Article Text

Download PDFPDF

A prospective study of punch biomechanics and cognitive function for amateur boxers
  1. S Stojsih1,
  2. M Boitano2,
  3. M Wilhelm1,
  4. C Bir1
  1. 1Wayne State University, Detroit, Michigan, USA
  2. 2McMaster University, Hamilton, Ontario, Canada
  1. Correspondence to Dr C Bir, Wayne State University, 818 W Hancock, Detroit, MI 48201-3719, USA; cbir{at}


Objective To evaluate several biomechanical factors of the head during a sparring session and their link to cognitive function.

Design Instrumented Boxing Headgear (IBH) was used for data collection during four 2 min sparring sessions. Neurocognitive assessment was measured using the ImPACT© Concussion management software. A baseline neurocognitive test was obtained from each athlete prior to sparring; two additional tests were obtained and compared with the baseline.

Setting Male and female amateur boxers.

Participants Data were collected from 30 male and 30 female amateur boxers.

Main outcome measurements Head accelerations (translational and rotational), injury severity indexes (Head Injury Criteria (HIC) and Gadd Severity Index (GSI)) and cognitive function scores.

Results Peak translational and rotational acceleration values were 191 g and 17 156 rad/s2, respectively, for males and 184 g and 13 113 rad/s2, respectively, for females. The peak HIC and GSI values for males were 1652 and 2292, respectively, and for females 1079 and 1487, respectively. There was no significant difference in the neurocognitive scores between genders. A decrease was exhibited in the delayed memory postbout scores. All other scores either increased or did not significantly decrease when compared with the baseline.

Conclusions The majority of impacts experienced by both genders were under the threshold for mild head injury. There was a statistically significant difference between peak translational and rotational acceleration, HIC and GSI when comparing genders. When analysing cognitive functions, there was no statistical difference between genders.

View Full Text

Statistics from

Repetitive head impacts sustained by boxers have become an important discussion in literature due to concern over acute and chronic injuries. A study found that over 70% of acute injuries related to the sport of boxing occur in the head region, with almost half of the injuries to this region being concussions, also referred to as mild traumatic brain injuries (mTBI).1 Although surrogates have been used to assess the mechanism that causes injuries,2,,4 real-time head acceleration data from the ring and correlation with cognitive function data have yet to be obtained.

Several studies have evaluated the biomechanics of a punch using different surrogate models.2,,4 Atha et al2 gathered data from a world-ranked British professional heavyweight boxer using a suspended targeted mass with instrumentation. The maximum translational acceleration recorded was 53 g, the peak contact force was 4096 N, and the impact velocity was 8.9 m/s. However, the purpose of the study was to determine the mechanical properties of a boxer's punch, not to observe the effects of a punch in creating an injurious effect. The use of a ballistic pendulum provides force data but does not correlate to head injury risk.

A relationship has been developed between impulse duration and acceleration to assess the risk of head injuries. This relationship is commonly known as the Wayne State Tolerance Curve (WSTC) and is the basis of most head-injury tolerance criteria. The curve indicated a decrease in the tolerable level of acceleration as pulse duration increased. Because the WSTC had various interpretive difficulties, the Gadd Severity Index (GSI) was introduced as a generalisation of the WSTC.5 A further refinement occurred in 1971 when Versace developed the Head Injury Criterion (HIC), based on the WSTC and GSI.5 6 Backaitis7 and Eppinger8 reported that HIC could be interpreted as a measure of the rate of change in specific kinetic energy imparted to the head, based on the resultant translational acceleration. The wellestablished HIC threshold of 1000 was developed to identify severe head injuries in automotive safety testing.9 Recently, an HIC value of 250 was suggested as a threshold for mTBI in professional football.10

Although HIC focuses on translational acceleration, forces produced by rapid acceleration or deceleration of the head can cause both translational and rotational movement of the brain, resulting in varying levels of injury. A study by Zhang et al11 proposed thresholds for translational and rotational acceleration based on a finite element model of the brain. The threshold established that a translational and rotational acceleration of 66 g and 4600 rad/s2 (25% probability), 82 g and 5900 rad/s2 (50% probability), and 106 g and 7900 rad/s2 (80% probability), respectively, would result in brain injury.

In a study correlating these injury parameters to impacts from Olympic boxers, Walilko et al4 evaluated the biomechanics of a punch to the jaw to determine the risk of head injury from translational and rotational accelerations using a Hybrid III anthropomorphic head. The translational acceleration was reported to be 58 (SD 13) g, the rotational acceleration was 6343 (1789) rad/s2, and the HIC was 71 (49). Although providing some preliminary data, this study is limited due to the lack of biofidelity of the jaw and Hybrid III headform for such an application.

Given advances in wireless technology, the collection of real-time data without game time interference Hampshire) system to is now developed possible. Simbex (Lebanon, New the Head Impact Telemetry (HIT) collect real-time head acceleration data during sporting activities. The system has been validated and utilised in various sports, specifically football.12 13 The use of such technology within the sport of boxing will allow for more precise analyses of both the impact and resulting outcome. The goal of the current study was to collect real-time head acceleration data from amateur boxers and document any resulting cognitive changes.


Sample population

A total of 30 male and 30 female amateur boxers consented to participate. The mean age of the male athletes was 22 years with a mean experience level of 6 years, and the mean age of the female athletes was 24 years with a mean experience of 3 years at the amateur level. Approval was garnered from the Wayne State University Human Investigation Committee prior to testing, with all subjects consenting to participate prior to testing.

Impact boxing headgear data collection

The Impact Boxing Headgear (IBH), a modification of the HIT system, is a wireless system that provides real-time data to a ringside receiver and laptop computer. The system was developed by Simbex and validated prior to testing.14 Each headgear consists of 12 single axis linear accelerometers, a wireless transceiver, onboard memory and data-acquisition capabilities (1000 Hz). For each impact the IBH data were collected over a 40 ms time window (8 ms pretrigger and 32 ms post-trigger) when any accelerometer exceeded 9.6 g. The acceleration data are time-stamped and wirelessly transmitted to the receiver interfaced with a laptop computer. If wireless communication is not present, the onboard memory within the headgear stores up to 100 impacts, and data are downloaded once communication has been restored.

Boxers were instructed to spar for four 2 min rounds with an opponent of the same weight class and gender. Impact data from the headgear were postprocessed by researchers at Simbex using an algorithm to calculate translational head acceleration, rotational head acceleration and impact location.14 Impact locations (front, back, top, right and left) were determined using the azimuth (θ) and elevation (α) recorded for each impact (fig 1).15 Two impact severity scores that relate head acceleration to risk of injury, HIC and GSI, were also calculated.5 6

Figure 1

General impact locations.

Neurocognitive assessment data collection

Within 24 h prior to the sparring session, each boxer completed baseline testing using the ImPACT© Version 5.0 cognitive software.16 Postbout tests were obtained within 30 min and 24 h following the sparring session. Each athlete performed the brief computerised test on a laptop computer with a mouse in a quiet area of the boxing facility.

The scores that resulted from the modules during the postbout tests were compared with the module results from the baseline test. From this comparison, several scores were generated to assess the boxer's memory, attention, problem solving and reaction time.

Statistical analysis

The IBH and cognitive function data were analysed using Statistical Package for Social Sciences (SPSS) version 16.0.2. A MANOVA was conducted when comparing the IBH data; however, violations of the normality assumption occurred, and non-parametric tests (Mann–Whitney) were used to make the comparison. The comparison of the number of impacts sustained between genders was analysed using a t test. A oneway ANOVA was used for the IBH impact location data to determine the equality of the means without grouping by gender. For the cognitive function data, a repeated-measures ANOVA was applied to analyse scores between and within gender groups. If a violation of the normality occurred, the non-parametric Mann–Whitney test for between-group comparisons and Wilcoxon signed ranks test for within-group comparisons were utilised. For all tests with multiple comparisons, Bonferroni adjustments were made to decrease the number of Type I errors. Pearson correlations were conductedonthe peak values forthe translationaland rotational acceleration, HIC, GSI and number of impacts versus ImPACT subscores to determine if there were any significant relationships.


Impact boxing headgear

Five participants withdrew from the study prior to the sparring session because cheek protectors were not an option on the instrumented headgear; therefore, the IBH data (tables 1, 2) included 55 participants with 1930 impacts. Twenty-seven of the participants were males with 1128 impacts (42 impacts/boxer), and the remaining 28 participants were females with 802 impacts (29 impacts/boxer).

Table 1

Descriptive statistics for mean Instrumented Boxing Headgear impact data

Table 2

Descriptive statistics for peak Instrumented Boxing Headgear impact data

No significant difference was found for the mean acceleration and injury severity scores when comparing gender groups using the non-parametric Mann–Whitney test. The peak HIC, GSI, translational acceleration and rotational acceleration values were also compared by gender groups. Male boxers experienced a significantly higher peak translational acceleration (p = 0.006), HIC (p = 0.005) and GSI (p = 0.005) when compared with female boxers with an adjusted alpha level of a= 0.0125. The peak rotational accelerations exhibited a mild level of significance (p = 0.013). Additionally, a t test was performed to compare the mean number of impacts between gender groups. The male boxers had a significantly higher number of impacts (41.8 impacts/boxer) when compared with the female boxers (28.6 impacts/boxer) (p = 0.04) (table 1).

Figures 24 illustrate that the majority of the impacts sustained by the boxers were below the 25% probability threshold for brain injury established by Zhang et al.11 Combining the male and female data, 94.8% of the impacts were under the translational acceleration threshold, and 89.6% of the impacts were under the rotational acceleration threshold. The percentage of impacts experienced by both genders that were under the mTBI threshold of 250, suggested by Pellman et al,10 was 97.8%.

Figure 2

Male and female peak translational acceleration (g).

Figure 3

Male and female peak rotational acceleration (rad/s2).

Figure 4

Male and female Head Injury Criteria scores. mTBI, Mild head injury or concussion; TBI, Severe head injury.

Figure 5 illustrates the total number of impacts that were sustained by all of the boxers grouped by location. A one-way ANOVA was used to compare the total number of impacts recorded for each region of the head. With an adjusted alpha value, α = 0.0125, the total number of impacts to the front region of the head (56%) (n = 1930) was significantly higher when compared with the other locations (p = 0.000): left (17%), right (14%), back (11%) and top (2%). Although the male boxers experienced a higher number of impacts, the impact location trend was comparable for both genders.

Figure 5

General locations for each impact sustained.

The mean accelerations and HIC scores were compared by general impact location using the non-parametric Mann– Whitney test. No significant difference was observed between gender groups. Impacts to the front and top locations of the head were statistically significant when compared with the other locations (table 3). The mean translational acceleration for the front was significantly lower compared with the back (p = 0.000), right (p = 0.009) and left (p = 0.003) regions, the mean rotational acceleration was significantly lower compared with the three regions (p = 0.000) as well, and the mean HIC score was significantly lower compared with the back (p = 0.011) and left (p = 0.007) regions. The mean translational acceleration and mean HIC score for the top location were significantly lower when compared with the back, right and left regions (p = 0.000). Additionally, the mean rotational acceleration was significantly lower than the back (p = 0.012) and left (p = 0.013) regions in comparison with the top location. The values remain significant with an adjusted alpha value of α = 0.0167 due to repeated comparisons.

Table 3

General locations and corresponding mean accelerations and injury severity

Figure 6 depicts a graph of the rotational and translational accelerations for each impact. The impacts are labelled by the general location. The impacts with the highest rotational and translational accelerations were front impacts. The shaded box in the lower left corner of the graph illustrates the impacts that are below the 25% probability of brain damage established by Zhang et al.11

Figure 6

Rotational and translational accelerations of each impact based on general location.

Neurocognitive assessment

Five athletes were unable to complete the testing series for various reasons (inability to locate the athletes following the session and cooperation issues); therefore, complete cognitive data sets were collected from 55 boxers, 26 males and 29 females (table 4).

Table 4

Mean ImPACT scores for baseline, post and 24 h postbout tests

Both parametric and non-parametric tests were utilised when appropriate and exhibited no significant difference when comparing gender groups. However, a significant decrease was reported for the delayed memory postscores when analysing the data using repeated-measures ANOVA. A pairwise comparison of the three time points for delayed memory revealed significance between the baseline test and post-test for both males and females (p = 0.000) with an adjusted alpha value for multiple tests, α = 0.025. The other scores either increased following the sparring session or did not significantly decrease when compared with the baseline tests. In addition, Pearson correlations between the head accelerations, HIC, number of impacts and cognitive scores were also investigated, but the results were not clinically significant.


The purpose of this study was to measure and analyse several biomechanical factors (translational acceleration, rotational acceleration, HIC, GSI and impact location) during sparring sessions and cognitive functioning of the athletes following the activity. Amateur boxers volunteered for the study from various gyms, clubs and training camps. The IBH system was utilised to collect real-time head impact data in the ring, while neurocognitive functions were assessed prior to sparring and at two time points following sparring sessions.

The current research revealed the mean magnitude of the impacts, obtained by the IBH system, to be lower than previously reported in the literature (table 5). Data from previously published literature were obtained when boxers impacted a surrogate at maximum impact speeds. The current study collected data during sparring sessions, and so the biomechanical factors observed may be lower than what would be recorded during competitive bouts. Further research should be conducted to determine whether competition would significantly increase the impacts experienced, and ultimately the cognitive function of the athletes. Although the sparring scenario provides the ability to have a more controlled study, competitive bouts would provide more realistic data. The majority of the impacts sustained in this study were subconcussive, and no concussions were noted in any of the participants following the sparring sessions. It is anticipated that more concussive events would be captured in a competition setting.

Table 5

Comparison of data found in literature for male boxers

Additionally, the general location of each impact was determined by the IBH system. Research has indicated that acceleration, duration and direction of the impact load may influence the outcome of an impact to the head.6 10 11 17,,20 More frequently, impacts in the lateral direction have been documented to result in brain injury.17 20 The current study revealed that impacts to either the side or back of the head exhibited, on average, significantly higher accelerations and HIC scores (p<0.05) (table 3). The impacts to the front region of the head produced significantly lower mean acceleration values which may be a result of specific training activities. Some athletes focus on building their cervical neck muscles to aid in the incidence or severity of concussion injury by potentially decreasing head acceleration.21 An additional explanation for the low head accelerations could be the frequency of the impacts to the front region of the head. Over half of all the impacts sustained were to the front region, indicating that athletes may be more prepared for those impacts resulting in reduced head accelerations.

While there are some data investigating the punch forces of male boxers,2,,4 the literature regarding female boxers is limited. This study provides one of the first analyses of real-time punch force data and cognitive function in female amateur boxers along with a comparison of the male and female data. When analysing the IBH impact data, the male boxers had significantly higher (p<0.05) peak values when compared with the females. In addition, the frequency of impacts sustained by the male boxers was significantly higher (p<0.05) when compared with the female boxers. However, a significant difference was not established when comparing mean translational and rotational accelerations, HIC and GSI between genders (table 1). On average, the male boxers experienced more impacts, but the magnitude of the impacts for both genders was similar. Comparison of the cognitive function data between genders proved that there was no significant difference in memory, attention, reaction time and problem solving following the sparring sessions.

Head accelerations and injury severity scores obtained from each impact during the sparring proved that the majority of the impacts were below suggested injury thresholds.11 The neurocognitive tests indicated that the athletes had difficulties retaining information during the postsparring test as indicated in a significant decrease in the delayed memory scores. This finding indicates that immediately after a boxer spars, retention of new information is compromised. Twenty four hours following the bout, the delayed memory score was still lower than the baseline; however, it was not significant. The present results compliment the results yielded in a study by Heilbronner et al22 where neuropsychological tests were administered to 23 amateur boxers. Comparison of test scores before and after a single sparring session showed a deficit in verbal memory and incidental memory in the amateur boxers.

In contrast, other studies have found no relationship between amateur boxing and cognitive dysfunction. Moriarity et al23 followed 82 amateur boxers in a 7-day tournament and found no evidence of cognitive dysfunction in the immediate postbout period. In a 9-year prospective study of amateur boxers, Porter24 found that the boxers exhibited no evidence of neuropsychological deterioration. Other studies also support this conclusion.25 26 Although Butler26 found no sign of neuropsychological dysfunction, it was suggested that a long amateur career might reduce fine motor movements. More extensive data are needed over a longer time line to determine the potential effects of repetitive concussive and subconcussive head impacts in amateur boxers.

This study represents the first in-ring collection of impact data for both males and females. These findings are an important contribution to both impact biomechanics and cognitive impairment research. Further efforts will enhance the body of knowledge related to the biomechanics of boxing and sports-related brain injury. It is critical to determine thresholds of injury so that protective measures, including headgear and gloves, may be properly evaluated.


The National Operating Committee on Standards for Athletic Equipment provided funding for this research project. Impact Boxing Headgear was developed for this study by Simbex, and the IBH data were postprocessed by J Beckwith of Simbex. K Podell assisted in the postprocessing of the neurocognitive data. Boxers from USA boxing, U of M, the Armed Forces, Ramos Boxing Gym, Zarzamora Street Gym, Windsor Boxing Club and Border City Boxing Gym volunteered their time to participate in the study. E Hanlon, C Schreiner, R Bolander and O Pinto Neto assisted in data collection for the study. Statistical analysis was completed by B Ozkan, from Wayne State University.


View Abstract


  • Funding The National Operating Committee on Standards for Athletic Equipment provided funding for this research project.

  • Competing interests None.

  • Ethics approval Ethics approval was provided by Wayne State University Human Investigation Committee.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.