- Email: [email protected]
Is the trunk movement more perturbed after an asymmetric than after a symmetric perturbation during lifting?
ARTICLE IN PRESS
Journal of Biomechanics 37 (2004) 1071–1077
Is the trunk movement more perturbed after an asymmetric than after a symmetric perturbation during lifting? J.C.E. van der Burg (Petra)*, Idsart Kingma, Jaap H. van Die.en Faculty of Human Movement Sciences, Institute for Fundamental and Clinical Human Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam 1081 BT, The Netherlands Accepted 19 November 2003
Abstract Low back injury is associated with sudden movements and loading. Trunk motion after sudden loading depends on the stability of the spine prior to loading and on the trunk muscle activity in response to the loading. Both factors are not axis-symmetric. Therefore, it was hypothesized that the effects on trunk dynamics would be larger after an asymmetric than after a symmetric perturbation. Ten subjects lifted a crate in which, prior to lifting, a mass was displaced to the front or to the side without the subjects being aware of this. Crate and subject movements, crate reaction forces and muscle activity were recorded. From this, the stability prior to the perturbation was estimated, and the trunk angular kinematics and moments at the lumbo-sacral joint were calculated. Both perturbations only minimally affected the trunk kinematics, although the stability of the spine prior to the lifting movement was higher in the sagittal plane than in the frontal plane. In both conditions the stability appeared to be sufficient to absorb the applied perturbation. r 2003 Elsevier Ltd. All rights reserved. Keywords: Perturbation; Lifting; Stability; Symmetric loading; Asymmetric loading
1. Introduction Sudden movements and loading are important factors in the etiology of low back pain (Manning et al., 1984; Magora, 1973). Sudden movements often occur in occupational settings (Hirvonen et al., 1996; Andersen et al., 2001). During lifting, sudden movements and loading may occur when people estimate the mass of the object incorrectly. The precise mass of the object to be lifted is often unknown and can hence only be estimated on the basis of, for instance, size cues (Johansson and Edin, 1993). As subjects anticipate the mass they are going to lift (Commissaris et al., 2001), an incorrect estimation may cause a perturbation to the lifting movement due to inadequate muscular effort. After a perturbation caused by overestimation of object mass, unplanned rotations of the trunk were found (Commissaris and Toussaint, 1997) and consequently, excessive rotations of spinal motion segments could occur. *Corresponding author. Tel.: +31-204448457; fax: +31-204448529. E-mail address: P van der [email protected] (J.C.E. van der Burg (Petra)). 0021-9290/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbiomech.2003.11.016
Trunk rotations after a perturbation depend on the stability of the spine prior to the perturbation. More stable conditions lead to smaller angular displacements of the trunk after a perturbation (Chiang and Potvin, 2001; Krajcarski et al., 1999). Furthermore, the response of the trunk muscles contributes to prevention of excessive motion after a perturbation (Cholewicki et al., 2000). An adequate muscle response minimizes the effects of the perturbation. The effectiveness of both factors, initial stability and muscle response, may depend on the direction of the perturbation. Ligaments, muscles and tendons and the neural control system determine spinal stability (Panjabi, 1992). Both the anatomy of the structures and the activation of the muscles prior to lifting are not axissymmetric. Granata and Wilson (2001) found that the stabilization of the spine in asymmetric postures required more muscle recruitment than in symmetrical postures for the same demand. In addition, generating an adequate muscle response to a sideward perturbation may be more challenging because no single trunk muscle generates purely lateral rotation moments. Thomas et al. (1998) found higher muscle activation, indicating a more
ARTICLE IN PRESS 1072
J.C.E. van der Burg (Petra) et al. / Journal of Biomechanics 37 (2004) 1071–1077
general stiffening of the trunk after an asymmetric perturbation, compared to a symmetric trunk perturbation. With respect to the kinematic response, Gagnon et al. (1995) found no effect of perturbation direction in trunk flexion and lateral bending. On the other hand, Thomas et al. (1998) described a smaller trunk displacement after a perturbation in the sagittal plane than after a perturbation in the frontal plane. In both studies the perturbations were not self-initiated. The perturbations were applied, respectively, by dropping an object into a box that was held by the subjects or by pulling on a rope that was attached to the trunk. Functional tasks, in which the subjects induce perturbations themselves, may result in responses that differ from situations where the perturbation is induced by the experimenter (van der Burg and van Die.en, 2001; van der Burg et al., 2003). We therefore decided to study a functional task, i.e. lifting an object with an unexpectedly changed center of mass, such that the perturbation was self-initiated by lifting the object. The aim of this study was to investigate whether the trunk movement is affected more after a sideward than after a forward perturbation during lifting. Both perturbations may occur commonly in lifting e.g. when an object in a box suddenly shifts to the side or to the front. The stability of the spine just prior to the start of the lifting movement and the responses after the perturbation, were examined to explain possible differences in response to the two perturbation directions.
2. Materials and methods The data were collected during a previously reported asymmetrical lifting experiment, in which the mass in the crate was unexpectedly displaced from the center to the left (van der Burg et al., 2003). In this experiment ten male subjects (age 22.1 years (SD=2.6), height 1.83 m (SD=0.06), body mass 73.8 kg (SD=4.1)) lifted a crate of 2.6 kg in which a mass of 10 kg was placed. The crate was placed 0.20 m in front of the subjects’ toes at a handle height of 0.91 m. The lifting movement was restricted to the arms. The subjects were asked to lift the crate symmetrically with an upright trunk until the flexion of the elbows was 90 . In addition, they were instructed not to rotate the crate in any direction while lifting. None of the subjects had a history of back pain. Prior to testing, all subjects signed an informed consent form approved by the local ethics committee. The present study used data collected during the normal trials, and data collected during two lifting series ending with a lift before which the mass had been displaced to the side or to the front in the crate by 0.065 m, without the subjects’ knowledge (Fig. 1) (in the previously reported experiment the displacement was 0.125 m (van der Burg et al., 2003)). The order of the
Symmetrical perturbation Series 1: 1
2
6
7
8
Series 2: 1
2
12
13
Asymmetrical perturbation
y
x z
Fig. 1. Experimental protocol. The normal condition is defined based on the last three lifting movements before the mass was unexpectedly replaced.
lifting series was varied across the subjects (Fig. 1). The normal condition was defined as the lifting movement in which the mass was centrally placed within the crate, according to the subjects’ expectations. In the asymmetrical perturbation condition, the 10 kg mass was placed to the left before the start of the lifting movement, while in the symmetrical perturbation condition the mass was placed to the front of the crate, without the subjects’ knowledge. It was arbitrarily chosen to displace the mass after seven lifting movements in the symmetrical condition and after 12 lifting movements in the asymmetrical condition. In both lifting series, the last three lifting movements, before the displacement of the mass occurred, were recorded. Before the start of both lifting series, the subjects were warned that the mass might be displaced. Between the lifting movements, the subjects wore non-transparent glasses, so that it was possible to move the mass in the crate without the subjects’ knowledge. After each lifting movement, the mass was lifted out of the crate to assure that auditory cues were not different between the perturbed and unperturbed trials. The movements of the upper body were recorded at 100 Hz using an automated video-based recording system (Optotrakt, Nothern Digital Inc., Canada). Fourteen LEDs were placed bilaterally at the upper part of the body to indicate the location of the following joints: hip, lumbo-sacral, spinous process of the first thoracic vertebra, shoulder, elbow and wrist joint, as described in detail by van der Burg et al. (2003). One LED was placed at the back of the hand, to indicate the position of the center of the hand. The location of the crate center of mass was calculated with the help of three LEDs at both sides of the crate. To describe the segment angles of the trunk and the crate, we determined Euler
ARTICLE IN PRESS J.C.E. van der Burg (Petra) et al. / Journal of Biomechanics 37 (2004) 1071–1077
computed starting from the hands. Net moments calculated from the hands have been shown to closely correspond to net moment calculated from the feet (r ¼ 0:99) (de Looze et al., 1992). Besides, the lifting movement was performed with an upright trunk, so that the estimation of the inertia and center of mass of the trunk was not of great importance. Net moments were calculated as a function of the relevant external forces and accelerations of the trunk and the crate according to the method described by Hof (1992). To quantify the perturbation to the trunk, the moment of the crate with respect to the L5–S1 joint (=crate moment) was calculated using the same method. Anthropometrical data (body mass, length of segments, standing height) were measured to estimate segment anthropometry (Plagenhoef et al., 1983). Each trial was synchronized in time to the instant the subjects started to exert a positive vertical force on the crate. This instant was defined as t ¼ 0: A total of 1.5 s was analyzed. At t ¼ 0 all angles were defined as zero to focus on the effects of the perturbation with respect to the initial posture. For each subject, the 95% confidence interval of the normal, unperturbed trials was calculated to estimate the normal range of variability for the crate and trunk angles and the net moments. To calculate the confidence interval, the three normal, unperturbed trials in both conditions were used, and also the data of three unperturbed lifting movements described in van der Burg et al. (2003) to increase the reliability of the confidence interval. When the curve of the perturbed condition lay outside the 95% confidence interval of the normal condition, the deviation was calculated. To test the difference between the perturbation directions, the variables were compared in the planes in which the perturbation occurred (sagittal versus frontal plane). The first instant the curve deviated from the confidence interval was defined as the onset of the effect of the perturbation (Fig. 2). The magnitude of the perturbation effects was defined as the peak deviation outside the confidence interval (Fig. 2). A paired samples-t-test was
95% Confidence interval Mean of the normal condition Peak deviation
Perturbed condition
Onset of perturbation effect
Variable
angles in the following sequence: sagittal (flexionextension), frontal (lateral bending) and transverse (torsion). Backward rotation, counter clockwise torsion and right lateral bending were defined as positive. Simultaneously with the movement registration, the reaction force on the crate was measured with a forceplatform placed below the crate (Kistler 9218B). All analogue force signals were amplified, sampled (100 Hz), filtered (10 Hz, fourth-order Butterworth filter), and stored. On both sides of the body, the electromyographic activation (EMG) of seven trunk muscles and the m. biceps brachii was recorded by means of disposable EMG-electrodes (Ag/AgCl) with a center to center spacing of 2.5 cm. The EMG signals were recorded from the erector spinae muscles at the level of L1 and T9, the latissimus dorsi, and the external and internal oblique muscles. All trunk muscles were monitored at positions described in van der Burg et al. (2003). The electrodes on the m. biceps brachii were positioned at the muscle belly. The EMG signals were amplified 20 times (Porti-17, Twente Medical Systems), band-pass filtered (10– 400 Hz), sampled (1600 Hz) and stored on disk with a 16 bit resolution. To reduce the influence of possible movement artefacts and electrocardiographic signals, the EMG signals were high-pass filtered (digital finite impulse response filter, 30 Hz) (Redfern et al., 1993). The signals were bandstop filtered between 49 and 51 Hz to reduce artifacts from the electric mains, rectified, and low-pass filtered (second-order Butterworth filter, 2.5 Hz (Potvin et al., 1996)). The EMG signals of the trunk muscles were normalized to the maximum value per muscle of seven tests of maximal isometric contractions as described by McGill (1991). All tests were repeated 3 times. The kinematic data and the normalized trunk EMG data at the instant of onset of m. biceps brachii activity served as input for the biomechanical model developed for quantifying stability of the lumbar spine (Cholewicki and McGill, 1996). The model consisted of a rigid pelvis and sacrum, five lumbar vertebrae connected by a lumped parameter, nonlinear disc and ligament, rigid ribcage and 90 muscle fascicles. Three axes of rotation were assigned to each intervertebral joint between T12 and S1. Muscle forces, necessary to balance the external load and the upper body weight, were estimated for all 90-muscle fascicles with the help of a distribution model and an EMG assisted optimization approach. Stability analyses were performed in accordance with the minimal potential energy principle for all five vertebrae for all three axes. The average curvature of the surface of the system’s potential energy in the vicinity of the static equilibrium served as the stability index for that plane (Cholewicki and VanVliet, 2002). To study the responses to the perturbation, the net moments at the lumbo-sacral (L5–S1) joint were
1073
time
Fig. 2. The approach that was used to calculate the onset and the peak deviation caused by the perturbation.
ARTICLE IN PRESS 1074
J.C.E. van der Burg (Petra) et al. / Journal of Biomechanics 37 (2004) 1071–1077
used to calculate the differences in maximum values between the conditions. Due to the asymmetric distribution of the difference in onset values, the differences between the conditions were tested with a non-parametric, Wilcoxon test. Effects were considered to be significant at po0:05:
3. Results The data of nine subjects were used for the analysis. The data set of one subject was incomplete, and was hence discarded. Prior to the perturbation, the stability of the lumbar spine, estimated from the biomechanical model, was higher in the sagittal plane than in the frontal plane for all subjects. The mean stability index was 405.0 N m/rad (SD=128.1 N m/rad) in the sagittal plane and 342.1 N m/rad (SD=92.7 N m/rad) in the frontal plane (po0:001). A similar difference between the planes was found when the minimal curvature instead of the average was used to calculate the stability index. At this instant, the net extension moment deviated from zero, 7.5 N m (SD=2.2 N m, p ¼ 0:000) whereas the lateral bending moment did not deviate from zero (0.1 N m, SD=0.3 N m, p ¼ 0:26). In both perturbed conditions, the displacement of the mass caused a similar increase in crate moment with
respect to L5–S1 (difference between conditions, p ¼ 0:23). The mean crate moment, starting from liftoff until 100 ms after lift-off, was increased, as was expected due to the displacement of 0.065 m of a mass of 10 kg (symmetric increase: 6.9 N m (SD=1.2 N m) asymmetric increase: 6.5 N m (SD=1.3 N m), Fig. 3). Before the crate was fully lifted from the surface, peaks in crate moments were seen, caused by displacements of the point of application of the force vector due to rotations of the crate (Fig. 3). After a symmetric perturbation, the difference between the perturbed and normal forward bending crate moment showed a temporary dip just after lift-off for all subjects due to a delayed increase in extension moment in the perturbed signal. Such a decrease was absent in lateral bending crate moments after an asymmetrical perturbation (Fig. 3). The net moments at the L5–S1 joint were increased in response to the additional moment caused by the displaced mass (Fig. 4). In both conditions the adaptations in net moment occurred significantly earlier than the deviations of the trunk angle outside the 95% confidence interval (asymmetric condition, p ¼ 0:017; symmetric condition, p ¼ 0:025) (Fig. 4). The deviations of the net moment outside the 95% confidence interval occurred on average of 238 ms (range 30–330 ms) after the first positive vertical force was exerted on the crate. Deviations in trunk angle outside the 95% confidence interval occurred on average 572 ms (range 20–960 ms;
Crate moment relative to L5S1 Symmetric condition
Asymmetric condition
-5
5 0 -5
Flexion moment (Nm)
Left lat. bending Moment (Nm)
60 40 20 0 -20
10
10
Difference between conditions
Flexion moment (Nm)
0
60 40 20 0 -20
Flexion moment (Nm)
Perturbed condition
5
Left lat. bending Moment (Nm)
Normal condition
Left lat. bending Moment (Nm)
10
0 -10 0
0.5 time (s)
1
1.5
10 0 -10 0
0.5
1
1.5
time (s)
Fig. 3. A typical example of the crate moments in the direction of the perturbation. The left panels represent the asymmetric perturbation condition and the right panels the symmetric perturbation condition. Time zero is when the subject starts to exert a positive vertical force on the crate. In the upper panels the crate moments of the normal condition are shown. The solid line represents the mean of the normal condition. In the middle panels the crate moments in the perturbed conditions and in the lower panels the difference in crate moments between the normal and the perturbed condition are shown. The vertical dotted lines in the middle panels represent the instant at which lift-off occurred.
ARTICLE IN PRESS J.C.E. van der Burg (Petra) et al. / Journal of Biomechanics 37 (2004) 1071–1077 Symmetric perturbation
10
Extension moment (in Nm)
Lateral bending moment (in Nm)
L5S1 moments
Asymmetric perturbation
1075
5 0 -5 -10
60 40 20 0 95% Confidence interval
0
Extension (in deg)
Lateral bending (in deg)
Trunk angles
Perturbed condition
0.5 0 -0.5 -1 0
0.5 1 time (in s)
1.5
2 0 -2 -4 -6 0
0.5 1 time (in s)
1.5
Fig. 4. A typical example of the deviation of the trunk angles and net moments in the direction of the perturbation in the both perturbed conditions for one subject. Time zero is when the subject starts to exert a positive vertical force to the crate. In the upper panels the L5–S1 moments are shown, and in the lower panels the trunk angle is shown. The left panels represent the asymmetric perturbation condition and the right panels the symmetric perturbation condition. The solid lines represent the perturbed condition, and the dotted lines the 95% confidence interval of the normal condition.
three subjects did not deviate outside the confidence interval) after the first positive force on the crate in both conditions. The onset of the deviations was not significantly different between the two conditions (moments: p ¼ 0:67; angles: p ¼ 0:40). The onset of the change in muscle activity was not significantly different from the onset of angular deviation (p ¼ 0:2 and higher for the various muscle groups). The variations in onset of the change in muscle activity between the subjects were very large. The peak trunk angular deviations after both perturbations were small and not significantly different between the two conditions (p ¼ 0:249). Peak deviations of the trunk angle outside the 95% confidence interval were 0.3 (SD=0.2 ) and 0.9 (SD=1.3 ) for, respectively, the asymmetric and the symmetric conditions. These deviations were significant in the asymmetric condition (p ¼ 0:004; deviations in 7 out of 9 subjects), but not significant in the symmetric condition (p ¼ 0:083; deviations in 5 out of 9 subjects). A typical example of the trunk deviations after both perturbations is shown in Fig. 4. In contrast to the deviations in trunk angles, the deviations of the crate trajectory were of considerable magnitude. The subjects were not able to lift the crate with the displaced mass without substantial crate rotations, although they were instructed to lift the crate horizontally (symmetric: 7.8 (SD=2.5 ), po0:001; asymmetric: 4.6 (SD=1.4 ), po0:001). The deviation of the crate angle in the perturbation direction was
significantly higher in the symmetric than in the asymmetric condition (p ¼ 0:030).
4. Discussion Subjects lifted a crate in which a 10 kg mass was unexpectedly displaced to the front or to the left side of the crate before lift-off. The displacement of the mass in the crate was large enough to cause a disturbance of the planned lifting movement. The deviations of the crate trajectory from the unperturbed condition were significant and larger moments were seen after the mass had been lifted from the surface to counteract the additional crate moments due to the displaced mass. Nevertheless trunk angular deviations were small (0.3 and 0.9 ) in both conditions. Two factors, initial stability and an accurate muscle response, could contribute to these small displacements. Based on the stability index of the trunk just prior to lift-off small angular displacements could be expected. (angular displacement of 6.5 N m/405 N m/rad=0.016 rad=0.9 in the saggital plane and 6.5 N m/342 N m/ rad=0.02 rad=1.1 in the frontal plane). The observed angular displacements were in the same range as the predicted deviations based on the stability index. This implies that the perturbation could be absorbed by stiffness alone. Nevertheless, the deviations in the net moments occurred earlier than the deviations in angular displacements. In an elastic and massless system,
ARTICLE IN PRESS 1076
J.C.E. van der Burg (Petra) et al. / Journal of Biomechanics 37 (2004) 1071–1077
changes in net moment and angular deviations should occur at the same time. However, in a system like the trunk inertia and damping contribute to changes in moment, which therefore can precede the changes in angular deviations. An adequate muscular response could be an additional explanation for the relatively early increase in moments. Such a response could be triggered by the perturbed arm movement prior to a perturbation to the trunk (Hodges et al., 2001). Given sufficient stability, the effect of responsive muscle activity may be less important. This might explain why no clear pattern of muscular response could be found across subjects. The relatively late occurrence of the moment and angular trunk deviations may be explained by joint rotations in the arms caused by the perturbation. As described in a simulation study of lifting of van der Burg (2003, submitted for publication), joint rotations in distal segments, due to low stiffness in these segments, delay the propagation of the perturbation to the trunk. In the present study, the propagation of the perturbation to the trunk appeared to be delayed, because rotations of the crate, and thus rotations of the upper extremity joints occurred, in both perturbed conditions. An additional effect of joint rotations was especially pronounced in the symmetric condition. Due to the forward rotation of the crate, a part of the mass of crate was longer on the platform, which attenuated the additional crate moment. By preserving the forward rotation of the crate, the moment arm of the crate would have been decreased during the whole lifting movement. However, due to the instructions to the subjects to keep the crate horizontal, subjects quickly corrected the forward rotation, and thus the crate moment was lower only temporarily. After an asymmetric perturbation, the center of mass of the crate was still in the middle of the hands in for-afterward direction. Therefore, no wrist rotations, but especially movements of the shoulder and elbow joints are to be expected. The stiffness of the elbow and shoulder joint is likely to be higher than the stiffness of the wrist, as is indicated by the smaller crate movements after an asymmetric perturbation than after a symmetric perturbation. The higher stability in the sagittal plane than in the frontal plane prior to the perturbation may be explained by the difference in net moments at that instant. The extension moments were increased due to the posture of the subject and likely also due to the anticipation to the lifting movement (van Die.en and de Looze, 1999; Bouisset and Zattara, 1987). The larger net moments increased the stability of the spine in this plane. So, the higher stability may be a positive effect of the task. In addition, the higher stability may also be caused part of by task specific directional stiffening of the spine, which Burdet et al. (2001) described for arm movements.
In contrast to the present findings, Thomas et al. (1998) reported a significantly smaller deviation of the trunk angle after an asymmetric perturbation as compared to a symmetric perturbation. The contrast between both studies might be due to the difference in method of applying the perturbing force. In their study the force was directly applied at the trunk, while in our study the perturbation was applied at the hands. As described above, the arms might reduce and delay the perturbation and thus influence the deviations of the trunk.
5. Conclusion The hypothesized difference between an asymmetric perturbation and a symmetric perturbation was not shown in spite of lower stability in the frontal plane. The kinematics of the trunk was minimally affected both by a symmetric and an asymmetric perturbation of a lifting movement. Likely the stability of the trunk was sufficient to counteract the applied perturbation in both conditions.
Acknowledgements We like to thank Dr. Jacek Cholewicki for his help in estimating spinal stability and Dr. Jim Potvin for the thoughtful comments on draft versions of this manuscript.
References Andersen, T., Bull, Schibye, B., Skotte, J., 2001. Sudden movements of the spinal column during health-care work. International Journal of Industrial Ergonomics 28, 47–53. Bouisset, S., Zattara, M., 1987. Biomechanical study of the programming of anticipatory postural adjustments associated with voluntary movement. Journal of Biomechanics 20, 735–742. Burdet, E., Osu, R., Franklin, D.W., Milner, T.E., Kawato, M., 2001. The central nervous system stabilizes unstable dynamics by learning optimal impedance. Nature 414, 446–449. Chiang, J., Potvin, J.R., 2001. The in vivo dynamic response of the human spine to rapid lateral bend perturbation: effects of preload and step input magnitude. Spine 26, 1457–1464. Cholewicki, J., McGill, S.M., 1996. Mechanical stability of the in vivo lumbar spine: implications for injury and chronic low back pain. Clinical Biomechanics 11, 1–15. Cholewicki, J., Simons, A.P., Radebold, A., 2000. Effects of external trunk loads on lumbar spine stability. Journal of Biomechanics 33, 1377–1385. Cholewicki, J., VanVliet, Iv Jj., 2002. Relative contribution of trunk muscles to the stability of the lumbar spine during isometric exertions. Clinical Biomechanics (Bristol, Avon) 17, 99–105. Commissaris, D.A., Toussaint, H.M., Hirschfeld, H., 2001. Anticipatory postural adjustments in a bimanual, whole-body lifting task seem not only aimed at minimising anterior-posterior centre of mass displacements. Gait Posture 14, 44–55.
ARTICLE IN PRESS J.C.E. van der Burg (Petra) et al. / Journal of Biomechanics 37 (2004) 1071–1077 Commissaris, D.A.C.M., Toussaint, H.M., 1997. Load knowledge affects low-back loading and control of balance in lifting tasks. Ergonomics 40, 559–575. de Looze, M.P., Bussmann, J.B.J., Kingma, I., Toussaint, H.M., 1992. Validation of a dynamic linked segment model to calculate joint moments in lifting. Clinical Biomechanics 7, 161–169. Gagnon, M., Plamondon, A., Gravel, D., 1995. Effects of symmetry and load absorption of a falling load on 3D trunk muscular moments. Ergonomics 38, 1156–1171. Granata, K.P., Wilson, S.E., 2001. Trunk posture and spinal stability. Clinical Biomechanics (Bristol, Avon) 16, 650–659. Hirvonen, M., Leskinen, T., Gronqvist, R., Viikari-Juntura, E.R.A., Riihimaki, H., 1996. Occurence of sudden movements at work. Safety Science 24, 77–82. Hodges, P.W., Cresswell, A.G., Thorstensson, A., 2001. Perturbed upper limb movements cause short-latency postural responses in trunk muscles. Experimental Brain Research 138, 243–250. Hof, A.L., 1992. An explicit expression for the moment in multibody systems. Journal of Biomechanics 25, 1209–1211. Johansson, R.S., Edin, B.B., 1993. Predictive feed-forward sensory control during grasping and manipulation in man. Biomedical Research 14 (Suppl. 4), 95–106. Krajcarski, S.R., Potvin, J.R., Chiang, J., 1999. The in vivo dynamic response of the spine to perturbations causing rapid flexion: effects of pre-load and step input magnitude. Clinical Biomechanics 14, 54–62. Magora, A., 1973. Investigation of the relation between low back pain and occupation. Part IV. Physical requirements: bending, rotation, reaching and sudden maximal effort. Scandinavian Journal of Rehabilitation Medicine 5, 186–190. Manning, D.P., Mitchell, R.G., Blanchfield, L.P., 1984. Body movements and events contributing to accidental and nonaccidental back injuries. Spine 9, 734–739. McGill, S.M., 1991. Electromyographic activity of the abdominal and low back musculature during the generation of isometric and
1077
dynamic axial trunk torque: implications for lumbar mechanics. Journal of Orthopaedic Research 9, 91–103. Panjabi, M.M., 1992. The stabilizing system of the spine. Part I. Function, dysfunction, adaptation, and enhancement. Journal of Spinal Disorders 5, 383–389 discussion 397. Plagenhoef, S., Evans, F.G., Abdelnour, T., 1983. Anatomical data for analyzing human motion. Research Quarterly for Exercise and Sport 54, 169–178. Potvin, J.R., Norman, R.W., McGill, S.M., 1996. Mechanically corrected EMG for the continuous estimation of erector spinae muscle loading during repetitive lifting. European Journal of Applied Physiology 74, 119–132. Redfern, M.S., Hughes, D.B., Chaffin, D.B., 1993. High-pass filtering to remove electrocardiographic interference from torso recordings. Clinical Biomechanics 8, 44–48. Thomas, J.S., Lavender, S.A., Corcos, D.M., Andersson, G.B.J., 1998. Trunk kinematics and trunk muscle activity during a rapidly applied load. Journal of Electromyography and Kinesiology 8, 215–225. van der Burg, J.C.E., 2003. Lifting objects: surprised by the mass. Ph.D. Thesis, Vrije Universiteit. van der Burg, J.C.E., Casius, L.J.R., Kingma, I., van Die.en, J.H., van Soest, A.J., submitted for publication. The contribution of muscle properties in controlling perturbations during lifting. Journal of Biomechanics. van der Burg, J.C.E., van Die.en, J.H., 2001. Underestimation of object mass in lifting does not increase the load on the low back. Journal of Biomechanics 34, 1447–1453. van der Burg, J.C.E., Kingma, I., van Die.en, J.H., 2003. Effects of unexpected lateral mass distribution in lifting on trunk movement and trunk loading. Spine 28, 764–770. van Die.en, J.H., de Looze, M.P., 1999. Directionality of anticipatory activation of trunk muscles in a lifting task depends on load knowledge. Experimental Brain Research 128, 397–404.
Journal of Biomechanics 37 (2004) 1071–1077
Is the trunk movement more perturbed after an asymmetric than after a symmetric perturbation during lifting? J.C.E. van der Burg (Petra)*, Idsart Kingma, Jaap H. van Die.en Faculty of Human Movement Sciences, Institute for Fundamental and Clinical Human Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam 1081 BT, The Netherlands Accepted 19 November 2003
Abstract Low back injury is associated with sudden movements and loading. Trunk motion after sudden loading depends on the stability of the spine prior to loading and on the trunk muscle activity in response to the loading. Both factors are not axis-symmetric. Therefore, it was hypothesized that the effects on trunk dynamics would be larger after an asymmetric than after a symmetric perturbation. Ten subjects lifted a crate in which, prior to lifting, a mass was displaced to the front or to the side without the subjects being aware of this. Crate and subject movements, crate reaction forces and muscle activity were recorded. From this, the stability prior to the perturbation was estimated, and the trunk angular kinematics and moments at the lumbo-sacral joint were calculated. Both perturbations only minimally affected the trunk kinematics, although the stability of the spine prior to the lifting movement was higher in the sagittal plane than in the frontal plane. In both conditions the stability appeared to be sufficient to absorb the applied perturbation. r 2003 Elsevier Ltd. All rights reserved. Keywords: Perturbation; Lifting; Stability; Symmetric loading; Asymmetric loading
1. Introduction Sudden movements and loading are important factors in the etiology of low back pain (Manning et al., 1984; Magora, 1973). Sudden movements often occur in occupational settings (Hirvonen et al., 1996; Andersen et al., 2001). During lifting, sudden movements and loading may occur when people estimate the mass of the object incorrectly. The precise mass of the object to be lifted is often unknown and can hence only be estimated on the basis of, for instance, size cues (Johansson and Edin, 1993). As subjects anticipate the mass they are going to lift (Commissaris et al., 2001), an incorrect estimation may cause a perturbation to the lifting movement due to inadequate muscular effort. After a perturbation caused by overestimation of object mass, unplanned rotations of the trunk were found (Commissaris and Toussaint, 1997) and consequently, excessive rotations of spinal motion segments could occur. *Corresponding author. Tel.: +31-204448457; fax: +31-204448529. E-mail address: P van der [email protected] (J.C.E. van der Burg (Petra)). 0021-9290/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbiomech.2003.11.016
Trunk rotations after a perturbation depend on the stability of the spine prior to the perturbation. More stable conditions lead to smaller angular displacements of the trunk after a perturbation (Chiang and Potvin, 2001; Krajcarski et al., 1999). Furthermore, the response of the trunk muscles contributes to prevention of excessive motion after a perturbation (Cholewicki et al., 2000). An adequate muscle response minimizes the effects of the perturbation. The effectiveness of both factors, initial stability and muscle response, may depend on the direction of the perturbation. Ligaments, muscles and tendons and the neural control system determine spinal stability (Panjabi, 1992). Both the anatomy of the structures and the activation of the muscles prior to lifting are not axissymmetric. Granata and Wilson (2001) found that the stabilization of the spine in asymmetric postures required more muscle recruitment than in symmetrical postures for the same demand. In addition, generating an adequate muscle response to a sideward perturbation may be more challenging because no single trunk muscle generates purely lateral rotation moments. Thomas et al. (1998) found higher muscle activation, indicating a more
ARTICLE IN PRESS 1072
J.C.E. van der Burg (Petra) et al. / Journal of Biomechanics 37 (2004) 1071–1077
general stiffening of the trunk after an asymmetric perturbation, compared to a symmetric trunk perturbation. With respect to the kinematic response, Gagnon et al. (1995) found no effect of perturbation direction in trunk flexion and lateral bending. On the other hand, Thomas et al. (1998) described a smaller trunk displacement after a perturbation in the sagittal plane than after a perturbation in the frontal plane. In both studies the perturbations were not self-initiated. The perturbations were applied, respectively, by dropping an object into a box that was held by the subjects or by pulling on a rope that was attached to the trunk. Functional tasks, in which the subjects induce perturbations themselves, may result in responses that differ from situations where the perturbation is induced by the experimenter (van der Burg and van Die.en, 2001; van der Burg et al., 2003). We therefore decided to study a functional task, i.e. lifting an object with an unexpectedly changed center of mass, such that the perturbation was self-initiated by lifting the object. The aim of this study was to investigate whether the trunk movement is affected more after a sideward than after a forward perturbation during lifting. Both perturbations may occur commonly in lifting e.g. when an object in a box suddenly shifts to the side or to the front. The stability of the spine just prior to the start of the lifting movement and the responses after the perturbation, were examined to explain possible differences in response to the two perturbation directions.
2. Materials and methods The data were collected during a previously reported asymmetrical lifting experiment, in which the mass in the crate was unexpectedly displaced from the center to the left (van der Burg et al., 2003). In this experiment ten male subjects (age 22.1 years (SD=2.6), height 1.83 m (SD=0.06), body mass 73.8 kg (SD=4.1)) lifted a crate of 2.6 kg in which a mass of 10 kg was placed. The crate was placed 0.20 m in front of the subjects’ toes at a handle height of 0.91 m. The lifting movement was restricted to the arms. The subjects were asked to lift the crate symmetrically with an upright trunk until the flexion of the elbows was 90 . In addition, they were instructed not to rotate the crate in any direction while lifting. None of the subjects had a history of back pain. Prior to testing, all subjects signed an informed consent form approved by the local ethics committee. The present study used data collected during the normal trials, and data collected during two lifting series ending with a lift before which the mass had been displaced to the side or to the front in the crate by 0.065 m, without the subjects’ knowledge (Fig. 1) (in the previously reported experiment the displacement was 0.125 m (van der Burg et al., 2003)). The order of the
Symmetrical perturbation Series 1: 1
2
6
7
8
Series 2: 1
2
12
13
Asymmetrical perturbation
y
x z
Fig. 1. Experimental protocol. The normal condition is defined based on the last three lifting movements before the mass was unexpectedly replaced.
lifting series was varied across the subjects (Fig. 1). The normal condition was defined as the lifting movement in which the mass was centrally placed within the crate, according to the subjects’ expectations. In the asymmetrical perturbation condition, the 10 kg mass was placed to the left before the start of the lifting movement, while in the symmetrical perturbation condition the mass was placed to the front of the crate, without the subjects’ knowledge. It was arbitrarily chosen to displace the mass after seven lifting movements in the symmetrical condition and after 12 lifting movements in the asymmetrical condition. In both lifting series, the last three lifting movements, before the displacement of the mass occurred, were recorded. Before the start of both lifting series, the subjects were warned that the mass might be displaced. Between the lifting movements, the subjects wore non-transparent glasses, so that it was possible to move the mass in the crate without the subjects’ knowledge. After each lifting movement, the mass was lifted out of the crate to assure that auditory cues were not different between the perturbed and unperturbed trials. The movements of the upper body were recorded at 100 Hz using an automated video-based recording system (Optotrakt, Nothern Digital Inc., Canada). Fourteen LEDs were placed bilaterally at the upper part of the body to indicate the location of the following joints: hip, lumbo-sacral, spinous process of the first thoracic vertebra, shoulder, elbow and wrist joint, as described in detail by van der Burg et al. (2003). One LED was placed at the back of the hand, to indicate the position of the center of the hand. The location of the crate center of mass was calculated with the help of three LEDs at both sides of the crate. To describe the segment angles of the trunk and the crate, we determined Euler
ARTICLE IN PRESS J.C.E. van der Burg (Petra) et al. / Journal of Biomechanics 37 (2004) 1071–1077
computed starting from the hands. Net moments calculated from the hands have been shown to closely correspond to net moment calculated from the feet (r ¼ 0:99) (de Looze et al., 1992). Besides, the lifting movement was performed with an upright trunk, so that the estimation of the inertia and center of mass of the trunk was not of great importance. Net moments were calculated as a function of the relevant external forces and accelerations of the trunk and the crate according to the method described by Hof (1992). To quantify the perturbation to the trunk, the moment of the crate with respect to the L5–S1 joint (=crate moment) was calculated using the same method. Anthropometrical data (body mass, length of segments, standing height) were measured to estimate segment anthropometry (Plagenhoef et al., 1983). Each trial was synchronized in time to the instant the subjects started to exert a positive vertical force on the crate. This instant was defined as t ¼ 0: A total of 1.5 s was analyzed. At t ¼ 0 all angles were defined as zero to focus on the effects of the perturbation with respect to the initial posture. For each subject, the 95% confidence interval of the normal, unperturbed trials was calculated to estimate the normal range of variability for the crate and trunk angles and the net moments. To calculate the confidence interval, the three normal, unperturbed trials in both conditions were used, and also the data of three unperturbed lifting movements described in van der Burg et al. (2003) to increase the reliability of the confidence interval. When the curve of the perturbed condition lay outside the 95% confidence interval of the normal condition, the deviation was calculated. To test the difference between the perturbation directions, the variables were compared in the planes in which the perturbation occurred (sagittal versus frontal plane). The first instant the curve deviated from the confidence interval was defined as the onset of the effect of the perturbation (Fig. 2). The magnitude of the perturbation effects was defined as the peak deviation outside the confidence interval (Fig. 2). A paired samples-t-test was
95% Confidence interval Mean of the normal condition Peak deviation
Perturbed condition
Onset of perturbation effect
Variable
angles in the following sequence: sagittal (flexionextension), frontal (lateral bending) and transverse (torsion). Backward rotation, counter clockwise torsion and right lateral bending were defined as positive. Simultaneously with the movement registration, the reaction force on the crate was measured with a forceplatform placed below the crate (Kistler 9218B). All analogue force signals were amplified, sampled (100 Hz), filtered (10 Hz, fourth-order Butterworth filter), and stored. On both sides of the body, the electromyographic activation (EMG) of seven trunk muscles and the m. biceps brachii was recorded by means of disposable EMG-electrodes (Ag/AgCl) with a center to center spacing of 2.5 cm. The EMG signals were recorded from the erector spinae muscles at the level of L1 and T9, the latissimus dorsi, and the external and internal oblique muscles. All trunk muscles were monitored at positions described in van der Burg et al. (2003). The electrodes on the m. biceps brachii were positioned at the muscle belly. The EMG signals were amplified 20 times (Porti-17, Twente Medical Systems), band-pass filtered (10– 400 Hz), sampled (1600 Hz) and stored on disk with a 16 bit resolution. To reduce the influence of possible movement artefacts and electrocardiographic signals, the EMG signals were high-pass filtered (digital finite impulse response filter, 30 Hz) (Redfern et al., 1993). The signals were bandstop filtered between 49 and 51 Hz to reduce artifacts from the electric mains, rectified, and low-pass filtered (second-order Butterworth filter, 2.5 Hz (Potvin et al., 1996)). The EMG signals of the trunk muscles were normalized to the maximum value per muscle of seven tests of maximal isometric contractions as described by McGill (1991). All tests were repeated 3 times. The kinematic data and the normalized trunk EMG data at the instant of onset of m. biceps brachii activity served as input for the biomechanical model developed for quantifying stability of the lumbar spine (Cholewicki and McGill, 1996). The model consisted of a rigid pelvis and sacrum, five lumbar vertebrae connected by a lumped parameter, nonlinear disc and ligament, rigid ribcage and 90 muscle fascicles. Three axes of rotation were assigned to each intervertebral joint between T12 and S1. Muscle forces, necessary to balance the external load and the upper body weight, were estimated for all 90-muscle fascicles with the help of a distribution model and an EMG assisted optimization approach. Stability analyses were performed in accordance with the minimal potential energy principle for all five vertebrae for all three axes. The average curvature of the surface of the system’s potential energy in the vicinity of the static equilibrium served as the stability index for that plane (Cholewicki and VanVliet, 2002). To study the responses to the perturbation, the net moments at the lumbo-sacral (L5–S1) joint were
1073
time
Fig. 2. The approach that was used to calculate the onset and the peak deviation caused by the perturbation.
ARTICLE IN PRESS 1074
J.C.E. van der Burg (Petra) et al. / Journal of Biomechanics 37 (2004) 1071–1077
used to calculate the differences in maximum values between the conditions. Due to the asymmetric distribution of the difference in onset values, the differences between the conditions were tested with a non-parametric, Wilcoxon test. Effects were considered to be significant at po0:05:
3. Results The data of nine subjects were used for the analysis. The data set of one subject was incomplete, and was hence discarded. Prior to the perturbation, the stability of the lumbar spine, estimated from the biomechanical model, was higher in the sagittal plane than in the frontal plane for all subjects. The mean stability index was 405.0 N m/rad (SD=128.1 N m/rad) in the sagittal plane and 342.1 N m/rad (SD=92.7 N m/rad) in the frontal plane (po0:001). A similar difference between the planes was found when the minimal curvature instead of the average was used to calculate the stability index. At this instant, the net extension moment deviated from zero, 7.5 N m (SD=2.2 N m, p ¼ 0:000) whereas the lateral bending moment did not deviate from zero (0.1 N m, SD=0.3 N m, p ¼ 0:26). In both perturbed conditions, the displacement of the mass caused a similar increase in crate moment with
respect to L5–S1 (difference between conditions, p ¼ 0:23). The mean crate moment, starting from liftoff until 100 ms after lift-off, was increased, as was expected due to the displacement of 0.065 m of a mass of 10 kg (symmetric increase: 6.9 N m (SD=1.2 N m) asymmetric increase: 6.5 N m (SD=1.3 N m), Fig. 3). Before the crate was fully lifted from the surface, peaks in crate moments were seen, caused by displacements of the point of application of the force vector due to rotations of the crate (Fig. 3). After a symmetric perturbation, the difference between the perturbed and normal forward bending crate moment showed a temporary dip just after lift-off for all subjects due to a delayed increase in extension moment in the perturbed signal. Such a decrease was absent in lateral bending crate moments after an asymmetrical perturbation (Fig. 3). The net moments at the L5–S1 joint were increased in response to the additional moment caused by the displaced mass (Fig. 4). In both conditions the adaptations in net moment occurred significantly earlier than the deviations of the trunk angle outside the 95% confidence interval (asymmetric condition, p ¼ 0:017; symmetric condition, p ¼ 0:025) (Fig. 4). The deviations of the net moment outside the 95% confidence interval occurred on average of 238 ms (range 30–330 ms) after the first positive vertical force was exerted on the crate. Deviations in trunk angle outside the 95% confidence interval occurred on average 572 ms (range 20–960 ms;
Crate moment relative to L5S1 Symmetric condition
Asymmetric condition
-5
5 0 -5
Flexion moment (Nm)
Left lat. bending Moment (Nm)
60 40 20 0 -20
10
10
Difference between conditions
Flexion moment (Nm)
0
60 40 20 0 -20
Flexion moment (Nm)
Perturbed condition
5
Left lat. bending Moment (Nm)
Normal condition
Left lat. bending Moment (Nm)
10
0 -10 0
0.5 time (s)
1
1.5
10 0 -10 0
0.5
1
1.5
time (s)
Fig. 3. A typical example of the crate moments in the direction of the perturbation. The left panels represent the asymmetric perturbation condition and the right panels the symmetric perturbation condition. Time zero is when the subject starts to exert a positive vertical force on the crate. In the upper panels the crate moments of the normal condition are shown. The solid line represents the mean of the normal condition. In the middle panels the crate moments in the perturbed conditions and in the lower panels the difference in crate moments between the normal and the perturbed condition are shown. The vertical dotted lines in the middle panels represent the instant at which lift-off occurred.
ARTICLE IN PRESS J.C.E. van der Burg (Petra) et al. / Journal of Biomechanics 37 (2004) 1071–1077 Symmetric perturbation
10
Extension moment (in Nm)
Lateral bending moment (in Nm)
L5S1 moments
Asymmetric perturbation
1075
5 0 -5 -10
60 40 20 0 95% Confidence interval
0
Extension (in deg)
Lateral bending (in deg)
Trunk angles
Perturbed condition
0.5 0 -0.5 -1 0
0.5 1 time (in s)
1.5
2 0 -2 -4 -6 0
0.5 1 time (in s)
1.5
Fig. 4. A typical example of the deviation of the trunk angles and net moments in the direction of the perturbation in the both perturbed conditions for one subject. Time zero is when the subject starts to exert a positive vertical force to the crate. In the upper panels the L5–S1 moments are shown, and in the lower panels the trunk angle is shown. The left panels represent the asymmetric perturbation condition and the right panels the symmetric perturbation condition. The solid lines represent the perturbed condition, and the dotted lines the 95% confidence interval of the normal condition.
three subjects did not deviate outside the confidence interval) after the first positive force on the crate in both conditions. The onset of the deviations was not significantly different between the two conditions (moments: p ¼ 0:67; angles: p ¼ 0:40). The onset of the change in muscle activity was not significantly different from the onset of angular deviation (p ¼ 0:2 and higher for the various muscle groups). The variations in onset of the change in muscle activity between the subjects were very large. The peak trunk angular deviations after both perturbations were small and not significantly different between the two conditions (p ¼ 0:249). Peak deviations of the trunk angle outside the 95% confidence interval were 0.3 (SD=0.2 ) and 0.9 (SD=1.3 ) for, respectively, the asymmetric and the symmetric conditions. These deviations were significant in the asymmetric condition (p ¼ 0:004; deviations in 7 out of 9 subjects), but not significant in the symmetric condition (p ¼ 0:083; deviations in 5 out of 9 subjects). A typical example of the trunk deviations after both perturbations is shown in Fig. 4. In contrast to the deviations in trunk angles, the deviations of the crate trajectory were of considerable magnitude. The subjects were not able to lift the crate with the displaced mass without substantial crate rotations, although they were instructed to lift the crate horizontally (symmetric: 7.8 (SD=2.5 ), po0:001; asymmetric: 4.6 (SD=1.4 ), po0:001). The deviation of the crate angle in the perturbation direction was
significantly higher in the symmetric than in the asymmetric condition (p ¼ 0:030).
4. Discussion Subjects lifted a crate in which a 10 kg mass was unexpectedly displaced to the front or to the left side of the crate before lift-off. The displacement of the mass in the crate was large enough to cause a disturbance of the planned lifting movement. The deviations of the crate trajectory from the unperturbed condition were significant and larger moments were seen after the mass had been lifted from the surface to counteract the additional crate moments due to the displaced mass. Nevertheless trunk angular deviations were small (0.3 and 0.9 ) in both conditions. Two factors, initial stability and an accurate muscle response, could contribute to these small displacements. Based on the stability index of the trunk just prior to lift-off small angular displacements could be expected. (angular displacement of 6.5 N m/405 N m/rad=0.016 rad=0.9 in the saggital plane and 6.5 N m/342 N m/ rad=0.02 rad=1.1 in the frontal plane). The observed angular displacements were in the same range as the predicted deviations based on the stability index. This implies that the perturbation could be absorbed by stiffness alone. Nevertheless, the deviations in the net moments occurred earlier than the deviations in angular displacements. In an elastic and massless system,
ARTICLE IN PRESS 1076
J.C.E. van der Burg (Petra) et al. / Journal of Biomechanics 37 (2004) 1071–1077
changes in net moment and angular deviations should occur at the same time. However, in a system like the trunk inertia and damping contribute to changes in moment, which therefore can precede the changes in angular deviations. An adequate muscular response could be an additional explanation for the relatively early increase in moments. Such a response could be triggered by the perturbed arm movement prior to a perturbation to the trunk (Hodges et al., 2001). Given sufficient stability, the effect of responsive muscle activity may be less important. This might explain why no clear pattern of muscular response could be found across subjects. The relatively late occurrence of the moment and angular trunk deviations may be explained by joint rotations in the arms caused by the perturbation. As described in a simulation study of lifting of van der Burg (2003, submitted for publication), joint rotations in distal segments, due to low stiffness in these segments, delay the propagation of the perturbation to the trunk. In the present study, the propagation of the perturbation to the trunk appeared to be delayed, because rotations of the crate, and thus rotations of the upper extremity joints occurred, in both perturbed conditions. An additional effect of joint rotations was especially pronounced in the symmetric condition. Due to the forward rotation of the crate, a part of the mass of crate was longer on the platform, which attenuated the additional crate moment. By preserving the forward rotation of the crate, the moment arm of the crate would have been decreased during the whole lifting movement. However, due to the instructions to the subjects to keep the crate horizontal, subjects quickly corrected the forward rotation, and thus the crate moment was lower only temporarily. After an asymmetric perturbation, the center of mass of the crate was still in the middle of the hands in for-afterward direction. Therefore, no wrist rotations, but especially movements of the shoulder and elbow joints are to be expected. The stiffness of the elbow and shoulder joint is likely to be higher than the stiffness of the wrist, as is indicated by the smaller crate movements after an asymmetric perturbation than after a symmetric perturbation. The higher stability in the sagittal plane than in the frontal plane prior to the perturbation may be explained by the difference in net moments at that instant. The extension moments were increased due to the posture of the subject and likely also due to the anticipation to the lifting movement (van Die.en and de Looze, 1999; Bouisset and Zattara, 1987). The larger net moments increased the stability of the spine in this plane. So, the higher stability may be a positive effect of the task. In addition, the higher stability may also be caused part of by task specific directional stiffening of the spine, which Burdet et al. (2001) described for arm movements.
In contrast to the present findings, Thomas et al. (1998) reported a significantly smaller deviation of the trunk angle after an asymmetric perturbation as compared to a symmetric perturbation. The contrast between both studies might be due to the difference in method of applying the perturbing force. In their study the force was directly applied at the trunk, while in our study the perturbation was applied at the hands. As described above, the arms might reduce and delay the perturbation and thus influence the deviations of the trunk.
5. Conclusion The hypothesized difference between an asymmetric perturbation and a symmetric perturbation was not shown in spite of lower stability in the frontal plane. The kinematics of the trunk was minimally affected both by a symmetric and an asymmetric perturbation of a lifting movement. Likely the stability of the trunk was sufficient to counteract the applied perturbation in both conditions.
Acknowledgements We like to thank Dr. Jacek Cholewicki for his help in estimating spinal stability and Dr. Jim Potvin for the thoughtful comments on draft versions of this manuscript.
References Andersen, T., Bull, Schibye, B., Skotte, J., 2001. Sudden movements of the spinal column during health-care work. International Journal of Industrial Ergonomics 28, 47–53. Bouisset, S., Zattara, M., 1987. Biomechanical study of the programming of anticipatory postural adjustments associated with voluntary movement. Journal of Biomechanics 20, 735–742. Burdet, E., Osu, R., Franklin, D.W., Milner, T.E., Kawato, M., 2001. The central nervous system stabilizes unstable dynamics by learning optimal impedance. Nature 414, 446–449. Chiang, J., Potvin, J.R., 2001. The in vivo dynamic response of the human spine to rapid lateral bend perturbation: effects of preload and step input magnitude. Spine 26, 1457–1464. Cholewicki, J., McGill, S.M., 1996. Mechanical stability of the in vivo lumbar spine: implications for injury and chronic low back pain. Clinical Biomechanics 11, 1–15. Cholewicki, J., Simons, A.P., Radebold, A., 2000. Effects of external trunk loads on lumbar spine stability. Journal of Biomechanics 33, 1377–1385. Cholewicki, J., VanVliet, Iv Jj., 2002. Relative contribution of trunk muscles to the stability of the lumbar spine during isometric exertions. Clinical Biomechanics (Bristol, Avon) 17, 99–105. Commissaris, D.A., Toussaint, H.M., Hirschfeld, H., 2001. Anticipatory postural adjustments in a bimanual, whole-body lifting task seem not only aimed at minimising anterior-posterior centre of mass displacements. Gait Posture 14, 44–55.
ARTICLE IN PRESS J.C.E. van der Burg (Petra) et al. / Journal of Biomechanics 37 (2004) 1071–1077 Commissaris, D.A.C.M., Toussaint, H.M., 1997. Load knowledge affects low-back loading and control of balance in lifting tasks. Ergonomics 40, 559–575. de Looze, M.P., Bussmann, J.B.J., Kingma, I., Toussaint, H.M., 1992. Validation of a dynamic linked segment model to calculate joint moments in lifting. Clinical Biomechanics 7, 161–169. Gagnon, M., Plamondon, A., Gravel, D., 1995. Effects of symmetry and load absorption of a falling load on 3D trunk muscular moments. Ergonomics 38, 1156–1171. Granata, K.P., Wilson, S.E., 2001. Trunk posture and spinal stability. Clinical Biomechanics (Bristol, Avon) 16, 650–659. Hirvonen, M., Leskinen, T., Gronqvist, R., Viikari-Juntura, E.R.A., Riihimaki, H., 1996. Occurence of sudden movements at work. Safety Science 24, 77–82. Hodges, P.W., Cresswell, A.G., Thorstensson, A., 2001. Perturbed upper limb movements cause short-latency postural responses in trunk muscles. Experimental Brain Research 138, 243–250. Hof, A.L., 1992. An explicit expression for the moment in multibody systems. Journal of Biomechanics 25, 1209–1211. Johansson, R.S., Edin, B.B., 1993. Predictive feed-forward sensory control during grasping and manipulation in man. Biomedical Research 14 (Suppl. 4), 95–106. Krajcarski, S.R., Potvin, J.R., Chiang, J., 1999. The in vivo dynamic response of the spine to perturbations causing rapid flexion: effects of pre-load and step input magnitude. Clinical Biomechanics 14, 54–62. Magora, A., 1973. Investigation of the relation between low back pain and occupation. Part IV. Physical requirements: bending, rotation, reaching and sudden maximal effort. Scandinavian Journal of Rehabilitation Medicine 5, 186–190. Manning, D.P., Mitchell, R.G., Blanchfield, L.P., 1984. Body movements and events contributing to accidental and nonaccidental back injuries. Spine 9, 734–739. McGill, S.M., 1991. Electromyographic activity of the abdominal and low back musculature during the generation of isometric and
1077
dynamic axial trunk torque: implications for lumbar mechanics. Journal of Orthopaedic Research 9, 91–103. Panjabi, M.M., 1992. The stabilizing system of the spine. Part I. Function, dysfunction, adaptation, and enhancement. Journal of Spinal Disorders 5, 383–389 discussion 397. Plagenhoef, S., Evans, F.G., Abdelnour, T., 1983. Anatomical data for analyzing human motion. Research Quarterly for Exercise and Sport 54, 169–178. Potvin, J.R., Norman, R.W., McGill, S.M., 1996. Mechanically corrected EMG for the continuous estimation of erector spinae muscle loading during repetitive lifting. European Journal of Applied Physiology 74, 119–132. Redfern, M.S., Hughes, D.B., Chaffin, D.B., 1993. High-pass filtering to remove electrocardiographic interference from torso recordings. Clinical Biomechanics 8, 44–48. Thomas, J.S., Lavender, S.A., Corcos, D.M., Andersson, G.B.J., 1998. Trunk kinematics and trunk muscle activity during a rapidly applied load. Journal of Electromyography and Kinesiology 8, 215–225. van der Burg, J.C.E., 2003. Lifting objects: surprised by the mass. Ph.D. Thesis, Vrije Universiteit. van der Burg, J.C.E., Casius, L.J.R., Kingma, I., van Die.en, J.H., van Soest, A.J., submitted for publication. The contribution of muscle properties in controlling perturbations during lifting. Journal of Biomechanics. van der Burg, J.C.E., van Die.en, J.H., 2001. Underestimation of object mass in lifting does not increase the load on the low back. Journal of Biomechanics 34, 1447–1453. van der Burg, J.C.E., Kingma, I., van Die.en, J.H., 2003. Effects of unexpected lateral mass distribution in lifting on trunk movement and trunk loading. Spine 28, 764–770. van Die.en, J.H., de Looze, M.P., 1999. Directionality of anticipatory activation of trunk muscles in a lifting task depends on load knowledge. Experimental Brain Research 128, 397–404.