A Bio Mechanical Study of Balance Recovery

8/2/2019 A Bio Mechanical Study of Balance Recovery

1/7

,. B,omrcknr cs Vol. 15. No. 12. pp 933-939. 1982. OO?l-9290/82 ,120933-07 SO3 000Prmred ,n Great Bntam Pergamon Press Ltd

A BIOMECHAN ICAL STUDY OF BALAN CE RECOVERYDURING THE FALL FORWARD*

M. C. Do, Y. BRENIERE, P. BRENGUIERLaboratoire de Physiologie du Mouvement. Universitt Paris-Sud. 91405 Orsay, France

Abstract-The study deals with the biomechanics of balance recovery of human subjects in a falling forwardsituation. Eight subjects took part in the experiment. The subject, held in the initial leaning forward position,was released without his knowledge. The instruction was to recover the induced disequilibrium by walking.The biomechanical analysis shows two phases in the balance recovery. The first phase-preparation phase--is characterized by three events at fixed timing whatever the initial inclination. (i) Dynamic reaction time,showing no significant inter-individual variation (mean value = 90.8 ms). (ii) Braking of the forward fall,between 184 ms and 237.2 ms,depending on the subject. (iii) Beginning of the swing phase-i.e. toe-offinstant-between 235.9 ms and 328.3 ms, depending on the subject. The second phase-gait executionphase---is characterized by the duration of the swing phase, the duratioqof the stance phase, the stride lengthand execution speed. The durations diminish whereas the stride length and the execution speed increase withrespect to the initial inclination. For the same execution speed, the stride length is shorter than in normalwalking.It has been concluded that balance recovery following an induced fall forward begins with an invariablepreparation process which is followed by an adaptable recovery one.

INTRODUCTIONFor several years now, a renewal of interest has beenshown in the neurophysiological mechanisms whichunderlie the postural activity in man. For example, theelectromyographic (EMG) responses following a freefall (Melvill Jones er a[., 1971; Greenwood andHopkins, 1976), a voluntary fall forward onto aninclined plane (Dietz and Noth, 1978), a forward orbackward push (Bussel et a[., 1980), a disturbance ofthe posture of subjects standing upon a movableplatform (Lestienne et al., 1977; Nashner et al., 1979)have been studied. Recently EMG analysis of muscleresponses characterizing postural reactions to a fall hasbeen the focus of inquiry (Do et a[., 1981). The aim ofthis article is to provide a biomechanical analysis ofthis forward falling movement.Among the different techniques used in the investi-gation of human movements, we chose a force plat-form, the advantages of which have been recognizedsince the first experimental research of Amar (1923)and Demeny (1924). Indeed one of the reasons is thatthis methodology reduces the study of a pluri-segmented system to a study of the dynamic process ofits center of gravity. Consequently, we can thus avoidresorting to a body model in the form of link-articulated systems which involves simplifying hypo-theses (Braune and Fischer, 1895 and many otherauthors since). In addition, the fact that the subject isnot constrained in his movement is important in thepresent experimental situation.

TECHNIQUE --PROCEDUREThe subject was secured in a wide pelvic belt which

._*Rrceiced 29 Drcemhrr 1980: in.finalform 13 July 1982.

was connected by a horizontal cable to a forcetransducer mounted on a fixed frame. The subjectsinclination, a, is defined by tan a which is equal to theratio of the restraining force, F,. (Fig. 1) to the subjectsweight, w tanu = F,IW

The force platform and its characteristics weredetailed in Breniere et al. (1981). In sum, it is anequilateral triangle structure, 2 m on each side, which issuspended by means of cables to three extensiometrygauge systems measuring in the range O-1000 N with a0.2 / linearity error. The gauge systems were adjustedto be sensitive to vertical force variations. Their rigiditywas 10 N/m, and the cables had a Youngs modulus,E = 8750 kg/mm. The cable platform suspensiontype left six degrees of freedom to the system. Thenatural frequency of the platform when loaded is ap-proximately S@Hz, which is higher than physiologi-cal frequencies of articulated systems (Payne, 1968). Inthis investigation, only the vertical resultant of force,AR,, and its point of application, I, i.e. its coordinatesx,, y,, were considered (see calculations in Appendix).The proportionality which exists between AR, and yG,the vertical acceleration of the center of gravity G,AR, = Myc, where M is the subjects mass, makes itpossible to consider that the reaction force AR, reflectsthe acceleration of the center of gravity. The forceplatforms axes have been set so that:

(1) a negative value of AR, corresponds to a verticalacceleration of G in the same direction as gravity g;(2) a positive value of x, corresponds to a displace-ment of the center of foot pressure towards the front ofthe supporting area;

(3) a positive value of yl corresponds to a displace-ment of the center of foot pressure towards the left sideof the supporting area, i.e. the left foot.Data processing and data acquisition were carried

933


2/7

934 M. C. Do, Y. BRENIERE nd P. BRENCUIERout by a PDPll/O4 mini-computer with 28,672 16-bitwords of central memory. The sampling period in dataacquisition was 5 ms per channel. Data acquisition wastriggered by the opening of an electrical switch whenthe first of the two heels left the ground. It gave timetHo. The duration of the recording prior to tHOcan bechosen by a special procedure. The results, i.e. AR,, x,,y,, F, were stored on a floppy disk.

The investigation was performed on 8 healthysubjects and repeated at least twice for each one. Foreach examination, between 20 and 50 trials wererecorded. The subjects were held in the initial forwardleaning position (Fig. 1) in which the body was straightand the arms were held alongside the body. The initialinclination, between 5 and 35 depending on thecapability of the subject, was controlled by changingthe length of the cable in a pseudo-random sequence soas to avoid habituation. When the subject was motion-less, the restraining connection was broken withoutwarning. His instruction was to recover his balance bywalking, which is the natural behavior, at least as soonas the inclination goes beyond several degrees. Due tothe platforms large size, the subject executed one stepor more on it, and then continued to walk, if necessary,on an additional contiguous wooden platform,situated at the same level. No instruction was givenabout which foot should initiate the movement.

RESULTS

The balance recovery consists in the execution ofone or more steps, and it is necessary, as for the gaitstudy, to determine the biomechanical parameterswhich allow for a convenient characterization of theprocess. We are chiefly considering in the present paperthe period covering the release of the subject and theexecution of the first step.1. Biom echanical param eters of balance recov ery

Figure 2 shows a typical recording of a balancerecovery movement. It can be seen that the verticalresultant of forces, AR,, presents large variation

around the reference line, i.e. the subjects body weight.These oscillations start 5 ms after the subjects release(which corresponds to the abrupt variation of F, at to)and then show successively a negative peak value attime t,, a positive peak value at time tt and twodifferent negative peak values at tie and tHC,and soon.

Because AR, is first negative, the movement starts bya fall of the bodys center of gravity. Because the systemof the external forces is limited to the body weight andthe ground reaction, the first change in the direction ofAR,, which occurs at tf , can be considered as aninitiated brake of Gs fall reaching a relative maximumat t,. This change is due to the action of muscularforces. A direct consideration of the EMG activity ofthe ankle extensors supports this opinion (Do et al.,1981). The time tHOwhen the contact between the heeland the ground is lost, is situated between tf and tl.The significance of the instants t+. and tHC s easy tounderstand when considering x, and y,. The variationof y, indicates a displacement of I towards the footwhich is going to oscillate, which is in this case the rightfoot, and then towards the stance foot, i.e. the left foot.From point M to point N, the center of foot pressurestays in a stable position which corresponds to MN ony, recording. During this time, the subject is in thesimple stance phase. As it can be noticed, M and Ncoincide with AR, negative peak values occurringrespectively at t i. and tHc. In other words t+. and t HCcorrespond to the beginning and to the end of theswing phase-i.e. the toe-off (TO) and the heel contact(HC). The variation of x, starts at the time when thesubjects heel leaves the ground-instant t,,-and itsamplitude is low until tHC,where an abrupt forwarddisplacement takes place. The amplitude of the initialx, variation makes it possible to know the initialdistribution of the pressure under the feet, which canbe, depending on the subject, located either more onthe heel or on the forefoot. The double stance phasebegins at the instant tHCand lasts until the instant tie,where a second swing phase begins which can bedetermined as before. Between these two instants,

Fig. 1. Diagram of the experimental situation. Subject is on the force platform (1) and is maintained in aleaning forward position by a restraining and releasing apparatus (2). A wooden platform (3) is added at thesame level contiguous to the force platform.


3/7

A biomechanical study of balance recovery during the fall forward

200msFig. 2. Typical recording of a balance recovery movement. Upper part: diagram of step execution phases.Left foot (LF), right foot (RF), time of releasing (f,).time of heel-off (t HO), time of swing phase (tfo). time ofheel-contact (tic), time of second swing phase (t&), mechanical reflex time (tf), time of the maximum verticalpush (ft ). Lower part: AR,: variations of the vertical resultant of force with respect to the weight, K of thesubject. x,: antero-posterior displacement of the point ofapplication of forces, forward (F), backward (B). J,:lateral displacement of the point ofapplication of forces, towards the left foot (L), towards the right foot (R).

F,: intensity of the restrainmg force. Time scale: 200ms.

935

there is a rapid displacement of the center of footpressure forward and toward the new stance foot. Theforward displacement allows the measurement of thestep length, L, which is represented by the amplitude ofpoint P. The step speed, V. can also be calculated asLo= L~(t~o-t~ojwhere(t~o-t~O)isthedurationofhalf of the step cycle.

The synthesis of these results permits the represen-tation of the balance recovery process. The subject,released at time t,, falls around the ankles with a yGwhich reaches a relative minima at tj A lateralizationof the fall can sometimes be observed, as in Fig. 2. AttHO, the heel contact ceases and is followed by animportant positive value of yG at time tf The subjectstarts executing the step at time rio. So, the balancerecovery step is preceded by a particular preparation,the events of which can be defined by tHO, ti, tz andt &. The step that follows is defined by tHc, ttO, L andV. According to the initial inclination, the balancerecovery needs one step or more. Each modification ofthe step phases is accompanied by correspondingvariations of yc, X, and y,.

2. E&t of the initial inclination on the biomechanicaiparameters of balance recovery

The effect of the initiaLinclination on the process ofbalance recovery was considered successively for thepreparation phase and the step execution phase, asdefined above.

(a) Preparation phase. Figure 3(a and b) shows,respectively the relationships between the times l:, rz,ti03 HO and the initial inclination tan 2. The generalresult is that none of these times depends on the initialinclination, i.e. each subject shows a set of values whichdo not depend on the initial inclination. Time t; , whichcorresponds to the first negative peak, takes placebetween 85.4 ms (SD = 7.5 ms) and 100.2 msSD = 10.6ms) after 1,. according to the subject (seeTable 1). The mean value of these averages is 90.8 ms(SD = 4.6 ms). The Student-Fischer t test applied totwo subjects at a time does not show any significantdifference between the individual means. Thus, it canbe assumed that the relationships between I: and theweight, w or the height, H, are not simple, as isconfirmed by Fig. 4(a and b).

The instant ti, which corresponds to a verticalmaximum push, occurs between 184 ms (SD = 7.6 ms)and 237.2ms (SD = 16.3 ms) after to, according to thesubject. Contrary to tj. the Student-Fischer t test.applied to two subjects at a time, indicates a significantdifference between the subjects. However, Fig. 4 (a andb) shows that these differences cannot be simplyexplained in terms of subjects weight or height.Time tko shows very small variations for a givensubject. The mean individual values vary between235.9 ms (SD = 7.3 ms) and 328.3 ms (SD = 15.2 ms).Just as for ti, the mean values of tie were significantlydifferent among the subjects, but are not simply relatedto weight or the height [Fig. 4 (a and b)].


4/7

936 M. C. Do, Y. BRENIERE and P. BRENGUIER

hlSl3 00

1

(b)

Fig. 3. Relationships between characteristic times and initial inclination. (a) Relationships between rl, cj, tieand the initial inclination tan a. t:, mechanical reflex time ( m ; :, instant of the maximum of the vertrcal push(0); tie, beginning of the swing phase (A). (b) Relationship between time of heel-off, tHO. and the initialinclination. (c) Relationships between times tHC and t$o, and the initial inclination. tH,-, heel contact (o),z$~, beginning of the swing phase (A). tan a, initial inclination in radians, equivalent in degrees is in brackets,

time scale, in milliseconds (ms).

Table 1. Grouped results of the eight subjects examined(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

H W rf r2(m) (kg) (ms) (r&

40 L =f(tan rf) V =f(tan z) L =f( CI (2ro- I +.o)=,f( an a) Cadence(ms) min: max

CM 9 1.62 56 92.5 211.1 292.2 1.26; .63 9.8; 0.81 0.12; 0.536 - 0.72; 0.465 156: 220(k8.1) (k7.5) (rt15.1) (0.71) (0.89) (0.78) ( - 0.75)

DJ 9 1.64 50 88.7 202.3 295.8 1.97; 0.48 14.3; 0.36 0.13; 0.437 - 0.85; 0.445 158; 247(k8.3) (t5.6) (k5.5) (0.89) (0.95) (0.92) (-0.89)

NS ? 1.75 73 100.2 231.2 328.2 1.05; 0.38 8.7; 0.03 0.11; 0.401 - 0.92; 0.536 130; 187(i 10.6) (& 16.3) (+ 15.2) (0.89) (0.97) (0.82) (-0.91)

NC s! 1.65 54 91.9 186.0 265.4 2.49; 0.26 12.3; 0.13 0.11; 0.280 - 1.09; 0.445 164; 317(k10.7) (k10.7) (k9.6) (0.71) (0.92) (0.72) (-0.78)

EJ c? 1.82 78 92.7 197.2 285.9 0.81; 0.56 6.6; 0.07 0.12: 0.588 - 0.59; 0.596 132; 215(k12.4) (k32.4) (k7.3) (0.91) (0.96) (0.93) ( - 0.95)FH 6 1.80 69 86.9 218.1 298.5 1.77; 0.49 10.8; 0.39 0.13; 0.512 -0.90; 0.516 132; 240

(k 10.2) (+ 12.1) (+ 14.6) (0.84) (0.88) (0.73) (-0.71)SM 6 1.85 80 85.4 184.0 235.9 1.38; 0.52 12.1; 0.13 0.12; 0.542 - 1.01; 0.566 134; 240

( + 7.5) ( k 7.6) ( + 8.2) (0.84) (0.94) (0.87) (-0.88)AH 6 1.78 76 88.0 219.9 318.4 0.23: 0.82 10.6; 0.68 0.01; 0.829 - 1.06; 0.534 139; 305

(k11.6) (k30.4) (k14.3) (0.19) (0.94) (0.12) (-0.89)Columns 1 to 3. Subjects initials and sex (1); height (H) in meters (2); weight (W) in kilograms (3).Columns 4 to 6. Means and SD of times t: (4), t: (5), tie (6).Columns 7 to 10. Slope (1st value); intercept (2nd value); correlation coefficient (3rd value in parentheses) of regression linesbetween, the step length (L) and the initial inclination tan a (7), the average speed ( V) and tan a (8). the step length (L) and theaverage speed ( V) (9), half cycle of the stride (rto - tiO) and tan a (10).Column 11. Minimum (min) and maximum (max) cadences in numbers of steps per minute.

Unlike the preceding times, the instant of heel-off, the feet flat while maintaining at the same time antHo, fluctuated appreciably. The individual means vary initial straightforward leaning position. It can bebetween 130 ms and 220 ms, but the amplitude of these stressed that the standard deviations, with the excep-fluctuations differs from one subject to another. Only tion of tHO, re small if the measurement error, which isfour out of eight subjects exhibited low variation. The about 10 ms, is taken into account. In other words, thefluctuations are explained by the difficulty in keeping balance recovery preparation phase follows a re-


5/7

A biomechanical study of balance recovery during the fall forward 931

,?!

(a),t+

+ I

t+ +t It

(b)

+*+++

+ +t[HDJK ? A:!Aso J I

50 60 70 5O Mb1 1.6 1.7 La 1.9 H ( ml

Fig. 4. Representation of the set of values, t: , t: , t &, ofeach subject in terms of weight ( W) (a) and height (H)(b). Each set of initials represents a different subject. Scales: time in milliseconds (ms), weight in kilograms(kg), height in meters (m).

markably fixed sequence, or at least it varies very littlefor each subject.

(b) Execution phase. The recovery step executionphase is characterized by time tHc and time fro, as wellas by the length L and the speed Y of the stepexecution. The general result is that all these para-meters depend on the initial inclination, tan a.Times tHc and t:. diminish with repect to tan a[Fig. 3(c)]. If we consider the correlation between t,,or tto and the initial inclination, a highly significant,(p < O.Ol), correlation coefficient is obtained for eachsubject: r(t,,-, tana) is between 0.72 and 0.95, and

LP-1-1 taj : ..9 I -.**.* .*.a - : ..7 *

1

r(tto, tan a) is between 0.72 and 0.92. The duration ofthe swing and double stance phases can be calculatedfrom times fro, tHC and tto. Respectively, they areequal to (tHC- tie) and @to- t&. The half cycle isequal to (tiO- ti,). The parameters of the linearregression equations are given in Table 1.

Figure S(a and b) shows the variations of the lengthof the step, and the average speed in relation to theinitial inclination. It is observed that L and V increaseas function of the inclination.

By approximating the variation of L as function ofthe inclination, by a linear regression, highly significantcorrelation coefficients between 0.71 and 0.91 areobtained, for seven subjects out of eight. Only onesubject presents a correlation which is not significant(0.19; see Table 1).

In the same way, highly significant correlationcoefficients ranging for a given subject from 0.88 to0.97 are obtained for V. It can be noticed thatcorrelations between V and tan a always show highercoefficients than those between L and tan a.

V(m/9 (b)1 . 7 - .0 .. .1 . 5 - . . * l

. . . .1 . 3 -

* : :l . l - * :

I-% I. 2 . 3 . 4 tan ofIll?) 116?) (2181

Fig. 5. Relationships between the step length (L) (a) and theaverage speed (V) (b) and the initial inclination tan 0~. cales:step length (L) in meters (m), average speed (v) in meters persecond (m/s).

DISCUSSIONThe above results make possible a differentiation of

the preparation phase from the execution phase,according to their dependence on the initialinclination.fnuariability of the prepururion phaseBalance recovery following the induced forward fallbegins by a phase which ends at the toe-off and may beconsidered as a preparation for the recovery move-ment. During balance recovery, the sequence of theprincipal instants t, , tf , rio is chronologically fixedand their values are not dependent on the initial incii-


6/7

938 M. C. Do, Y. BRENIERE and P. BRENGU IERnation. Each of them corresponds to an important stepin the series of events? which results in the execution ofthe balance recovery process. The first occurrenceis comprised of the negative deflection of AR,whose relative maximum occurred at mean timet, = 90.8 ms and shows a low intra-individual scatter-ing. Electromyographic activity of the ankle extensormuscles starts approximately 50 ms on average aftert,, irrespective of the initial inclination, which cor-responds to a reflex time (Do et al., to be published).Thus, taking into account the electro-mechanical delaywhich is known to be 3@-40ms, time t, may beconsidered as a mechanical reflex time. Then, thisdynamic reaction develops in the form of a verticalpush whose maximum was at time tz. This constitutedthe second occurrence. The vertical push correspondsin this case to a braking of the fall. Time tf wasdifferent for each subject and it would be reasonable tosuppose that braking was changed according to theactivities of the ankle extensor muscles, especially, andalso according to the activities of back muscles whichchange the trunk position. These muscle activitiesshould lead to the step execution. During these firsttwo sequences, the only apparent movement is the heellift. This heel lift fulfills a biomechanical necessity.Indeed, on the one hand, the subject has somedifficulties in keeping the heels on the ground, and onthe other hand the leaning position may be comparedto the situation preceding the double stance in thenormal gait where the subjects weight is supported bythe fore part of the stance foot. In other words, and interms of motor program, the step execution is preparedsoon after the release.

The preparation process ends at the toe-off, fro,which depends on the preceding events and, as tz,shows inter-individual variations, depending on indi-vidual initial position and strategies. But it does notshow important intra-individual variations. This timetie may be considered as a physiological time whosemechanical expression is the time which is necessary torealize the displacement of the swing foot. On theassumption that one of the purposes of the posturaladjustment associated with movement is . . . to makepossible the displacement ofa limb that was previouslysupporting a part of the body weight (Massion andGahery, 1979), then time fro, which does not dependon the initial inclination, is the minimal durationnecessary to complete this function. As a matter of fact,if the subject did not have this minimal time, balancecould not be recovered by executing a step. In such acase, the subject would probably adopt a differentstrategy which would consist in breaking the fall withhis arms. We have observed that when subjects leanedmarkedly forward, there was considerable elbowflexion just after the release. This strategy is identical tothe one observed by Dietz and Noth (1978). Theredundancy in the balance recovery strategies remindsus that the sensory messages can participate either in ahierarchical manner or in a joint manner in thetriggering of the balance recovery program. This

triggering seems to be regulated automatically on thebasis of sensory cues (Massion and Gahery, 1979).Adaptability of the recovery phase

The recovery step is carried out according to thespecific dynamics mentioned above. Its kinematiccharacteristics, contrary to the preparation phase, arerelated to the initial inclination: tHCand t to decreasewith increasing inclination, while L and V increase. Inother words, the recovery phase depends on the initialinclination, i.e. it adapts to the initial disequilibriumtorque.

The fact that the double stance always exists nomatter what the initial inclination is, makes it possibleto compare the balance recovery step with the gait stepat different speeds. On the assumption that in balancerecovery, the length of the first step is the step length,the relationship between the speed and the stride canbe outlined (Fig. 6). It is noted that the aspect of therelationship between L and V is similar to what hasalready been pointed out by Marey (1891), i.e. thelength of the step increases with the execution speed.However, compared with the values obtained byCavagna and Margaria (1966), the step length isrelatively shorter for the same speed (Fig. 6). Thislength rarely goes beyond one meter, even for subjectsas tall as 1.85 m. Conversely, for the same step, theexecution speed in our experimentation is faster. Thedifference observed in execution speeds is possiblebecause of the high execution frequency combinedwith a relatively short step length. For all the subjects,cadence based on the first step is between 130 steps perminute and 317 steps per minute (see Table l), whereasnormal walking cadence rarely goes beyond 1.50stepsper minute (Grieve and Gear, 1966).Another characteristic can be deduced from therelationship between L and V. The linearity of therelationship, also verified in normal walking (Cavagnaand Margaria, 1966), shows that the variables Land V

LmlI ,.s ,,,&Fig. 6. Gathering of relationships between step length (L)and average speed (V).The broken straight line is taken fromCavagna and Margaria (1966). The other straight linesconcern the eight subjects examined. Note that the balancerecovery step length is shorter than in normal gait.Scales: step length (t) in meters (m), average speed (V) inmeters per second (m/s).


7/7

A biomechanical study of balance recovery during the fall forward 939are biomechanically linked. However, it seems difficultto decide what variable is relevant. All the same, anelement of reflection can be provided by the fact that inone of the eight subjects, a non-significant correlationbetween the step length and the initial inclination, aswell as a highly significant correlation between thespeed and the initial inclination were observed. Giventhat the speed is equal to L/(t$,- rio), it means that Land (t f. - t io) are related in a special way in order togive a constant relationship to V. In other words, sincethe relationship between the speed and the initialinclination is the most consistent for all the subjects,speed would be the relevant kinematic variable.

CONCLUSION

Balance recovery while walking is a dynamic processwhich consists of two phases. The purpose of the firstphase, which is invariable, is to create the conditionsnecessary for step execution, i.e. to reach time tie. Itimplies a special strategy: the subject lifts his heels andstands on the soles of his forefeet. Given the dynamicsof the system at the end of the first phase, the executionof the second phase-i.e. the step execution-which isadapted to the initial conditions, is carried out with arelatively short stride but with a high stride frequency.

REFERENCES

Amar, J. (1923). Le Moteur Humain. (Edited by Dunod),Paris. p. 690.Braune, W. und Fischer, 0. (1895) Der Gang des Menschen.Ges. Wk. Math. Phys. T.1, abh. K. S&hi, p. 153.Breniere. Y.. Do. M. C. and Sanchez. J. (1981) A bio-~mechanical study of the gait initiation process. J. 1;.Biophy s. M &d. Nucl., 5, 197-206.Bussel, B., Katz, R., Pierrot-Deseilligny, E., Bergego, C. andHayat, A. (1980) Vestibular and proprioceptive influenceson the postural reactions to a sudden body displacement inman. In Spinnl and Supraspinal Mechanisms of VotuntaryMot or Control and Locom otio n: Progress in ClinicalNeurophysiology (Edited by Desmedt, E.), Vol. 8,pp. 3l(t322. Karger, Basel.

Cavagna, G. A. and Margaria, R. (1966) Mechanics ofwalking. J. appl. Physiol. 21, 271-278.Demeny, G. (1924).Mk anismeet Education des Mouoements.(Edited by Alcan, F.). Paris.Do, M. C., Breniere, Y. and Bouisset, S. (1981) Is the balancerecovery following an induced fall-forward, programmed?5th European Neuroscience Congress, Liege, Belgium,Sept. 14-18. Neuroscience Lett. Suppl. 7, 11%.Dietz, V. and Noth, J., (1978) Pre-innervation and stretchresponses of triceps brachii in man falling with and withoutvisual control. Brain Res. 142, 576-579.Greenwood, R. and Hopkins, A. (1976) Muscle responsesduring sudden falls in man. J. Physiol., Lond. 254,507-518.Greenwood, R. and Hopkins, A., (1976) Landing from an

unexpected fall and a voluntary step. Brain, 99, 375--386.Grieve, D. W. and Gear, R. J. (1966) The relationshipsbetween length of stride, step frequency, time of swing andspeed of walking for children and adults. Ergonomics 5,379-399.Lestienne, F., Soechting, J. F. and Berthoz, A. (1977) Posturalreadjustments induced by linear motion of visual scenes.Expl. Brain Res. 28, 363-384.Massion, J. and Gahery, Y. (1979) Diagonal stance inquadrupeds: a postural support for movement. In ReflexControl of Posture and Mov ement (Edited by Granit. R. andPompeiano, O.), pp. 219-226. Elsevier, Amsterdam.Marey, E. J. (1891). La Machine Animale. (Edited by Alcan,F.). Paris, p. 331.Melvill Jones, G. and Watt, D. G. D. (1971) Muscular controlof landing from unexpected falls in man. J. Physiol., Lond.219, 729-737.Nashner, L. M., Woollacott. M. and Yuma, G. (1979)Organization of rapid responses to postural andlocomotor-like pertubations of standing man. Expl. BrainRes. 36, 463-476.

Payne, A. H. (1968)The use of force platforms for the study ofphysical activity. Biomechanics 1. 1st Int. Seminar. Zurich1967, pp. 83-86. Karger, Base].

APPENDIX

The vertical transducers of the force platform, N,, N,. N,and the axes have been set so that:R,(t) = N,(r) + Nz(t) + N,(t) N.x _ c,bJMt)+N,(t)2 R,(t)y,(t) = ;.Ns(t)- N,(t)R,(t) N2 ; N3

where c is the triangle side and t the time.To increase the precision of the measurements we pro-ceeded as follows: before each trial the static values of R,, x,,y t , which define the initial position of the subject at rest, werecomputed:

R,(t,) = N,(t,) + Nz(to) + Na(t0) = Wx,(t,) = @, N&)+N&)

2 Wc N,(b) -N,(b)Y/(b) = 2 W

where W is the subjects weight and t,, the instant when theinitial values were computed.Then these values were subtracted from the instantaneousvalues of the trialR,(t) - R,(t,) = AR,x,(t) - x1(t,) = xlyr(r)-Y&) = Yr.

In other words, the shift of the origin has cancelled the offsetwhich would exist for a fixed origin hecause the initialposition of the subject can be different between trials.

A Bio Mechanical Study of Balance Recovery

Documents

Transcript of A Bio Mechanical Study of Balance Recovery