Human Factors: The Journal of the Human Factors and Ergonomics Society http://hfs.sagepub.com/
Coactivation of the Trunk Muscles during Asymmetric Loading of the Torso Steven A. Lavender, Yang-Hwei Tsuang, Ali Hafezi, Gunnar B. J. Anderson, Don B. Chaffin and Richard E. Hughes Human Factors: The Journal of the Human Factors and Ergonomics Society 1992 34: 239 DOI: 10.1177/001872089203400209 The online version of this article can be found at: http://hfs.sagepub.com/content/34/2/239
Published by: http://www.sagepublications.com
On behalf of:
Human Factors and Ergonomics Society
Additional services and information for Human Factors: The Journal of the Human Factors and Ergonomics Society can be found at: Email Alerts: http://hfs.sagepub.com/cgi/alerts Subscriptions: http://hfs.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014 Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://hfs.sagepub.com/content/34/2/239.refs.html
>> Version of Record - Apr 1, 1992 What is This?
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
HUMAN
FACTORS,
1992,34(2),239-247
Coactivation of the Trunk Muscles during Asymmetric Loading of the Torso STEVEN A. LAVENDER/ YANG-HWEI TSUANG,2 ALI HAFEZI, and GUNNAR B. J. ANDERSSON, Rush-Presbyterian-St. Luke's Medical Center, and DON B. CHAFFIN and RICHARD E. HUGHES, University of Michigan
Materials handling tasks in industry are rarely performed in the midsagittal plane. Often these tasks, labeled nonsagittally symmetric or asymmetric lifting tasks, can be expected to lead to an unequal distribution of forces between the left and right sides of the body. Because of the large number of muscles capable of resisting loads in the torso, researchers are forced to make simplifications when using biomechanical models to estimate mechanical loading of the spine during such tasks. Simplifications and assumptions regarding the coactivation of antagonistic muscles are frequently used because sufficient experimental data do not exist. The present study was designed to quantify coactivation of the trunk musculature in response to applied asymmetric loads. This load was varied in direction from an anterior midsagittal plane orientation to a posterior midsagittal plane orientation in 15-deg increments. The results showed little coactivation when the applied load directions were anterior and within 45 deg of the midsagittal orientation. When load directions were greater than 45 deg, coactivation was quantifiable in ipsilateral and posterior muscle groups.
INTRODUCTION Occupational low-back injuries cost U.S. industry between $4.5 and $38 billion per year (Andersson, Pope, Frymoyer, and Snook, 1990). Several epidemiological studies have suggested that repetitive lifting, bending, and twisting play a role in the development of low-back injuries (Andersson, 1981, 1991; Bergquist-Ullman and Larsson, 1977; Fry1 Requests for reprints should be sent ender, Department of Orthopedic Presbyterian-St. Luke's Medical Center, Parkway, Chicago, IL 60612. 2 Now at National Taiwan University Taiwan.
to Steven A. LavSurgery, Rush1653 W. Congress Hospital,
Taipei,
moyer, Pope, Clements, Wilder, MacPherson, and Ashikaga, 1983; Magora, 1973; Mitchell, Blanchfield, and Manning, 1983). Still, understanding of the biomechanical factors associated with mechanical loading of the torso is limited. Many studies have described and modeled mechanical loading of the torso under conditions in which the load is applied or resisted in the midsagittal plane (Chaffin and Baker, 1970; Freivalds, Chaffin, Garg, and Lee, 1984; Marras, King, and Joynt, 1984; Reilly and Marras, 1989; Schultz, Andersson, Ortengren, Bjork, and Nordin, 1982). However, when loads from the body itself and/or externally
© 1992, The Human Factors atSociety, Inc.LAVAL All on rights Downloaded from hfs.sagepub.com UNIVERSITE June 16,reserved. 2014
240-April
1992
HUMAN
applied loads are not symmetric in relation to the midsagittal plane, the internal muscle force requirements necessary for stabilizing the torso are not shared evenly between the left and right sides of the body. This type of loading is often referred to as asymmetric loading. The biomechanical consequences of asymmetric loading are not well understood, yet much of the manual materials handling performed in industrial settings could be characterized as asymmetric. Previous research has investigated the strength decrements that occur as loading becomes increasingly asymmetric. For example, Garg and Badger (1986) showed that maximum voluntary isometric strengths and the psychophysically determined maximum acceptable loads decreased by 7% and 10% per 30 deg of asymmetry, respectively. Isokinetic trunk strengths have been shown to decrease at an even higher rate between 0 and 30 deg of asymmetry (Marras and Mirka, 1989). Biomechanical models have been developed to estimate the mechanical loading of the spine during lifting tasks (Chaffin, 1982). Such models emphasize the mechanical disadvantage endured by the musculoskeletal system relative to the point at which an externalload is applied (Basmajian, 1976). This implies that when a load is applied externally, the mechanical forces acting on the spine will be generated primarily by the muscles supporting the torso rather than the load itself. In the case of loadings in the midsagittal plane, the modeling task is simplified. Investigators often assume the muscle forces to be symmetric with regard to the left and right sides of the body, thereby reducing the number of unknown quantities in the model's formulation. Under asymmetric loading conditions the internal force contributions are complex. Biomechanical models of the torso need to account for both the ipsilateral and contralat-
FACTORS
eral muscles with respect to the load position. Seroussi and Pope (1987), for example, demonstrated the sensitivity of four trunk muscles to moments that were experienced during asymmetric loading. Furthermore, their data showed increased activation of the anterior and posterior muscles as the loads became more asymmetric. This simultaneous activation of opposing or antagonistic muscles is commonly referred to as coactivation. The determination of when muscles will be coactivated is closely linked to the function of the coactivation. For example, triphasic coactivation patterns are typically found in ballistic motions, in which the antagonistic muscle is used to decelerate the joint (Meinck, Benecke, Meyer, Hohne, and Conrad, 1984). Under more constrained conditions, coactivation has been found to occur during exertions requiring compensatory force adjustments (De Luca and Mambrito, 1987; Smith, 1981). In isometric tests coactivation has been shown to increase the stiffness of a joint, thereby allowing increased stability (Humphery and Reed, 1983). However, the coactivation of muscles surrounding a joint increases the mechanical loading of the joint. Describing the conditions under which the trunk muscles are coactivated is essential for the development of accurate biomechanical models. Ladin, Murthy, and De Luca (1989) defined switching curves, which predict when trunk muscles should be active in response to static, externally applied moments. However, they did not discuss the amplitudes of muscles coactivated. Marras and Mirka (1992) described the normalized electromyographic (NEMG) response of the trunk muscles under both isometric and isokinetic conditions as trunk rotation was increased to 30 deg. Additional data are needed to describe the muscular coactivation levels with even greater load asymmetries. No attempt was made in these prior investigations to quantify the level of coactivation.
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
April 1992-241
TRUNK MUSCLE COACTIVATION
The objectives of the current study were twofold. First, this study was designed to quantify the activities of eight primary trunk muscles resisting asymmetrically applied loads. Second, we sought to quantify the relative coactivation in the trunk musculature in response to these external loads. METHODS Subjects Eleven healthy male subjects between 20 and 40 years old with no history of back pain were recruited for this study. Subjects were informed of the nature of the experimental task and signed a consent form approved by the Human Investigation Committee at RushPresbyterian-St. Luke's Medical Center. The mean and range of the anthropometric and weight measurements for these subjects were as follows: height, 175.9 cm (165.1-188.9 cm); weight, 78.5 kg (58.2-94.3 kg); trunk breadth, 29.7 cm (18'(~26.0 cm); trunk depth, 21.9 cm (24.0-33.2 cm). Experimental
Design
The experiment was designed to detect electromyographic outputs of eight trunk muscles while the torso was loaded by a force vector of 100 N from variable directions. The directions employed included sagittally symmetric loading (both front and rear) and all asymmetric angles to the subject's right side in 15-deg increments between these two conditions. The experiment employed a randomized block design wherein each subject served as a block and experienced each of the 13 loading conditions in a randomized sequence. The dependent measures in this experiment were the EMG activity in eight of the primary trunk muscles crossing the transverse plane at the L3/L4 level (Schultz and Andersson, 1981). These included the left and right erector spinae (ERSR and ERSL), latis-
simus dorsi (LATL and LATR), external obliques (EXOL and EXOR), and rectus abdomini (RABL and RABR). Apparatus Subjects were fitted with a harness constructed from aluminum plates and webbing material. The harness had three attachment points: front, at the center of the chest at the nipple line; rear, over the spine at the lower tip of the scapula, and side. on a cable connecting the front and rear plates. The subject stood in a reference frame that allowed loading via dynamometers and pulleys, data collection in a specified posture during maximal voluntary contraction (MVC), and constant magnitude loadings. The 100 N loadings were obtained by hanging either a 100 N load or the vector components that would create a 100 N resultant vector in the desired direction on the torso (see Figure 1). The EMG data were collected using surface electrodes. The raw signals were amplified and rectified with a band pass frequency range of between 20 and approximately 600 Hz. From these signals, root mean square (RMS) data were obtained with a sampling frequency of 10 Hz. The RMS data were stored on the hard disk of a personal computer for later analysis. Procedure Upon entering the laboratory for testing, subjects were prepared for surface electromyography by identifying the muscles of interest, cleaning the skin, and applying the electrodes. The harness was adjusted for each individual so as to be at the anatomical landmarks described earlier. The reference frame was adjusted to each individual so that the force vectors would be parallel to the floor. The subject's pelvis was strapped against a board to minimize hip involvement during the testing procedure and provide overall stability. While in a static, upright posture, sub-
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
HUMAN
242-April 1992
FACTORS
VECTOR SUM OF THE TWO WEIGHTS
pelvis support
pulleys
/ SUBJECT
weight
/ weights
Figure 1. Schematic diagrams showing the reference frame and subjects' position relative to the applied loads (left: side view; right: top view). The direction of the resultant load was controlled by adjusting the magnitude of the forces applied in the midsagittal plane (either anterior or posterior) and the midcoronal plane.
jects were asked to provide MVCsin attempting to flex, extend, and twist the torso to the left and right. Following the MVCtrials, data were collected while subjects stood upright for a short period in order to obtain baseline values for each muscle sampled. Testing was done using the same posture while the subject was loaded with 100 N force in the direction specified for the particular trial. The 100 N force vector was created by weights constituting the vector components of the 100 N load in the sagittal and coronal planes which were linked, via cables, to a harness worn by the subjects. EMG data were collected for 10 s on each trial after the subject stabilized his posture under the load. Two-minute rest periods were given between trials. of which there were 13. Data Treatment
The RMS data were averaged over the 10-s data collection period for each muscle in each
condition. Each of these averaged EMG values was normalized relative to the maximal values observed during the MVC trials and the baseline value as follows: NEMGj =
test valuej - resting valuej I . I' max va uej - restmg va uej
(1)
where i = individual muscles 1 to 8, test value = the averaged RMS value in muscle i for each trial; resting value = the baseline value for muscle i obtained while standing upright in apparatus; and max value; = the greatest observed value for muscle i during the four MVCtrials. This normalization procedure assigned negative values to observations with values lower than that observed during the resting posture (standing upright and relaxed). These normalized EMG data were used in the statistical analysis designed to evaluate which muscles' activity changed as a function of load angle. j
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
j
April 1992-243
TRUNK MUSCLE COACTIVATION
RESULTS The NEMG data were first used in a multivariate analysis of variance (MANOVA) designed to evaluate whether the eight trunk muscles were sensitive to the direction of loading. This test was significant using Wilks' criterion (p < 0.001), so follow-up univariate analysis of variance (ANOVA) procedures were performed for each of the eight muscles. These tests indicated that all muscles, with the exception of the LATL, were significantly affected by load direction (p < 0.001 for all muscles). Figures 2 and 3 show the mean levels of NEMG activity across subjects as a function of the applied force direction for the posterior and anterior muscles, respectively. From these figures it can be seen that the primary
Normalized
muscles resisting the load with force directions between 0 and 4S deg were the erector spinae muscles. Within this sector the ERSL's response peaked at 30 deg whereas the ERSR's response continually decreased with increased asymmetry. Conversely, with force directions between 13S and 180 deg, the primary muscles resisting the load were the external oblique muscles and the rectus abdomini muscles. Coactivation of the eight trunk muscles was most apparent with force directions between 60 and 120 deg (see Figures 2 and 3). Coactivation of the trunk muscles was quantified for each applied force direction as follows. First, each muscle's activity level was statistically tested to determine whether the muscle could be characterized as active or inactive for the given force direction. Each
EMG (% MVC)
16 14
""*
12
-B- ERSR
LATR ---b- LATL
~
10
ERSL
8
6 4
2
o -2
o
15
30
45
60
75
90
105 120 135 150 165 180
DIRECTION OF APPLIED FORCE (DEGREES) Figure 2. Mean NEMG activities across subjects for the four posterior trunk muscles as a function of applied force direction. Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
24~April1992
Normalized
HUMAN
FACTORS
EMG (% MVC)
16 14
12 10
-*-
EXOR
-b-
EXOL
-B-
RABR
-+-
RABL
8 6
2
o -2
o
15
30
45
60
75
90
105 120 135 150 165 180
DIRECTION OF APPLIED FORCE (DEGREES) Figure J. Mean NEMG activities across subjects for the four anterior trunk muscles as a function of applied force direction.
one-tailed t test evaluated the muscle's mean activity level, taken across subjects, against the null hypothesis that its mean activity was equal to a value of zero. If the null hypothesis was rejected, the muscle was considered to be active and is represented by a shaded ellipse in Figure 4. In addition, each active muscle's mean value, standard error, and t test probability are presented in the adjacent table. Second, the relative muscle activities were computed for each force direction. Relative activity for a given muscle was defined as that muscle's proportion of the summed muscular activities obtained from the active muscles under the specific force direction condition. These relative levels, expressed as percentages, are shown in the rows marked "percent" in Figure 4. For example, with a
force direction of 90 deg the LATR, LATL, ERSR, ERSL, and the EXOL were determined to be active. By examining the percentages, it can be seen that the primary muscles active under this condition were ERSL, LATR, and EXOL. These muscles represented 29.08%, 27.13%, and 24.63% of the total muscle activity, respectively. Also active, but at a lower level, were LATR (15.45%) and ERSR (3.58%). This analysis provides a means whereby coactivation between active muscles can be quantified; however, the reader is warned that the value of the totaled EMG across muscles is different under each force direction and has no physiologic basis. Even so, this analysis is useful for obtaining information about the relative muscle activations.
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
TRUNK MUSCLE COACTIVATION
ANGLE
CROSS-SECTION
April 1992-245
PARAMETER
LATR
LATL
~
SlD
ERR
0·
.•.• 11,1.
00.
PERCENT
hoIEAN
15
SlO ERR p· •••• lu.
~
•0"
PERCENT
255P
co,
"'EAN
30
510
EAA
023
p'value
~
0" 30.
PERCENT
0.0
•••eAN
45
510
Et:Hl
0'3 026
0· .•.• 1••••
~
PERCENT
3" • O' 0.0
MeAN
60
STO
ERR
p- ••••Iu.
~
0'. • O.
PEACENT
3 11 0 ••
o .e
00' 2112
03' teo
WEAN
75
STD
~
ERR
PERCENT
• 0' "3 00' 2113
MEAN
90
STD
~'ERCENT
105
SrD
~
120
135
ERR
O-v,IIl'
~
H' 0.' 003
•••eAN ERR
PERCEN~
c;;) c;;)
2990
MeAN
sro
ERA
totE""" 510 ERR P'vIIUO
150
STC 0·
~
ERR
,.,
•••• 11.111
hlE"'N
STD
ERR
0' .•. 11101.
~
PERCENT !I.lEAN STO ERR
180
p""llue
~
0 •• 03. 0'.
• ••
'"
1333 33. 003 144\
U.3 3" 00' 82 18 13 e; 3 ;~ 00. 8698 1005 322 00' 8521
072 00. 21 20 &j
o
'C
0" .20
PERCENT
30. 0 •• 00' 2037
'3' 0 ••
0" 030
•• 3 0.'
02'
0" 3 ••
.QOl
,
..
..
032
• o.
0" 003 2' 63
'"
o
2680
. ,.
0.' 00' 3505
."
0>1
•• 0 , O.
, oe
00' 2990
0.3
1,t.
. ,.
00' 3052
5.39
56. I 2'
00' 2509
00' 2838
8.43 I ., 00' 2766
,"
0.' 00' 32 13
, 90
25.5
0"
."
4.82
001
0.21 00.
28
3 ••
0.0 29 De
,
0'. 03. , 03
033 15
'3' , .3
0.3 0'0 363
'"
'" o ••
00' 55 11
o ••
RA8L
,or
'"
0'0 10n
RA8R
00' 5340
." 0" 0" 012 ~ !OS
EXOL
10 .~
0 .• 3
PERCENT
165
, .,
00. 03' 10 ,e
039
, ,.
1 ,.
T
0 •• 00. :!.J!.
0>2 039
0' 3
lotEAN
,,,
o. ,
'33 0"
200 0.' 0" u 23
MUSCLES ERSL EXOR
'"
,,. 033 1266
PERCENT
PERCEN
., 40'. ·..•• •,O.
MEAN
0
ERSR
,"
00' 3\ 66
0"
••• 03< ,001
2,06
0>3 003 1172
..
,
0>1 001
11.20
1363
''3 , O.
''3
00.
I " 0'3 1930
..,., ... ...
1930
• '0 , 0'
00 • 2046
0.3 00 • 2 I 26
,
2\
0'. 74
, O. 00'
1(1 IP
Figure 4. The figures under the heading "cross-section" represent a schematic of a transverse cut to the trunk at the L3 level with the anterior side facing up. The eight ellipses represent the muscles sampled in the current study, and the circle represents the approximate location of the vertebral body. Active muscles for each force direction are indicated by shaded ellipses. Means, standard errors (STD ERR), outcome probabilities for each muscle t test (p-value), and the relative percentages of the totaled NEMG activity for each force direction are presented.
DISCUSSION The present study was developed to quantify coactivation of the trunk musculature as a function of the direction of the external
load. As would be anticipated, the experiment demonstrated that the primary muscle activations in the torso were those of agonist muscles contralateral to the direction of the external load. Quantitative data were also ob-
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
246-April 1992
HUMAN
FACTORS
tained regarding the relative activities of the muscles that were coactivated during these isometric exertions. Coactivation of the eight trunk muscles differed as a function of load direction. The musculoskeletal system can be divided into contralateral and ipsilateral muscles with respect to the midsagittal plane as a function of load orientation. In all experimental conditions, except those with force directions in the midsagittal plane (0 and 180 deg), the
ceeding physiological limitations and inconsistent with experimental data (Bean, Chaffin, and Schultz, 1988). In essence, the problem of describing the muscle tensions is statically indeterminate; a unique solution is not possible without further constraining the model. Historically, investigators have further constrained their models by assuming the muscular coactivation to be negligible (Schultz and Andersson, 1981). Although this simplifies the modeling
muscles on the right side would be classified
effort and allows computation
as ipsilateral muscles. Likewise, under these same experimental conditions, the muscles on the left side would be classified as the contralateral muscles. Any action on the part of the ipsilateral muscles should be considered at least in part to be antagonistic, considering that such action ultimately creates internal moments that must be compensated for by activities in the contralateral muscles. In all conditions in which the applied force was not symmetric with regard to the midsagittal plane (15 deg-165 deg), ipsilateral muscles were active. For example, with the applied force direction between 15 and 60 deg, the continued activation of the ERSR, while providing an additional trunk extension moment, increases the bending moment in the coronal plane, which must be countered by the contralateral musculature. Several authors have discussed this coactivation in terms of the stabilization of the torso. Seroussi and Pope (1987) modeled the lumbar spine as a "multi-degree-of-freedom pivot" wherein the muscles act as guy wires to stabilize the "joint." Biomechanically, this is a difficult situation to model, given that the anatomical arrangement of the trunk muscles allows numerous combinations of muscular activations that are capable of stabilizing the torso and the external load. Furthermore, if not constrained, biomechanical models would primarily select those muscles with the longest moment arms in a manner ex-
lution, the data collected in the current study indicate that such an assumption leads to inaccurate predictions of mechanical loading of the vertebral joints. Bean et al. (1988) used a double linear programming method whereby the predicted spinal compression is obtained by sequentially solving two linear programming problems. The first problem selects feasible solutions that minimize the maximum muscle intensity. The second selects the solution that minimizes compression on the spine. Based on this formulation muscle forces were predicted for a 222 N load with directions ranging from 0 to 90 deg of asymmetry. Qualitative comparison between the prediction of Bean et al. and our data show good correspondence for the posterior muscle activations. However, the anterior muscle activations did not show as good a correspondence. For example, the EXOL was predicted to be active with only 30 deg of asymmetry, whereas our experimental data did not find this muscle to be active until the load direction was at 75 deg. In part, the differences between their model and our observed activations may be attributable to the differences in external load magnitude. We anticipate that the relative activation data presented in this paper will allow the formulation of additional constraints in linear programming models. In this manner negligible coactivation will not have to be assumed because proportional activation con-
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
of a unique so-
April 1992-247
TRUNK MUSCLE COACTIVATION
straints could be employed. Such an approach should alleviate the statically indeterminate problem while increasing the accuracy of the spinal load predictions. The data presented in this paper reflect small loads of approximately 100 N with subjects in upright postures. Additional work is needed to quantify how these coactivation levels change with increased load magnitudes and varying postures. CONCLUSION For the eight trunk muscles sampled, load direction affected coactivation as follows: With load directions of 45 deg or less, coactivation was mainly between the ipsilateral and contralateral posterior muscle groups. With load directions between 60 and 120 deg, muscular coactivation occurred in both the anterior and posterior ipsilateral and contralateral muscles. With load directions between 135 and 180 deg, posterior muscle activations became more variable and less pronounced, thereby limiting coactivation to the four anterior muscles. ACKNOWLEDGMENTS Financial support for this research was provided by NIH Grant AR39350.
REFERENCES Andersson, G. B. J. (1981). Epidemiologic aspects of lowback pain in industry. Spine, 6, 53-60. Andersson, G. B. J. (1991) The epidemiology of spinal disorders. In J. W. Frymoyer (Ed.), The adult spine: Principles and Practice (pp. 107-146). New York: Raven. Andersson, G. B. J., Pope, M. H., Frymoyer, J. W., and Snook, S. H. (1990). Epidemiology and cost. In M. H. Pope, G. B. J. Andersson, M. H. Frymoyer, and D. B. Chaffin (Eds.), Occupational low back pain (pp. 95-113). St. Louis: Mosby. Basmajian, J. V. (1976). Primary anatomy (7th ed.). Baltimore: Williams & Wilkins. Bean, J. C., Chaffin, D. B., and Schultz, A. B. (1988). Biomechanical model calculation of muscle contraction forces: A double linear programming method. Journal of Biomechanics, 21, 59-66. Bergquist-Ullman, M., and Larsson, U. (1977). Acute low back pain in industry. Acta Orthopaedica Scandinavica (Suppl.),170,1-1I7. Chaffin, D. B. (1982). Occupational biomechanics of low back injury. In A. A. White III and S. Gordon (Eds.), American Academy of Orthopaedic Surgeons Symposium
on Idiopathic Low Back Pain (pp. 323-330). St. Louis: Mosby. Chaffin, D. B., and Baker, W. H. (1970). A biomechanical model for analysis of symmetric sagittal plane lifting. AIlE Transactions, 2, 16-27. De Luca, C. J., and Mambrito, B. (1987). Voluntary control of motor units in human antagonist muscles: Coactivation and reciprocal activation. Journal ofNeurophysiology, 58, 525-542. Freivalds, A., Chaffin, D. B., Garg, A., and Lee, K. S. (1984). A dynamic biomechanical evaluation of lifting maximum acceptable loads. Journal of Biomechanics, 17, 251-262. Frymoyer, J. W., Pope, M. H., Clements, J. H., Wilder, D. G., MacPherson, B., and Ashikaga, T. (1983). Risk factors in low-back pain. Journal of Bone and Joint Surgery, 65A, 213-218. Garg, A., and Badger, D. (1986). Maximum acceptable weights and maximum voluntary isometric strengths for asymmetric lifting. Ergonomics, 29, 879-892. Humphery, D. R., and Reed, D. J. (1983). Separate cortical systems for control of joint movement and joint stiff· ness: Reciprocal activation and coactivation of antagonist muscles. In J. E. Desmedt (Ed.), Motor control mechanisms in health and disease (pp. 347-372). New York: Raven. Ladin, Z., Murthy, K. R., and De Luca, C. J. (1989). Me· chanical recruitment of low-back muscles: Theoretical predictions and experimental validation. Spine, 9, 927938. Magora, A. (1973). Investigation of the relation between low back pain and occupation: 4. Physical requirements: Bending, rotation, reaching and sudden maximal effort. ScandanavianJournal of Rehabilitation Medicine,5, 191-196. Marras, W. S., King, A. I., and Joynt, R. 1. (1984). Measurements of load on the lumbar spine under isometric and isokinetic conditions. Spine, 9, 176-187. Marras, W. S., and Mirka, G. (1989). Trunk strength during asymmetric trunk motion. Human Factors, 31, 667677. Marras, W. S., and Mirka, G. (1992). A comprehensive evaluation of trunk response to asymmetric trunk motion. Spine, 17, 318-326. Meinck, H. M., Benecke, R., Meyer, W., Hohne, J., and Conrad, B. (1984). Human ballistic finger flexion: Uncoupling of the three-burst pattern. Experimental Brain Research, 55, 127-133. Mitchell, R. G., Blanchfield, 1. P., and Manning, D. P. (1983). Back pain-Not always due to lifting. Occupational Health, 35, 316-321. Reilly, C. H., and Marras, W. S. (1989). SIMULIFr: A simulation model of human trunk motion. Spine, 14, 5-11. Schultz, A. B., and Andersson, G. B. J. (1981). Analysis of loads on the lumbar spine. Spine, 6, 76-82. Schultz, A., Andersson, G. B. J., Ortengren, R., Bjork, R., and Nordin, M. (1982). Analysis and quantitative myoelectric measurements of loads on the lumbar spine when holding weights in standing postures. Spine, 7, 390--396. Seroussi, R. E., and Pope, M. H. (1987). The relationship between trunk muscle electromyography and lifting moments in the sagittal and frontal planes. Journal of Biomechanics, 20, 135-146. Smith, A. M. (1981). The coactivation of antagonist muscles. Canadian Journal of Physiology and Pharmacology, 59,733-747.
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
Coactivation of the Trunk Muscles during Asymmetric Loading of the Torso Steven A. Lavender, Yang-Hwei Tsuang, Ali Hafezi, Gunnar B. J. Anderson, Don B. Chaffin and Richard E. Hughes Human Factors: The Journal of the Human Factors and Ergonomics Society 1992 34: 239 DOI: 10.1177/001872089203400209 The online version of this article can be found at: http://hfs.sagepub.com/content/34/2/239
Published by: http://www.sagepublications.com
On behalf of:
Human Factors and Ergonomics Society
Additional services and information for Human Factors: The Journal of the Human Factors and Ergonomics Society can be found at: Email Alerts: http://hfs.sagepub.com/cgi/alerts Subscriptions: http://hfs.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014 Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://hfs.sagepub.com/content/34/2/239.refs.html
>> Version of Record - Apr 1, 1992 What is This?
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
HUMAN
FACTORS,
1992,34(2),239-247
Coactivation of the Trunk Muscles during Asymmetric Loading of the Torso STEVEN A. LAVENDER/ YANG-HWEI TSUANG,2 ALI HAFEZI, and GUNNAR B. J. ANDERSSON, Rush-Presbyterian-St. Luke's Medical Center, and DON B. CHAFFIN and RICHARD E. HUGHES, University of Michigan
Materials handling tasks in industry are rarely performed in the midsagittal plane. Often these tasks, labeled nonsagittally symmetric or asymmetric lifting tasks, can be expected to lead to an unequal distribution of forces between the left and right sides of the body. Because of the large number of muscles capable of resisting loads in the torso, researchers are forced to make simplifications when using biomechanical models to estimate mechanical loading of the spine during such tasks. Simplifications and assumptions regarding the coactivation of antagonistic muscles are frequently used because sufficient experimental data do not exist. The present study was designed to quantify coactivation of the trunk musculature in response to applied asymmetric loads. This load was varied in direction from an anterior midsagittal plane orientation to a posterior midsagittal plane orientation in 15-deg increments. The results showed little coactivation when the applied load directions were anterior and within 45 deg of the midsagittal orientation. When load directions were greater than 45 deg, coactivation was quantifiable in ipsilateral and posterior muscle groups.
INTRODUCTION Occupational low-back injuries cost U.S. industry between $4.5 and $38 billion per year (Andersson, Pope, Frymoyer, and Snook, 1990). Several epidemiological studies have suggested that repetitive lifting, bending, and twisting play a role in the development of low-back injuries (Andersson, 1981, 1991; Bergquist-Ullman and Larsson, 1977; Fry1 Requests for reprints should be sent ender, Department of Orthopedic Presbyterian-St. Luke's Medical Center, Parkway, Chicago, IL 60612. 2 Now at National Taiwan University Taiwan.
to Steven A. LavSurgery, Rush1653 W. Congress Hospital,
Taipei,
moyer, Pope, Clements, Wilder, MacPherson, and Ashikaga, 1983; Magora, 1973; Mitchell, Blanchfield, and Manning, 1983). Still, understanding of the biomechanical factors associated with mechanical loading of the torso is limited. Many studies have described and modeled mechanical loading of the torso under conditions in which the load is applied or resisted in the midsagittal plane (Chaffin and Baker, 1970; Freivalds, Chaffin, Garg, and Lee, 1984; Marras, King, and Joynt, 1984; Reilly and Marras, 1989; Schultz, Andersson, Ortengren, Bjork, and Nordin, 1982). However, when loads from the body itself and/or externally
© 1992, The Human Factors atSociety, Inc.LAVAL All on rights Downloaded from hfs.sagepub.com UNIVERSITE June 16,reserved. 2014
240-April
1992
HUMAN
applied loads are not symmetric in relation to the midsagittal plane, the internal muscle force requirements necessary for stabilizing the torso are not shared evenly between the left and right sides of the body. This type of loading is often referred to as asymmetric loading. The biomechanical consequences of asymmetric loading are not well understood, yet much of the manual materials handling performed in industrial settings could be characterized as asymmetric. Previous research has investigated the strength decrements that occur as loading becomes increasingly asymmetric. For example, Garg and Badger (1986) showed that maximum voluntary isometric strengths and the psychophysically determined maximum acceptable loads decreased by 7% and 10% per 30 deg of asymmetry, respectively. Isokinetic trunk strengths have been shown to decrease at an even higher rate between 0 and 30 deg of asymmetry (Marras and Mirka, 1989). Biomechanical models have been developed to estimate the mechanical loading of the spine during lifting tasks (Chaffin, 1982). Such models emphasize the mechanical disadvantage endured by the musculoskeletal system relative to the point at which an externalload is applied (Basmajian, 1976). This implies that when a load is applied externally, the mechanical forces acting on the spine will be generated primarily by the muscles supporting the torso rather than the load itself. In the case of loadings in the midsagittal plane, the modeling task is simplified. Investigators often assume the muscle forces to be symmetric with regard to the left and right sides of the body, thereby reducing the number of unknown quantities in the model's formulation. Under asymmetric loading conditions the internal force contributions are complex. Biomechanical models of the torso need to account for both the ipsilateral and contralat-
FACTORS
eral muscles with respect to the load position. Seroussi and Pope (1987), for example, demonstrated the sensitivity of four trunk muscles to moments that were experienced during asymmetric loading. Furthermore, their data showed increased activation of the anterior and posterior muscles as the loads became more asymmetric. This simultaneous activation of opposing or antagonistic muscles is commonly referred to as coactivation. The determination of when muscles will be coactivated is closely linked to the function of the coactivation. For example, triphasic coactivation patterns are typically found in ballistic motions, in which the antagonistic muscle is used to decelerate the joint (Meinck, Benecke, Meyer, Hohne, and Conrad, 1984). Under more constrained conditions, coactivation has been found to occur during exertions requiring compensatory force adjustments (De Luca and Mambrito, 1987; Smith, 1981). In isometric tests coactivation has been shown to increase the stiffness of a joint, thereby allowing increased stability (Humphery and Reed, 1983). However, the coactivation of muscles surrounding a joint increases the mechanical loading of the joint. Describing the conditions under which the trunk muscles are coactivated is essential for the development of accurate biomechanical models. Ladin, Murthy, and De Luca (1989) defined switching curves, which predict when trunk muscles should be active in response to static, externally applied moments. However, they did not discuss the amplitudes of muscles coactivated. Marras and Mirka (1992) described the normalized electromyographic (NEMG) response of the trunk muscles under both isometric and isokinetic conditions as trunk rotation was increased to 30 deg. Additional data are needed to describe the muscular coactivation levels with even greater load asymmetries. No attempt was made in these prior investigations to quantify the level of coactivation.
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
April 1992-241
TRUNK MUSCLE COACTIVATION
The objectives of the current study were twofold. First, this study was designed to quantify the activities of eight primary trunk muscles resisting asymmetrically applied loads. Second, we sought to quantify the relative coactivation in the trunk musculature in response to these external loads. METHODS Subjects Eleven healthy male subjects between 20 and 40 years old with no history of back pain were recruited for this study. Subjects were informed of the nature of the experimental task and signed a consent form approved by the Human Investigation Committee at RushPresbyterian-St. Luke's Medical Center. The mean and range of the anthropometric and weight measurements for these subjects were as follows: height, 175.9 cm (165.1-188.9 cm); weight, 78.5 kg (58.2-94.3 kg); trunk breadth, 29.7 cm (18'(~26.0 cm); trunk depth, 21.9 cm (24.0-33.2 cm). Experimental
Design
The experiment was designed to detect electromyographic outputs of eight trunk muscles while the torso was loaded by a force vector of 100 N from variable directions. The directions employed included sagittally symmetric loading (both front and rear) and all asymmetric angles to the subject's right side in 15-deg increments between these two conditions. The experiment employed a randomized block design wherein each subject served as a block and experienced each of the 13 loading conditions in a randomized sequence. The dependent measures in this experiment were the EMG activity in eight of the primary trunk muscles crossing the transverse plane at the L3/L4 level (Schultz and Andersson, 1981). These included the left and right erector spinae (ERSR and ERSL), latis-
simus dorsi (LATL and LATR), external obliques (EXOL and EXOR), and rectus abdomini (RABL and RABR). Apparatus Subjects were fitted with a harness constructed from aluminum plates and webbing material. The harness had three attachment points: front, at the center of the chest at the nipple line; rear, over the spine at the lower tip of the scapula, and side. on a cable connecting the front and rear plates. The subject stood in a reference frame that allowed loading via dynamometers and pulleys, data collection in a specified posture during maximal voluntary contraction (MVC), and constant magnitude loadings. The 100 N loadings were obtained by hanging either a 100 N load or the vector components that would create a 100 N resultant vector in the desired direction on the torso (see Figure 1). The EMG data were collected using surface electrodes. The raw signals were amplified and rectified with a band pass frequency range of between 20 and approximately 600 Hz. From these signals, root mean square (RMS) data were obtained with a sampling frequency of 10 Hz. The RMS data were stored on the hard disk of a personal computer for later analysis. Procedure Upon entering the laboratory for testing, subjects were prepared for surface electromyography by identifying the muscles of interest, cleaning the skin, and applying the electrodes. The harness was adjusted for each individual so as to be at the anatomical landmarks described earlier. The reference frame was adjusted to each individual so that the force vectors would be parallel to the floor. The subject's pelvis was strapped against a board to minimize hip involvement during the testing procedure and provide overall stability. While in a static, upright posture, sub-
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
HUMAN
242-April 1992
FACTORS
VECTOR SUM OF THE TWO WEIGHTS
pelvis support
pulleys
/ SUBJECT
weight
/ weights
Figure 1. Schematic diagrams showing the reference frame and subjects' position relative to the applied loads (left: side view; right: top view). The direction of the resultant load was controlled by adjusting the magnitude of the forces applied in the midsagittal plane (either anterior or posterior) and the midcoronal plane.
jects were asked to provide MVCsin attempting to flex, extend, and twist the torso to the left and right. Following the MVCtrials, data were collected while subjects stood upright for a short period in order to obtain baseline values for each muscle sampled. Testing was done using the same posture while the subject was loaded with 100 N force in the direction specified for the particular trial. The 100 N force vector was created by weights constituting the vector components of the 100 N load in the sagittal and coronal planes which were linked, via cables, to a harness worn by the subjects. EMG data were collected for 10 s on each trial after the subject stabilized his posture under the load. Two-minute rest periods were given between trials. of which there were 13. Data Treatment
The RMS data were averaged over the 10-s data collection period for each muscle in each
condition. Each of these averaged EMG values was normalized relative to the maximal values observed during the MVC trials and the baseline value as follows: NEMGj =
test valuej - resting valuej I . I' max va uej - restmg va uej
(1)
where i = individual muscles 1 to 8, test value = the averaged RMS value in muscle i for each trial; resting value = the baseline value for muscle i obtained while standing upright in apparatus; and max value; = the greatest observed value for muscle i during the four MVCtrials. This normalization procedure assigned negative values to observations with values lower than that observed during the resting posture (standing upright and relaxed). These normalized EMG data were used in the statistical analysis designed to evaluate which muscles' activity changed as a function of load angle. j
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
j
April 1992-243
TRUNK MUSCLE COACTIVATION
RESULTS The NEMG data were first used in a multivariate analysis of variance (MANOVA) designed to evaluate whether the eight trunk muscles were sensitive to the direction of loading. This test was significant using Wilks' criterion (p < 0.001), so follow-up univariate analysis of variance (ANOVA) procedures were performed for each of the eight muscles. These tests indicated that all muscles, with the exception of the LATL, were significantly affected by load direction (p < 0.001 for all muscles). Figures 2 and 3 show the mean levels of NEMG activity across subjects as a function of the applied force direction for the posterior and anterior muscles, respectively. From these figures it can be seen that the primary
Normalized
muscles resisting the load with force directions between 0 and 4S deg were the erector spinae muscles. Within this sector the ERSL's response peaked at 30 deg whereas the ERSR's response continually decreased with increased asymmetry. Conversely, with force directions between 13S and 180 deg, the primary muscles resisting the load were the external oblique muscles and the rectus abdomini muscles. Coactivation of the eight trunk muscles was most apparent with force directions between 60 and 120 deg (see Figures 2 and 3). Coactivation of the trunk muscles was quantified for each applied force direction as follows. First, each muscle's activity level was statistically tested to determine whether the muscle could be characterized as active or inactive for the given force direction. Each
EMG (% MVC)
16 14
""*
12
-B- ERSR
LATR ---b- LATL
~
10
ERSL
8
6 4
2
o -2
o
15
30
45
60
75
90
105 120 135 150 165 180
DIRECTION OF APPLIED FORCE (DEGREES) Figure 2. Mean NEMG activities across subjects for the four posterior trunk muscles as a function of applied force direction. Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
24~April1992
Normalized
HUMAN
FACTORS
EMG (% MVC)
16 14
12 10
-*-
EXOR
-b-
EXOL
-B-
RABR
-+-
RABL
8 6
2
o -2
o
15
30
45
60
75
90
105 120 135 150 165 180
DIRECTION OF APPLIED FORCE (DEGREES) Figure J. Mean NEMG activities across subjects for the four anterior trunk muscles as a function of applied force direction.
one-tailed t test evaluated the muscle's mean activity level, taken across subjects, against the null hypothesis that its mean activity was equal to a value of zero. If the null hypothesis was rejected, the muscle was considered to be active and is represented by a shaded ellipse in Figure 4. In addition, each active muscle's mean value, standard error, and t test probability are presented in the adjacent table. Second, the relative muscle activities were computed for each force direction. Relative activity for a given muscle was defined as that muscle's proportion of the summed muscular activities obtained from the active muscles under the specific force direction condition. These relative levels, expressed as percentages, are shown in the rows marked "percent" in Figure 4. For example, with a
force direction of 90 deg the LATR, LATL, ERSR, ERSL, and the EXOL were determined to be active. By examining the percentages, it can be seen that the primary muscles active under this condition were ERSL, LATR, and EXOL. These muscles represented 29.08%, 27.13%, and 24.63% of the total muscle activity, respectively. Also active, but at a lower level, were LATR (15.45%) and ERSR (3.58%). This analysis provides a means whereby coactivation between active muscles can be quantified; however, the reader is warned that the value of the totaled EMG across muscles is different under each force direction and has no physiologic basis. Even so, this analysis is useful for obtaining information about the relative muscle activations.
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
TRUNK MUSCLE COACTIVATION
ANGLE
CROSS-SECTION
April 1992-245
PARAMETER
LATR
LATL
~
SlD
ERR
0·
.•.• 11,1.
00.
PERCENT
hoIEAN
15
SlO ERR p· •••• lu.
~
•0"
PERCENT
255P
co,
"'EAN
30
510
EAA
023
p'value
~
0" 30.
PERCENT
0.0
•••eAN
45
510
Et:Hl
0'3 026
0· .•.• 1••••
~
PERCENT
3" • O' 0.0
MeAN
60
STO
ERR
p- ••••Iu.
~
0'. • O.
PEACENT
3 11 0 ••
o .e
00' 2112
03' teo
WEAN
75
STD
~
ERR
PERCENT
• 0' "3 00' 2113
MEAN
90
STD
~'ERCENT
105
SrD
~
120
135
ERR
O-v,IIl'
~
H' 0.' 003
•••eAN ERR
PERCEN~
c;;) c;;)
2990
MeAN
sro
ERA
totE""" 510 ERR P'vIIUO
150
STC 0·
~
ERR
,.,
•••• 11.111
hlE"'N
STD
ERR
0' .•. 11101.
~
PERCENT !I.lEAN STO ERR
180
p""llue
~
0 •• 03. 0'.
• ••
'"
1333 33. 003 144\
U.3 3" 00' 82 18 13 e; 3 ;~ 00. 8698 1005 322 00' 8521
072 00. 21 20 &j
o
'C
0" .20
PERCENT
30. 0 •• 00' 2037
'3' 0 ••
0" 030
•• 3 0.'
02'
0" 3 ••
.QOl
,
..
..
032
• o.
0" 003 2' 63
'"
o
2680
. ,.
0.' 00' 3505
."
0>1
•• 0 , O.
, oe
00' 2990
0.3
1,t.
. ,.
00' 3052
5.39
56. I 2'
00' 2509
00' 2838
8.43 I ., 00' 2766
,"
0.' 00' 32 13
, 90
25.5
0"
."
4.82
001
0.21 00.
28
3 ••
0.0 29 De
,
0'. 03. , 03
033 15
'3' , .3
0.3 0'0 363
'"
'" o ••
00' 55 11
o ••
RA8L
,or
'"
0'0 10n
RA8R
00' 5340
." 0" 0" 012 ~ !OS
EXOL
10 .~
0 .• 3
PERCENT
165
, .,
00. 03' 10 ,e
039
, ,.
1 ,.
T
0 •• 00. :!.J!.
0>2 039
0' 3
lotEAN
,,,
o. ,
'33 0"
200 0.' 0" u 23
MUSCLES ERSL EXOR
'"
,,. 033 1266
PERCENT
PERCEN
., 40'. ·..•• •,O.
MEAN
0
ERSR
,"
00' 3\ 66
0"
••• 03< ,001
2,06
0>3 003 1172
..
,
0>1 001
11.20
1363
''3 , O.
''3
00.
I " 0'3 1930
..,., ... ...
1930
• '0 , 0'
00 • 2046
0.3 00 • 2 I 26
,
2\
0'. 74
, O. 00'
1(1 IP
Figure 4. The figures under the heading "cross-section" represent a schematic of a transverse cut to the trunk at the L3 level with the anterior side facing up. The eight ellipses represent the muscles sampled in the current study, and the circle represents the approximate location of the vertebral body. Active muscles for each force direction are indicated by shaded ellipses. Means, standard errors (STD ERR), outcome probabilities for each muscle t test (p-value), and the relative percentages of the totaled NEMG activity for each force direction are presented.
DISCUSSION The present study was developed to quantify coactivation of the trunk musculature as a function of the direction of the external
load. As would be anticipated, the experiment demonstrated that the primary muscle activations in the torso were those of agonist muscles contralateral to the direction of the external load. Quantitative data were also ob-
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
246-April 1992
HUMAN
FACTORS
tained regarding the relative activities of the muscles that were coactivated during these isometric exertions. Coactivation of the eight trunk muscles differed as a function of load direction. The musculoskeletal system can be divided into contralateral and ipsilateral muscles with respect to the midsagittal plane as a function of load orientation. In all experimental conditions, except those with force directions in the midsagittal plane (0 and 180 deg), the
ceeding physiological limitations and inconsistent with experimental data (Bean, Chaffin, and Schultz, 1988). In essence, the problem of describing the muscle tensions is statically indeterminate; a unique solution is not possible without further constraining the model. Historically, investigators have further constrained their models by assuming the muscular coactivation to be negligible (Schultz and Andersson, 1981). Although this simplifies the modeling
muscles on the right side would be classified
effort and allows computation
as ipsilateral muscles. Likewise, under these same experimental conditions, the muscles on the left side would be classified as the contralateral muscles. Any action on the part of the ipsilateral muscles should be considered at least in part to be antagonistic, considering that such action ultimately creates internal moments that must be compensated for by activities in the contralateral muscles. In all conditions in which the applied force was not symmetric with regard to the midsagittal plane (15 deg-165 deg), ipsilateral muscles were active. For example, with the applied force direction between 15 and 60 deg, the continued activation of the ERSR, while providing an additional trunk extension moment, increases the bending moment in the coronal plane, which must be countered by the contralateral musculature. Several authors have discussed this coactivation in terms of the stabilization of the torso. Seroussi and Pope (1987) modeled the lumbar spine as a "multi-degree-of-freedom pivot" wherein the muscles act as guy wires to stabilize the "joint." Biomechanically, this is a difficult situation to model, given that the anatomical arrangement of the trunk muscles allows numerous combinations of muscular activations that are capable of stabilizing the torso and the external load. Furthermore, if not constrained, biomechanical models would primarily select those muscles with the longest moment arms in a manner ex-
lution, the data collected in the current study indicate that such an assumption leads to inaccurate predictions of mechanical loading of the vertebral joints. Bean et al. (1988) used a double linear programming method whereby the predicted spinal compression is obtained by sequentially solving two linear programming problems. The first problem selects feasible solutions that minimize the maximum muscle intensity. The second selects the solution that minimizes compression on the spine. Based on this formulation muscle forces were predicted for a 222 N load with directions ranging from 0 to 90 deg of asymmetry. Qualitative comparison between the prediction of Bean et al. and our data show good correspondence for the posterior muscle activations. However, the anterior muscle activations did not show as good a correspondence. For example, the EXOL was predicted to be active with only 30 deg of asymmetry, whereas our experimental data did not find this muscle to be active until the load direction was at 75 deg. In part, the differences between their model and our observed activations may be attributable to the differences in external load magnitude. We anticipate that the relative activation data presented in this paper will allow the formulation of additional constraints in linear programming models. In this manner negligible coactivation will not have to be assumed because proportional activation con-
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
of a unique so-
April 1992-247
TRUNK MUSCLE COACTIVATION
straints could be employed. Such an approach should alleviate the statically indeterminate problem while increasing the accuracy of the spinal load predictions. The data presented in this paper reflect small loads of approximately 100 N with subjects in upright postures. Additional work is needed to quantify how these coactivation levels change with increased load magnitudes and varying postures. CONCLUSION For the eight trunk muscles sampled, load direction affected coactivation as follows: With load directions of 45 deg or less, coactivation was mainly between the ipsilateral and contralateral posterior muscle groups. With load directions between 60 and 120 deg, muscular coactivation occurred in both the anterior and posterior ipsilateral and contralateral muscles. With load directions between 135 and 180 deg, posterior muscle activations became more variable and less pronounced, thereby limiting coactivation to the four anterior muscles. ACKNOWLEDGMENTS Financial support for this research was provided by NIH Grant AR39350.
REFERENCES Andersson, G. B. J. (1981). Epidemiologic aspects of lowback pain in industry. Spine, 6, 53-60. Andersson, G. B. J. (1991) The epidemiology of spinal disorders. In J. W. Frymoyer (Ed.), The adult spine: Principles and Practice (pp. 107-146). New York: Raven. Andersson, G. B. J., Pope, M. H., Frymoyer, J. W., and Snook, S. H. (1990). Epidemiology and cost. In M. H. Pope, G. B. J. Andersson, M. H. Frymoyer, and D. B. Chaffin (Eds.), Occupational low back pain (pp. 95-113). St. Louis: Mosby. Basmajian, J. V. (1976). Primary anatomy (7th ed.). Baltimore: Williams & Wilkins. Bean, J. C., Chaffin, D. B., and Schultz, A. B. (1988). Biomechanical model calculation of muscle contraction forces: A double linear programming method. Journal of Biomechanics, 21, 59-66. Bergquist-Ullman, M., and Larsson, U. (1977). Acute low back pain in industry. Acta Orthopaedica Scandinavica (Suppl.),170,1-1I7. Chaffin, D. B. (1982). Occupational biomechanics of low back injury. In A. A. White III and S. Gordon (Eds.), American Academy of Orthopaedic Surgeons Symposium
on Idiopathic Low Back Pain (pp. 323-330). St. Louis: Mosby. Chaffin, D. B., and Baker, W. H. (1970). A biomechanical model for analysis of symmetric sagittal plane lifting. AIlE Transactions, 2, 16-27. De Luca, C. J., and Mambrito, B. (1987). Voluntary control of motor units in human antagonist muscles: Coactivation and reciprocal activation. Journal ofNeurophysiology, 58, 525-542. Freivalds, A., Chaffin, D. B., Garg, A., and Lee, K. S. (1984). A dynamic biomechanical evaluation of lifting maximum acceptable loads. Journal of Biomechanics, 17, 251-262. Frymoyer, J. W., Pope, M. H., Clements, J. H., Wilder, D. G., MacPherson, B., and Ashikaga, T. (1983). Risk factors in low-back pain. Journal of Bone and Joint Surgery, 65A, 213-218. Garg, A., and Badger, D. (1986). Maximum acceptable weights and maximum voluntary isometric strengths for asymmetric lifting. Ergonomics, 29, 879-892. Humphery, D. R., and Reed, D. J. (1983). Separate cortical systems for control of joint movement and joint stiff· ness: Reciprocal activation and coactivation of antagonist muscles. In J. E. Desmedt (Ed.), Motor control mechanisms in health and disease (pp. 347-372). New York: Raven. Ladin, Z., Murthy, K. R., and De Luca, C. J. (1989). Me· chanical recruitment of low-back muscles: Theoretical predictions and experimental validation. Spine, 9, 927938. Magora, A. (1973). Investigation of the relation between low back pain and occupation: 4. Physical requirements: Bending, rotation, reaching and sudden maximal effort. ScandanavianJournal of Rehabilitation Medicine,5, 191-196. Marras, W. S., King, A. I., and Joynt, R. 1. (1984). Measurements of load on the lumbar spine under isometric and isokinetic conditions. Spine, 9, 176-187. Marras, W. S., and Mirka, G. (1989). Trunk strength during asymmetric trunk motion. Human Factors, 31, 667677. Marras, W. S., and Mirka, G. (1992). A comprehensive evaluation of trunk response to asymmetric trunk motion. Spine, 17, 318-326. Meinck, H. M., Benecke, R., Meyer, W., Hohne, J., and Conrad, B. (1984). Human ballistic finger flexion: Uncoupling of the three-burst pattern. Experimental Brain Research, 55, 127-133. Mitchell, R. G., Blanchfield, 1. P., and Manning, D. P. (1983). Back pain-Not always due to lifting. Occupational Health, 35, 316-321. Reilly, C. H., and Marras, W. S. (1989). SIMULIFr: A simulation model of human trunk motion. Spine, 14, 5-11. Schultz, A. B., and Andersson, G. B. J. (1981). Analysis of loads on the lumbar spine. Spine, 6, 76-82. Schultz, A., Andersson, G. B. J., Ortengren, R., Bjork, R., and Nordin, M. (1982). Analysis and quantitative myoelectric measurements of loads on the lumbar spine when holding weights in standing postures. Spine, 7, 390--396. Seroussi, R. E., and Pope, M. H. (1987). The relationship between trunk muscle electromyography and lifting moments in the sagittal and frontal planes. Journal of Biomechanics, 20, 135-146. Smith, A. M. (1981). The coactivation of antagonist muscles. Canadian Journal of Physiology and Pharmacology, 59,733-747.
Downloaded from hfs.sagepub.com at UNIVERSITE LAVAL on June 16, 2014
