Journal of Electromyography and Kinesiology 21 (2011) 270–275

Contents lists available at ScienceDirect

Journal of Electromyography and Kinesiology

journal homepage: www.elsevier .com/locate / je lek in

Effects of submaximal fatiguing contractions on the components of dynamicstability control after forward falls

Mark Walsh a,⇑, Andreas Peper b,c, Stefanie Bierbaum b,c, Kiros Karamanidis c, Adamantios Arampatzis b

a Department of Kinesiology and Health, Miami University, Oxford, OH 45056, USAb Department of Training and Movement Sciences, Humboldt-Universität zu Berlin, Philippstr. 13, Haus 11, 10115 Berlin, Germanyc Institute of Biomechanics and Orthopaedics, German Sport University Cologne, Carl-Diem-Weg 6, D-50933 Cologne, Germany

a r t i c l e i n f o a b s t r a c t

Article history:Received 5 February 2010Received in revised form 7 December 2010Accepted 10 December 2010

Keywords:Margin of stabilityFallingFatigueMuscle strength

1050-6411/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.jelekin.2010.12.005

⇑ Corresponding author. Tel.: +1 513 529 2708; faxE-mail address: walshms@muohio.edu (M. Walsh)

The present study aimed to investigate the effect of lower extremity muscle fatigue on the dynamic sta-bility control of physically active adults during forward falls. Thirteen participants (body mass: 70.2 kg,height: 175 cm) were instructed to regain balance with a single step after a sudden induced fall from aforward-leaning position before and after the fatigue protocol. The ground reaction forces were collectedusing four force plates at a sampling rate of 1080 Hz. Kinematic data were recorded with 12 vicon cam-eras operating at 120 Hz. Neither the reaction time nor the duration until touchdown showed any differ-ences (p > 0.05). The ability of the subjects to prevent falling did not change after the fatigue protocol. Inthe fatigued condition, the participants demonstrated an increase in knee flexion during the main stancephase and an increased time to decelerate the horizontal CM motion (both p < 0.05). Significant (p < 0.05)decreases were seen post-fatigue in average horizontal and vertical force and maximum knee and anklejoint moments. The fatigue related decrease in muscle strength did not affect the margin of stability, theboundary of the base of support or the position of the extrapolated centre of mass during the forwardinduced falls, indicating an appropriate adjustment of the motor commands to compensate the deficitin muscle strength.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction the effects of muscular strength on forward falls we can control

Falls are an inherent part of life for most everyone. A number offactors may contribute to falls. Many studies have examined staticand dynamic balance situations in an attempt to learn more abouthow the human system controls stability and prevents falls.

Although aging is a common focus of balance literature, a loss ofmuscular strength can also be induced through means other thanaging, such as fatigue (Yeung et al., 1999; Moritani et al., 1990).Several studies reported that loss of muscle strength may alterthe capacity of the human system to generate rapid force for bal-ance corrections after sudden perturbations (Granacher et al.,2008; Pijnappels et al., 2005; Simoneau and Corbeil, 2005).Karamanidis and Arampatzis (2007) reported that muscle–tendoncapacities of the lower limbs contribute about 33% to balancerecovery after forward induced falls. However, postural correctionsafter a sudden perturbation involve sensorimotor adaptationalresponses which include mechanisms responsible for maintainingthe dynamic stability and thus muscle weakness may be partlycompensated for by proper planning and execution of the usedlocomotion strategy. By using fatigued younger subjects to test

ll rights reserved.

: +1 513 529 5006..

for the other changes that occur with aging. Three mechanismshave been presented (Hof, 2007) by which stability may be main-tained; by (1) increasing the base of support with relation to theextrapolated centre of mass, (2) counter rotating segments aroundthe CM, and (3) applying an external force other than the groundreaction force (e.g. grasping a handrail or secure object).

Examining the human’s capabilities to regain balance after sud-den release from a forward inclined body position is a commonway to assess dynamic stability (Hsiao and Robinovitch, 1999; The-len et al., 1997; Wojcik et al., 2001). The margin of stability duringlocomotion can be quantified by the position of the extrapolatedcentre of mass (CM) in relation to the base of support (Hof et al.,2005). When the position of the extrapolated CM exceeds the ante-rior boundary of the base of support, the result is a loss of stability(Hof et al., 2005). Although studies regarding either fatigue or bal-ance are numerous, to date only one study we found has examinedthe effects of fatigue on dynamic stability control (Mademli et al.,2008). This study found that fatigue had no effect on the compo-nents of dynamic stability. However, the fatigue protocol for thisstudy used the knee extension exercise and therefore only fatiguedthe knee extensor muscles. Preventing a forward fall requires amotion from the front leg that is similar to a lunge and requiresnot only the knee extensor muscles, but also hip and anklemuscles. The knee extension is a single joint movement requiring

M. Walsh et al. / Journal of Electromyography and Kinesiology 21 (2011) 270–275 271

force production from one muscle group and has little relevance tonormal human movement. Additionally, it has been reported thatto regain balance during forward falls, extensor muscles from themajor lower extremity joints (hip, knee and ankle) are all impor-tant (Pijnappels et al., 2005; Madigan and Lloyd, 2005).

Therefore the purpose of the present study is to experimentallyinvestigate the effect of fatigue of the lower extremity, induced bysubmaximal fatiguing contractions, on the dynamic stability con-trol of young adults during forward falls. We hypothesize that dy-namic stability will be compromised when the lower extremity isfatigued. This could be seen in a decrease in the ability to regainbalance after an induced forward fall.

2. Methods

Thirteen participants (7 men and 6 women), age 27.7 ± 4.5years, body mass: 70.2 ± 13.1 kg, body height: 175 ± 11.1 cm, par-ticipated in this study. Before the initiation of this experiment,the procedures were explained to the subjects and informed con-sent protocol was followed in accordance with the rules of the Uni-versity Human Subjects Board.

Subjects were instructed to regain balance with a single stepafter a sudden induced fall from a forward-leaning position bothbefore and after a fatiguing protocol targeting the lower extremity.The experiment was performed at the maximal individual lean an-gle of each participant (maximal forward incline position waswhere the subjects were able to regain balance with a single stepafter the sudden release of a support cable which supported themfrom behind). The force on the cable at maximum lean positionwas between 25–40% bodyweight. The maximal individual leanangle was determined experimentally before the main experiment.After a short rest (�5 min) subjects were tested at their previouslymeasured maximum lean angle. The experimental design of thebalance recovery task has been previously described (Karamanidisand Arampatzis, 2007). Following the determination of the sub-ject’s maximum lean angle from which they can be released andstill catch themselves, all subjects performed a maximum volun-tary isometric squat. Subjects performed the isometric squat witha knee angle of approximately 110 �C. The knee angle was con-trolled using a goniometer. The isometric squats were performedon a force platform with an adjustable squat bar that was fixedat a height so that the subject’s knee angle was approximately110 �C. The maximum measured force as measured by the forceplatform was considered the isometric squat maximum.

For the fatigue protocol, subjects were instructed to perform dy-namic squats from an extended knee angle of approximately180 �C to a lower position of approximately 90 �C. The weight usedduring the fatiguing protocol was 30% of their maximum isometricsquat. The dynamic squats were performed with free weights.Once they reached the extended leg position they immediatelystarted moving down again to eliminate the chance of ‘resting’ dur-ing the fatigue protocol. Three spotters were present during the fa-tigue protocol to help the subject with the last repetitions and tohelp place the bar on the squat rack. The dynamic squat sets werecompleted until voluntary exhaustion. After 1 min of rest anotherset of squats was performed and this was repeated until each givensubject performed four sets. Directly after the fatigue protocol, themaximum isometric squat of each subject was tested again to mea-sure the extent of the fatigue. Then, immediately after the postmaximum isometric squat (�30 s) the ability of the subject’s to re-gain balance with a single step after a sudden induced fall wastested again at their maximum lean angle to see if fatigue affectedtheir dynamic stability control. Following the post induced falltrial, the maximum isometric squat was again tested to documentany recovery that may have occurred.

3. Equipment

Kinematic data were tracked using a Vicon Motion Capture Sys-tem using 12 vicon cameras operating at 120 Hz. Thirty-eightreflective markers (diameter 14 mm) were fixed to landmarks ofeach subject to aid in motion capture. This included a headbandthat had four markers on it. The whole body model was used to cal-culate joint angles, the parameters of dynamic stability and theresultant joint moments through inverse dynamics. We used dataprovided by Zatsiorsky and Selujanov (1983) to calculate massesand moments of inertia of the body segments.

For the trials of each subject, the release, touchdown and min-imum knee joint angle were identified. The time of release was de-fined as the moment the cable holding the subjects in a forwardlean was released. Touchdown was defined as the moment therecovery leg contacted the ground after the step. This event wasidentified using a Kistler force plate (60 � 90 cm, Kistler, Winter-hur, Switzerland, trigger was force exceeding 20 N). Minimumknee angle corresponded with the vertical velocity of the subjectreaching a value close to 0 m/s. Using those three points, tworecovery phases were identified: (1) phase until touchdown and(2) main stance phase. The reaction time was defined as the timefrom moment of release to the moment the midpoint of the footexceeded an acceleration of 1.5 m/s/s.

The ground reaction force of both legs was collected using fourKistler force plates at a sampling rate of 1080 Hz. At the beginningof each trial subjects stood on two force plates and when the fallwas induced they used a one step strategy and stepped forwardonto another force plate. Ground reaction forces of both feet wereinspected before each ‘fall’ to confirm that subjects were standingsymmetrically with their weight evenly distributed between theirright and left legs.

The margin of stability in the anteroposterior direction was cal-culated according to Hof et al. (2005) as follows:

bx ¼ Ux max � XCM;where XCM ¼ PXCM þVXCMffiffiffiffiffiffiffi

g=lp

bx is the margin of stability in the anteroposterior direction, Ux max isthe anterior boundary of the base of support (i.e. horizontal compo-nent of the projection of the toe from the recovery limb to theground; a zero value represents the position of the toe before re-lease), and XCM is the position of the extrapolated CM in the anter-oposterior direction. PXCM is the horizontal (anteroposterior)component of the projection of the CM to the ground, VXCM is thehorizontal (anteroposterior) CM velocity, g is the acceleration ofgravity, and l is the distance between CM and centre of the anklejoint in the sagittal plane. The term

pg/l is known as the eigenfre-

quency. Hof et al. (2005) used the eigenfrequency for much smallerangles, but more recently it has also been shown valid for larger an-gles of the inverted pendulum (Arampatzis et al., 2008; Curtze et al.,2010; Hof, 2008; Hof et al., 2007). The eigenfrequency is used to cal-culate the extrapolated center of mass, which tells us where thebase of support needs to be to attain stability in a moving (falling)subject. Postural stability is maintained in circumstances where theposition of the extrapolated CM is within the base of support (posi-tive values of margin of stability) while stability is lost in caseswhere the extrapolated CM surpasses the anterior boundary ofthe base of support.

4. Statistics

Paired T-tests were performed to determine pre-fatigue/post-fatigue differences in the components of dynamic stability at bothtouchdown and minimum knee angle. Additionally, paired t-testswere used to compare, the margin of stability, the maximum hip,

Table 2Kinematic and kinetic parameters during the main stance phase means ± SDminimum knee joint angle of the recovery limb after touchdown (Knee anglemin),duration of the main stance phase, average of vertical and horizontal ground reactionforces (Average GRFvertical, Average GRFhorizontal), vertical and horizontal force integralover time (vertical force integral, horizontal force integral) and resultant joint

272 M. Walsh et al. / Journal of Electromyography and Kinesiology 21 (2011) 270–275

knee and ankle joint torques, reaction time and time to touchdown,maximum vertical and horizontal forces, and time to joint torquesbetween the pre and post measurements. For this study differenceswere considered significant if p 6 0.05. No differences in the resultswere found regarding subject height

moments, maximum moment at the ankle (Max momentankle), knee (Maxmomentknee) and Hip (Max momenthip) joints, of the recovery limb during the mainphase, Boundary of base of support, extrapolated centre of mass, projected centre ofmass, centre of mass velocity,

pg/l at minimum knee angle of the stance phase.

Parameter Pre-fatigue Post-fatigue

Knee anglemin (�) 112 ± 14 96 ± 24Duration main stance phase (ms)* 253 ± 75 340 ± 85Average GRFvertical (N/kg)* 14.87 ± 1.17 13.28 ± 3.28Average GRFhorizontal (N/kg)* �3.79 ± 1.17 �2.72 ± 0.95Vertical force integral (N s/kg)* 3.76 ± 0.77 4.47 ± 1.18Horizontal force integral (N s/kg) �0.90 ± 0.24 �0.84 ± 0.24Max momentankle (Nm/kg)* �2.17 ± 0.63 �1.74 ± 0.37Max momentknee (Nm/kg)* 1.76 ± 0.58 1.54 ± 0.66Max momenthip (Nm/kg) �4.49 ± 1.77 �4.19 ± 1.32Boundary of base of support (cm) 1.07 ± 0.16 1.04 ± 0.14Extrapolated centre of mass (cm) 0.99 ± 0.14 1.00 ± 0.20Projected centre of mass (cm) 0.86 ± 0.11 0.91 ± 0.21Centre of mass velocity (m/s) 0.52 ± 0.32 0.35 ± 0.30p

g/l 3.85 ± 0.85 3.90 ± 0.18

* Statistically significant fatigue effect (p < 0.05).

5. Results

5.1. Fatigue protocol

Subjects performed the 4 sets of fatiguing dynamic squats last-ing 104 ± 43, 61 ± 21, 57 ± 29, 44 ± 15 s, respectively. Because thetempo of the fatiguing squats varied with state of fatigue, the re-searcher decided to report the times of the trials. Since the subjectswere not allowed to rest at the top, the reported times are approx-imately the amount of time the lower extremity of the subjectswere being loaded during each set.

The results of the pre and post isometric squats showed a de-crease in the maximum voluntary squat induced by the fatigueprotocol. The Isometric squat maximums for the group of subjectsbefore and directly after the fatigue protocol were 27.3 ± 9.5 N/kgbody mass and 21.0 ± 7.0 N/kg body mass. This was a reductionof 23 ± 8% of the original maximum. After the last dynamic stabilitytest the isometric maximum recovered to 18 ± 9% of the originalmaximum value (22.4 ± 9.2 N/kg). The final dynamic stability testwas after the 2nd isometric max and before the 3rd isometricmax indicating that the fatigue experience by the subjects at thetime of the post dynamic stability test was between 77% and 82%of their maximum. At release during the forward falls, the marginof stability did not show any significant differences between thepre and post-fatigue condition (pre: �32.1 ± 5.6 cm, post:�32.0 ± 6.8 cm) indicating that the participants started the fallsfrom the same incline position.

5.2. Phase until touchdown

Although the reaction time showed a slight increase after thefatigue protocol compared to the non-fatigued value, the differencewas insignificant (p > 0.05) (Table 1). The fatigue protocol had noeffect on duration until touchdown (Table 1). Results of the depen-dent T-tests showed no significant changes (p > 0.05) in the ante-rior boundary of the base of support, horizontal component ofthe projection of the CM to the ground, horizontal CM velocity,the term

ffiffiffiffiffiffiffig=l

por the extrapolated centre of mass between the

Table 1Duration and kinetic parameters at phase until touchdown means ± SD of reactiontime, duration until touchdown, maximum of vertical and horizontal ground reactionforces (Max GRFvertical, Max GRFhorizontal) of both legs in the phase until touchdownand maximum moment at the ankle (Max momentankle), knee (Max momentknee) andHip (Max momenthip) joints of the support limb, Boundary of base of support attouchdown, extrapolated centre of mass at touchdown, projected centre of mass attouchdown, centre of mass velocity at touchdown and

pg/l at touchdown. None of the

pre-post differences were significant.

Parameter Pre-fatigue Post-fatigue

Reaction time (ms) 104 ± 25 121 ± 53Duration until touchdown (ms) 405 ± 47 410 ± 51Max GRFvertical (N/kg) 11.71 ± 1.07 11.59 ± 1.66Max GRFhorizontal (N/kg) 6.86 ± 0.93 6.73 ± 0.75Max momentankle (Nm/kg) 1.36 ± 0.28 1.43 ± 0.33Max momentknee (Nm/kg) 1.13 ± 0.29 1.03 ± 0.39Max momenthip (Nm/kg) 0.99 ± 0.49 0.82 ± 0.63Boundary of base of support (cm) 1.07 ± 0.16 1.06 ± 0.13Extrapolated centre of mass (cm) 1.05 ± 0.14 1.04 ± 0.15Projected cenre of mass (cm) 0.67 ± 0.08 0.67 ± 0.09Centre of mass velocity (m/s) 1.34 ± 0.21 1.35 ± 0.27p

g/l 3.67 ± 0.21 3.66 ± 0.27

pre and post-fatigue conditions at the instant of touchdown (Table2). Likewise no significant difference (p > 0.05) was calculated be-tween the pre and post margin of stability at touchdown (Fig. 1).

The complete ground reaction force curves of the support andrecovery legs can be seen in Fig. 2. The complete joint momentcurves for both the support and recovery legs can be seen inFig. 3. The maximum values of the hip knee and ankle joint mo-ments of the support leg as well as the maximum values for thehorizontal and vertical ground reaction force did not differ signifi-cantly (p > 0.05) between fatigued and non-fatigued conditions inthe phase until touchdown (Table 1).

5.3. Main stance phase

The components of dynamic stability control showed no signif-icant differences (p > 0.05) between the pre and post-fatigue condi-tions during the main stance phase (Table 2). Likewise, nosignificant difference (p > 0.05) was calculated between the preand post margin of stability during the main stance phase(Fig. 1). The duration of the main stance phase after the fatiguewas significantly (p < 0.05) greater compared to the non-fatigued

TD END-30

-20

-10

0

10

20

30

40 Pre

Mar

gin

of S

tabi

lity

[cm

]

TD END

Post

Fig. 1. Individual and mean (SD) values of margin of stability for the pre and postconditions at the instant of touchdown (TD) and at the end of the main stance phase(end).

-9

-6

-3

0

3

0 200 400 600 8000

5

10

15

20

25

Release

Pre Post

ENDTD

F X (N

/kg)

F Z (N

/kg)

Time (ms)

0

2

4

6

0 100 200 300 4000

2

4

6

8

TDRelease Pre Post

F X (N

/kg)

F Z (N

/kg)

Time (ms)

Fig. 2. Mean and SEM curves of the horizontal (Fx) and vertical (Fz) ground reaction force of the recovery leg (left) from release until 800 ms after touchdown and of thesupport leg (right) from release until touchdown before (pre) and after (post) the fatiguing task.

-3

-2

-1

0

1

-1

0

1

2

3

0 200 400 600 800-6

-4

-2

0

2

ENDTD

Dorsal

PlantarAnkl

e (N

m/k

g)

Extension

FlexionKnee

(Nm

/kg)

Release

Flexion

Extension

Hip

(Nm

/kg)

Time (ms)

-2

-1

0

-2

-1

0

0 100 200 300 400-2

-1

0

1

Pre Post

Release TDDorsal

Plantar

Ankl

e (N

m/k

g)

Extension

Flexion

Knee

(Nm

/kg)

Flexion

Extension

Hip

(Nm

/kg)

Time (ms)

Fig. 3. Mean and SEM curves of the sagittal plane joint moments (internal) of the ankle, knee and hip joints of the recovery leg (left) from release until 800 ms aftertouchdown and of the support leg (right) from release until touchdown before (pre) and after (post) the fatiguing task.

M. Walsh et al. / Journal of Electromyography and Kinesiology 21 (2011) 270–275 273

condition (Table 2). In the fatigued condition, the participant’sminimum lead leg knee angle went into significantly more flexionduring the main stance phase leading to an increased increment intime available to decelerate the horizontal CM motion. The averagehorizontal and vertical forces of the recovery leg decreased signif-icantly with fatigue (Table 2). Although the average horizontalforce decreased, the increase in time led to the same horizontalforce integral of the recovery leg (Table 2). After the fatiguing pro-tocol the participants did not show any differences in the maximalhip joint moment of the recovery leg during the main stance phasecompared to the unfatigued condition (Table 2). At the ankle and

knee joints the maximal moment values were lower after the fati-gue than before (Table 2).

6. Discussion

The purpose of the present study was to investigate the effect offatigue of the lower extremity on the dynamic stability control ofyoung adults during forward falls. The fatigue related decrease inmuscle strength induced by our fatigue protocol did not affectthe margin of stability, the horizontal component of the projection

274 M. Walsh et al. / Journal of Electromyography and Kinesiology 21 (2011) 270–275

of the CM, the boundaries of the base of support, the horizontalCM-velocity and the term

ffiffiffiffiffiffiffig=l

pduring the forward induced falls.

In short, the fatigue caused by our fatiguing protocol did not neg-atively affect the subject’s dynamic stability control to the extentthat the subjects’ ability to prevent falling was impaired. This goesagainst our hypothesis.

Our muscle fatigue induced by repetitive squats appeared tocause a reduction in the force generating capacity of the leg exten-sors muscles in the post-fatigue test of about 20% as measured by amaximum isometric squat. The motor behaviour of the examinedparticipants was not affected in the phase until touchdown dueto the reduction in force generating capacity. In the phase untiltouch down, the maximum values of the resultant moments atthe ankle, knee and hip joints as well as of the ground reactionforces in the post-fatigue condition did not show any differencesto the pre-fatigue values in both legs (i.e. support and recoveryleg). In a similar manner the increase of the base of support inthe phase until touchdown as well as the needed time were equalbetween the two conditions. The participants were able to attainthe same boundary of the base of support at the same velocityand thus the same margin of stability at touchdown as in thenon-fatigued condition despite of the reduced force generationcapacity of the leg extensors. An explanation of this phenomenonmay be the relative lower resultant joint moment values neededduring the phase until touchdown compared to the main stancephase. This means that the main mechanism (i.e. increased baseof support in relation to extrapolated CM) during the forward fallsto regain balance does not require high level joint moments andthus high level muscle forces from the human system. Further-more, the fact that all participants who showed positive values inmargin of stability at touchdown achieved a stable position inthe end of the main phase indicates that the state of stability attouchdown determines the ability of the human system to recoverbalance with a single step after forwards falls. Therefore the abilityof the participants to increase the base of support in relation toextrapolated CM during the phase until touchdown seems to bethe most important event for dynamic stability control after sud-den induced forward falls. One alternate hypothesis is that fatiguedhip flexors would impair the ability to regain balance during a for-ward fall. This alternate hypothesis is supported by the fact thatsuccessful balance recovery was based on the margin of stabilityat touchdown. In other words, if they got their foot out far enoughin front of them, they were able to stop their fall and regain bal-ance. Because they are falling the amount of time to place theirlead foot out in front of them is limited. Therefore it seems plausi-ble the ability to quickly flex at the hip, likely influenced by hipflexor strength, may be a deciding factor in achieving an acceptablemargin of stability and regaining balance during forward falls.Knee flexion and ankle dorsiflexion could possibly also help recov-ery by shortening the moment arms of the leg around the hip jointand aiding in toe clearance.

These findings have important implications for interventionsstrategies to prevent falls during daily life. It is accepted that mus-cle strength and tendon stiffness contribute significantly (�30–40%) to the capacity of the human system to regain balance afterforward falls (Wojcik et al., 2001; Grabiner et al., 2005; Karamani-dis and Arampatzis, 2007; Karamanidis et al., 2008). Our resultsindicate that increased strength of the lower extremity extensorsmay not be the only effective intervention strategy for preventingfalls. The ability of humans to immediately generate proper motorbehaviour for successful postural corrections including mecha-nisms responsible to maintain stability seems to be very importantfor fall preventions. Increasing the base of support after a perturba-tion does not require high levels of muscle force. Therefore, we canargue that, practicing motor tasks including mechanisms responsi-ble for dynamic stability control, such as a quick lunge step, may

contribute in an improvement in the ability to recover balancewithout falling.

During the main stance phase several significant modificationsin the way to achieve balance were documented indicating thatthe fatiguing protocol did affect certain aspects of the dynamic sta-bility control. The decrease in the maximum ankle and knee jointmoments during the main stance phase show that the fatiguingprotocol affected the moment generation in the lower extremitiesjoints. As a consequence the horizontal and vertical ground reac-tion forces decreased during the main stance phase after the fati-gue. Nevertheless, the participants showed similar margins ofstability in the end of the main stance phase in the pre- andpost-fatigue conditions. They were able to counteract the de-creased average horizontal ground reaction forces after the fatigueby flexing their lead leg knee to a greater extent thereby increasingthe deceleration time of the horizontal CM velocity. This adjust-ment in the recovery leg resulted to similar integral of the horizon-tal ground reaction force during the main stance as that in non-fatigued condition. The results indicate an appropriate adjustmentof the human system to compensate for the deficit in musclestrength.

The concept of the extrapolated centre of mass used in the cur-rent study is based on the inverted pendulum model. Although theformulation has been progressed from standing balance to dy-namic conditions (i.e. when CM has an initial velocity) the modelis a simplification of the human body as a pendulum with a mass(Arampatzis et al., 2008; Hof, 2008; Hof et al., 2007, 2005; Curtzeet al., 2010). During dynamic situations, for example the simulatedforward falls, several segments of the human body are movingaround the CM. The non consideration of these movements inthe model may reveal some inaccuracies regarding the state of sta-bility. Another limitation is that because our fatigue protocol useda multi-joint exercise performed to self determined exhaustion theexact fatigue level of any of the given muscles is unknown. We be-lieve that this loss of control was justified by the increase in eco-logical validity of the dynamic squat exercise over, for example,an isometric leg extension.

We concluded that the decrease in lower extremity extensormuscle strength after the fatiguing submaximal contractions didnot lead to functional deficits related to the capacity of the humansystem to regain balance after forward induced falls. This demon-strates that this specific impairment of the musculoskeletal systemalone cannot predict the postural performance. The main mecha-nism responsible for regaining balance after a forward fall from aforward inclined position is increasing the base of support duringthe swing phase of the recovery limb and the using of this mecha-nism does not require high level muscle forces from the humans.We suggested that exercise interventions including mechanismsresponsible for dynamic stability control may allow humans tolearn and effectively use these mechanisms during sudden pertur-bations and, hence, reduce the risk of falls during daily life.

References

Arampatzis A, Karamanidis K, Mademli L. Deficits in the way to achieve balancerelated to mechanisms of dynamic stability control in the elderly. J Biomech2008;41:1754–61.

Curtze C, Hof AL, Otten B, Postema K. Balance recovery after an evoked forward fallin unilateral transtibial amputees. Gait Posture 2010;3:336–41.

Grabiner MD, Owings TM, Pavol MJ. Lower extremity strength plays only a smallrole in determining the maximum recoverable lean angle in older adults. JGerontol A-Biol Sci Med Sci 2005;60:M1447–50.

Granacher U, Zahner L, Gollhofer G. Strength, power, and postural control inseniors; considerations for functional adaptations and for fall prevention. Eur JSports Sci 2008;8(6):325–40.

Hof AL. The equations of motion for a standing human reveal three mechanisms forbalance. J Biomech 2007;40:451–7.

Hof AL. The ‘extrapolated center of mass’ concept suggests a simple control ofbalance in walking. Hum Mov Sci 2008;27:112–25.

M. Walsh et al. / Journal of Electromyography and Kinesiology 21 (2011) 270–275 275

Hof AL, Gazendam MG, Sinke WE. The condition for dynamic stability. J Biomech2005;38:1–8.

Hof AL, van Bockel RM, Schoppen T, Postema K. Control of lateral balance in walking,experimental findings in normal subjects and above-knee amputees. GaitPosture 2007;25:250–8.

Hsiao ET, Robinovitch SN. Biomechanical influences on balance recovery bystepping. J Biomech 1999;32:1099–106.

Karamanidis K, Arampatzis A. Age-related degeneration in leg extensor muscle-tendon units decreases recovery performance after a forward fall:compensation with running experience. Eur J Appl Physiol 2007;99:73–85.

Karamanidis K, Arampatzis A, Mademli L. Age-related deficit in dynamic stabilitycontrol after forward falls is affected by muscle strength and tendon stiffness. JElectromyogr Kinesiol 2008;18:980–9.

Mademli Lida, Arampatzis Adamantios, Karamanidis Kiros. Dynamic stabilitycontrol in forward falls: postural corrections after muscle fatigue in youngand older adults. Eur J Appl Physiol 2008;103(3):295–306.

Madigan, M.L., Lloyd, E.M., Age related differences in peak joint torquesduring thesupport phase of single step recovery from a forward fall. J Gerontol A-Biol SciMed Sci 2005;60:910–4.

Moritani T, Oddson L, Thorstensson A. Electromyographic evidence of selectivefatigue during the eccentric phase stretch/shortening cycles in man. Eur J ApplPhysiol 1990(60):425–9.

Pijnappels M, Bobbert MF, van Dieen JH. Push-off reactions in recovery aftertripping discriminate young subjects, older non-fallers and older fallers. GaitPosture 2005;21:388–94.

Simoneau M, Corbeil P. The effect of time to peak ankle torque on balance stabilityboundary: experimental validation of a biomechanical model. Exp Brain Res2005;165:217–28.

Thelen DG, Wojcik LA, Schultz AB, Ashton-Miller JA, Alexander NB. Age differencesin using a rapid step to regain balance during a forward fall. J Gerontol A-Biol SciMed Sci 1997;52(1):M8–M13.

Wojcik LA, Thelen DG, Schultz AB, Ashton-Miller JA, Alexander NB. Age and genderdifferences in peak lower extremity joint torques and ranges of motion usedduring single-step balance recovery from a forward fall. J Biomech2001;34(1):67–73.

Yeung SS, Au AL, Chow CC. Effects of Fatigue on the temporal neuromuscular controlof the vastus medialis muscles in humans. Eur J Appl Physiol 1999;8:379–85.

Zatsiorsky VM, Selujanov VN. The mass and inertia characteristics of the mainsegments of the human body. In: Matsui H, Kabayashi K, editors. BiomechanicsVIIIB. Champaign, IL: Human Kinetics; 1983. pp. 1152–9.

Mark Walsh is an associate professor at Miami Uni-versity in the United States. He received his Ph.D. inbiomechanics at the German Sport University Cologne,Germany. His research interests include the mechanicsof human movement, balance and sport biomechanics.

Andreas Peper graduated in sport science at the Ger-man Sport University of Cologne. There he started hisPh.D. in the field of dynamic stability control of elderlypersons in conjunction with different training inter-ventions and continues his thesis since May 2009 at thedepartment of training and movement science of theHumboldt-University of Berlin.

Stefanie Bierbaum graduated 2006 in sport science atthe German Sport University of Cologne. She started herPh.D. in biomechanics at the Institute for Biomechanicsof the German Sport University and continues with herwork since 2009 at the department of training andmovement science of the Humboldt-University of Ber-lin. She is doing her Ph.D. thesis on locomotor adapta-tion (short- and long-term) and dynamic stability of theelderly.

Kiros Karamanidis, received his Ph.D. at the GermanSport University of Cologne in 2006. His main researchinterests are in the field of adaptation of aging musclesand its effect on gait mechanics focusing on the pre-vention of falls in the elderly.

Adamantios Arampatzis is professor and chair of theDepartment Training and Movement Sciences at theHumboldt-University of Berlin. Among his researchinterests are the plasticity of the musculoskeletal sys-tem to exercise and the influence of the neuromuscularcapacity of the human system on motor task behaviourduring daily and sport activities. His research workconcentrates on muscle–tendon unit adaptation, neu-romuscular control of locomotion, dynamic stability andjoint mechanics.