- Email: [email protected]
Coordination of the leg muscles in backlift and leglift
3. Biomechanics Vol. 25, No. II, pp. 1279-1289,
COZl-9290/92
1992.
COORDINATION Huue
IS.oO+.OO
Pergamon Press Ltd
Printed in Great Britain
OF THE LEG MUSCLES IN BACKLIFT AND LEGLIFT
M. TOWSAINT,*CORINNEE. VAN BAAR,PAULP. VAN LANGEN, MICHIELP. DE LOOZEand JAAPH. VAN DIE~N
Department of HealthScience, Faculty of Human Movement Sciences, Vrije Universiteit and Universiteit van Amsterdam, Amsterdam, The Netherlands Ahstrae-Net joint moments are often used to quantify the loading of structures (e.g. the intervertebral disc at LSSl) during lifting, This quantification method is also used to evaluate the loading of the knee, for instance, to determine the effect of backlifting as opposed to leglifting. However, the true loading of the joint as derived from net joint moments can be obscured by a possible co-contraction of antagonists. To unravel
the mechanisms that determine the net joint moments in the knee, the leglift was compared to the backlift. Although a completely different net knee moment curve was found when comparing the two lifting techniques, it appeared to be closely related to the ground reaction force vector and its orientation with respect to the joint centre of rotation (R>0.995). This close relation was established by co-contraction of both flexors and extensors of the knee. Furthermore, a close relation appeared to exist between the joint moment difference between hip and knee and the activity difference between rectus femoris muscle and hamstring (R = 0.72 and 0.83 in leglift and backlift, respectively). The knee-ankle joint moment difference and the gctivity of the gastrocnemius showed a close relation as well (R= -0.89 and 0.96 in Ieglift and backlift, respectively). These relations can be interpreted as a mechanism to distribute net moments across joints. It is concluded that during lifting tasks the intermuscular coordination is aimed at coupling of joint moments, such that the ground reaction force points in a direction that provides balance during the movement. The use of net joint moments as direct indicators for joint loading (e.g. knee) seems, therefore, questionable.
INTRODUCTION Manual materials handling and especially lifting can be hazardous and has often been associated with the occurrence of low-back pain (Andersson, 1981; Troup, 1965). It is therefore not surprising that most studies of lifting reported in the literature have been concerned mainly with the trunk. The effect of lifting on the spine has been evaluated using different techniques and methods. Especially, studies in which biomechanical models were used to calculate moments and forces acting on the lower back do abound (Ekholm et al., 1982; Freivalds et al., 1984; Gagnon and Smyth, 1991; Garg et al., 1982; Grieve, 1974; Hall, 1985; Kumar and Davis, 1983; Leskinen, 1985; Leskinen et al., 1983a, b; Martin and Chaffin, 1972; McGill and Norman, 1986; Schipplein et al., 1990; Schultz and Andersson, 1981; Schultz et al., 1983; Troup, 1977; Troup et al., 1983). However, the function of the legs in the lifting process has received scant attention. The two extremes of lifting technique frequently discussed in the literature are leglifting, where the knees are flexed and the trunk is vertical, and backlifting, where the knees are straight and the trunk is flexed. It has been widely advocated that all lifting should be carried out with a leglifting technique
Received infinalform 10 February 1992. * Author to whom correspondence should be addressed at: Department of Health Sciences, Faculty of Human Movement Sciences, van der Boechorststraat 9, 1081 BT Amsterdam, The Netherlands.
(Bendix and Eid, 1983; Miller, 1980). The presumed basis for this has been to reduce the compressive force on the low back and to shift the stresses on the body from the low back to the legs. The latter assumption has been substantiated by Bejiani et al. (1984), who developed a model relating the loads at the knee to the relative participation of the back in lifting. The results of their study indicate a strong inverse relation between joint reaction forces occurring in back and knee. This suggests that it is indeed possible to distribute the load during lifting over the knee and back. The question then arises to what extent this load sharing can be realised. This question was examined by Schipplein et al. (1990) who studied the load sharing during leglifting. They observed that when handling increasing weights using the leglifting technique, the net knee moment remained constant, whereas the moment in the back increased. Their interpretation of this observation was that the quadriceps muscle strength limits the subjects’ ability to lift with their knees flexed. However, this conclusion was based on the assumption that the net knee moment is determined by the activity of the quadriceps muscle only, ignoring the possible effect of other muscles that might influence the knee moment. Recently an alternative hypothesis concerning the regulation of the net joint moments was proposed (Ingen Schenau, 1989, 1990). The often observed cocontractions of monoarticular and biarticular muscles around joints (e.g. the knee) was placed in the perspective of the control of the direction of an external force (during lifting; i.e. the ground reaction force). It
1279
H. M. TOUSSAINTet al.
1280
was stressed that the relation between net joint moments and external forces is causal. In slow and static movements and ignoring the influence of gravity and inertial effects one can approximate these moments directly from the magnitude and direction of the external force. For an example presented in Fig. 1 this can be performed by multiplying the smallest distance between the line of action and the location of the centre of rotation of the joint and the magnitude of the ground reaction force. For the knee and hip joint this leads to M ,_
= d * F,, and M,i, = g - F,,.
(1)
It was argued that biarticular muscles can adjust the distribution of net moments over joints that are crossed. Activation of the rectus femoris muscle, for example, will cause an increase of the net knee moment and a decrease of the net moment in the hip, while hamstring activity causes a shift of net moments in opposite direction. Based on an inverse dynamical analysis of cycling it was demonstrated that the net joint moments necessary to control a certain direction of the external force are largely independent from the required joint displacements associated with the task (Ingen Schenau, 1989; Ingen Schenau et al., 1992). This apparent contradiction can be solved by coactivation of monoarticular and biarticular muscles. The biarticular muscles are able to change the distribution of net moments over the joints that are crossed. Furthermore, biarticular muscles can act as a stiff tendon. Extension of the knee by the knee extensors leads ,to displacement of the origin of the gastrocnemius. This will induce a pull on the insertion of this muscle when the muscle is not lengthened. This means
that the extension of the knee will also cause a plantar flexion by means of the gastrocnemius. This was defined by Cleland (1867) as a ‘ligamentous action’(see Ingen Schenau, 1990). The work done in the knee extensors can (partially) be used for plantar flexion. Hence, using the ligamentous action of biarticular muscles one can transport energy. Combined with the previous described capacity to distribute net moments, biarticular muscles allow the monoarticular muscles to contribute to positive power irrespective of the net moments required in the joints crossed by these latter muscles (Ingen Schenau, 1990). The question arises whether an analysis of lifting in the perspective of the control of the direction of an external force would enhance the understanding of the function of the legs during lifting. In this approach the relation between net joint moments and activity of synergistic muscles surrounding the joint as employed by Schipplein et al. (1990) is extended. It is taken into account that during the lifting movement the direction of the ground reaction force should be such that no net rotation of the subject leading to balance loss occurs. In this sense the lifting movement is performed within the constraint of balance (i.e. the subject does not fall forward or backward). Furthermore, if the latter hypothesis (net joint moments explained by an external force) is valid it is interesting to raise the question whether during lifting tasks the intermuscular coordination can be understood on the basis of the described cooperation between monoarticular and biarticular muscles. The characteristic difference between the two extremes of lifting technique, leglifting and backlifting, is the position of the centre of rotation of the knee joint with respect to line of action of the ground reaction force as illustrated in Fig. 2. This will induce a rather high compensatory net knee flexion moment in the
M knee
Fig. 2. Direct estimation of net joint moment based on the ground reaction force and its moment arm with respect to the Fig. 1. A quasi-static analysis reveals that the ground reaction force requires a net knee and hip extending moment to ensure equilibrium.
centre of rotation in the knee. In the backlift (left panel) the effect of the ground reaction force with respect to the knee must be compensated by a knee flexing moment, in the leglift (right panel) a knee extending moment is required to ensure equilibrium.
Coordination in lifting backlift and in the leglift a net moment fluctuating around zero. The aim of the present study is to compare the leglift to the backlift to find out whether the net joint moments are determined by the required direction of the ground reaction force vector. If this would appear to be true then the intermuscular coordination leading to the required distribution of net joint moments will be investigated. Especially the activation patterns of the monoarticular knee extensor (vastus medialis muscle) and biarticular knee flexors and extensors (hamstrings, rectus femoris muscle) in relation to the net moment are of interest.
1281
was 20% of the body height and the vertical distance between the load and the floor was 10% of body height. For the leglift the extent to which the buttocks had to be lowered was indicated too using a wire frame. This lowering depth, representing the minimum distance between the greater trochantor and the floor was 66% of the leg length, quantified as the distance between the greater trochanter and the lateral malleolus when standing upright. Practice lifts were encouraged until the subjects were comfortable with the technique to be executed in the required frequency. Anthropometry
METHODS
Subjects
Eight healthy male students (age 22.9k2.4 yr, height 1.80 + 0.06 m, body mass 70.5 $- 4.5 kg) participated in the experiment. All subjects signed informed consent prior to the experiment. None of the subjects reported a history of low-back disorders. Test protocol
Subjects were asked to lift and lower a barbell (mass 15.3 kg) in two sessions, using different lifting techniques. Each session consisted of seven lifting/ lowering cycles. The duration of the lifting cycle remained fixed and was partitioned in four phases of 1 s each: standing in a standardized starting position in which the barbell was to be maintained in a standardized position also (dependent on the lifting technique, see below), lifting of the barbell, standing upright, and finally lowering the barbell. The subjects were aided in keeping a constant lifting pace by a metronome. During the last two cycles subjects were filmed while ground reaction forces and muscle activity were recorded. Between the sessions subjects were allowed to rest for at least 5 min, during which the lifting technique of the next session was explained to them. Lifting techniques
Two lifting techniques (back and leg) were prescribed. The backlift was performed with extended knees. During the leglift the subjects were instructed to keep the spine as erect as possible. A stick figure representing the sequence of the two lifting techniques is presented in Fig. 3. The lifts were performed as sagittally symmetric as possible. Care was taken to further standardize the execution of the lifting movements in order to reduce inter subject variability. A framework to which tiny wires were attached was used to indicate the location from which the load was to be lifted in the two lifting conditions. This location remained the same for the two lifting conditions and depended on the anthropometry of the subject: the horizontal distance between the load and the distal end of the fifth metatarsal
Markers (diameter 0.5 cm) were placed on the skin to indicate the location of the fifth metatarsophalangeal joint, the ankle joint (the distal part of the iateral malleolus), the knee joint (lateral collateral ligament at the height of the joint cleft), the greater trochanter, L5-Sl (according to Looze et al., in press), and the spinous process of the first thoracic vertebra. The coordinates defined five body segments: the foot, lower leg, upper leg, pelvis and upper trunk/head/ arms/load. Anthropometric data (standing height, total body mass, length of segments) were measured. Positions of segmental centres of gravity and moments of inertia were estimated on the basis of tables-devised by Dempster (1955) and revised by Winter (1979)-and the measured segmental lengths and body weights of the subjects. Kinematics and kinetics
During the sessions the subjects were filmed at 50 frames per second using a 16 mm high-speed film camera (Teledyne type DBM 55) piaced on the coronal axis of the body at a distance of 6.5 m, so that movements in the sagittal plane could be recorded. The coordinates of the anatomical landmarks, of three markers placed on the force plate (used as a point of reference), and of two markers with a known intermarker distance were determined for each tine frame using a motion analyser (Reinka). Absolute coordinates, after proper scaling, were low-pass filtered by a digital filter (zero-phase lag, bidirectional application of a fifth approximation, second-order Butterworth filter with an effective cut-off frequency of 5 Hz). The angles of each segment were calculated relative to the horizontak joint angles were defined as the angles between adjacent body segments. Numerical differentiation of the time histories of the angles and the positions of the segmental centres of gravity (Lanczos 5-point differentiation filter; Lees, 1980) yielded (angular) velocities and accelerations. Vertical (F,) and foreaft (F,,) components of the ground reaction forces were recorded by means of a force platform (Kistler, type 9281B). The analog force signals were amplified, low-pass filtered (30 Hz, fourth order at 24 dB/oct), sampled (100 Hz, 12 bits) and
1282
H. M. TOUSSAINT~~ al.
Backlift
Fig. 3. Stick figures representing a sequence.of the two lifting techniques employed in the present study. Top row represents the ‘leglift’,bottom row represents the ‘backlift’.The ground reaction force is indicated. For more details the reader is referred to the text.
stored on disk. From the distribution of the force. components, the centre of pressure of the force vector was calculated. The synchronization of the film and force data was achieved by means of electrical pulses. Biomechanical model
Net moments and net forces were calculated by means of an inverse dynamical analysis (Elftman, 1939) using a dynamical two-dimensional linked-segment model as described by Looze et al. (in press). The model comprises eight segments (feet, lower legs, upper legs, pelvis, upper trunk/head, upper arms, forearms and hands/load) which were assumed to be rigid and connected to each other by intersegmental joints Moments having a knee, hip, and L5-Sl extending influence were defined positive as well as moments having an ankle plantar flexing effect. In this study the results of the first five segments were analysed. Additionally, using the direct calculation method the moments based on the ground reaction force vector and the position of the joints were calculated. Electromyography
The electromyographic activities (EMG) of six muscles [gastrocnemius mediali, vastus medialis, rectus femoris, semitendinosus, biceps femoris (caput longum) and gluteus maximus] were recorded by means of bipolar direct measurements @ISA 15COl). Pairs of Ag/AgCl surface electrodes (Sentry Medical Products, CA, lead-off area 1 cm2, centre to centre electrode distance 4.5 cm) were applied to the muscles after standard skin preparation (Basmajian, 1978). The positions of the electrodes are described by Gregoire et al. (1984). The EMG signals were stored on tape (TEAC SR70) and plotted on a writer (Gould ES 1OOQ)
for visual inspection during the experiment. The electrical signal was filtered (band pass 20-200 Hz), sampled (400 Hz) and stored on disk. To obtain a linear envelope the signals were rectifield and low-pass filtered (zero-phase lag, bidirectional application of a fifth approximation, second-order Butterworth filter with an effective cut-off frequency of 5 Hz). The subjects performed three standard isometric contractions for each muscle group under five standardized tasks, while EMG was recorded (see Gregoire et al., 1984; Jacobs and Ingen Schenau, in press): -for gluteus maximus: maximal hip extension against external resistance (hip joint W’), -for biceps femoris (caput longum) and semitendinosus: maximal knee flexion against external -resistance (knee joint 9W), -for rectus femoris and vasti: maximal knee extension against external resistance (knee joint 90”), -for gastrocnemius: maximal plantar flexion standing on one leg, -for erector spinae: maximal extension against external resistance (L5-Sl joint 135”). Contractions lasted 3 s. The standard isometric contraction level for each muscle was determined by taking the mean of the rectified EMG of 1 s in which the signal remained constant. The EMGs obtained during the lifting experiments were normalized to 100% standard isometric contraction level (SIC). Treatment of data
Film, force and EMG data were synchronized on the basis of pulses generated by the experimentator which were marked on the film and recorded simultaneously with the force and EMG signals. For the interpretation of the EMG data, a phase lag of 90 ms between the change in EMG and change in mechanical output, electromechanical delay (EMD), was
Coordination in Wing
extended from about 1lo” to about 180”. In the leglift the rotation of lower and upper leg is (apart from the effect on the knee joint) reflected in an ankle plantar flexion and hip extension. The movement range in the LS-Sl joint is during the leglift smaller than during the backlift. The net moment curves in the knee (Fig. 5) show the expected difference between the leglift and backlift; in the backlift a considerable knee flexing moment is required during the first 50% of the lift, whereas in the beginning of the leglift a relatively small knee extending moment is required that oscillates around zero during the major part of the movement. It is interesting to note that both for the hip and L5-Sl joint no difference in net peak moment was found between the two lifting techniques. This is in line with Anderson et al. (1976) who found the intradiscal pressure more closely related to the distance between the load and the body than to the lifting technique used.
taken into account (Cavanagh and Komi, 1979; Olney
and Winter, 1985; Vos et al., 1990, 1991). The two lifting cycles incorporated in the analysis were divided into four parts; two lifting and two lowering movements. Only the lifting phases were analysed. The time period of these individual curves were normalized to 100% of the mean movement time (Winter, 1983). From these curves the mean curves per subject, per technique, and finally for the total group ($- S.E.M., standard error of the means) were obtained. For the sake of completeness, data regarding the L5-Sl joint were also presented, although they are not incorporated into the discussion. RESULTS Kinematics and kinetics
The differences with respect to kinetics and kinematics between the leglift and backlift become apparent from the joint angle and net joint moment curves as presented in Figs 4 and 5. Figure 4 illustrates the difference in movement pattern between leglift and backlift. In the backlift the knee remains extended (about 170”), whereas in the leglift the knees extend from loo” to about 170”. In the backlift the legs are tilted slightly backwards to enable the subject to keep the combined body + load centre of gravity above his feet. During the lift this backward tilt is slowly reduced resulting in a slight dorsal flexion of the ankle. The major part of the movement during the backlift occurs (according to the model) in the L5-Sl joint which is
Ground reaction force
To answer the question to what extent the net joint moments were determined by the magnitude of the ground reaction force and the distance between its line of action and the location of the centre of rotation of the joint, the net moments computed by the linkedsegment model were compared to the outcome of this direct calculation. The results are presented in Figs 6 (backlift) and 7 (leglift). In the backlift a high degree of similarity is obtained (Pearson’s coefficient of correlation exceeded 0.998 for all joints), although in the hip
ioint angle in ankle
1
3 50
iolnt am&:inknee , I::i++T::;T. I-AI-* _=160Z’ X” I’ 140I / .’ 120P 100 __*_/-’ 180
h
0
800A
100
ioint anRle in hip
50
100
timein%
time in 56 200,
1283
ioint an&e in L5-S 1
1
I
180h
I % 50 timein% Mean joint angle curves for baclclift(solid line) and leglift (dashed line). Standard error of the mean is indicated.
1284
H. M. TOUSSAINT et al. moment (lsm) in knee
moment (lsm) in ankle
E - -lOO-
50 time in %
100
0
50
100
time in %
50 time in % Fig. 5. Mean net joint moment curves for backlift (solid line) and leglift (dashed line). The net moments are calculated by means of a linked-segment model (lsm). Standard error of the mean is indicated.
time in %
timein%
hip
time in %
time in %
Fig. 6. Comparison of net joint moments as calculated by linked-segment model (lsm, solid line) to net joint moments as calculated by the magnitude of the ground reaction force and the distance between its line of action and the location of the centre of rotation of the joint (direct, dashed line). Results are presented for the bacldift.
1285
Coordination in lifting and L5-Sl joint an increasing effect of the ‘ignored weight of the segments below the joint can be observed. Note that for the sake of clarity all extending moments are presented positive irrespective of their absolute direction. In the leglift (Fig. 7) the moments calculated by the direct method deviate more from the net moments as calculated by the linked-segment model (Pearson’s R still better than 0.995). This can be explained from the fact that in the starting position of the leglift the centres of mass horizontal positions of the legs and pelvis with respect to the joint centres of rotation is relatively large. The effect of this increased moment arm on the net joint moments is ignored in the direct calculation method leading to an overestimation of the net joint moment. Taking this into account it can be concluded that the global patterns of the net joint moments can to a large extent be explained on the basis of the ground reaction force vector relative to the position of the joint centres of rotation. Electromyography
The mean EMG signals of the muscles were expressed as a percentage of the activity levels measured during maximal standard isometric contractions. The activity of muscles having a joint extending effect are presented as positive, whereas activity levels of flexing muscles are presented as negative. The signals are presented in relation to percentage of movement time and to the net moment in knee and or hip joint (Figs 8
50
100
and 9). In the backlift the kneW&t%gmuscles (gastrocnemius, biceps femoris atid s&t&dinosus) reach activity levels up to 100% SIC, whereas the knee extending vastus medialis and rectus femoris show only low levels of activity. This corresponds to a considerable flexing moment in the knee joint during the backlift (see Fig. 5). During the leglift a considerable level of activity is recorded in both the knee extensors and knee flexors. This co-contraction of socalled agonists and antagonists becomes especially apparent in the graphs in which the EMG-levels are presented as a function of net knee moment. At roughly the same knee moment (especially at about 30 Nm, beginning of the li!?) both groups of muscle demonstrate a range of EMG amplitudes indicating an interaction of activity of both groups of muscle such that the required net knee moment is produced. Both in backlift and leglift a low I=1 of activity of the rectus femoris was found. The activity of muscles around the hip joint (Fig. 9) demonstrates cooperation of synergists. Note that in the leglift the activity of the biartictrlar semitendinosus and biceps femoris is higher than in the backlift. The activity level of the gluteus muscleis about the same in both the conditions. Work
The work produced in the ankle, knee, hip and L5-Sl joint is presented in Fig. 10. In both lifting techniques the major part of the work is produced in
0
time in % hip
0
50
time in %
50
100
time in % I
I 100
300,
OL
0
L5-Sl
I
I 50
100
timein%
Fig. 7. Comparison of net joint moments as calculated by linked-segment model (lsm,solid&@ to net joint moments as calculated by the magnitude of the ground reaction force and the distaua.hdtipecrtita Iine of action and the location of the centre of rotation of the joint (direct, dashed line). Results ,far the . i ?r leglift. .
1286
H. M. TOUSSAWTet al. backlift
50 __-.---
baclclift
50
. -.- _
- -.-.- -.-._ _ _.__.__
VaS --,_--._____,______
’ - .._
0
_
ILL! J
I
I’
._‘,;_---’
:’
.’
_LL---
__-
,”
.”
-150
-200 time
_._
,..
.*‘___,,,
%movement
-.-._
.“,/
At
.l(-J(J
-‘-
*I’
I--
-50
--
-50
-100
0
knee moment (Nm)
leglift _ _,_._..-.-.-._. ; , - _.- ’ .F ., I
50
O‘. . * .\I \ \ _--.
‘\
0
50 %movement
LJ 0
-50 -20
100
‘.,, .____--___-- L-
time
40
20
knee moment (Nm)
Fig. 8. Activity of the vastus medialis (Vas), rectus femoris (Ref), gastrocnemius (Gas), biceps femoris (Bif) and semitendinosus muscle (Sem) presented as percentage of the standard isometric contraction (o/SIC). Data are presented dependent on percentage movement time (left panels) and knee moment (right panels) for both the backlift (upper panels) and leglift (lower panels).
backlift
100
u
V
50
5;
5; @
t3 0 1 -50: 0
50 %movement
hip moment (Nm)
leglift
leglift
60, 40
5;
20
4 50 %movement
time
-20
.. 0
)
_------
:
I’
#’
________- * - - - ___-.-_
--_ I 100
_____
_I’:
E8 0
I/“:
,,‘.._
,,-
v
-201 0
200
100
100 time
100
200
300
hip moment (Nm)
Fig. 9. Activity of the gluteus maximus (Glu), rectus femoris (Ref), biceps femoris (Bif) and semitendinosus muscle (Sem) presented as percentage of the standard isometric contraction (o/SIC). Data are presented dependent on percentage movement time (left panels) and hip moment (right panels) for both the backlift (upper panels) and leglift (lower panels).
1287
Coordination in lifting
30 20 h 3
lo-
Work in ankle _______-------_,/ I’ ,/’ ,’ I’,’ I’
Work in knee _*----______ _.= 5
50
50
time in %
time in % Work in L5-S 1
Work in hip ,/
,’
__*----------
II’
5
-50; 0 time in %
50
100
time in %
Fig. 10. Work (J) produced in ankle, knee, hip and L5-Sl joint during the backlift (solid line) and leglift
(dashed line).
the hip and in LS-Sl. Especially the low work values in the knee are noteworthy. When comparing the backlift to the leglift it is interesting that during execution of the latter technique work deliverance in the hip is about 50% higher. DISCUS!SION
In this study the leglift was compared to the backlift first to address the question whether net joint moments are determined by the difference in the position of the centre of rotation of the knee joint with respect to the line of action of the ground reaction force (Figs 2 and 3). The effect of the variation in joint position on the net knee moment is illustrated in Fig. 5, where the net knee moment during the leglift and backlift is given with respect to prcentage of movement time. In the backlift a considerable knee flexing moment is found, whereas in the leglift the net knee moment starts at about 30 Nm having a knee extending effect, and slowly reduces to values oscillating around zero. When using the method of analysis employed by Schipplein et al. (1990) each joint is analysed separately in which the net joint moments are coupled directly to required muscle forces. For the present study this approach would lead to the conclusion that the low values found for the net knee moment during the execution of the leglift implies a reduction in knee loading. This may be a surprising result, since it has been advocated to use the leglift to reduce the compressive force on the low back and to shift the stresses
on the body from the low back to the legs (Bendix and Eid, 1983; Miller, 1980). Apparently, the shift of stresses from the back to the knee cannot be deduced from the net moments. The question arises what interpretation can be given to the net moments. As pointed out in the introduction, these net moments may be approximated directly from the magnitude and direction of the external force (Ingen Schenau, 1989, 1990), ignoring the influence of gravity and inertial effects. The results of the direct calculation of the net knee moment are presented together with the actually obtained values for the net knee moment by means of inverse dynamics (linked-segment model, lsm) in Figs 6 and 7. The virtually identical curves in the backlift illustrate that in the rather slow lifting movement a close relation exists between the net joint moments and the direction and magnitude of the ground reaction force, and consequently, that the effects due to angular acceleration play a minor role in determining the magnitude of the net joint moments. The differences that occur between the outcome of the linked-segment model and the direct method in the leglift can be attributed to the moment effect of the mass of the segment(s) below the joint for which the net moment is calculated. In the direct calculation method this effect is ignored, resulting in an underestimation of the net moment in the knee joint (centre of mass of feet and lower leg is positioned posterior of the knee joint, resulting in an knee extending effect), while in the hip and LS-Sf joint the combined centre of mass of the segments below the
1288
H. M.
TOUSSAINT et al.
observed joint are positioned anterior to the joint. However, the differences are slight and it can be concluded that. a’ clase relation exists between the moment caused. by the ground reaction force and the net joint moments. Apparently, a process of coordination endorses a distribution of net joint moments across joints, such that the ground reaction force points in a direction that provides balance during the movement. This conclusion raises the question to the intermuscular coordination of the leg muscles during lifting. The fact that the net knee moment oscillates around zero during the leglift is a consequence of the necessity to keep balance during the movement. However, this seems rather inefficient because the quadriceps muscles are in aprfect position to deliver useful work. The activation’ patterns of the monoarticular knee extensor (vastus mddialis) and biarticular knee flexors and extensors (biceps femoris, semitendinosus, and gastrocnemius) shed some light on the solution of this apparent contradiction. The vastus medialis appeared to be rather active (about 50% SIC) and furthermore, coactivated with its antagonistically acting semitendinosus, biceps femoris and gastrocnemius as indicated in Fig. 7: The quadriceps is active and shortens, thus producing useful work. The undesired ‘side effect’ which an increased knee moment would have on the direction of the ground reaction force is compensated by the increased activity of the knee flexors (gastroenemius &nd hamstrings). The latter biarticular must& .group lias a hip extending effect also, hence, the activation of the quadriceps pulls the hamstrings in order to facilitate hip extension. This socalled ligamentous action of biarticular muscles (see Introduction se&on.and, for an extensive description, Ingen Schenau, 1990) facilitates transport of work delivered by the quadriceps to the hip joint. From Fig. 10 it can be seen that the work produced in the hip joint during the leglift is considerably higher than the value obtained durhig thetbacklift. Furthermore, the direction and magnitude of the ground reaction force has to be such that the subject executing the lift keeps his balance during the movement. This coupling between the net joint moments and the control of an external force in leg extensions was described by Jacobs and Ingen Schenau (in pressb). They demonstrated that biarticular muscles (hamstrings and rectusfemoris) play an important role in the distribution of the net moment over the joint crossed in a static lcgLejrtegsion task in which the subject was asked to deliver a specific external force. In their experiments high correlation coefficients (0.96) were found between the difference in activity of the rectus femoris and the hamstrings and the difference in net moments of the&ip and knee joints. In the present study a significant correlation between the activity difference (RF-HAM) and the net joint moment difference (Mtnec - M,,& was found as well, Pearson’s R being 0.72 and 0.83 for the leg and backlift, respectively. The shortening of the muscles during the move-
ment has a disturbing effect on the EMG-force relation (Perry and Bekey, 1981). This can explain the higher correlation coefficient in the backlift in which the shortening of the leg muscles is likely to be much smaller and the relatively low correlation when compared to the results obtained by Jacobs and Ingen Schenau (in pressb). Correlation coefficients approaching these values were found in the present study between the activity level of the gastrocnemius and the net joint moment difference between knee and ankle, Pearson’s R being-O.89 and 0.96 for the leglift and backlift, respectively. Thus, although the coefficients of correlation are not as high as in static exercise, the conclusion seems warranted that during lifting tasks the intermuscular coordination is aimed at coupling of joint moments,‘such that the ground reaction force points in a direction that provides balance during the movement. Acknowledgements-The authors gratefully acknowledge the cooperation of our students Jos Twisk, Ad van de Ven and Carien Thissen in the process of data collection. Idsart Kingma provided us with part of the software to analyse the data. and reviewed the manuscript. We thank Han Kemper and Gerrit-Jan van Ingen Schenau for guidance and advice, for critically reviewing the manuscript, and for providing facilities to pursue,this project.
REFERENCES
Andersson, G. B. J. (1981) Epidemiological aspects on low back pain in industry. Spine 6, 53-59. Andcrsson, G, B. J., Ortengren, R. and Nachemson, A. (1976) Quantitative studies of back loads in lifting. Spine 1, 178-185. Basmajian, J. V. (1978) Muscles Alive; Their Functions Reoealed by Electromyogrophy. Williams and Wilkins, Baltimore. Bejjani, F. J., Gross, C. M. and Pugh, J. W. (1984) Model for static lifting: relationship of loads on the spine and the knee. J. Biomechanics 17,281-286. Bendix,T. and Eid, S. E. (1983)The distance between the load and the body with three bi-manual lifting techniques. Appl. Ergonomics 14, 185-192. Cavanagh, P. R. and Komi, P. V. (1979) Electromechanical delay in human skeletal muscle under concentric and eccentric contractions. Eur. .I. appl. Physiol. 42, 159-163. Cleland, J. (1867) On the actions of muscles passing over more than one joint. J. Anat. Physiol. 1, 85-93. Dempster, W. T. (1955) Space Requirements of the Seated Operator. Wright Patterson Air Force Base, Dayton, OH. Ekholm, J., Arborelius, U. P. and Nemeth, G. (1982) The load on the lumbo-sacral joint and trunk muscle activity during lifting. Ergonomics 25, 145-161. Elftman, H. (1939) Forces and energy changes in the leg during walking. Am. J. Physiol. 125, 339-356. Freivalds, A., Chaffin, D. B., Garg, A. and Lee, K. S. (1984) A dynamic biomechanical evaluation of lifting maximum acceptable loads. J. Biomechanics 17, 251-262. Gagnon, M. and Smyth, G. (1991) Muscular mechanical energy expenditure as a process for detecting potential risks in manual materials handling. J. Biomechanics 24, 191-203. Garg, A., Chaffin, D. B. and Freivalds, A. (1982) Biomechanical stresses from manual load lifting: a static vs dynamic evaluation. IIE Transactions 14, 272-281.
Coordination in lifting Gregoire, L., Veeger, H. E., Huijing, P. A. and Ingen Schenau, G. J. van (1984) The role of mono- and biarticular muscles in explosive movements. Int. J. Sports Med. 5, 301-305. Grieve, D. W. (1974) Dynamic characteristics of man during crouch- and stoop lifting. In Eiomechanics IV (Edited by Nelson R. C. and Morehouse C. A.), pp. 19-29. University Park Press, New York. Hall, S. J. (1985) Effect of attempted lifting speed on forces and torque exerted on the lumbar spine. Med. Sci. Sports Exert. 17, 44@444. Ingen Schenau, G. J. van (1989) From rotation to translation: Constraints on multi-joint movements and the unique action of biarticular muscles. Hum. Mvmt Sci. 8. 301-337. Ingen Schenau, G. J. van (1990) On the action of bi-articular muscles. A review. Neth. J. 2001. 40, 521-540. Ingen Schenau, G. J. van, Boots, P. J. M., Groot, G. de, Snackers, R. J. and Woenzel, W. W. L. van (1992) The constrained control of force and position in multi-joint movements. Neuroscience 46, 197-207. Jacobs, R. and Ingen Schenau, G. J. van. Intermuscular coordination in sprint running. J. Biomechanics (in pressa). Jacobs, R. and Ingen Schenau, G. J. van. Control of an external force in leg extensions. J. Physiol. (in pressb). Kumar, S. and Davis, P. R. (1983) Spinal loading in static and dynamic postures: EMG and intraabdominal pressure study. Ergonomics 26,913-922. Lees, A. (1980) An optimized film analysis method based on finite difference techniques. J. Hum. Mumt Stud. 6, 165-180. Leskinen, T. P. J. (1985) Comparison of static and dynamic biomechanical models. Ergonomics 26, 595-604. Leskinen, T. P. J., Stalhammar, H. R., Kuorinka, I. A. A. and Troup, J. G. D. (1983a) A dynamic analysis of spinal compression with different lifting techniques. Ergonomics 26, 595-604.
Leskinen, T. P. J., Stllhammar, H. R., Kuorinka, I. A. A. and Troup, J. G. D. (1983b) The effect of inertial factors on spinal stress when lifting. Engng Med. 12, 87-89. Looze, M. P. D., Kingma, I., Bussmann, J. B. J. and Toussaint, H. M. Validation of a dynamic linked segment model to calculate joint moments in lifting. Clin. Biomechanics (in press). Martin, J. B. and Chaffin, D. B. (1972) Biomechanical computerized simulation of human strength in sagittal plane activities. AIIE Transactions 4, 19-28.
1289
McGill, S. M. and Norman, R. W. (1986) Partitioning of the LA-L5 dynamic moment into disc, ligamentous, and muscular components during lifting. Spine 7, 666-678. Miller, R. L. (1980) When you lift, bend your knees. Occup. Health Safety 45, 46-47.
Olney, S. J. and Winter, D. A. (1985) Predictions of knee and ankle moments of force in walking from EMG and kinematic data. J. Biomechanics 18, 9-20. Perry, J. and Bekey, G. A. (1981) EMG-force reltionships in skeletal muscle. CRC Crit. Rev. Biomed. Engng 4, l-22. Schipplein, 0. D., Trafimow, J. H., Andersson, G. B. J. and A&acchi, T. P. (1990) Relationship between moments at the L5/Sl level, hip and knee joint when lifting. J. Biomechanics 23, 907-912.
Schultz, A. B. and Andersson, G. B. J. (1981) Analysis of loads on the lumbar spine. Spine 6, 76-82. Schultz, A. B., Haderspeck, K., Warwick, D. and Pontillo, D. (1983) Use of lumbar trunk muscles in isometric performance of mechanically complex standing tasks. J. Orthop. Res. 1, 77-91.
Troup, J. D. G. (1965) Relation of lumbar spine disorders to he&y manual work and lifting. Lancet i, 857-861. Trout. J. D. G. (1977) Dvnamic factors in the analvsis of s&p and croich IGtini methods: a methodologiial approach to the development of safe materials handling standards. Orthop. Clin. N. Am. 8,201-209. Troup, J. D. G., Leskinen, T. P. J., StBlhammar, H. R. and Kuorinka, I. A. A. (1983) A comparison ofintra-abdominal pressure increases, hip torque, and lumbar intervertebral compression in different lifting techniques. Hum. Factors 25,517-525.
Vos, E. J., Harlaar, J. and Ingen Schenau, G. J. van (1991) Elcctro-mechanical delay during knee extensor contractions. Med. Sci. Sports Exert. 23, 1187-1193. Vos, E. J., Mullender, M. G. and Ingen Schenau, G. J. van (1990) Electra-mechanical delay in the vastus lateralis muscles during dynamic contractions. Eur. J. appl. Physiol. 60,467471.
Winter, D. A. (1979) Biomechanics of Human Movement. Wiley, New York. Winter, D. A. (1983) Moments of force. and mechanical power in jogging. J. Biomechanics 16, 91-97.
COZl-9290/92
1992.
COORDINATION Huue
IS.oO+.OO
Pergamon Press Ltd
Printed in Great Britain
OF THE LEG MUSCLES IN BACKLIFT AND LEGLIFT
M. TOWSAINT,*CORINNEE. VAN BAAR,PAULP. VAN LANGEN, MICHIELP. DE LOOZEand JAAPH. VAN DIE~N
Department of HealthScience, Faculty of Human Movement Sciences, Vrije Universiteit and Universiteit van Amsterdam, Amsterdam, The Netherlands Ahstrae-Net joint moments are often used to quantify the loading of structures (e.g. the intervertebral disc at LSSl) during lifting, This quantification method is also used to evaluate the loading of the knee, for instance, to determine the effect of backlifting as opposed to leglifting. However, the true loading of the joint as derived from net joint moments can be obscured by a possible co-contraction of antagonists. To unravel
the mechanisms that determine the net joint moments in the knee, the leglift was compared to the backlift. Although a completely different net knee moment curve was found when comparing the two lifting techniques, it appeared to be closely related to the ground reaction force vector and its orientation with respect to the joint centre of rotation (R>0.995). This close relation was established by co-contraction of both flexors and extensors of the knee. Furthermore, a close relation appeared to exist between the joint moment difference between hip and knee and the activity difference between rectus femoris muscle and hamstring (R = 0.72 and 0.83 in leglift and backlift, respectively). The knee-ankle joint moment difference and the gctivity of the gastrocnemius showed a close relation as well (R= -0.89 and 0.96 in Ieglift and backlift, respectively). These relations can be interpreted as a mechanism to distribute net moments across joints. It is concluded that during lifting tasks the intermuscular coordination is aimed at coupling of joint moments, such that the ground reaction force points in a direction that provides balance during the movement. The use of net joint moments as direct indicators for joint loading (e.g. knee) seems, therefore, questionable.
INTRODUCTION Manual materials handling and especially lifting can be hazardous and has often been associated with the occurrence of low-back pain (Andersson, 1981; Troup, 1965). It is therefore not surprising that most studies of lifting reported in the literature have been concerned mainly with the trunk. The effect of lifting on the spine has been evaluated using different techniques and methods. Especially, studies in which biomechanical models were used to calculate moments and forces acting on the lower back do abound (Ekholm et al., 1982; Freivalds et al., 1984; Gagnon and Smyth, 1991; Garg et al., 1982; Grieve, 1974; Hall, 1985; Kumar and Davis, 1983; Leskinen, 1985; Leskinen et al., 1983a, b; Martin and Chaffin, 1972; McGill and Norman, 1986; Schipplein et al., 1990; Schultz and Andersson, 1981; Schultz et al., 1983; Troup, 1977; Troup et al., 1983). However, the function of the legs in the lifting process has received scant attention. The two extremes of lifting technique frequently discussed in the literature are leglifting, where the knees are flexed and the trunk is vertical, and backlifting, where the knees are straight and the trunk is flexed. It has been widely advocated that all lifting should be carried out with a leglifting technique
Received infinalform 10 February 1992. * Author to whom correspondence should be addressed at: Department of Health Sciences, Faculty of Human Movement Sciences, van der Boechorststraat 9, 1081 BT Amsterdam, The Netherlands.
(Bendix and Eid, 1983; Miller, 1980). The presumed basis for this has been to reduce the compressive force on the low back and to shift the stresses on the body from the low back to the legs. The latter assumption has been substantiated by Bejiani et al. (1984), who developed a model relating the loads at the knee to the relative participation of the back in lifting. The results of their study indicate a strong inverse relation between joint reaction forces occurring in back and knee. This suggests that it is indeed possible to distribute the load during lifting over the knee and back. The question then arises to what extent this load sharing can be realised. This question was examined by Schipplein et al. (1990) who studied the load sharing during leglifting. They observed that when handling increasing weights using the leglifting technique, the net knee moment remained constant, whereas the moment in the back increased. Their interpretation of this observation was that the quadriceps muscle strength limits the subjects’ ability to lift with their knees flexed. However, this conclusion was based on the assumption that the net knee moment is determined by the activity of the quadriceps muscle only, ignoring the possible effect of other muscles that might influence the knee moment. Recently an alternative hypothesis concerning the regulation of the net joint moments was proposed (Ingen Schenau, 1989, 1990). The often observed cocontractions of monoarticular and biarticular muscles around joints (e.g. the knee) was placed in the perspective of the control of the direction of an external force (during lifting; i.e. the ground reaction force). It
1279
H. M. TOUSSAINTet al.
1280
was stressed that the relation between net joint moments and external forces is causal. In slow and static movements and ignoring the influence of gravity and inertial effects one can approximate these moments directly from the magnitude and direction of the external force. For an example presented in Fig. 1 this can be performed by multiplying the smallest distance between the line of action and the location of the centre of rotation of the joint and the magnitude of the ground reaction force. For the knee and hip joint this leads to M ,_
= d * F,, and M,i, = g - F,,.
(1)
It was argued that biarticular muscles can adjust the distribution of net moments over joints that are crossed. Activation of the rectus femoris muscle, for example, will cause an increase of the net knee moment and a decrease of the net moment in the hip, while hamstring activity causes a shift of net moments in opposite direction. Based on an inverse dynamical analysis of cycling it was demonstrated that the net joint moments necessary to control a certain direction of the external force are largely independent from the required joint displacements associated with the task (Ingen Schenau, 1989; Ingen Schenau et al., 1992). This apparent contradiction can be solved by coactivation of monoarticular and biarticular muscles. The biarticular muscles are able to change the distribution of net moments over the joints that are crossed. Furthermore, biarticular muscles can act as a stiff tendon. Extension of the knee by the knee extensors leads ,to displacement of the origin of the gastrocnemius. This will induce a pull on the insertion of this muscle when the muscle is not lengthened. This means
that the extension of the knee will also cause a plantar flexion by means of the gastrocnemius. This was defined by Cleland (1867) as a ‘ligamentous action’(see Ingen Schenau, 1990). The work done in the knee extensors can (partially) be used for plantar flexion. Hence, using the ligamentous action of biarticular muscles one can transport energy. Combined with the previous described capacity to distribute net moments, biarticular muscles allow the monoarticular muscles to contribute to positive power irrespective of the net moments required in the joints crossed by these latter muscles (Ingen Schenau, 1990). The question arises whether an analysis of lifting in the perspective of the control of the direction of an external force would enhance the understanding of the function of the legs during lifting. In this approach the relation between net joint moments and activity of synergistic muscles surrounding the joint as employed by Schipplein et al. (1990) is extended. It is taken into account that during the lifting movement the direction of the ground reaction force should be such that no net rotation of the subject leading to balance loss occurs. In this sense the lifting movement is performed within the constraint of balance (i.e. the subject does not fall forward or backward). Furthermore, if the latter hypothesis (net joint moments explained by an external force) is valid it is interesting to raise the question whether during lifting tasks the intermuscular coordination can be understood on the basis of the described cooperation between monoarticular and biarticular muscles. The characteristic difference between the two extremes of lifting technique, leglifting and backlifting, is the position of the centre of rotation of the knee joint with respect to line of action of the ground reaction force as illustrated in Fig. 2. This will induce a rather high compensatory net knee flexion moment in the
M knee
Fig. 2. Direct estimation of net joint moment based on the ground reaction force and its moment arm with respect to the Fig. 1. A quasi-static analysis reveals that the ground reaction force requires a net knee and hip extending moment to ensure equilibrium.
centre of rotation in the knee. In the backlift (left panel) the effect of the ground reaction force with respect to the knee must be compensated by a knee flexing moment, in the leglift (right panel) a knee extending moment is required to ensure equilibrium.
Coordination in lifting backlift and in the leglift a net moment fluctuating around zero. The aim of the present study is to compare the leglift to the backlift to find out whether the net joint moments are determined by the required direction of the ground reaction force vector. If this would appear to be true then the intermuscular coordination leading to the required distribution of net joint moments will be investigated. Especially the activation patterns of the monoarticular knee extensor (vastus medialis muscle) and biarticular knee flexors and extensors (hamstrings, rectus femoris muscle) in relation to the net moment are of interest.
1281
was 20% of the body height and the vertical distance between the load and the floor was 10% of body height. For the leglift the extent to which the buttocks had to be lowered was indicated too using a wire frame. This lowering depth, representing the minimum distance between the greater trochantor and the floor was 66% of the leg length, quantified as the distance between the greater trochanter and the lateral malleolus when standing upright. Practice lifts were encouraged until the subjects were comfortable with the technique to be executed in the required frequency. Anthropometry
METHODS
Subjects
Eight healthy male students (age 22.9k2.4 yr, height 1.80 + 0.06 m, body mass 70.5 $- 4.5 kg) participated in the experiment. All subjects signed informed consent prior to the experiment. None of the subjects reported a history of low-back disorders. Test protocol
Subjects were asked to lift and lower a barbell (mass 15.3 kg) in two sessions, using different lifting techniques. Each session consisted of seven lifting/ lowering cycles. The duration of the lifting cycle remained fixed and was partitioned in four phases of 1 s each: standing in a standardized starting position in which the barbell was to be maintained in a standardized position also (dependent on the lifting technique, see below), lifting of the barbell, standing upright, and finally lowering the barbell. The subjects were aided in keeping a constant lifting pace by a metronome. During the last two cycles subjects were filmed while ground reaction forces and muscle activity were recorded. Between the sessions subjects were allowed to rest for at least 5 min, during which the lifting technique of the next session was explained to them. Lifting techniques
Two lifting techniques (back and leg) were prescribed. The backlift was performed with extended knees. During the leglift the subjects were instructed to keep the spine as erect as possible. A stick figure representing the sequence of the two lifting techniques is presented in Fig. 3. The lifts were performed as sagittally symmetric as possible. Care was taken to further standardize the execution of the lifting movements in order to reduce inter subject variability. A framework to which tiny wires were attached was used to indicate the location from which the load was to be lifted in the two lifting conditions. This location remained the same for the two lifting conditions and depended on the anthropometry of the subject: the horizontal distance between the load and the distal end of the fifth metatarsal
Markers (diameter 0.5 cm) were placed on the skin to indicate the location of the fifth metatarsophalangeal joint, the ankle joint (the distal part of the iateral malleolus), the knee joint (lateral collateral ligament at the height of the joint cleft), the greater trochanter, L5-Sl (according to Looze et al., in press), and the spinous process of the first thoracic vertebra. The coordinates defined five body segments: the foot, lower leg, upper leg, pelvis and upper trunk/head/ arms/load. Anthropometric data (standing height, total body mass, length of segments) were measured. Positions of segmental centres of gravity and moments of inertia were estimated on the basis of tables-devised by Dempster (1955) and revised by Winter (1979)-and the measured segmental lengths and body weights of the subjects. Kinematics and kinetics
During the sessions the subjects were filmed at 50 frames per second using a 16 mm high-speed film camera (Teledyne type DBM 55) piaced on the coronal axis of the body at a distance of 6.5 m, so that movements in the sagittal plane could be recorded. The coordinates of the anatomical landmarks, of three markers placed on the force plate (used as a point of reference), and of two markers with a known intermarker distance were determined for each tine frame using a motion analyser (Reinka). Absolute coordinates, after proper scaling, were low-pass filtered by a digital filter (zero-phase lag, bidirectional application of a fifth approximation, second-order Butterworth filter with an effective cut-off frequency of 5 Hz). The angles of each segment were calculated relative to the horizontak joint angles were defined as the angles between adjacent body segments. Numerical differentiation of the time histories of the angles and the positions of the segmental centres of gravity (Lanczos 5-point differentiation filter; Lees, 1980) yielded (angular) velocities and accelerations. Vertical (F,) and foreaft (F,,) components of the ground reaction forces were recorded by means of a force platform (Kistler, type 9281B). The analog force signals were amplified, low-pass filtered (30 Hz, fourth order at 24 dB/oct), sampled (100 Hz, 12 bits) and
1282
H. M. TOUSSAINT~~ al.
Backlift
Fig. 3. Stick figures representing a sequence.of the two lifting techniques employed in the present study. Top row represents the ‘leglift’,bottom row represents the ‘backlift’.The ground reaction force is indicated. For more details the reader is referred to the text.
stored on disk. From the distribution of the force. components, the centre of pressure of the force vector was calculated. The synchronization of the film and force data was achieved by means of electrical pulses. Biomechanical model
Net moments and net forces were calculated by means of an inverse dynamical analysis (Elftman, 1939) using a dynamical two-dimensional linked-segment model as described by Looze et al. (in press). The model comprises eight segments (feet, lower legs, upper legs, pelvis, upper trunk/head, upper arms, forearms and hands/load) which were assumed to be rigid and connected to each other by intersegmental joints Moments having a knee, hip, and L5-Sl extending influence were defined positive as well as moments having an ankle plantar flexing effect. In this study the results of the first five segments were analysed. Additionally, using the direct calculation method the moments based on the ground reaction force vector and the position of the joints were calculated. Electromyography
The electromyographic activities (EMG) of six muscles [gastrocnemius mediali, vastus medialis, rectus femoris, semitendinosus, biceps femoris (caput longum) and gluteus maximus] were recorded by means of bipolar direct measurements @ISA 15COl). Pairs of Ag/AgCl surface electrodes (Sentry Medical Products, CA, lead-off area 1 cm2, centre to centre electrode distance 4.5 cm) were applied to the muscles after standard skin preparation (Basmajian, 1978). The positions of the electrodes are described by Gregoire et al. (1984). The EMG signals were stored on tape (TEAC SR70) and plotted on a writer (Gould ES 1OOQ)
for visual inspection during the experiment. The electrical signal was filtered (band pass 20-200 Hz), sampled (400 Hz) and stored on disk. To obtain a linear envelope the signals were rectifield and low-pass filtered (zero-phase lag, bidirectional application of a fifth approximation, second-order Butterworth filter with an effective cut-off frequency of 5 Hz). The subjects performed three standard isometric contractions for each muscle group under five standardized tasks, while EMG was recorded (see Gregoire et al., 1984; Jacobs and Ingen Schenau, in press): -for gluteus maximus: maximal hip extension against external resistance (hip joint W’), -for biceps femoris (caput longum) and semitendinosus: maximal knee flexion against external -resistance (knee joint 9W), -for rectus femoris and vasti: maximal knee extension against external resistance (knee joint 90”), -for gastrocnemius: maximal plantar flexion standing on one leg, -for erector spinae: maximal extension against external resistance (L5-Sl joint 135”). Contractions lasted 3 s. The standard isometric contraction level for each muscle was determined by taking the mean of the rectified EMG of 1 s in which the signal remained constant. The EMGs obtained during the lifting experiments were normalized to 100% standard isometric contraction level (SIC). Treatment of data
Film, force and EMG data were synchronized on the basis of pulses generated by the experimentator which were marked on the film and recorded simultaneously with the force and EMG signals. For the interpretation of the EMG data, a phase lag of 90 ms between the change in EMG and change in mechanical output, electromechanical delay (EMD), was
Coordination in Wing
extended from about 1lo” to about 180”. In the leglift the rotation of lower and upper leg is (apart from the effect on the knee joint) reflected in an ankle plantar flexion and hip extension. The movement range in the LS-Sl joint is during the leglift smaller than during the backlift. The net moment curves in the knee (Fig. 5) show the expected difference between the leglift and backlift; in the backlift a considerable knee flexing moment is required during the first 50% of the lift, whereas in the beginning of the leglift a relatively small knee extending moment is required that oscillates around zero during the major part of the movement. It is interesting to note that both for the hip and L5-Sl joint no difference in net peak moment was found between the two lifting techniques. This is in line with Anderson et al. (1976) who found the intradiscal pressure more closely related to the distance between the load and the body than to the lifting technique used.
taken into account (Cavanagh and Komi, 1979; Olney
and Winter, 1985; Vos et al., 1990, 1991). The two lifting cycles incorporated in the analysis were divided into four parts; two lifting and two lowering movements. Only the lifting phases were analysed. The time period of these individual curves were normalized to 100% of the mean movement time (Winter, 1983). From these curves the mean curves per subject, per technique, and finally for the total group ($- S.E.M., standard error of the means) were obtained. For the sake of completeness, data regarding the L5-Sl joint were also presented, although they are not incorporated into the discussion. RESULTS Kinematics and kinetics
The differences with respect to kinetics and kinematics between the leglift and backlift become apparent from the joint angle and net joint moment curves as presented in Figs 4 and 5. Figure 4 illustrates the difference in movement pattern between leglift and backlift. In the backlift the knee remains extended (about 170”), whereas in the leglift the knees extend from loo” to about 170”. In the backlift the legs are tilted slightly backwards to enable the subject to keep the combined body + load centre of gravity above his feet. During the lift this backward tilt is slowly reduced resulting in a slight dorsal flexion of the ankle. The major part of the movement during the backlift occurs (according to the model) in the L5-Sl joint which is
Ground reaction force
To answer the question to what extent the net joint moments were determined by the magnitude of the ground reaction force and the distance between its line of action and the location of the centre of rotation of the joint, the net moments computed by the linkedsegment model were compared to the outcome of this direct calculation. The results are presented in Figs 6 (backlift) and 7 (leglift). In the backlift a high degree of similarity is obtained (Pearson’s coefficient of correlation exceeded 0.998 for all joints), although in the hip
ioint angle in ankle
1
3 50
iolnt am&:inknee , I::i++T::;T. I-AI-* _=160Z’ X” I’ 140I / .’ 120P 100 __*_/-’ 180
h
0
800A
100
ioint anRle in hip
50
100
timein%
time in 56 200,
1283
ioint an&e in L5-S 1
1
I
180h
I % 50 timein% Mean joint angle curves for baclclift(solid line) and leglift (dashed line). Standard error of the mean is indicated.
1284
H. M. TOUSSAINT et al. moment (lsm) in knee
moment (lsm) in ankle
E - -lOO-
50 time in %
100
0
50
100
time in %
50 time in % Fig. 5. Mean net joint moment curves for backlift (solid line) and leglift (dashed line). The net moments are calculated by means of a linked-segment model (lsm). Standard error of the mean is indicated.
time in %
timein%
hip
time in %
time in %
Fig. 6. Comparison of net joint moments as calculated by linked-segment model (lsm, solid line) to net joint moments as calculated by the magnitude of the ground reaction force and the distance between its line of action and the location of the centre of rotation of the joint (direct, dashed line). Results are presented for the bacldift.
1285
Coordination in lifting and L5-Sl joint an increasing effect of the ‘ignored weight of the segments below the joint can be observed. Note that for the sake of clarity all extending moments are presented positive irrespective of their absolute direction. In the leglift (Fig. 7) the moments calculated by the direct method deviate more from the net moments as calculated by the linked-segment model (Pearson’s R still better than 0.995). This can be explained from the fact that in the starting position of the leglift the centres of mass horizontal positions of the legs and pelvis with respect to the joint centres of rotation is relatively large. The effect of this increased moment arm on the net joint moments is ignored in the direct calculation method leading to an overestimation of the net joint moment. Taking this into account it can be concluded that the global patterns of the net joint moments can to a large extent be explained on the basis of the ground reaction force vector relative to the position of the joint centres of rotation. Electromyography
The mean EMG signals of the muscles were expressed as a percentage of the activity levels measured during maximal standard isometric contractions. The activity of muscles having a joint extending effect are presented as positive, whereas activity levels of flexing muscles are presented as negative. The signals are presented in relation to percentage of movement time and to the net moment in knee and or hip joint (Figs 8
50
100
and 9). In the backlift the kneW&t%gmuscles (gastrocnemius, biceps femoris atid s&t&dinosus) reach activity levels up to 100% SIC, whereas the knee extending vastus medialis and rectus femoris show only low levels of activity. This corresponds to a considerable flexing moment in the knee joint during the backlift (see Fig. 5). During the leglift a considerable level of activity is recorded in both the knee extensors and knee flexors. This co-contraction of socalled agonists and antagonists becomes especially apparent in the graphs in which the EMG-levels are presented as a function of net knee moment. At roughly the same knee moment (especially at about 30 Nm, beginning of the li!?) both groups of muscle demonstrate a range of EMG amplitudes indicating an interaction of activity of both groups of muscle such that the required net knee moment is produced. Both in backlift and leglift a low I=1 of activity of the rectus femoris was found. The activity of muscles around the hip joint (Fig. 9) demonstrates cooperation of synergists. Note that in the leglift the activity of the biartictrlar semitendinosus and biceps femoris is higher than in the backlift. The activity level of the gluteus muscleis about the same in both the conditions. Work
The work produced in the ankle, knee, hip and L5-Sl joint is presented in Fig. 10. In both lifting techniques the major part of the work is produced in
0
time in % hip
0
50
time in %
50
100
time in % I
I 100
300,
OL
0
L5-Sl
I
I 50
100
timein%
Fig. 7. Comparison of net joint moments as calculated by linked-segment model (lsm,solid&@ to net joint moments as calculated by the magnitude of the ground reaction force and the distaua.hdtipecrtita Iine of action and the location of the centre of rotation of the joint (direct, dashed line). Results ,far the . i ?r leglift. .
1286
H. M. TOUSSAWTet al. backlift
50 __-.---
baclclift
50
. -.- _
- -.-.- -.-._ _ _.__.__
VaS --,_--._____,______
’ - .._
0
_
ILL! J
I
I’
._‘,;_---’
:’
.’
_LL---
__-
,”
.”
-150
-200 time
_._
,..
.*‘___,,,
%movement
-.-._
.“,/
At
.l(-J(J
-‘-
*I’
I--
-50
--
-50
-100
0
knee moment (Nm)
leglift _ _,_._..-.-.-._. ; , - _.- ’ .F ., I
50
O‘. . * .\I \ \ _--.
‘\
0
50 %movement
LJ 0
-50 -20
100
‘.,, .____--___-- L-
time
40
20
knee moment (Nm)
Fig. 8. Activity of the vastus medialis (Vas), rectus femoris (Ref), gastrocnemius (Gas), biceps femoris (Bif) and semitendinosus muscle (Sem) presented as percentage of the standard isometric contraction (o/SIC). Data are presented dependent on percentage movement time (left panels) and knee moment (right panels) for both the backlift (upper panels) and leglift (lower panels).
backlift
100
u
V
50
5;
5; @
t3 0 1 -50: 0
50 %movement
hip moment (Nm)
leglift
leglift
60, 40
5;
20
4 50 %movement
time
-20
.. 0
)
_------
:
I’
#’
________- * - - - ___-.-_
--_ I 100
_____
_I’:
E8 0
I/“:
,,‘.._
,,-
v
-201 0
200
100
100 time
100
200
300
hip moment (Nm)
Fig. 9. Activity of the gluteus maximus (Glu), rectus femoris (Ref), biceps femoris (Bif) and semitendinosus muscle (Sem) presented as percentage of the standard isometric contraction (o/SIC). Data are presented dependent on percentage movement time (left panels) and hip moment (right panels) for both the backlift (upper panels) and leglift (lower panels).
1287
Coordination in lifting
30 20 h 3
lo-
Work in ankle _______-------_,/ I’ ,/’ ,’ I’,’ I’
Work in knee _*----______ _.= 5
50
50
time in %
time in % Work in L5-S 1
Work in hip ,/
,’
__*----------
II’
5
-50; 0 time in %
50
100
time in %
Fig. 10. Work (J) produced in ankle, knee, hip and L5-Sl joint during the backlift (solid line) and leglift
(dashed line).
the hip and in LS-Sl. Especially the low work values in the knee are noteworthy. When comparing the backlift to the leglift it is interesting that during execution of the latter technique work deliverance in the hip is about 50% higher. DISCUS!SION
In this study the leglift was compared to the backlift first to address the question whether net joint moments are determined by the difference in the position of the centre of rotation of the knee joint with respect to the line of action of the ground reaction force (Figs 2 and 3). The effect of the variation in joint position on the net knee moment is illustrated in Fig. 5, where the net knee moment during the leglift and backlift is given with respect to prcentage of movement time. In the backlift a considerable knee flexing moment is found, whereas in the leglift the net knee moment starts at about 30 Nm having a knee extending effect, and slowly reduces to values oscillating around zero. When using the method of analysis employed by Schipplein et al. (1990) each joint is analysed separately in which the net joint moments are coupled directly to required muscle forces. For the present study this approach would lead to the conclusion that the low values found for the net knee moment during the execution of the leglift implies a reduction in knee loading. This may be a surprising result, since it has been advocated to use the leglift to reduce the compressive force on the low back and to shift the stresses
on the body from the low back to the legs (Bendix and Eid, 1983; Miller, 1980). Apparently, the shift of stresses from the back to the knee cannot be deduced from the net moments. The question arises what interpretation can be given to the net moments. As pointed out in the introduction, these net moments may be approximated directly from the magnitude and direction of the external force (Ingen Schenau, 1989, 1990), ignoring the influence of gravity and inertial effects. The results of the direct calculation of the net knee moment are presented together with the actually obtained values for the net knee moment by means of inverse dynamics (linked-segment model, lsm) in Figs 6 and 7. The virtually identical curves in the backlift illustrate that in the rather slow lifting movement a close relation exists between the net joint moments and the direction and magnitude of the ground reaction force, and consequently, that the effects due to angular acceleration play a minor role in determining the magnitude of the net joint moments. The differences that occur between the outcome of the linked-segment model and the direct method in the leglift can be attributed to the moment effect of the mass of the segment(s) below the joint for which the net moment is calculated. In the direct calculation method this effect is ignored, resulting in an underestimation of the net moment in the knee joint (centre of mass of feet and lower leg is positioned posterior of the knee joint, resulting in an knee extending effect), while in the hip and LS-Sf joint the combined centre of mass of the segments below the
1288
H. M.
TOUSSAINT et al.
observed joint are positioned anterior to the joint. However, the differences are slight and it can be concluded that. a’ clase relation exists between the moment caused. by the ground reaction force and the net joint moments. Apparently, a process of coordination endorses a distribution of net joint moments across joints, such that the ground reaction force points in a direction that provides balance during the movement. This conclusion raises the question to the intermuscular coordination of the leg muscles during lifting. The fact that the net knee moment oscillates around zero during the leglift is a consequence of the necessity to keep balance during the movement. However, this seems rather inefficient because the quadriceps muscles are in aprfect position to deliver useful work. The activation’ patterns of the monoarticular knee extensor (vastus mddialis) and biarticular knee flexors and extensors (biceps femoris, semitendinosus, and gastrocnemius) shed some light on the solution of this apparent contradiction. The vastus medialis appeared to be rather active (about 50% SIC) and furthermore, coactivated with its antagonistically acting semitendinosus, biceps femoris and gastrocnemius as indicated in Fig. 7: The quadriceps is active and shortens, thus producing useful work. The undesired ‘side effect’ which an increased knee moment would have on the direction of the ground reaction force is compensated by the increased activity of the knee flexors (gastroenemius &nd hamstrings). The latter biarticular must& .group lias a hip extending effect also, hence, the activation of the quadriceps pulls the hamstrings in order to facilitate hip extension. This socalled ligamentous action of biarticular muscles (see Introduction se&on.and, for an extensive description, Ingen Schenau, 1990) facilitates transport of work delivered by the quadriceps to the hip joint. From Fig. 10 it can be seen that the work produced in the hip joint during the leglift is considerably higher than the value obtained durhig thetbacklift. Furthermore, the direction and magnitude of the ground reaction force has to be such that the subject executing the lift keeps his balance during the movement. This coupling between the net joint moments and the control of an external force in leg extensions was described by Jacobs and Ingen Schenau (in pressb). They demonstrated that biarticular muscles (hamstrings and rectusfemoris) play an important role in the distribution of the net moment over the joint crossed in a static lcgLejrtegsion task in which the subject was asked to deliver a specific external force. In their experiments high correlation coefficients (0.96) were found between the difference in activity of the rectus femoris and the hamstrings and the difference in net moments of the&ip and knee joints. In the present study a significant correlation between the activity difference (RF-HAM) and the net joint moment difference (Mtnec - M,,& was found as well, Pearson’s R being 0.72 and 0.83 for the leg and backlift, respectively. The shortening of the muscles during the move-
ment has a disturbing effect on the EMG-force relation (Perry and Bekey, 1981). This can explain the higher correlation coefficient in the backlift in which the shortening of the leg muscles is likely to be much smaller and the relatively low correlation when compared to the results obtained by Jacobs and Ingen Schenau (in pressb). Correlation coefficients approaching these values were found in the present study between the activity level of the gastrocnemius and the net joint moment difference between knee and ankle, Pearson’s R being-O.89 and 0.96 for the leglift and backlift, respectively. Thus, although the coefficients of correlation are not as high as in static exercise, the conclusion seems warranted that during lifting tasks the intermuscular coordination is aimed at coupling of joint moments,‘such that the ground reaction force points in a direction that provides balance during the movement. Acknowledgements-The authors gratefully acknowledge the cooperation of our students Jos Twisk, Ad van de Ven and Carien Thissen in the process of data collection. Idsart Kingma provided us with part of the software to analyse the data. and reviewed the manuscript. We thank Han Kemper and Gerrit-Jan van Ingen Schenau for guidance and advice, for critically reviewing the manuscript, and for providing facilities to pursue,this project.
REFERENCES
Andersson, G. B. J. (1981) Epidemiological aspects on low back pain in industry. Spine 6, 53-59. Andcrsson, G, B. J., Ortengren, R. and Nachemson, A. (1976) Quantitative studies of back loads in lifting. Spine 1, 178-185. Basmajian, J. V. (1978) Muscles Alive; Their Functions Reoealed by Electromyogrophy. Williams and Wilkins, Baltimore. Bejjani, F. J., Gross, C. M. and Pugh, J. W. (1984) Model for static lifting: relationship of loads on the spine and the knee. J. Biomechanics 17,281-286. Bendix,T. and Eid, S. E. (1983)The distance between the load and the body with three bi-manual lifting techniques. Appl. Ergonomics 14, 185-192. Cavanagh, P. R. and Komi, P. V. (1979) Electromechanical delay in human skeletal muscle under concentric and eccentric contractions. Eur. .I. appl. Physiol. 42, 159-163. Cleland, J. (1867) On the actions of muscles passing over more than one joint. J. Anat. Physiol. 1, 85-93. Dempster, W. T. (1955) Space Requirements of the Seated Operator. Wright Patterson Air Force Base, Dayton, OH. Ekholm, J., Arborelius, U. P. and Nemeth, G. (1982) The load on the lumbo-sacral joint and trunk muscle activity during lifting. Ergonomics 25, 145-161. Elftman, H. (1939) Forces and energy changes in the leg during walking. Am. J. Physiol. 125, 339-356. Freivalds, A., Chaffin, D. B., Garg, A. and Lee, K. S. (1984) A dynamic biomechanical evaluation of lifting maximum acceptable loads. J. Biomechanics 17, 251-262. Gagnon, M. and Smyth, G. (1991) Muscular mechanical energy expenditure as a process for detecting potential risks in manual materials handling. J. Biomechanics 24, 191-203. Garg, A., Chaffin, D. B. and Freivalds, A. (1982) Biomechanical stresses from manual load lifting: a static vs dynamic evaluation. IIE Transactions 14, 272-281.
Coordination in lifting Gregoire, L., Veeger, H. E., Huijing, P. A. and Ingen Schenau, G. J. van (1984) The role of mono- and biarticular muscles in explosive movements. Int. J. Sports Med. 5, 301-305. Grieve, D. W. (1974) Dynamic characteristics of man during crouch- and stoop lifting. In Eiomechanics IV (Edited by Nelson R. C. and Morehouse C. A.), pp. 19-29. University Park Press, New York. Hall, S. J. (1985) Effect of attempted lifting speed on forces and torque exerted on the lumbar spine. Med. Sci. Sports Exert. 17, 44@444. Ingen Schenau, G. J. van (1989) From rotation to translation: Constraints on multi-joint movements and the unique action of biarticular muscles. Hum. Mvmt Sci. 8. 301-337. Ingen Schenau, G. J. van (1990) On the action of bi-articular muscles. A review. Neth. J. 2001. 40, 521-540. Ingen Schenau, G. J. van, Boots, P. J. M., Groot, G. de, Snackers, R. J. and Woenzel, W. W. L. van (1992) The constrained control of force and position in multi-joint movements. Neuroscience 46, 197-207. Jacobs, R. and Ingen Schenau, G. J. van. Intermuscular coordination in sprint running. J. Biomechanics (in pressa). Jacobs, R. and Ingen Schenau, G. J. van. Control of an external force in leg extensions. J. Physiol. (in pressb). Kumar, S. and Davis, P. R. (1983) Spinal loading in static and dynamic postures: EMG and intraabdominal pressure study. Ergonomics 26,913-922. Lees, A. (1980) An optimized film analysis method based on finite difference techniques. J. Hum. Mumt Stud. 6, 165-180. Leskinen, T. P. J. (1985) Comparison of static and dynamic biomechanical models. Ergonomics 26, 595-604. Leskinen, T. P. J., Stalhammar, H. R., Kuorinka, I. A. A. and Troup, J. G. D. (1983a) A dynamic analysis of spinal compression with different lifting techniques. Ergonomics 26, 595-604.
Leskinen, T. P. J., Stllhammar, H. R., Kuorinka, I. A. A. and Troup, J. G. D. (1983b) The effect of inertial factors on spinal stress when lifting. Engng Med. 12, 87-89. Looze, M. P. D., Kingma, I., Bussmann, J. B. J. and Toussaint, H. M. Validation of a dynamic linked segment model to calculate joint moments in lifting. Clin. Biomechanics (in press). Martin, J. B. and Chaffin, D. B. (1972) Biomechanical computerized simulation of human strength in sagittal plane activities. AIIE Transactions 4, 19-28.
1289
McGill, S. M. and Norman, R. W. (1986) Partitioning of the LA-L5 dynamic moment into disc, ligamentous, and muscular components during lifting. Spine 7, 666-678. Miller, R. L. (1980) When you lift, bend your knees. Occup. Health Safety 45, 46-47.
Olney, S. J. and Winter, D. A. (1985) Predictions of knee and ankle moments of force in walking from EMG and kinematic data. J. Biomechanics 18, 9-20. Perry, J. and Bekey, G. A. (1981) EMG-force reltionships in skeletal muscle. CRC Crit. Rev. Biomed. Engng 4, l-22. Schipplein, 0. D., Trafimow, J. H., Andersson, G. B. J. and A&acchi, T. P. (1990) Relationship between moments at the L5/Sl level, hip and knee joint when lifting. J. Biomechanics 23, 907-912.
Schultz, A. B. and Andersson, G. B. J. (1981) Analysis of loads on the lumbar spine. Spine 6, 76-82. Schultz, A. B., Haderspeck, K., Warwick, D. and Pontillo, D. (1983) Use of lumbar trunk muscles in isometric performance of mechanically complex standing tasks. J. Orthop. Res. 1, 77-91.
Troup, J. D. G. (1965) Relation of lumbar spine disorders to he&y manual work and lifting. Lancet i, 857-861. Trout. J. D. G. (1977) Dvnamic factors in the analvsis of s&p and croich IGtini methods: a methodologiial approach to the development of safe materials handling standards. Orthop. Clin. N. Am. 8,201-209. Troup, J. D. G., Leskinen, T. P. J., StBlhammar, H. R. and Kuorinka, I. A. A. (1983) A comparison ofintra-abdominal pressure increases, hip torque, and lumbar intervertebral compression in different lifting techniques. Hum. Factors 25,517-525.
Vos, E. J., Harlaar, J. and Ingen Schenau, G. J. van (1991) Elcctro-mechanical delay during knee extensor contractions. Med. Sci. Sports Exert. 23, 1187-1193. Vos, E. J., Mullender, M. G. and Ingen Schenau, G. J. van (1990) Electra-mechanical delay in the vastus lateralis muscles during dynamic contractions. Eur. J. appl. Physiol. 60,467471.
Winter, D. A. (1979) Biomechanics of Human Movement. Wiley, New York. Winter, D. A. (1983) Moments of force. and mechanical power in jogging. J. Biomechanics 16, 91-97.