SPINE Volume 34, Number 24, pp 2626–2633©2009, Lippincott Williams & Wilkins

Neck Muscle Load Distribution in Lateral, Frontal,and Rear-End ImpactsA Three-Dimensional Finite Element Analysis

Sofia Hedenstierna, PhD,* Peter Halldin, PhD,* and Gunter P. Siegmund, PhD†

Study Design. A finite element (FE) model of the hu-man neck was used to study the distribution of neckmuscle loads during multidirectional impacts. The com-puted load distributions were compared to experimentalelectromyography (EMG) recordings.

Objective. To quantify passive muscle loads in nonac-tive cervical muscles during impacts of varying directionand energy, using a three-dimensional (3D) continuum FEmuscle model.

Summary of Background Data. Experimental and nu-merical studies have confirmed the importance of mus-cles in the impact response of the neck. Although EMGhas been used to measure the relative activity levels inneck muscles during impact tests, this technique has notbeen able to measure all neck muscles and cannot di-rectly quantify the force distribution between the mus-cles. A numerical model can give additional insight intomuscle loading during impact.

Methods. An FE model with solid element musculaturewas used to simulate frontal, lateral, and rear-end vehicleimpacts at 4 peak accelerations. The peak cross-sectionalforces, internal energies, and effective strains were calcu-lated for each muscle and impact configuration. The com-puted load distribution was compared with experimentalEMG data.

Results. The load distribution in the cervical musclesvaried with load direction. Peak sectional forces, internalenergies, and strains increased in most muscles with in-creasing impact acceleration. The dominant musclesidentified by the model for each direction were spleniuscapitis, levator scapulae, and sternocleidomastoid in lat-eral impacts, splenius capitis, and trapezoid in frontalimpacts, and sternocleidomastoid, rectus capitis poste-rior minor, and hyoids in rear-end impacts. This corre-sponded with the most active muscles identified by EMGrecordings, although within these muscles the distribu-tion of forces and EMG levels were not the same.

Conclusion. The passive muscle forces, strains, and en-ergies computed using a continuum FE model of the cervi-cal musculature distinguished between impact directionsand peak accelerations, and on the basis of prior studies,isolated the most important muscles for each direction.

Key words: finite element model, cervical muscula-ture, impact biomechanics, muscle load, EMG. Spine 2009;

34:2626–2633

Neck injuries due to car collisions are a common prob-lem in the western world.1,2 In the prevention and treat-ment of neck injuries, it is important to understand theinjury mechanisms of the cervical spine. During the last10 years, the importance of the cervical musculature hasbeen identified as both a supportive structure and a sitefor soft tissue injuries.3 The significance of the muscula-ture as a stabilizer during loading has been reported inexperimental studies4 –7 and numerical analyses.8 –13

However, the loading mechanism and load distributionbetween the neck muscles remains unclear. This is ofinterest for an improved understanding of how the headand neck is protected by the musculature and what rolemuscle strain plays in impact-induced neck pain.

The response and relative contribution of the cervicalmuscles has been studied experimentally in volunteersusing electromyography (EMG). Although EMG andmuscle force are not directly related, EMG amplitude hasbeen used to infer relative muscle force under isometricvoluntary contractions. For instance, Kumar et al14

found that the sternocleidomastoid muscle was the larg-est contributor in flexion and anterolateral flexion con-tractions whereas trapezius was the largest contributorin extension and splenius capitis was the largest contrib-utor in lateral flexion. Schuldt and Harms-Ringdahl15

also used EMG to show that trapezius and cervical erec-tor spinae made maximal contributions during voluntaryextension, whereas splenius and levator scapulas did soduring lateral flexion.

During the imposed neck motion that develops in avehicle impact, the cervical muscles may be loaded dif-ferently than during voluntary motion. The experimentalresults have shown that muscle activation depends onload direction and peak impact acceleration.16–20 Ac-cording to Kumar et al16–18 the main load bearer is ster-nocleidomastoid in rear-end impacts, trapezius in frontalimpacts, and splenius capitis in lateral impacts. For rear-end impacts Siegmund et al19 reported high EMG activ-ity in the inferior sternohyoid and sternocleidomastoidmuscles, as well as some superficial and deep posteriormuscles during the early phase of head extension.

One difficulty with EMG is that it has not been possibleto measure all cervical muscles. Surface electrodes capturesignals from relatively larger volumes and give rise to cross-talk if electrode pairs are placed too closely. With wire elec-

From the *Division of Neuronic Engineering, School of Technologyand Health, Royal Institute of Technology, Stockholm, Sweden; and†MEA Forensic Engineers & Scientists Ltd, Richmond, BC, Canada.Acknowledgment date: November 2, 2008. Revision date: April 20,2009. Acceptance date: April, 21, 2009.The manuscript submitted does not contain information about medicaldevice(s)/drug(s).Institutional and Foundation funds were received in support of thiswork. One or more of the author(s) has/have received or will receivebenefits for personal or professional use from a commercial party re-lated directly or indirectly to the subject of this manuscript: e.g., hon-oraria, gifts, consultancies, royalties, stocks, stock options, decisionmaking position.Address correspondence and reprint requests to Sofia Hedenstierna,PhD, Neuronik KTH, Alfred Nobels Alle 10, 141 52 Huddinge, Swe-den; E-mail: sofiah@kth.se

2626

trodes it is possible to measure deeper muscles,19,21,22 butthey record local signals that can potentially be specific to asingle compartment within a muscle and still it is difficult toreach all muscles due to their location. Volunteer studiesare also restricted with regard to the magnitude of the im-pact, since injury risk must be minimized.

In contrast to EMG studies, numerical models can beused to study all cervical muscles simultaneously and can besubjected to any range of acceleration. Moreover, muscleload can be quantified directly using different measures(e.g., force, energy, and strain) depending on the aim of theanalysis. One concern in muscle load analyses is the indi-vidual active muscle response. Large variations in activa-tion schemes and the lack of activation data for some mus-cles increase the uncertainty of these analyses. One way toaddress this uncertainty is to exclude activation and to clar-ify the individual muscle loads using a purely passive re-sponse. This type of analysis will highlight the muscles bestable to control the induced dynamics and should corre-spond closely with the muscles enlisted by the neuromus-cular system in vivo. If so, passive numerical models couldbe an important complement to the experimental analysesof muscle mechanics.

The objectives of this study were to quantify muscleloads in the nonactivated cervical musculature using a3D continuum Finite Element muscle model, to analyzethe relation between load and both impact direction andseverity, and to compare the numerical model with ex-perimental EMG results.

Materials and Methods

The numerical model used in the study (Figure 1) represents a50th percentile male and was developed at the Department of

Neuronic Engineering, KTH,8,23,24 using LS-DYNA soft-ware.25 The model includes the cervical vertebrae, facet joints,cervical ligaments, intervertebral discs, a rigid skull, rigidshoulders, and an elastic torso. The ligamentous spine modelhas been validated both at the motion segment level and as anentire cervical spine.8 The newly developed solid element mus-culature includes nonlinear viscoelastic material properties(Ogden rubber) derived from experimental data26 and a 3Dgeometry on the basis of magnetic resonance images.24,27 Thekinematics of the final model have been compared against vol-unteer experiments with good correlation, in rear-end, frontal,and lateral impacts24 and includes 22 separate pairs of muscleswith mass and length according to Table 1.

FE SimulationsThe neck model was subjected to sinusoidal acceleration im-pulses with a duration of 100 milliseconds applied to the firstthoracic vertebra (T1). Three different impact directions (sim-ulating rear-end, frontal, and lateral impacts) and 4 peak ac-celerations (1, 5, 10, and 15 g), reaching from very low tomoderate/severe impacts, were studied (Table 2). T1 was con-strained in all other displacements and all rotations. No muscleactivation was included during the simulations.

Figure 1. The KTH neck model including the continuum elementmusculature.

Table 1. The Muscles Included in the KTH MuscleModel, With Mass and Length*

Muscles AbbreviationMass

(g)Length(mm)

S Rectus capitis posterior major RCapMaj 3.46 61S Rectus capitis posterior minor RcapMin 1.45 33S Obliquus capitis superior OblSup 1.6 51S Obliquus capitis inferior OblInf 3.33 51P Semispinalis capitis SsCap 44.4 285P Semispinalis cervicis SsCerv 24.2 200P Longissimus capitis LongCap 16.6 237P Longissimus cervicis LongCervP Iliocostalis cervicis IlioCerv 4.4 150P Multifidus Mult 55 440P Splenius capitis SplCap 32.1 260P Splenius cervicis SplCerv 15.5 290P Levator scapula LevScap 47.7 160P Trapezius (cervical part) TZ 105.4 252.5S Rectus capitis anterior RCapAnt 0.61 33S Rectus capitis lateralis RCapLat 1.0 29A Longus colli LongColli 10.2 188A Longus capitis LongCap 8.6 115A Scalenus anterior ScalAnt 8.5 115A Scalenus medius ScalMed 14.4 139P Scalenus posterior (not entire

muscle)ScalPost 6.7 84

A Sternocleidomastoid SCM 49.2 229A Sternohyoid/sternothyroid SH 15.2 105A Suprahyoid SupHyoid Springs

*The letters P, A, and S denote if it is grouped as a posterior, anterior, orsuboccipital muscle. The abbreviations for each muscle are used in subse-quent figures, tables, and the text.

Table 2. The 12 Simulations Performed With 3 DifferentDirections and 4 Peak Accelerations

1 g 5 g 10 g 15 g

Rear-end (�x) 1R 5R 10R 15RFrontal (x) 1F 5F 10F 15FLateral (y) 1L 5L 10L 15L

2627FE Neck Muscle Load Analysis • Hedenstierna et al

Load Distribution AnalysisFor each simulation, the peak values were determined for 3different muscle load measurements: cross-sectional force (N),Green-Lagrange effective strain (%),28 and internal energy (J).Cross-sectional force represents the stabilizing force exerted onthe neck by a muscle and was computed for each muscle at asection close to the superior insertions but below any branchingof the muscle. Force was calculated as the sum of elementstresses multiplied by element cross-sectional area.25 Musclestrain reflects the load distribution within a muscle and hasbeen related to muscle injury potential29 and thus the peakvalue of the Green-Lagrange effective strain for all elementswithin a muscle was used for further analysis. Elements nearthe muscle insertions were removed from the strain analysisbecause the material stiffness in these regions did not reflect theincreased percentage of tendinous tissue. Internal energy quan-tifies how much energy is stored in each muscle and how themuscle dampens the forces in the neck. Internal energy includesthe energy absorbed in the muscles by deformation and vis-coelastic effects.

To compare different areas of the neck, the muscles weredivided into 3 groups, anterior, posterior, or suboccipital(�rectus capitis anterior and lateralis), depending on their lo-cation relative to the spinal column (Table 1). The total musclemass was divided as anterior, 22%; posterior, 75.5%; and sub-occipital, 2.5%.

Comparison of FE Model Muscle Load andExperimental EMG Recordings During Impact

Neck muscle EMG activity during impacts has been measuredin human subjects.16–19 Kumar et al used surface electrodesover 3 muscles and studied 3 impact-directions (Table 3). Sieg-mund used wire electrodes and measured 8 superficial and deepmuscles in frontal, lateral, and rear-end impacts, although onlythe rear-impact data have been published.19 For this study, the

frontal and lateral data were processed the same way as thepublished rear impact data.19 All EMG data were rectified,averaged, and normalized against maximum voluntary con-tractions. To enable comparison between the different entitiesof EMG, strain, force, and energy, each unit was normalized byits maximum value.

Results

The maximum head displacement increased with in-creasing impact severity (Figure 2). At higher impact ac-celeration the head motion was restricted by contactwith the torso.

Load Distribution AnalysisThe load distribution in the cervical muscles varied withboth impact direction and peak acceleration, and a fewmuscles dominated the sectional forces for each direction(Figure 3). In lateral impacts the cross-sectional forcewas greatest in levator scapulae and sternocleidomastoid(SCM), and in frontal impacts the force was greatest insplenius capitis (SplCap) and trapezius (TZ). In rear-endimpacts, SCM, rectus capitis posterior minor (RCap-Min), and at higher collision accelerations SH were thedominant muscles. This corresponds well with the mus-cles described as important voluntary force producers byKumar et al and by Schuldt and Harms-Ringdahl, exceptthat cervical erector spinae and TZ were grouped to-gether in the latter study.14,15

Strain AnalysisOf the 3 load parameters, strain varied least with musclesize. The highest strains were typically found at the in-sertion areas of the muscle and the time of maximum

Table 3. The Muscles and Directions Included in the EMG Experimental Studies by Kumar et al16–18 and Siegmund et al19

Impact directionsKumar et al16–18 Peak acc 1.4 g Frontal 10 subjects Rear-end 7 subjects Lateral 20 subjectsSiegmund et al19 Peak acc 1.55 g Frontal 5 subjects Rear-end 5 subjects Lateral 5 subjects

MusclesKumar et al16–18 TZ SCM SplCapSiegmund et al19 TZ SCM SplCap SsCerv SsCap Multif LevScap

Figure 2. Maximum head dis-placement during impact analysisfor different peak accelerations.

2628 Spine • Volume 34 • Number 24 • 2009

strain generally occurred when the head reached maxi-mum displacement, except for the rear-end impact wherethe suboccipital muscle strains peaked during the headretraction before maximum extension.

Strains were highest in lateral impacts and lowest inrear-end impacts (Figure 4). During lateral impacts highstrains were seen in LevScap, SCM, TZ, and the Scale-nus. During frontal impacts, high strains occurred in theextensor muscles, whereas for rear-end impacts highstrains were found in the hyoid and suboccipital muscles.Strain generally increased with peak acceleration and inmany cases increased almost linearly with T1 accelera-tion. Between 1 and 5 g, however, there was a clearchange in strain distribution between muscles in all loaddirections. In rear-end impacts the strain increased rap-idly and the highest values shifted from deep to superfi-cial muscles. In frontal impacts the strain in the suboc-cipital muscles even decreased between 1 and 5 g.

Internal Energy AnalysisThe internal energy (Figure 5) increased with appliedpeak acceleration for all muscles except between 1 and5 g, where it decreased for a few of the suboccipitalmuscles (RCapMin, OblSup, RCapAnt, and RCapLat).The sum of muscle internal energy (Table 4) was highestduring frontal and lowest in rear-end impacts. As a per-centage of the whole neck internal energy, rear-end im-pacts generated the lowest muscle energy. Internal en-ergy increased with increasing peak T1 acceleration,especially between 1 and 5 g. Nevertheless, in relation tothe total internal energy, the proportion of internal en-ergy in the muscles decreased with increasing accelera-tion for all impact directions.

In Figure 6, the muscles were divided into 3 groupsand the energy in each group normalized by the total, sothat the sum of all muscles equaled 1. When grouped intoanterior, posterior, and suboccipital muscles, most of the

Figure 3. Sectional forces in 3 di-rections and 4 accelerations; andmain force producers in corre-sponding directions.14,15 Muscleswith * were measured jointly asTZ/cervical erector spinae.15

Figure 4. The peak effectivegreen strain for each muscleduring impacts of 1, 5, 10, and15 g in lateral, frontal, and rear-end impact simulations.

2629FE Neck Muscle Load Analysis • Hedenstierna et al

energy was absorbed in the posterior muscles and theleast energy in the suboccipital muscles. For the 1 g rear-end impact, the suboccipital muscles absorb almost 20%of the total muscle energy though they makeup only2.5% of the total muscle volume.

Comparison of Muscle Load Distribution and EMGDuring Impact

The EMG data from the experimental studies differedbut showed similar tendencies between muscles exceptfor TZ in frontal impacts (Figure 7). The muscle loadspredicted by the model for 1 g, did not correlate very wellwith the EMG data. The strain and energy was high inTZ and low in semispinalis cervicis compared to the nor-malized EMG levels. However, the muscles with thehighest EMG values predicted high muscle loads andsimilar trends in load patterns could be seen. The straindistribution is similar to Siegmund EMG in lateral im-pact, except for LevScap and TZ; and in frontal impact

except the TZ. The Kumar EMG was closer to the sec-tional forces in lateral and rear-end impacts.

Discussion

The present study has shown that passive muscle force,strain, and internal energy can be analyzed during im-pact loading using a 3D FE model of the cervical muscu-lature. All measurements distinguished between impactdirections and peak accelerations, and on the basis ofprevious studies, isolated the most important muscles foreach impact direction. The muscles with highest EMG

Figure 5. Peak internal energyfor each muscle during frontalimpacts with increasing acceler-ations from 1 to 15 g. At the topleft the maximum head displace-ment in m is plotted.

Table 4. The Max Total Neck Muscle InternalEnergy in Joule (J), and as Percentage of Total NeckInternal Energy, for the Different Directions andPeak Accelerations

Lateral Frontal Rear-End

Muscles% ofTotal Muscles

% ofTotal Muscles

% ofTotal

1 g 1.1 0.66 1.5 0.63 0.2 0.585 g 7.9 0.59 10.0 0.57 5.8 0.4510 g 20.1 0.43 21.0 0.47 13.6 0.2815 g 27.0 0.30 29.8 0.39 20.5 0.21

Figure 6. Muscle internal energy in posterior, anterior and suboc-cipital muscles during impact. The energy is normalized so that thetotal energy in all muscles equals one.

2630 Spine • Volume 34 • Number 24 • 2009

recordings during impact also exhibited the highest mus-cle load parameters in the simulations.

The current model has been validated for the materialresponse in a single muscle27 and the kinematic re-sponses of head and vertebrae during impact.24 Becauseof the complexity of the human neck, the model is nec-essarily a simplification of the human neck muscle sys-tem. The primary limitations of the model concern itsmaterial properties, the geometry of the tendinous inser-tions, and the boundary conditions between surroundingtissues.24,27 Despite these limitations, the complexity ofthe current model was considered sufficient to study theload distribution between muscles during various impactconditions. Since the myotendinous junctions at the in-sertions are a potentially important area for injury anal-ysis,30 these could be improved in the model to givegreater insight into explicit muscle injury mechanisms.Further, as the muscles with highest sectional force inthis study corresponded to the muscles with proposedhighest strength in the counteracting direction accordingto experimental studies (Figure 3), it was assumed thatthe model predicts a fairly accurate load distribution.Another simplification of the current study was the ab-sence of muscle activation. As a first step, we avoided theinclusion of active forces due to uncertainties related tothe activation schemes. Future work, however, will needto incorporate muscle activations and assess how theyalter the muscle load distribution.

Load MeasurementsThe distribution between muscles was different for the 3computed variables, though they showed similar varia-tions with impact direction and acceleration. The strainresults were refined during postprocessing by removinginsertion areas. A few elements were also discarded dueto penetration in the contact interfaces between muscles.This occurred locally at the edge of some muscles and didnot affect the kinematics. Though, this procedure im-

proved the reasonability of the strain results, it madethem more subjective. In contrast, the energy includedthe entire muscle and the forces considered the pre-defined cross section.

The FE model predicts muscle strains that exceed thesuggested failure strains of muscle (25%–60%)31,32 evenat the 5 g level. This may be related to the lack of muscleactivation and to the fixed boundary condition of T1. Inany case it highlights the need for more data on musclemechanics during impact to validate a detailed musclemodel.

Impact Direction and Peak AccelerationThe difference in muscle load distributions is clearly seenbetween 1 and 5 g while for accelerations above 5 g theloads are increasing almost linearly in some muscles. Thechange in load distribution for increased impact acceler-ations is mainly due to larger head kinematics. At lowimpact accelerations, a large part of the motion takesplace between the skull and the first cervical vertebra,resulting in relatively high loads in the suboccipital mus-cles. At higher impact accelerations, the head moves fur-ther and rotates so that the superficial and larger musclestake the main load at maximum head displacement. Inrear-end impacts the model predicts, as expected fromthe literature,33 an S-shaped cervical column in the initialphase of the impact, with flexion in the upper cervicalspine and extension in the lower cervical spine. This re-sults in high initial strains in the suboccipital musclesrepresenting a high percentage of the total muscle energy(Figure 6).

Muscle Load and Neck InjuriesNeck injuries occur in all directions of impact, thoughthe risk is higher in rear-end impacts.2 The energy ab-sorbed by the muscles was highest during frontal im-pacts. In rear-end impacts, the absorbed energy and mus-cle strain was comparatively small (Table 4 and Figure

Figure 7. Normalized EMG datafrom impact loading experimentsby Kumar et al16 –18 and Sieg-mund et al19 together with 3 dif-ferent load measurements from1 g FE simulations.

2631FE Neck Muscle Load Analysis • Hedenstierna et al

4). This could mean that the ligamentous spine bears alarger part of the load during rear-end impacts, andtherefore, could be injured at lower peak accelerationsand speed changes than during frontal impacts.

Further, the difference seen in load distribution be-tween low and moderate/high accelerations, as well asthe difference in the time of peak load between suboccip-ital and superficial muscles, could affect our ability toprotect the cervical spine by muscle activation.

FE Muscle Load and Experimental EMG RecordingsThe most common method used in neck muscle analysistoday is EMG. Muscle activity (EMG) is not only a func-tion of the applied external forces, but also of anticipa-tion levels and cognitive cues such as sudden loudsounds.34 This means that the activity does not corre-spond solely to the actual physical load and is thereforeproblematic to use in comparison with muscle strain(Figure 7). Another uncertainty in EMG is the depen-dency on normalization and the ability to generate truemaximum voluntary contractions. Because there is a re-dundant system of cervical muscles and a limited numberof measured directions, it may not be possible to volun-tarily evoke maximal contractions in each muscle. Thismay also explain why it is difficult to compare resultsfrom the experimental studies16–19 with different exper-imental protocols and different electrodes.

In conclusion, there is a great need for more experi-mental data, both for input to numerical models and toincrease the understanding of the muscular functionsduring impact loading. An interesting measurement forfuture study could be a combination of muscle strain andactivation (EMG), because the usual failure mechanismof muscle is an eccentric contraction.29,35

Conclusion

Using a continuum FE model of the cervical musculatureexposed to 4 levels of impact severity in 3 directions, thisstudy has shown that muscle force, strain, and energygenerally increase with increasing acceleration. More-over, we have shown that different muscles contribute toresisting the impact-induced motion under different im-pact directions (SplCap, LevScap, and SCM in lateral;SplCap and TZ in frontal; SCM, RcapMin, and SH inrear-end) and impact severities (a larger part of the loadin the superficial muscles at high accelerations).

Key Points

● An FE neck model with passive continuum mus-culature was used to identify important musclesin impacts from different directions.

● The model showed that the loads in most mus-cles increase with increasing impact energy.

● The model predicted that load distribution is de-pendent on impact direction and impact severity.

● The muscle loads did not have the same distribu-tion between muscles as the experimental EMGdata though the muscles with peak EMG record-ing during impact also predicted high muscleloads.

AcknowledgmentThe authors thank Jean-Sebastien Blouin of UBC Schoolof Human Kinetics for his assistance with the EMG data.

References

1. Kullgren A, Krafft M, Nygren A, et al. Neck injuries in frontal impacts:influence of crash pulse characteristics on injury risk. Accid Anal Prev 2000;32:197–205.

2. Jakobsson L, Lundell B, Norin H, et al. WHIPS–Volvo’s whiplash protectionstudy. Accid Anal Prev 2000;32:307–19.

3. Vasavada AN, Brault JR, Siegmund GP. Musculotendon and fascicle strainsin anterior and posterior neck muscles during whiplash injury. Spine 2007;32:756–65.

4. Brault JR, Siegmund GP, Wheeler JB. Cervical muscle response during whip-lash: evidence of a lengthening muscle contraction. Clin Biomech 2000;15:426–35.

5. Kumar S, Ferrari R, Narayan Y. Kinematic and electromyographic responseto whiplash loading in low-velocity whiplash impacts–a review. Clin Bio-mech 2005;20:343–56.

6. Ono K, Ejima S, Kaneoka K, et al. Biomechanical responses of head/neck/torso to lateral impact loading on shoulders of male and female volunteers.In: IRCOBI Conference Prague (Czech Republic); 2005.

7. Siegmund GP, Sanderson DJ, Myers BS, et al. Rapid neck muscle adaptationalters the head kinematics of aware and unaware subjects undergoing mul-tiple whiplash-like perturbations. J Biomech 2003;36:473–82.

8. Brolin K, Halldin P, Leijonhufvud I. The effect of muscle activation on neckresponse. Traffic Inj Prev 2005;6:67–76.

9. Brolin K, Hedenstierna S, Halldin P, et al. The importance of muscle tensionon the outcome of impacts with a major vertical component. Int J Crashwor-thiness 2008;13:487–98.

10. Chancey VC, Nightingale RW, Van Ee CA, et al. Improved estimation ofhuman neck tensile tolerance: reducing the range of reported tolerance usinganthropmetrically correct muscles and optimized physiologic initial condi-tions. Stapp Car Crash J 2003;47:135–53.

11. Ejima S, Ono K, Kaneoka K, et al. Development and validation of the humanneck muscle model under impact loading. In: IRCOBI Conference Prague(Czech Republic); 2005.

12. van der Horst MJ, Thunnissen JG, Happee R, et al. Influence of muscleactivity on head-neck response during impact. In: Stapp Car Crash Confer-ence Proceedings; 1997.

13. Wittek A, Kajzer J, Haug E. Hill-type muscle model for analysis of mechan-ical effect of muscle tension on the human body response in a car collisionusing an explicit finite element code. JSME Int J Ser A 2000;43:8–18.

14. Kumar S, Narayan Y, Amell T, et al. Electromyography of superficial cervicalmuscles with exertion in the sagital, coronal and oblique planes. Eur Spine J2002;11:27–37.

15. Schuldt K, Harms-Ringdahl K. Activity levels during isometric test contrac-tions of neck and shoulder muscles. Scand J Rehabil Med 1988;20:117–27.

16. Kumar S, Ferrari R, Narayan Y. Electromyographic and kinematic explora-tion of whiplash-type neck perturbations in left lateral collisions. Spine 2004;29:650–9.

17. Kumar S, Narayan Y, Amell T. An electromyographic study of low-velocityrear-end impacts. Spine 2002;27:1044–55.

18. Kumar S, Narayan Y, Amell T. Analysis of low velocity frontal impacts. ClinBiomech 2003;18:694–703.

19. Siegmund GP, Blouin JS, Brault JR, et al. Electromyography of superficialand deep neck muscles during isometric, voluntary, and reflex contractions.J Biomech Eng 2007;129:66–77.

20. Siegmund GP, Sanderson DJ, Inglis JT. Gradation of neck muscle responsesand Head/neck kinematics to acceleration and speed change in rear-endcollisions. Stapp Car Crash J 2004;48:419–30.

2632 Spine • Volume 34 • Number 24 • 2009

21. Gabriel DA, Matsumoto JY, Davis DH, et al. Multidirectional neck strengthand electromyographic activity for normal controls. Clin Biomech 2004;19:653–8.

22. Magnusson ML, Pope MH, Hasselquist L, et al. Cervical electromyographicactivity during low-speed rear impact. Eur Spine J 1999;8:118–25.

23. Halldin P. Prevention and Prediction of Head and Neck Injury in TrafficAccidents–Using Experimental and Numerical Methods [PhD-thesis at De-partment of Aeronautics]. Stockholm, Sweden: Royal Institute of Technol-ogy, KTH; 2001.

24. Hedenstierna S, Halldin P. How does a three-dimensional continuum musclemodel affect the kinematics and muscle strains of a finite element neck modelcompared to a discrete muscle model in rear-end, frontal, and lateral im-pacts. Spine 2008;33:E236–45.

25. Hallquist JO. LS-DYNA Keyword User’s Manual. Livermore, CA: Liver-more Software Technology Corporation; 2007.

26. Myers BS, Van Ee CA, Camacho DL, et al. On the structural and materialproperties of mammalian skeletal muscle and its relevance to human cervicalimpact dynamics. In: Stapp Car Crash Conference Proceedings; 1995.

27. Hedenstierna S, Halldin P, Brolin K. Evaluation of a combination of contin-uum and truss finite elements in a model of passive and active muscle tissue.Comput Methods Biomech Biomed Engin 2008;11:627–39.

28. Holzapfel GA. Nonlinear Solid Mechanics: A Continuum Approach for En-gineering. Chichester, WS, United Kingdom: John Wiley; 2000.

29. Lieber RL, Friden J. Muscle damage is not a function of muscle force butactive muscle strain. J Appl Physiol 1993;74:520–6.

30. Garrett WE Jr, Safran MR, Seaber AV, et al. Biomechanical comparison ofstimulated and nonstimulated skeletal muscle pulled to failure. Am J SportsMed 1987;15:448–54.

31. Best TM. A Biomechanical Study of Skeletal Muscle Strain Injuries. Durham,NC: Department of Biomedical Engineering, Duke University; 1993.

32. Davis J, Kaufman KR, Lieber RL. Correlation between active and passiveisometric force and intramuscular pressure in the isolated rabbit tibialisanterior muscle. J Biomech 2003;36:505–12.

33. Ono K, Inami S, Kaneoka K, et al. relationship between spine deformationand cervical vertebral motions for low speed rear impacts using humanvolunteers. In: IRCOBI conference Sitges; 1999; Spain.

34. Siegmund GP, Inglis JT, Sanderson DJ. Startle response of human neck mus-cles sculpted by readiness to perform ballistic head movements. J Physiol2001;535:289–300.

35. McCully KK, Faulkner JA. Injury to skeletal muscle fibers of mice followinglengthening contractions. J Appl Physiol 1985;59:119–26.

2633FE Neck Muscle Load Analysis • Hedenstierna et al