1HREE DIMENSIONAL ANALYSIS OF LINEAR AND ANGULAR ACCELERATIONS OF 1HE HEAD EXPERIENCED IN BOXING Terry A Smith Biokinetics and Associates Ltd. Ottawa, Ontario Canada Patrick J. Bishop University of Waterloo Waterloo, Ontario Canada Richard P. Wells University of Waterloo Waterloo, Ontario Canada ABSTRACT Te mechanical response characteristics of a Hybrid m headform and mechanical neck to the left jab and left hook pnches have been stdied. Three Canadian amater boxers participated in the stdy by delivering sch blows to the head and neck of the Hybrid m which was solidly monted to a concrete table. Anglar and linear acceleration measres were taken simltaneosly sing a specifically designed nine accelerometer array monted to the headform. Both peak resltant linear accelerations and peak resltant anglar accelerations were higher for the left hook pnch than for the left jab pnch. Te reslts for both the left jab and the left hook demonstrated that neither pnch was capable of casing head accelerations which were of concssive strength sing pblished tolerance data. Te reslts demonstrate that the probabili of receiving a concssive injry from a single pnch of the types delivered in this stdy is qite low; however, the acceleration levels may be high enogh to case mild cerebral concssion in some instances. IN1RODUCI1N At the present time, only a few stdies on boxing exist and very little is known abot the forces developed from a boxing blow or the sbseqent head acceleration effects following the blow. Dplication of the boxing ring environment presents both ethical and physical problems, thereby making research in this field very difficlt. 271
As a sitable alternative, many researchers have sed the srrogate headforms developed for atomotive research instead of hman sbjects. Newman (1) employed amater boxers and a Hybrid m head and neck system to assess the protective vale of a variety of boxing helmets which were worn by the Hybrid m. The head and neck were monted on a pedestal and a linear accelerometer was placed at the centre of gravity of the headform. The peak headform acceleration was 85 g with a pnch impact velocity of 9.1 m/s. The assmption was made that all head accelerations remain prely translational and in the plane of measrement of the accelerometer. However, in general, when a boxer strikes the headform, the approach, the contact, and the followthrogh tend to force the headform in several different directions. These directional accelerations wold not be detected or measred by the niaxial accelerometer and therefore, the resltant headform accelerations may actally be higher. A stdy by Koey (2) acconted for this by sing a triaxial accelerometer monted at the centre of gravity of a Hodgson-WSU headform which was in trn monted secrely to an atomotive spring in order to simlate the neck response of the boxer being strck. Both amater and professional boxers participated in the stdy. Their weights ranged between 132 to 181 ponds (lightweight to heavyweight) and the mean peak triaxial acceleration was 84.2 g with a mean GSI of 162.2. Te athor reported a variability between boxers and between individal pnches as high as 3 percent in some cases. As well, the effect of a freely moving "neck" pon the resltant linear accelerations is not known. lt may be sggested that this is not a tre representation of a boxer's neck since most boxers ndergo intensive training in order to develop a very stiff neck, probably in an effort to redce the amont of head movement which may occr following a blow. The most recent stdy involving measrement of boxing blows with anthropomorphic test dmmies was done by Schwart et al. (3) sing a Hybrid II headform monted on a niversal joint which pennitted movement abot three orthogonal axes. The Hybrid II headform was also covered with a 5 cm thick layer of medim density foam as well as a 1 cm layer of Plastiote to simltate the friction properties of the skin. The overall damping effect that this wold have on the acceleration signal is not known. Following application of the padding, the apparats was then monted onto a stiff steel colmn which was bolted directly to the floor. Forteen black belt karate sbjects were then asked to either pnch or kick the headform apparats. Te headform acceleration reslts from the boxing pnches are not well reported, however, the athors mention that peak linear accelerations of 9 g were recorded on several occasions and one vale of 12 g was also recorded. lt is apparent from the boxing data presented to date, that there is not eno evidence to prove that linear acceleration effects alone are sfficient to case acte head injry in boxing. However, the linear acceleration levels may be enogh to case some form of head injry if they were copled with significant anglar acceleration effects ( 4 ). 272
Therefore, adeqate srrogate test devices mst be capable of measring both linear and anglar acceleration simltaneosly. Unfortnately, only a limited amont of research has been done in the area of anglar acceleration measrements. This is larely de to limitations in the development of accrate anglar acceleration measrement systems (5). This paper describes the development of a system for measring three dimensional linear and anglar accelerations of the head experienced in boxing and the reslts of a pnching stdy sing three Canadian amater boxers. The tolerance limits selected from the literatre were 2 g for linear head acceleration (6) and 45 rad/s/s for anglar head acceleration effects (7). The reslts of the pnch stdy will be compared to these tolerance vales and discssed in hght of difficlties encontered when attempting to simlate the boxing ring environment. ME1HOIX>LOGY All impacts were directed towards a Hybrid m headform with an accompanying Hybrid m mechanical neck. A boxing helmet was placed on the headform for all pnching trials. In addition to the head and neck model, an accelerometer monting system was located at the back of the skll. This device was fabricated from alminim, steel, and polyvinylchloride (PVC). The alminim skll plate was tooled to resemble the original Hybrid m rear skll plate. Two protrsions were made in the skll plate so that it was possible to attach two steel rods which travelled from the posterior of the dmmy head to the front of the dmmy head in direct line with the centre of gravity of the headform. The location of the centre of gravity was taken directly from the reference markings fond on the magnesmm Hybrid m skll. A PVC cbe was then threaded onto the end of each steel rod sch that the cbe was located over the centre of gravity of the headform. A third cbe was monted directly onto the back of the skll plate and oriented to be directly in line with the centre of gravity as well. The cbes located at the end of the steel rods were fashioned to rest pon the latex skin of the dmmy headform. This was done to redce any high freqency oscillations which might have occred following impact. A triaxial accelerometer was located at the head centre of gravity and pon each arm was monted a pair of niaxial accelerometers onented sch that they were positioned orthogonally along the axes originating from the head centre of gravity. Another pair of niaxial accelerometers was monted on the rear cbe in a similar manner ( see Figre 1 ). This prodced the typical 3-2-2-2 configration with the resltant linear acceleration calclated from the triaxial accelerometer and the anglar accelerations abot all three planes of motion calclated sing the eqations: 273
CtX = a 1 = Ct2 = (ai - ao) 21 1 - ( a,3 - a,o) 213 - axo) ( ad - ao) (axl - axo) ( a2 - ao) (ax3 213 211 212 21, ths, the anglar acceleration abot each axis is calclated by taking differences between pairs of accelerometers within the nine accelerometer clster and dividing by the distance between the two accelerometers (1). Figre 2 illstrates the axis system for measring translational and rotational accelerations in relation to the Hybrid III headform. Figre 1 : Hybrid III Headform with Rear Skll Plate 274
Figre 2: Accelerometry System Monted on the Hybrid ill Headform Te advantages of the se of this nine accelerometer array for impact biomechanics have been explained previosly by Padgoankar et al. (8). Te above eqations were written into a compter program which performed all linear and anglar acceleration calclations and provided graphical otpt for each trial. VALIDATION OF TIIE HEAD IMPACT MODEL Prior to testing, the head accelerometry system was properly validated to ensre accracy dring testing with the boxers. Te validat1on procedre inclded comparing the physical parameters of the new headform apparats with those of the original Hybrid m headform as weil as validation of the nine accelerometer system. Te reslts of this validation indicated that the new rear skll plate cased minimal movement of the original head centre of gravity (3 mm) and the addition of the rear skll plate and boxing helmet cased a net 29 % increase in the mass moment of inertia. Te implications of this increase moment of inertia shall be discssed relative to the anglar acceleleration findings. Following the validation tests, the anglar acceleration measrement system was fond to be very accrate, indicating its sitability for measring headform accelerat1ns generated from boxing pnches. A detailed discssion of these validation tests and validation reslts may be fond elsewhere (9). 275
BOXING PUNCH S1UDY Following validation of the accelerometry system, a boxing stdy was condcted sing three Canadian amater boxers. Sbject data may be fond in Table 1. Prior to any testing, the boxer's hands were taped with the reglation length of band wrapp ing and all boxers sed the same pair of Everlast 12 o boxing gloves. Following a warmp, the boxers were allowed to take several random pnches at the instrmented headform. This was done to accomodate the boxer to the stiffness and response of the dmmy head and neck system. Boxers were then instrcted that two pnches were going to be sed in this exp eriment and they were the left Jab and the left hook. The boxers were frther instrcted to contact the headform sing these pnches and to make an effort to hit the headform either at its centre of mass (on-centre) or away from its centre of mass (off-centre ). Three trials were reqired for each pnch type and pnch location for a minimm of twelve pnches for each boxer. Table 1 Sbject Data Sbject Age Mass (kg) Boxing Division wr 17 57 Featherweight GJ 2 6 Lightweight ll 21 1 Sperheavyweight All sbjects were instrcted to stand in a boxing ready position and pon command from the experimenter, they were instrcted to hit the headform with the reqired pnch. Data collection was initiated by the triggering of a reflected light beam. Once the beam was broken by the boxer's fist, all nine accelerometer signals were A/D converted at a rate of 23 h per channel sing the Watscope Data Acqisition System. Once the trial window was selected, the HYB3D program removed bias, calibrated the signal, filtered each channel at 1 5 h sfög a 4th order Btterworth low pass digital filter and calclated the resltant linear and anglar headform accelerations. RESULTS I.JNEAR HEADFORM ACCELERATIONS Typical resltant linear headform accelerations for the left jab and left hook are presented in Figres 3 and 4. These plots are cons1dered to be typical of all the data observed. A minimm of three trials per condition were averaged for each sbject and the mean linear headform accelerations for all sbjects are presented in Table 2. Across all sbjects, the left hook accelerations were greater than the left jab acceleration vales. Average headform acceleration across sbjects was 276
22 g for the left jab and 21 g for the off centre left jab. The average on centre left hook headform acceleration was 44 g while the off centre average headform acceleration for the left hook was 58 g. Comparison of the on centre and off centre pnch locations showed only a marginal increase in on centre headform accelerations for the left jab while the left hook showed higher mean headform accelerations for the off centre pnches. Table 2 Mean Peak Linear Resltant Headform Accelerations ( + /- SD) For the Left Jab and Left Hook Pnches Sbject Lef t Jab Lef t Hook On Centre Off Centre On Centre Off Centre.g g g GJ LL WT Across Ss 22.1 24.4 35.8 58.6 (2.) (1.6) (7.6) (3.1) 25.4 23.9 62.2 78.4 (3.9) (.9) (5.9) (9.8) 16.9 13.4 32.8 (2.7) (2.5) (6.) (6.5) 35.5 21. 5 2.6 43.6 57.5 (4.6) (5.6) (15.6) (18.9) Anglar acceleration measrements were taken in all three planes of motion, frontal, sagittal, and transverse. Figres 5 and 6 show typical resltant anglar headform accelerations as measred by the accelerometer array. The planar anglar acceleration patterns remained qite consistent both withm and between all sbjects. Mean resltant anglar acceleration data ( + /- 1 SD) are shown in Table 3. 277
U!Jl'T JA8 ON! - (-.JECT UJ.! ö!. 1 3 c i 2 s i 1 1 TIME (IN N&J LEFT JA8 OFF 1 (8l.8JECT &JJ.4Q.! ö!. ; 3 c i! 2. 1 1 TIME (IN Mill Figre 3: Linear Headform Accelerations for the Left Jab. 278
LEFT HOOK ON CENTRE (SUB.JECT LLI TRIAL 1 7 iii 1H :::> 6!I!:! H 1 a:: w..j w 5 4 a:: w H..J 3 5 1-2 III w a: 1 9 29 19 TIME 49 (IN MS) LEFT HOOK ON CENTRE iii 39 (SUB.JECT LLI TRIAL 2 6 1H 5 C!) 5!:! H 1 a: III 41 ra er III H..J 1 3 2 5 III w er 1 29 19 TIME 39 (IN MS) Figre 4: Linear Headform Accelerations for the Left Hook. 279 49
Table 3 Mean Peak Anglar Resltant Headform Accelerations ( + /- SD) For the Left Jab and Left Hook Pnches Sbj ect Lef t Jab Lef t Ho ok On Centre Off Centre On Centre Off Centre rad/s/s rad/s/s rad/s/s rad/s/s GJ 311 352 497 697 (35.1) (38.8) (19.1) (33.9) LL 344 324 874 859 (55.8) (2.3) (119.) (36.8) WT 26 154 45 497 (26.) (39.7) (8.5) (65.7).Across Ss 292.7 276.6 675.9 644.1 (72.2) (97.5) (23.6) (159.8) As with the resltant linear data, the resltant anglar accelerations for the left hook are higher than those for the left jab. Groping across sbjects shows a large difference between the two pnches. Average resltant anglar headform acceleration for the left jab pnches are 293 rad/s/s ancl 277 rad/s/s for on centre and off celltre pnches respectively. The left hook pnches have mch higher average resltant anglar headform accelerations for both Oll celltre and off centre pnches. The Oll celltre left hook average is 676 rad/s/s while the average resltant anglar headform acceleration for the off centre left hook is 644 rad/s/s. Althogh the mean resltant anglar headform acceleratioll for th e on centre pnches is higher than the mean resltallt anglar acceleratioll for the off celltre pnches, the high standard deviations make it difficlt to determine any differences between sbjects and any actal differellces between pnch locations. 28
LEFT JAS OFF (SUBJECT so äi ' ). TIME (IN MS) LEFT JJ.8 ON CENTRE OO " ls 3 19 2SI 39 9 (SUBJECT LU. ffi crl Cl 2!5 2 1!5 1.. so t9 TIME 2SI (IN MSl 9 Figre 5: Resltant Anglar Accelerations For the Left Jab. 281!59
left HOOK OFF CENTRE lsubject WT) iii ' II) ' c!:! : aoo 5!5!5 eo ffi.coo s 3 "" 2 1!5 ri:l i t- 2!5 1!5 9 19 TIME lih MS) l.eft HOOK ON CEHTRE iii ' cn.. c!:! w..l w ""!5 3 2 cn (SUBJECT WT) 3!5 "" ""..cg.coo 2!5 C!I 39 eo "" a: 29 1!5 1 5 9 TIME lin MS) 39 9 Figre 6: Resltant Anglar Accelerations For the Left Hook. 282!59
DISCUSSION The complex fnctionin of the brain and the diverse natre of head injry makes it very difficlt to nderstand the different mechanisms and effects of head injry. Perhaps the greatest problem is in the design of kinematic experiments which will not interfere with the physical system being monitored, yet provide complete and accrate reslts that will describe the major effects of head impact. This becomes nearly impossible when considering the fact that there may be several different injry mechanisms present dring any given impact. The development of hman srrogates for impact testing has greatly improved this sitation by providing accrate and repeatable kinematic reslts for a variety of impact sitations and a variety of experimental techniqes. Care shold be taken however, in interpreting the reslts of these tests becase mch of the misnderstanding with regards to impact kinematics stems from the assmptions which were generated becase of experimental techniqes that do not ( or cannot) accrately measre the parameters reqired. This is not to say that the hman model simlations shold not be sed for head injry research. Rather, the limitations of sing these impact models and the limitations of the experimental techniqes sed shold be well explained with specific reference to the hman system which is being modelled. This step in the dedctive reasoning process will then allow for a mch better explanation of the potential mechanisms which may or may not be present dring head impact. The present stdy was ndertaken to examine the linear and anglar accelerations experienced by the head when sbjected to selected pnches that are typically sed in a boxing match. The pnches were delivered to a Hybrid - m accelerometer array from which both resltant linear and resltant anglar acceleration measres were calclated. Prior to the actal testing, the battery of tests perf ormed on the fabricated three dimensional accelerometry system provided consistent evidence with regards to the system's validity and accracy in recording head impacts. lt was initially sggested in this stdy that the head accelerations seen by the Hybrid m headform dring typical boxing pnches wold exceed both the linear ( 6) and anglar (7) acceleration limits previosly established for hman tolerance to concssion or brain injry. Both the resltant anglar headform acceleration vales and the resltant linear headform accelerations are too low to spport this hypothesis. As has been sggested, the interpretation of this reslt is dependant pon a thoro knowledge of the limitations of the test eqipment and the difficlties m the development of accrate measrement systems. 283
Table 4 demonstrates some of the factors which may limit the interpretation of the headform acceleration levels prodced dring this stdy. Table 4 Factors Affecting Headform Acceleration Net Affect Helmeted Headform Increased Mass Moment of Inertia Neck Stiffness Non-maximal Pnch Effort Decrease Headform Acceleration Decrease Headform Acceleration Decrease Headform Acceleration Decrease Headform Acceleration Althogh the helmet fnctions largely to revent facial lacerations, it does possess some im.pact absorbing capabilities. The net increase in the mass moment of inertia de to the boxing helmet and the rear skll plate will reslt in a net decrease in anglar acceleration becase the net moment abot the point of rotation in the neck is eqal to the prodct of the moment of inertia and the anglar acceleration. Therefore, it the inertia tends to increase, then the anglar acceleration will tend to decrease given the net joint moment. This may redce overall anglar acceleration that the headform experiences dring a boxing pnch. lt is also evident from this table that one of the major limitations in head injry research is the development of an accrate model for head impact which dplicates the hman response as closely as possible. The sitation created in the laboratory was one of ideal cond1tions for the boxer. Actal bots involve mainly glancing blows, deflected blows and off balance blows. lt is only if the boxer is seriosly fatiged or slightly stnned that he will be open to receive a pnch wh1ch wold be identical to those seen in this stdy. A better approach may be to directly record the boxing events as they occr in the ring as was done by Chamoard et al. (1). Their stdy involved a pair of boxers who were fitted directly with headgear designed to measre both linear and anglar accelerations and asked to box. Altho this approach does present many physical limitations, ie. restrict1n to physical movement, it may be a better representation of the actal boxing environment. In conclsion, it can be seen that althogb boxing does possess inherent danger in the form of high levels of head acceleration, the probability of 284
receiving an acte dangeros blow does not appear to be that great. The reslts of this stdy tend to sggest that cmlative trama, in the form of repeatd sb-concssive level pnches, may be a factor in the development of nerological impairment in boxers. Therefore, it may be the gradal development of a series of minte anatomical disrpt1ns, throgh constant low level l'nching, that predisposes a boxer to knockots or even sdden death m the boxing ring. Unfortnately, techniqes for the accrate measrement of the head accelerations experienced in boxing are only crrently being developed and little is known with regards to hman tolerance to the cmlative trama experienced dring boxing impacts. lt is hoped that the apparats developed for this stdy may be sed for frther boxing research involving a greater nmber and wider variety of boxers. REFERENCES 1. 2. Newman, J. (1982). The protective capabilities of boxing helmets nder life-like conditions - a progress report. Biokinetics and Associates Limited., Ottawa, Ontario. Koey, J.W. and Stanish, W.D. (1985). The measrement of headform accelerations from boxing pnches. Medicine and Science in Sports and Exercise, 17(2):222. 3. Schwart, M.L, Hdson, AR., Fernie, G.R., Hayashi, K., and Coleclogh, AA (1986). Biomechanical stdy of fll contact karate contrasted with boxing. Jornal of Nerosrgery, 64:248-252. 4. 5. 6. 7. 8. Goldsmith, W., and Ommaya, AK. (1984). Head and neck injr y criteria and tolerance levels. In B. Aldman and A Chapon ( eds) The Biomechanics of Impact Trama, New York: Elsevier Science Pblishers. Cho, C.C., and Sinha, S.C. (1976). On the kinematics of the head sing linear acceleration measrements. Jornal of Biomechanics, 9,67-613. Patrick, L.M., Llssner, H.R., and Grdjian, E.S. (1963). Srvival by design - head protection. In 7th Stapp Car Crash Conference Proceedings, SAE, 171-182. Ommaya, AK., and Hirsch, AE. (1971). Tolerance for cerebral concssion from head impact and whiplash in primates. Jornal of Biomechanics, 4:13-21. Padgoankar, A.J., Krieger, K.W., and King, Al (1975). Measrement of anglar acceleration of a rigid body sing linear accelerometers. Jornal of Applied Mechanics, 42(3):552-556. 285
9. Smith, T.A (1987). Three dimensional analys is of linear and anglar accelerations of the head experienced in boxing. M.Sc. Dissertation, U niversity of Waterloo, Waterloo, Ontario. 1. Cham.oard, F., Trosseille, X., Pincemaille, Y. and Tarriere, C. (1987). Methodological aspects of an experimental research on cerebral tolerance on the basis of boxers' training fights. In 3 lst Stapp Car Crash Conference Proceedings, SAE, 15-28. 285