately 10 minutes each. The outcome variables in the study were change in multiplanar isometric neck strength and performance on the neuromuscular training device after training. Strength changes in the transverse plane (axial rotation) were of particular interest as strengthening in this direction is clinically challenging.
Participants for this pilot study were selected from a list of current players who had participated in spring training and performed baseline strength testing. There were 38 eligible players for this study in total. For feasibility, selection was based on players who were locally available to train during the 7 weeks leading up to the start of the fall football season. The control group was selected to match the intervention group on height (threshold ± 5 cm), body mass (±8 kg), age (±2 years), and neck girth (±3 cm) (Table 1). All subjects were between 18 and 23 years of age. Subjects were excluded if, at the time of recruitment, team medical staff indicated that there were any concussion symptoms or musculoskeletal issues that prevented them from participation in their team prescribed preseason training. Figure 2 presents a flow diagram of the participants through each stage of the study. Formal written informed consent was obtained from all subjects before participation in the study. The Western University Human Research Ethics Board approved the project before recruitment.
Self-reported age, height, and body mass were collected from each player at the start of the study. Neck girth was measured using a flexible measuring tape at the level just below the thyroid cartilage. Isometric neck strength (recorded in newtons) was measured using a handheld dynamometer according to a previously described assessment protocol (28). This protocol uses self-generated resistance from the subject's upper extremity to press into the dynamometer to evaluate isometric neck strength in flexion, extension, and right and left side flexion, side-flexion/rotation, and axial rotation. This protocol has shown good to excellent test-retest reliability (intersession intraclass correlation coefficient range from 0.87 to 0.95 and SEM range from 12.7 to 20 N for all tested directions). The strength values for the 3 cardinal planes of motion were used for analysis (sagittal plane: flexion/extension, frontal plane: right and left side flexion, and transverse plane: right and left axial rotation). The average between the right and left sides for side-flexion and rotation was used for their respective single plane isometric strength analyses. A single metric termed “composite neck strength” was also calculated as the arithmetic mean of all 6 movements. This allowed easier comparison of findings against those reported by Collins et al. (5) and was intended to facilitate interpretation of results for clinical and conditioning personnel and to determine if the training protocol could also be expected to improve strength across all planes when considered together.
Both the intervention and control groups participated in their team prescribed off-season training program that included 2 neck-training days per week. One day would involve training on a 4-way uniplanar (flexion/extension and right/left side-flexion) isotonic neck-strengthening machine, 2 sets of 8–12 repetitions in each direction. The second day involved “manual neck” strengthening with a workout partner, wherein the partner applied manual resistance to neck movement in each of the same 4 directions, 1 set of 5–8 repetitions. The 2 different protocols were separated by 3–4 days and were administered under the guidance of the team's strength and conditioning coach.
The intervention group players were fitted with a secure football helmet with flange-mounted bearing attached to the top (Figure 1). A 25-cm rod extended from this bearing such that the rod was perpendicular to the bearing and parallel to the floor. A small mass (125 g) was located at the distal end of the rod (weighted arm). With the helmet tightly secured on the head, the players were seated on a bench with their back unsupported and feet flat on the ground. Players then created coordinated circumduction movements of the head using the neck muscles to start the weighted arm spinning about its axis while keeping the rest of the trunk as motionless as possible. As spin speed increased, the small weight provided increased resistance to the neck muscles through centripetal force. Once the subject felt comfortable with the movement pattern, they completed 3 sets of 50 revolutions in each direction (clockwise and counterclockwise), for a total of 6 sets. The weight selection and protocol of 50 revolutions in each direction was informed by prior in-laboratory testing with select members of the target population, feasibility of training time, and as best as possible matched to the theory of intensity needed for neuromuscular adaptations and coordination of neck muscle activity. Each set was timed with a stopwatch and recorded (time clockwise [Tcw50] and time counterclockwise [Tccw50]). A portable cycling computer was used to count the revolutions and calculate the velocity of each revolution of the set. The peak velocity achieved in kilometers per hour was then stored on the cycling computer and recorded for analysis. Given speed (km·h−1) and the preset distance per revolution (200 cm) from the cycling computer, the peak revolutions per minute (RPMpeak) was calculated. The best Tcw50, Tccw50, and the best RPMpeak of all 6 sets were used as outcomes.
Because this type of training has not been previously examined, an effective training protocol has not yet been established. Therefore, the meta-analysis of Peterson et al. (22) on maximizing strength development in athletes was used as a rough guide for creating the intervention training protocol. The intervention consisted of 2 high-intensity training sessions per week with an average of 8 sets per training session. Each session lasted between 8 and 12 minutes and was separated by 2–3 days of rest. In weeks 1–3, the players performed 3 timed sets of 50 revolutions in each direction. For each training session, the best RPMpeak of the 6 sets was recorded. In weeks 4–7, the players performed 5 sets of 50 revolutions in each direction, with the best of the 10 sets RPMpeak achieved recorded for each session.
On the final training session, the players again completed 3 timed sets of 50 revolutions in each direction. The best Tcw50 and Tccw50 were recorded along with the best RPMpeak of the 6 sets. After completing the neuromuscular evaluation and a short rest (approximately 3–5 minutes), the isometric neck strength protocol was repeated using the handheld dynamometer. The control group also performed the follow-up isometric neck strength testing.
Adherence was measured as the proportion of neuromuscular training sessions that each subject attended over the maximum offered (n = 14). Dropout rate was defined as subjects who completed baseline (pre) testing for the intervention group with the neuromuscular training device but did not complete the final follow-up (post) testing. Questions about adverse events from the previous session were asked at each subsequent session (3). Acute head or neck pain associated with the use of the neuromuscular training device was of particular interest. Because this method of training involves a novel method of exercising the neck muscles, it was expected that subjects might experience delayed onset muscle soreness (4). If the pain or duration were greater than the subjects had experienced with other neck training programs, they were to inform the primary investigator. Other adverse events, regardless of whether they were clearly because of the training regimen (e.g., headache, dizziness), were collected through direct questioning at the beginning of each subsequent session.
Subject characteristics were explored descriptively (mean, 95% confidence interval [CI]), along with recruitment rate, retention rate, adherence rates, dropouts, and adverse events reported as proportions. Raw change along with 95% CIs and percentage change were calculated for pre and post values for isometric strength and for the neuromuscular training device performance metrics. An important function of pilot research is to provide an estimate for the magnitude of effect of the intervention, for purposes of calculating the sample size of a fully powered study. Accordingly, effect sizes (Hedge's d; 95% CI) (20) were calculated for the differences in the primary dependent variables of isometric neck strength (N) for each direction of motion and the composite value between the intervention and control groups. Effect sizes were also calculated for the differences in the secondary dependent variables of performance on the neuromuscular device in the intervention group only (TCW50, TCCW50, RPMpeak). The independent variables were time (pre and post) and group (intervention and control). Hedge's d was used for effect size calculations because this value is considered unbiased and more accurate than Cohen's d when the sample size is less than 20 (20). Effect sizes and 95% CIs were calculated in Microsoft Excel (Microsoft; Redmond, WA, USA) using the formulas described by Nakagawa and Cuthill (20), whereas all other statistical analyses were conducted in SPSS (v24.0, IBM, Armonk, NY).
Of the 24 eligible participants, 21 consented to participate for a recruitment rate of 88%. Of those, 18 completed the pre- and posttest measures with a retention rate of 86%. Of those entered into the intervention arm, 8 of 12 (67%) completed at least 11 (79%) of the 14 training sessions over 7 weeks. Reasons for noncompletion were that 3 participants declined to participate in the study and 1 participant was involved in an unrelated motor vehicle collision before the initiation of the training protocol. Although some did report the experience of delayed onset muscle soreness for up to 24 hours after the training, none rated it as any more intense than that experienced routinely after other types of training sessions. No participant in the intervention group reported any other adverse events at any time during the protocol. Participants in the intervention group attended an average of 85% of the 14 training sessions (mean = 11.9; range = 11–14). Of the 12 enrolled into the control group, 2 participants were lost to follow-up (1 missed posttesting and 1 was no longer on the team at the end of the study period).
The mean isometric strength values, raw and percent change, and observed effect sizes between the intervention and control groups pre and post testing are presented in Table 2. The point estimates indicated that composite neck strength improvement favored the intervention group. Mean change in composite strength of the intervention group was 32 N (95% CI, 13–50), whereas in the control group, it was 12 N (95% CI, −10 to 34). Change in axial rotation strength, the direction of most interest, demonstrated the largest mean difference between the control and intervention cohorts of 46 N (95% CI, 9–83) and the largest effect size with 95% CIs that do not include zero (Hedge's d Δrotation = 1.24; 95% CI, 0.23–2.26). As retention was not consistent between groups, a sensitivity analysis in which only 8 control subjects were selected as matched to the intervention group revealed nearly identical findings to the full sample (not shown).
Pre- and postneuromuscular performance parameters over the 7 weeks of training, along with mean change scores, percent change, and effect sizes, are presented in Table 3. The RPMpeak over each training session during the 7 weeks of training is displayed in Figure 3. Consistent with a training effect, all performance parameters showed a qualitative improvement over the course of the 7 weeks of training protocol, evident as early as training session 3.
The purpose of this pilot study was to investigate the feasibility and anticipated training effect of a novel neuromuscular training device in a cohort of highly trained and otherwise healthy athletes. Two-thirds of the subjects approached for involvement in the intervention arm of the study completed the training program. Subjects who trained on the device demonstrated an 85% adherence rate with no dropouts, and there were no adverse events reported to the investigators by the subjects in the duration of the study. This is slightly higher than the compliance rates seen in other neuromuscular training programs in athletic populations ranging from 52 to 79% (12,21,25,26). Adherence has previously been shown to significantly influence the effectiveness of injury prevention programs (27), so this is an important consideration in sample size calculations.
This new 7-week training protocol was demonstrated to be potential for improving isometric composite neck strength (increase of 20 N, 95% CI −8 to 48 N, over control) with a moderate to large effect size (Hedge's d = 0.68). Additionally, training on the device in conjunction with traditional neck strengthening may be an effective means of improving isometric neck axial rotation strength above traditional neck strengthening alone (mean increase of 46 N; 95% CI, 9–83 N, over control group). As both the intervention group and control group continued their standard preseason training that involved using the 4-way neck machine, it is not surprising that there was evidence of improvement over time for flexion and side-flexion in both groups. The 4-way neck machine trains the neck isotonically in these directions and is known to improve isometric neck strength in those planes (14). However, axial rotation strength is not trained with the 4-way neck machine. There are few methods available for training neck rotation strength owing largely to the shape of the head that is not conducive to adding rotational resistance. Yet forces in the horizontal (rotational) plane have previously been associated with more vulnerability to postconcussive syndrome (19). Although purely speculative, this or any other protocol for training dynamic contractions of the neck in a rotation direction may have value for mitigating risk of long-term problems from head-neck trauma.
Predictably, the results also indicate that training on the neuromuscular device improves performance on the device; however it is valuable to note the extent of change. RPMpeak more than doubled, and both Tcw50 and Tccw50 times were roughly cut in half after the training, suggesting that this change may be to the result of improved neuromuscular control or coordination. It is interesting to the authors that the average RPMpeak after training was more than 250 RPM. This represents more than 4 revolutions per second and would suggest that the neck muscles involved in the training were contracting at a rate of more than 4 contractions per second to achieve this speed (or less than 250 milliseconds per contraction). In trials involving an unanticipated head perturbation in a group of healthy adults, the neck muscles responded with a peak latency of 224 milliseconds (1). Training on the device approached a similar neuromuscular latency. It is also notable that the training effect did not appear to reach a plateau after 7 weeks, meaning that with continued training, it may be possible to achieve higher speeds and further approach these potentially important contraction latencies. The aim of this training approach was to emphasize high-velocity muscle contractions and facilitate the short latency rate of force development, as described by Gilchrist et al. (11). This training approach is in line with the type of training, as suggested by prior researchers, that should be investigated as a means of training the neck to prevent concussion (11,16,17,24).
There were several limitations to this study. First, the study population was a group of highly trained male athletes (university football players) and not representative of the general population. Different populations may respond differently to the dynamic training presented here, and other sporting populations that include both male and female subjects who may benefit from dynamic neck strengthening should also be investigated. Second, measurement bias may have occurred because the subjects were not blinded to the training they received and those in the intervention group may have put more effort into their neck strength assessments post training. The use of the 4-way neck machine by control and intervention groups and other training techniques was also not documented or controlled, preventing the exploration of combined training. This, along with other potentially unknown variables, may have been confounding variables. Although potential differences were seen with the 7-week training program, it is likely that the protocol could yet be refined to optimize the training effect. For instance, Conley et al. (6) demonstrated an increase in head extension strength of 34% after a 12-week training program involving 3 training sessions per week. Different training programs, such as longer duration, more training sessions or more sets per session, greater or fewer revolutions per set, and using a heavier or lighter weight on the spinning arm, may influence the dose response and produce greater differences and results. It is also recognized that CIs are influenced by sample size; therefore, larger sample sizes will produce more accurate results. Future studies can consider variations to the training program parameters and also consider monitoring the cross-sectional area, fiber composition, EMG response of key muscles, and kinematic responses to head perturbation trials before and after training on this device to help define the physiological response to training.
In summary, this pilot study has demonstrated that the protocol seems to be feasible in that all 8 participants completed at least 79% of the training, and no adverse events were reported. It has also provided important information for conducting a fully powered study. For example, a trial comparing improvements in composite neck strength using the device over and above a control group using traditional neck strengthening could be conducted by recruiting a minimum of 67 football players. With an expected 88% recruitment rate and 86% retention rate, this would provide 25 subjects per arm and would achieve 80% power to detect differences of 20 N or more in composite neck strength (6). Recognizing that neck strength is only one of many factors that influence concussion and whiplash risk, longer and larger multivariate trials will be needed to determine whether this type of training program does contribute to reducing the sequelae of head impacts.
Prior research suggests that neck strength may play a significant protective role in the incidence of concussion. It has also been suggested that research into neck-training programs with a goal of decreasing concussion risk should focus on multiplanar dynamic, plyometric-type contractions that increase the short-latency rate of force development. The results of this pilot study provide a method of neck training that is consistent with the suggested approach and has demonstrated effectiveness at improving dynamic neck strength as measured by performance on this device. For the strength and conditioning coach, physical therapist, or clinician, the type of neuromuscular training described in this study provides a time-efficient method of multiplanar neck training that includes axial rotation. This is a direction of strengthening that is very difficult to achieve through traditional methods. The presented method also offers these professionals a novel means of dynamic neck strengthening for their clients with a more ballistic or plyometric-type of approach to neck muscle training than traditional neck strengthening alone.
The authors thank all the players who participated in this study and trained on the device during their off-season. The authors also confirm that no financial support was received to conduct this research. The author T. H. Versteegh is the developer and owner of the patent of the training device used for the intervention group in this research and cofounder in the company that owns the rights to this patent. The author L. Fischer’s spouse is also a cofounder in this company. The other authors have no personal, financial, or institutional interest in the device described in this article. The results of this study do not constitute endorsement by the National Strength and Conditioning Association.
This research received no external funding. J. MacDermid was supported by a Canadian Institutes of Health Research Chair in Gender, Work and Health and the Dr. James Roth Chair in Musculoskeletal Measurement and Knowledge Translation.
1. Alsalaheen B, Bean R, Almeida A, Eckner J, Lorincz M. Characterization of cervical neuromuscular response to head-neck perturbation in active young adults. J Electromyogr Kinesiol 39: 70–76, 2018.
2. Benson BW, McIntosh AS, Maddocks D, Herring SA, Raftery M, Dvorak J. What are the most effective risk-reduction strategies in sport concussion? Br J Sports Med 47: 321–326, 2013.
3. Carlesso LC, Macdermid JC, Santaguida LP. Standardization of adverse event terminology and reporting in orthopaedic physical therapy: Application to the cervical spine. J Orthop Sports Phys Ther 40: 455–463, 2010.
4. Cheung K, Hume PA, Maxwell L. Delayed onset muscle soreness. Sports Med 33: 145–164, 2003.
5. Collins CL, Fletcher EN, Fields SK, Kluchurosky L, Rohrkemper MK, Comstock RD, Cantu RC. Neck strength: A protective factor reducing risk for concussion in high school sports. J Prim Prev 35: 309–319, 2014.
6. Conley MS, Stone MH, Nimmons M, Dudley GA. Specificity of resistance training responses in neck muscle size and strength. Eur J Appl Physiol Occup Physiol 75: 443–448, 1997.
7. Dirisala V, Karami G, Ziejewski M. Effects of neck damping properties on brain response under impact loading. Int J Numer Method Biomed Eng 28: 472–494, 2012.
8. Eckner JT, Oh YK, Joshi MS, Richardson JK, Ashton-Miller JA. Effect of neck muscle strength and anticipatory cervical muscle activation on the kinematic response of the head to impulsive loads. Am J Sports Med 42: 566–576, 2014.
9. Emery CA, Black AM, Kolstad A, Martinez G, Nettel-Aguirre A, Engebretsen L, et al. What strategies can be used to effectively reduce the risk of concussion in sport? A systematic review. Br J Sports Med 51: 978–984, 2017.
10. Frémont P, Bradley L, Tator CH, Skinner J, Fischer LK; Canadian Concussion Collaborative. Recommendations for policy development regarding sport-related concussion prevention and management in Canada. Br J Sports Med 49: 88–89, 2015.
11. Gilchrist I, Storr M, Chapman E, Pelland L. Neck muscle strength training in the risk management of concussion in contact sports: Critical appraisal of application to practice. J Athl Enhancement 04: 1–19, 2015.
12. Hägglund M, Atroshi I, Wagner P, Waldén M. Superior compliance with a neuromuscular training programme is associated with fewer ACL injuries and fewer acute knee injuries in female adolescent football players: Secondary analysis of an RCT. Br J Sports Med 47: 974–979, 2013.
13. Harmon KG, Drezner JA, Gammons M, Guskiewicz KM, Halstead M, Herring SA, et al. American Medical Society for Sports Medicine position statement: Concussion in sport. Br J Sports Med 47: 15–26, 2013.
14. Hrysomallis C. Neck muscular strength, training, performance and sport injury risk: A review. Sports Med 46: 1111–1124, 2016.
15. Jin X, Feng Z, Mika V, Li H, Viano DC, Yang KH. The role of neck muscle activities on the risk of mild traumatic brain injury in American football. J Biomech Eng 1: 139, 2017.
16. Lisman P, Signorile JF, Del Rossi G, Asfour S. Investigation of the effects of cervical strength training on neck strength, EMG, and head kinematics during a football tackle. Int J Sports Sci Eng 06: 131–140, 2012.
17. Mansell J, Tierney RT, Sitler MR, Swanik KA, Stearne D. Resistance training and head-neck segment dynamic stabilization in male and female collegiate soccer players. J Athl Train 40: 310–319, 2005.
18. McCrory P, Meeuwisse W, Dvorák J, Aubry M, Bailes J, Broglio S, et al. Consensus statement on concussion in sport-the 5(th) international conference on concussion in sport held in Berlin, 2016. Br J Sports Med 51: 838–847, 2017.
19. Meaney DF, Smith DH. Biomechanics of concussion. Clin Sports Med 30: 19–31, 2011, vii.
20. Nakagawa S, Cuthill IC. Effect size, confidence interval and statistical significance: A practical guide for biologists. Biol Rev Camb Philos Soc 82: 591–605, 2007.
21. Pasanen K, Parkkari J, Pasanen M, Hiilloskorpi H, Mäkinen T, Järvinen M, et al. Neuromuscular training and the risk of leg injuries in female floorball players: Cluster randomised controlled study. BMJ 337: a295, 2008.
22. Peterson MD, Rhea MR, Alvar BA. Maximizing strength development in athletes: A meta-analysis to determine the dose-response relationship. J Strength Cond Res 18: 377–382, 2004.
23. Rössler R, Donath L, Verhagen E, Junge A, Schweizer T, Faude O. Exercise-based injury prevention in child and adolescent sport: A systematic review and meta-analysis. Sports Med 44: 1733–1748, 2014.
24. Schmidt JD, Guskiewicz KM, Blackburn JT, Mihalik JP, Siegmund GP, Marshall SW. The influence of cervical muscle characteristics on head impact biomechanics in football. Am J Sports Med 42: 2056–2066, 2014.
25. Steffen K, Bakka HM, Myklebust G, Bahr R. Performance aspects of an injury prevention program: A ten-week intervention in adolescent female football players. Scand J Med Sci Sports 18: 596–604, 2008.
26. Steffen K, Emery CA, Romiti M, Kang J, Bizzini M, Dvorák J, et al. High adherence to a neuromuscular injury prevention programme (FIFA 11+) improves functional balance and reduces injury risk in Canadian youth female football players: A cluster randomised trial. Br J Sports Med 47: 794–802, 2013.
27. van Reijen M, Vriend I, van Mechelen W, Finch CF, Verhagen EA. Compliance with sport injury prevention interventions in randomised controlled trials: A systematic review. Sports Med 46: 1125–1139, 2016.
28. Versteegh T, Beaudet D, Greenbaum M, Hellyer L, Tritton A, Walton D. Evaluating the reliability of a novel neck-strength assessment protocol for healthy adults using self-generated resistance with a hand-held dynamometer. Physiother Can 67: 58–64, 2015.
29. Viano DC, Casson IR, Pellman EJ. Concussion in professional football: Biomechanics of the struck player—Part 14. Neurosurgery 61: 313–318, 2007.