Main force directions of trunk muscles: A pilot study in healthy male subjects

Human Movement Science 60 (2018) 214–224

Contents lists available at ScienceDirect

Human Movement Science journal homepage: www.elsevier.com/locate/humov

Full Length Article

Main force directions of trunk muscles: A pilot study in healthy male subjects

T

⁎

Christoph Andersa, , Beatrice Steinigerb a

Clinic for Trauma, Hand and Reconstructive Surgery, Division of Motor Research, Pathophysiology and Biomechanics, Jena University Hospital, Bachstrasse 18, 07743 Jena, Germany Institute of Diagnostic and Interventional Radiology, Jena University Hospital, Am Klinikum 1, 07747 Jena, Germany

b

A R T IC LE I N F O

ABS TRA CT

Keywords: Trunk muscles Main force directions Human Surface EMG

Muscles work most effectively along their anatomically defined action vector(s) which has implications in training and therapeutics. Action vectors can easily be identified in extremity muscles and smaller muscles of the trunk, but are less clear in larger trunk muscles. Trunk muscle exercises and diagnostics have traditionally relied on tasks in the sagittal plane – a practice that is being reconsidered. Therefore, this study aimed at identifying main force directions (MFDs) of major trunk muscles expressed in terms of deviation from the sagittal plane. 20 healthy male subjects underwent graded isometric submaximal static load applications on their trunk by application of simultaneous and incremental tilting and rotating from vertical to horizontal at rotational angles of 45° starting from 0° (forward tilting) around 360° with only the lower body secured. Surface EMG (SEMG) from six trunk muscles on each body side was recorded. The MFD of each trunk muscle was estimated by considering SEMG amplitudes of all rotational angles, separately for all tilt angles, and was expressed as angular deviation from sagittal plane. The calculated MFDs of trunk muscles deviated from sagittal plane to differing extents. Mean MFD angle was smallest (more parallel to sagittal plane) for rectus abdominis muscle ( ± 14°), becoming more lateral for external oblique (OE, ± 32°) and internal oblique abdominal muscles (OI, ± 47°). As tilt angle increased, MFD angles increased for OE, but decreased for OI. Iliocostalis muscle showed an almost laterally directed MFD with systematic dependency on body side (−90° for left and +75° for right side). Both paravertebral muscles (longissimus and multifidus muscles) showed almost identical MFD angles of about ± 145° and varied the least with tilt angle. All trunk muscles’ MFDs deviate from sagittal plane and, in addition to flexing and extending, have both bending and/or rotational capabilities. MFDs of oblique abdominal muscles are systematically altered by tilt angle in accordance with their more divergent fiber directionality. The results provide a basis for specifically targeted diagnostics and training of trunk muscles.

1. Introduction Muscular force is most effective along the main force vector of a muscle, muscle part, or muscle group. These vectors are defined as the main force directions and are described in relation to body perspective. For most muscles, main force directions (MFDs) are defined anatomically by fiber direction, anatomical location, and insertion points of the respective muscles (Gray, 1942; Kendall,

⁎

Corresponding author. E-mail addresses: [email protected] (C. Anders), [email protected] (B. Steiniger).

https://doi.org/10.1016/j.humov.2018.06.012 Received 23 October 2014; Received in revised form 21 June 2018; Accepted 22 June 2018

Available online 20 July 2018 0167-9457/ © 2018 Elsevier B.V. All rights reserved.

Human Movement Science 60 (2018) 214–224

C. Anders, B. Steiniger

Kendall Mc Geary, Provance, Rodgers, & Romani, 2005), best seen in the extremity muscles and smaller muscles of the trunk as they have clearly localized insertion spots. In the case of the larger muscles of the trunk and especially abdominal trunk muscles, the MFD picture becomes less clear because these muscles have multiple insertions that may span over large distances as well as divergent fiber directions. Because of this, their mechanical action cannot be defined as accurately through an anatomical study (Gray, 1942; Kendall et al., 2005). Correspondingly, large trunk muscles are also activated less predictably in terms of a main direction, dependent upon load level and angle, which could have implications therapeutically and diagnostically. Strengthening and therapeutic exercises for the trunk have traditionally focused on tasks in sagittal and / or frontal plane, a practice that precludes more global tasks to improve muscle coordination and provide better spinal support (Ekstrom, Donatelli, & Carp, 2007; Willett, Hyde, Uhrlaub, Wendel, & Karst, 2001). New methodological approaches and normative data are needed to experiment in the development of functionally more adequate improvements of diagnostic methods and biofeedback control. To get a better picture of how the trunk muscles manipulate and support the spine, extensive work has been done, starting in the late 1950’s and still continuing to date, aimed at quantifying trunk muscle’s effort during specific tasks (Bressel, Dolny, & Gibbons, 2011; Godfrey, Kindig, & Windell, 1977; Guimaraes, Vaz, De Campos, & Marantes, 1991; Juker, McGill, Kropf, & Steffen, 1998; Lipetz & Gutin, 1970; Noble, 1981; Walters & Partridge, 1957; Willett et al., 2001), or – the other way around – to identify tasks that activate the respective muscles most effectively (Shamsi, Sarrafzadeh, Jamshidi, Zarabi, & Pourahmadi, 2016). Several studies have already tried to systematically examine trunk muscle function with respect to activation characteristics by applying specifically directed force vectors on the trunk (Lavender, Tsuang, Andersson, Hafezi, & Shin, 1992; Perez & Nussbaum, 2002). These latter investigations aimed at the identification of co-contraction characteristics of trunk muscles when force vectors were altered systematically during bending to the side and twisting. However, the individual action vectors of trunk muscles were not yet examined during loading from different angles. The aim of our study was therefore to add to the existing knowledge of large trunk muscle activation by determination of main force directions using surface EMG: we precisely defined MFDs by using surface EMG signals measured at the SENIAM recommended spots (Hermens et al., 1999; Ng, Kippers, & Richardson, 1998; SENIAM) of individual trunk muscles during static loading at different angles. In addition we asked, if MFD directions were influenced by load level and body side. To this end we applied graded isometric submaximal forces on the trunk in transversal plane by both tilting and rotating the trunk, held erect with lower body fixed. Increased understanding of the most effective force direction of trunk muscles could inform general training methods, functional diagnostics, and targeted rehabilitation approaches. 2. Methods For this study 22 healthy male subjects from the university campus were recruited (subject data are given in Table 1). They underwent a brief clinical examination and were questioned about their medical history to exclude possible relevant orthopedic pathologies like chronic or acute back pain, past injuries or surgeries of the spine. Two subjects had to be excluded. As part of a larger study, ethical approval from the local ethics committee was given (0558-11/00). Every participant was informed prior to the study and gave his written informed consent. As a first fundamental approach, this investigation was conducted on slender healthy male subjects to avoid sex-related and adipose related variations. 2.1. Device The investigation was performed in a computerized test and training device (CTT Centaur, BfMC, Germany) enabling the application of graded forces on the trunk by incrementally tilting the person to horizontal position (Fig. 1). In addition to whole body tilt in sagittal plane, the device can rotate, enabling application of lateral force vectors. This meant that the load applications had both a horizontal and a vertical component: A certain “tilt angle” imposes a specific angle (degrees) as a measure of load (i.e. moment along the gravitational field). The term “rotation angle” describes the load vectors’ directionality in three-dimensional space. In the device, the person’s lower body was fixed while the upper body above the iliac crest maintained its freedom of movement. Subjects, while positioned in the device, had to counteract simultaneous tilt and rotation angles, during which their only task was to maintain erect body position. The device was equipped with an open harness positioned over the subject’s shoulders. Strain gauges together with a crosshair display, visible to the subjects, provided biofeedback to ensure erect body position: as long as the biofeedback display’s control point remained in the center of the crosshair no force was being applied to the harness and erect body position was achieved and maintained. 2.2. Procedure Subjects were tilted to tilt angles of 5°, 10°, 20°, 30°, 45°, 60°, and 90° that are equal to portions of their upper body weight of 9%, Table 1 Subject data (mean values ± SD). Age [years]

Weight [kg]

Height [cm]

BMI [kg/m2]

36.0 ± 7.5

75.1 ± 8.3

180 ± 5.9

23.1 ± 1.9

215

Human Movement Science 60 (2018) 214–224

C. Anders, B. Steiniger

Fig. 1. Subject performing a 20° tilt at −135° rotation task in the CTT Centaur. As long as no force is applied to the harness the control point in the feedback monitor in front of the person remains in the center of the crosshairs.

17%, 34%, 50%, 71%, 87%, and 100%. During tilting, rotation angles were also applied with a resolution of 45° for a complete 360° rotation. By definition, left-sided rotation angles were indicated with a negative sign and right sided rotation angles with a positive sign, with 0° indicating straight forward and 180° referring to straight backward. The predetermined rotation and tilt combinations resulted in 56 load situations. All were applied isometrically, i.e. during taxiing subjects remained relaxed, but after reaching the next test position they were asked to hold their upper body away from the harness and remain in an upright position for approximately 10 s. This was controlled by the investigator. Between test positions the subjects were taxied to upright position to rest for at least 30 s. The 90° tilts at all rotation angles, the most strenuous submaximal situations, were performed at the beginning and held for only approximately 5 s. All other load situations were randomized individually to control for systematic order dependent effects and fatigue. 2.3. Surface EMG measurements Surface EMG (sEMG) utilizing a standard bipolar montage was simultaneously taken from six paired trunk muscles on both body sides (12 in total): rectus abdominis muscle (RA), external and internal oblique abdominal muscles (EO, IO), multifidus muscle (MF), longissimus muscle (LO), and ilicostalis muscle (ICO, see Table 2). Electrode positions were chosen according to international recommendations (Hermens et al., 1999; SENIAM), and if not available, according to Ng et al. (1998). Electrode positions were all identified by the same experienced examiner (CA) and marked. These positions were then cleaned with abrasive paste and shaved if Table 2 SEMG Electrode Positions. Muscle

Electrode localization/orientation

Rectus abdominis (RA) Obliquus internus (OI) Obliquus externus (OE) Iliocostalis (ICO) Longissimus (LO) Multifidus (MF) ECG

caudal electrode at umbilical height, 4 cm from midline, vertical medial from inguinal ligament, at anterior superior iliac spine (ASIS) height, horizontal cranial electrode directly below lowest point of the costal arch, on line from there to contralateral pubic tubercle cranial electrode at L2 height, medial from line from SIPS to lowest point of the costal arch caudal electrode at L1 height, over palpable bulge of muscle (approx. 2 fingers lateral from midline), vertical caudal electrode at L5 height, 1 cm medial and parallel to line between posterior superior iliac spine line (SIPS) and L1 along heart axis, above heart

216

Human Movement Science 60 (2018) 214–224

C. Anders, B. Steiniger

a)

b)

c)

rms [μV] 0° 140

120

120 -45°

+

80 rms [μV]

80 60 40 20 -90°

-

45°

100

d)

40 0

90°

0

-120

-80

-40

0

40

80

120

-40 -80 -135°

135° -120 180°

rms [μV]

Fig. 2. Schematic illustration of how the main force directions (MFD) were determined (example: OE at 90° tilt angle): a) polar plot of gathered rms amplitudes for all applied rotation angles; b) transfer of these data into Cartesian co-ordinates, including determination of the centroid position (open dot); c) calculation of the MFD as is the angle between zero and the position of the centroid; d) mirroring of the gathered angle along frontal plane to get MFD angles from body perspective (subplots a to c were related to the rotation angles).

necessary for noise reduction (Hermens, Freriks, Disselhorst-Klug, & Rau, 2000). Disposable solid gel Ag-AgCl electrodes (H93SG, Covidien) with a circular uptake area of 2 cm2 and an inter-electrode distance of 2.5 cm were attached. Also the force channels in sagittal and frontal planes were recorded to ensure subjects were maintaining the desired erect posture. In addition, one channel along heart axis was applied and measured cardiac activity for subsequent artifact elimination. Electrodes were connected to amplifiers (Biovision: gain 1000, −3 dB at 10 and 700 Hz, CMRR: 120 dB, input impedance: 12 GΩ) and these analog signals were then converted into digital ones (Tower of Measurement, DeMeTec) at 2048 samples/s with a resolution of 24 bit (anti-aliasing filter at 1024 Hz). Data were stored on hard drive (ATISArec, GJB) for further off-line analysis. 2.4. Data analysis During the isometric situations sEMG was quantified with a time-lag of 0.1 s according to every detected R-wave (i.e. the largest peak of the cardiac ventricular activity) for a time window of 0.4 s. Depending on the actual heart rate up to twelve sections were then used to obtain mean root mean square (rms) values in the frequency range between 20 Hz and 400 Hz for every muscle and load situation, respectively. MFDs of the muscles were calculated under the assumption that the largest rms values for the respective muscle or muscle part would be evoked along this direction, i.e. they were defined on a functional basis. For every muscle at each tilt angle an rms profile was calculated from the rms values at all rotation angles (Fig. 2a). This profile was calculated separately for every subject. Since these rms profiles were obtained with an accuracy of 45°, 8-corner polynomials of these profiles were constructed by transferring the polar into Cartesian coordinates (Fig. 2b). From each of these polynomials the geometric center, or centroid position, could be determined whose x and y coordinates were extracted. In Fig. 2c the vector from zero position to the centroid position contains two parameters: the vector amplitude and its angle. However, the tilts always provoke compensatory muscle activation on the opposite side of the body, i.e. forward tilts (at 0° rotation angle) activate back muscles, tilts to the right (−45° to −135°) muscles on the left body side and so on. Therefore, the Centroid coordinates were mirrored along frontal plane to get MFD angles from body perspective (Fig. 2d). These coordinates were further used to define MFDs and to determine the group variability of centroid coordinates. As can be taken from Fig. 2d, MFD angles were defined as deviation angles from sagittal plane where positive numbers refer to the right body side and negative numbers to the left body side. To determine if the respective MFDs were systematically influenced by body side and tilt angle, a repeated measures analysis of variance (ANOVA) was calculated (body side (2 levels) × tilt angle (7)). Pairwise post hoc tests for tilt angles and body side differences were applied. 3. Results ANOVA results indicated significant body side differences for RA and ICO. There was a tilt angle dependency for oblique abdominal muscles, ICO, and MF (Table 3). There was no interaction between both parameters. However, post hoc tests revealed that for most muscles differences occurred at slight tilt angles, except for OI and OE for which MFD changed continuously with tilt angles occurred (Table 4). 3.1. Rectus abdominis muscle RA showed an almost ideally forward directed MFD with deviations from sagittal plane of only about 10° (see Fig. 3 and Table 5). Variations of centroid position values were mainly caused by variations in sagittal plane (i.e. individual sEMG amplitude differences) rather than variations in frontal plane (Fig. 4). 217

Human Movement Science 60 (2018) 214–224

C. Anders, B. Steiniger

Table 3 ANOVA results for influences of tilt angle and body side on main force directions. muscle

body side

tilt angle

body side*tilt angle

0.012 0.698 0.621 0.001 0.806 0.613

0.692 0.027 < 0.001 0.009 0.168 < 0.001

0.513 0.620 0.867 0.508 0.071 0.368

RA OI OE ICO LO MF

Table 4 Post hoc pairwise differences of main force directions.

RA

OI

OE

ICO

LO

MF

Tilt angle

5°

10°

20°

30°

45°

60°

10° 20° 30° 45° 60° 90°

. . . . . .

. . . . .

. . . .

. . .

. .

.

10° 20° 30° 45° 60° 90°

. . . . 0.042 .

0.017 0.005 < 0.001 < 0.001 0.001

. 0.002 < 0.001 0.027

. . .

. .

.

10° 20° 30° 45° 60° 90°

. . . . . .

. . . . 0.036

0.003 < 0.001 < 0.001 < 0.001

0.004 < 0.001 < 0.001

< 0.001 < 0.001

.

10° 20° 30° 45° 60° 90°

. . . . . .

0.002 0.004 0.002 < 0.001 0.009

. . . .

. . .

. .

.

10° 20° 30° 45° 60° 90°

. . . . . .

. . . . .

. . . .

. . .

. .

.

10° 20° 30° 45° 60° 90°

. < 0.001 < 0.001 . . .

. . . . .

. . . .

. . .

. .

.

3.2. Obliquus internus abdominal muscle As could be expected from the fiber orientation, MFD of OI deviated considerably from sagittal plane (Fig. 3, Table 5). Anyhow, the observed MFD angles did not reach 90° by far. Further, they were subject to tilt angle: starting at large MFD angles of −82° (i.e. almost in frontal plane) for the left OI at 5° tilt angle contrariwise with increasing tilt angles absolute MFD angle values declined continuously (i.e. moved towards the sagittal plane), finally reaching values of −37°, and +33° for left and right body sides. Fig. 4 shows that centroid position variations whilst almost evenly allocated for deviations in sagittal and frontal planes but the larger whisker always pointing to the front and side. Functionally spoken, these variations represent individual amplitude differences (deviations to the front) and deviations from the mean MFD towards more sideward directed MFD angles to similar extents.

218

Human Movement Science 60 (2018) 214–224

C. Anders, B. Steiniger

5°

oe l ra l

ra l raoer l

oe r

oi r

ra r

10°

oe r

oi r

oi l

oi l ico r

ico r ico l ico l lo l

lo r mf l

20°

lo l

lo r

mf r

mf l

oe ra l l ra r

oe l

oe r

mf r

oe r

oi l

oi r

oi r

oi l

30°

ra l ra r

ico r

ico r

ico l

ico l

lo r

lo l

45°

oe l

lo l

mf r

mf l

mf l

ra l ra roe r

oi l

lo r mf r

oe l oi l

oi r

ra l ra r

60°

oe r oi r

ico r ico l

ico r ico l

lo r

lo l mf r

mf l

90°

oe l

ra l ra r

lo r

lo l

oe l

oe r oi r

oi l

mf r

mf l

ra l

mean

ra roe r

oi l

oi r ico r

ico r ico l

ico l

lo r

lo l mf l

ra

oe

lo r

lo l

mf r

mf r

mf l

oi

ico

lo

mf

Fig. 3. Main force directions (MFD) of all investigated trunk muscles and all applied tilt angles. Diagrams are arranged in the way that top corresponds with front and bottom with back. Colors indicate muscles. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

219

Human Movement Science 60 (2018) 214–224

C. Anders, B. Steiniger

Table 5 Main force directions of trunk muscles. Forward direction equals 0°, left sided directions indicated by negative values, right sided with positive. Data are displayed as median values and ± quartile ranges. Significant body side differences of absolute values (Wilcoxon signed rank test) are indicated by bold characters.

RA l RA r OI l OI r OE l OE r ICO l ICO r LO l LO r MF l MF r

5°

10°

20°

30°

45°

60°

90°

median

−39.3 (87.8/60) −79.7 (123.5/56.7)

−9.4 (63.5/67.2) −25.6 (33.9/108.7)

−10.5 (6.5/12.0) 5.9 (3.9/7.0)

−6.5 (2.3/7.3) 6.0 (9.0/4.8)

−6.7 (2.0/1.8) 7.3 (4.0/2.0)

−8.5 (5.4/3.8) 9.0 (3.0/4.1)

−10.1 (8.0/3.8) 6.3 (4.8/3.0)

−13.8 (3.5/5.9) 14.0 (7.6/1.6)

−82.3 (38.8/35.5) 62.8 (30.4/31.8)

−71.5 (23.8/14.2) 67.3 (32.8/7.0)

−50.7 (21.8/11.0) 46.5 (14.5/21.3)

−29.2 (9.9/25.8) 39.0 (21.4/22.4)

−29.0 (13.6/13.8) 39.4 (12.8/24.6)

−24.9 (8.5/25.0) 26.8 (17.7/7.4)

−37.4 (16.3/6.6) 33.3 (13.9/13.5)

−45.7 (10.4/8.9) 48.4 (10.6/23.6)

−19.1 (122.1/69.9) 30.5 (102.7/72.6)

−21.8 (12.1/12.2) 17.6 (11.9/27.9)

−20.9 (8.9/10.2) 10.2 (9.4/5.8)

−30.2 (18.5/4.1) 15.6 (8.7/6.8)

−27.6 (7.8/9.3) 19.4 (9.9/7.4)

−32.2 (11.4/3.1) 27.4 (6.7/7.4)

−31.5 (7.2/5.0) 31.6 (5.7/6.7)

−32.3 (4.4/14.6) 31.2 (3.1/8.6)

−115.6 (16.2/13.8) 95.3 (16.6/44.2)

−104.4 (20.6/11.7) 92.8 (16.1/15.4)

−81.0 (4.7/9.5) 67.0 (17.0/12.6)

−78.5 (7.8/19.4) 70.8 (7.3/12.8)

−80.1 (18.3/10.0) 64.7 (12.2/12.4)

−82.5 (16.9/5.3) 69.6 (10.8/6.0)

−83.7 (5.7/7.8) 74.4 (14.2/12.3)

−89.7 (10.5/9.9) 74.7 (14.7/7.0)

−122.9 (3.4/21.1) 127.9 (12.7/9.8)

−136.7 (10.4/5.5) 145.4 (5.2/21.8)

−146.3 (6.8/3.2) 144.8 (5.4/13.0)

−145.4 (6.6/8.3) 147.3 (2.8/8.0)

−144.2 (9.5/5.4) 142.4 (7.9/5.2)

−147 (9.9/9.3) 141.2 (8.0/10.7)

−142.3 (7.8/18.4) 132.9 (16.7/23.5)

−146.8 (8.6/5.4) 144.8 (7.4/10.3)

−130.8 (3.7/11.5) 131.6 (8.3/8.7)

−140.5 (4.8/6.4) 146.3 (4.0/5.8)

−146.8 (4.3/4.5) 144.7 (5.3/5.1)

−149.7 (7.9/1.1) 143.0 (5.1/1.8)

−144.7 (4.2/3.1) 143.3 (4.4/2)

−143.8 (3.3/2.5) 143.4 (1.2/5.6)

−144.6 (6.3/3.4) 143.7 (1.9/5.8)

−144.9 (4.0/4.9) 143.4 (3.8/2.3)

3.3. Obliquus externus abdominal muscle Comparable with OI results, Centroid positions of OE were dependent on tilt angle, but even larger (Fig. 4). This was apparent in the results of the ANOVA (Tables 3 and 4): for most tilt angles, OE MFDs differed significantly from each other (Fig. 3, Table 5). The variability of individual centroid positions for OE was comparable to what was found for OI (Fig. 4). 3.4. Iliocostalis muscle ICO muscle showed the most unexpected results: its MFD swayed around −80° and +70° (Fig. 3) for left and right body sides with a tendency towards lowest absolute values at medium tilt angles. For slight and high tilt angles absolute values of MFDs increased (Fig. 3, Table 5). Again, the variability of individual centroid positions was quite large for both, deviations from sagittal and frontal planes (i.e. individual amplitude differences and deviations from frontal plane, Fig. 5). 3.5. Longissimus muscle Together with RA, LO did not show any systematic dependency of MFD relating to tilt angle (Fig. 3, Tables 4 and 5). Its centroid position variability was more pronounced in sagittal plane than in frontal plane (Fig. 5). Unexpectedly, the MFD deviated from sagittal plane to about 35° for both body sides (Table 5). 3.6. Multifidus muscle By looking at the centroid position data (Fig. 5) it is surprising, that according to the calculated ANOVA, MF muscle’s MFD should be characterized by a significant tilt angle dependency, but this was due to the values at slightest tilt angles (Tables 4 and 5). Its centroid position variability was almost evenly distributed between sagittal and frontal planes (Fig. 5). MF’s main force directions were virtually identical with those of LO, but were even less variable (Fig. 3, Table 5). 4. Discussion This study revealed two main results: first, all measured trunk muscles have lateral directed main force directions and therefore combine bending and rotational capabilities, and second, the investigated main force directions of the oblique abdominal muscles were systematically altered by tilt angle. 220

Human Movement Science 60 (2018) 214–224

C. Anders, B. Steiniger

RA l

130

130

110

110

90

90

70

70

50

50

30

30

10

10

-10 -140 -120 -100

-80

-60

-40

-20

-10

au

0

OI l

-10

120

120

100

100

80

80

60

60

40

40

20

20

0 -120

-70

-20

5°

0

10°

20°

30

50

70

90

110

130

0 0

60

60

50

50

40

40

30

30

20

20

10

10

0 -40

10

OI r

au

-20

OE l

-60

RA r

50

100

OE r

0

au

30°

0

20

45°

60°

40

60

90°

Fig. 4. Centroid positions of all investigated abdominal muscles of both body sides. Colored dots indicate median values of centroid positions, error bars stand for quartiles. Axes units are arbitrary units, since the centroid positions simultaneously represent sEMG amplitude and location information. Positive values are directed to the subject’s front and right body side respectively. Negative values vice versa. The sEMG amplitude information corresponds with the dot’s distances from zero; x- and y-axes values correspond with the perpendicular projection of the centroid positions to these axes. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

At slight tilt angles the abdominal muscles and ICO showed an extremely large MFD variability. This finding can be explained by the low signal to noise ratio at slight tilt angles where individual subject’s postural patterns were more evident. This may have added some “sway noise” to an extremely low signal level of these muscles, potentially influencing the calculated polar pattern (Tokuno, Carpenter, Thorstensson, & Cresswell, 2006). Last but not least, ventilation alters abdominal muscle’s activity level to a certain extent (Caron, Fontanari, Cremieux, & Joulia, 2004). For the slight tilt angles this may have altered the MFD additionally. With increasing tilt angles the MFD values became more predictable and their variability dropped, indicating a load-dependent activation of the abdominal muscles with an improved signal to noise ratio. The back muscles on the other hand displayed less MFD variability already at slight tilt angles, which provides some indirect evidence for the validity of the effect of task specificity. In contrast to the abdominal muscles and ICO, MF and LO at slight tilt angles already showed relevant amplitude levels, i.e. improved signal to noise ratio (compare with Figs. 4 and 5). The mentioned lateral main force directions for all investigated trunk muscles were somehow surprising since one would not

221

Human Movement Science 60 (2018) 214–224

C. Anders, B. Steiniger

ICO l

-50

-40

-30

-20

-10

-25

-20

-15

-10

-5

45

45

40

40

35

35

30

30

25

25

20

20

15

au 15

10

10

5

5

0

0

-5

-5

0

0

au

0

LO l

-40

-20

0

0

0

-5

-5

-10

-10

-15

-15

-20

-20

-25 -60

10°

20°

20

30

40

50

5

10

15

20

25

LO r

0

0

5°

10

-25

au

0

MF l

ICO r

20

40

0

-10

-10

-20

-20

-30

-30

-40

-40

-50

-50

-60

-60

30°

60

MF r

45°

60°

90°

Fig. 5. Centroid positions of all investigated back muscles of both body sides. Colored dots indicate median values of centroid positions, error bars stand for quartiles. Axes units are arbitrary units, since the centroid positions simultaneously represent sEMG amplitude and location information. Positive values are directed to the subject’s front and right body side respectively. Negative values vice versa. The sEMG amplitude information corresponds with the dot’s distances from zero; x- and y-axes values correspond with the perpendicular projection of the centroid positions to these axes. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

suspect RA and the paravertebral muscles to evoke any relevant bending and rotational force vectors around the vertical axis as their fiber directions are ideally oriented in cranio-caudal direction (Gray, 1942). Some unexpected results may relate to stability issues and require further study to be sorted out from MFD. There was even a lateral tendency for RA, though it was quite small, occurring with low variability of about two to twelve degrees from sagittal plane (with exception of 5° and 10° tilt angles that varied more for both sides). The rotational and or lateral bending potential of the paravertebral muscles was much larger, indicated by an MFD deviation from sagittal plane to about 35° at either body side. Since the paravertebral muscles are all located in the space limited by the spinal processes and the angles of the ribs, i.e. they are located dorsal and lateral with respect to the vertebral bodies, their spinal rotation effect is expected to be more prominent (Gray, 1942). These results are in accordance with the results of Lavender, Tsuang, Hafezi, et al. (1992). There already do exist studies that evaluated trunk muscle activation patterns at different external moment directions and levels (Lavender, Tsuang, Hafezi, et al., 1992; Perez & Nussbaum, 2002). These similar studies used a harness, attached to the subject’s upper body and applied the respective moments by attaching weights via cables and pulleys to the harness. By using combinations of 222

Human Movement Science 60 (2018) 214–224

C. Anders, B. Steiniger

weights, different net force vectors could be applied with an even narrower angular resolution of 30° (Lavender, Tsuang, Andersson, et al., 1992) and even 20° (Lavender, Tsuang, Hafezi, et al., 1992). In our investigation we used an angular resolution of 45° in accordance with Perez and Nussbaum (2002). However, by transferring the polar amplitude profile information into Cartesian coordinates we were able to calculate trunk muscle’s MFDs with increased precision, since the MFD was determined by identifying the centroid position for the 8-corner polynomials, i.e. by considering the amplitude levels of all rotation angles and was therefore not limited to just use the direction that evoked the largest sEMG amplitude. This enabled the estimation of systematic load dependent changes of small amounts. In contrast to the mentioned investigations (Lavender, Tsuang, Andersson, et al., 1992; Perez & Nussbaum, 2002), which were mainly aimed at the identification of co-contractions of trunk muscles as a function of external moment vector direction and magnitude, our study was focused on determining MFDs, and, if the detected values were systematically influenced by tilt angle (i.e. load level) and body side. As already stated earlier, although our setup was quite different from the Lavender investigation (Lavender, Tsuang, Andersson, et al., 1992) our sEMG activation data for LO, OE and RA match their results quite well. Anyhow, at this point our data are not an appropriate source of information about co-contraction of trunk muscles for the applied tilt angles unless they are further analyzed in a way yet to be determined. Both external and internal oblique abdominal muscles were the only ones whose MFDs were systematically influenced by tilt angle (Fig. 4), although they behaved in opposite ways: For OE with increasing tilt angles the absolute MFD values increased, i.e. the more load was applied the more its MFD tended to lateral direction. In contrast, for OI increasing tilt angles were accompanied by considerably more sagittal MFDs. Since both muscles showed very different MFD angles at slight tilt angles, being more laterally directed for OI, with increasing tilt angles they ended up reaching similar MFD angles of about 32° (OE) and 37° (left OI). The more lateral orientation of OI’s MFD somehow mirrors its fiber direction, which at the applied electrode position is horizontal (Ng et al., 1998). This is in accordance with the largest of OI’s MFD angles that reached values of even more than 80°. Anyhow, with increasing tilt angles this deviation from sagittal plane showed a continuous reduction. However, these values were obtained for the left-sided OI and only at slight tilt angles and have therefore to be interpreted with caution. Moreover, these results argue for expanded electrode arrangements to systematically evaluate the effect of the diverging fiber directions on the respective MFDs of both oblique abdominal muscles. Furthermore, our results contribute to the discussions about spinal stability: Studies have shown both increased (Vera-Garcia, Brown, Gray, & McGill, 2006) and decreased (Brown, Vera-Garcia, & McGill, 2006) spinal stability from differing preactivation levels preceding quickly applied external loads. Considering our results, it appears that any increased preload level would result in a more sagittal oriented MFD direction with the consequence of an active flexion tension force that during quick release might cause an untimed muscle imbalance, independent of whether the back muscles are activated evenly or not (Brown et al., 2006). The provided data build a picture of normal muscle function in terms of main force vectors that, with respect to muscle action, includes the bending and rotational capacity of individual trunk muscles according to body side and applied load. This normative data would be suitable to identify coordination deficits of trunk muscles. The identified MFDs of trunk muscles bear the potential to improve trunk muscle diagnostics, with an eye on stabilizing the spine, since trunk muscle’s force capacity can now be determined more accurately during static loading. Normative data of intramuscular dynamics can also aid in the development of specifically targeted training. According to the present results all trunk muscles would most effectively be trained if they are being activated separately per body side and the applied force vectors would combine sagittal and frontal vector components to different amounts, specifically for every muscle and further considering the influence of varying load levels. 5. Limitations The current results were gathered at only one single electrode position per muscle, which due to muscle overlap are often the only muscle area with reduced cross talk issues. These positions were chosen in accordance with the accepted recommendations (Hermens et al., 1999; Ng et al., 1998; SENIAM). Particularly abdominal muscles show diverging fiber directions, bearing the potential for various MFDs at different positions. Moreover, there are hints from other investigations pointing at regionally different activities of trunk muscles during natural activities that would likely result in different MFDs as well (Anders, Reimann, Gotthardt, & Hofmann, 2017). In addition, we applied the varying forces in a three-dimensional space, i.e. we changed the subject’s position systematically with respect to the gravitational force by tilting them. Therefore, independent from the actual rotation angle the acting force on the trunk has always to be fractioned into two parts: the vertical and the horizontal (i.e. transversal) components. By tilting the subjects these fractions changed from 100% vertical force and 0% horizontal force at 0° tilt angle towards 0% vertical force and 100% horizontal force at 90° tilt angle. This change of the respective load components, i.e. the continuously raising horizontal component with increasing tilt angle may have influenced the results. The similar investigations by Lavender, Tsuang, Andersson, et al. (1992), Perez and Nussbaum (2002) added horizontal forces to an unchanged vertical postural situation as they applied the respective forces on upright standing subjects. Both the present and the mentioned investigations were performed as the pelvis was fixed. Such a setup enables systematic investigations but is an artificial situation. 6. Summary The data showed at which tilt and rotation the electrodes had the highest and lowest sEMG signals. The highest signals did not result from sagittal plane load vectors. Therefore, to properly strengthen or therapy the trunk muscles, exercise needs to involve 223

Human Movement Science 60 (2018) 214–224

C. Anders, B. Steiniger

loading along vectors that are not all parallel to the sagittal plane. Because different muscle (or muscle portions) required different load angles where they were the most activated it seems correct to assume that a wide range of loading vectors, i.e. tilt and rotation angles, are necessary to exercise all the trunk muscles adequately. Acknowledgements The laboratory facility was kindly provided by the Center of Interdisciplinary Prevention of Diseases related to Professional Activities (KIP) funded by Friedrich- Schiller- University Jena and the Berufsgenossenschaft Nahrungsmittel und Gastgewerbe. This manuscript was edited for English language by Marcie Matthews of Polishedwords. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.humov.2018. 06.012. References Anders, C., Reimann, R., Gotthardt, L., & Hofmann, G. O. (2017). Activation patterns of human abdominal muscles during walking: Electrode positions make a difference. Medical Research Archives, 5(7), 1–20. Bressel, E., Dolny, D. G., & Gibbons, M. (2011). Trunk muscle activity during exercises performed on land and in water. Medicine and Science in Sports and Exercise, 43(10), 1927–1932. Brown, S. H., Vera-Garcia, F. J., & McGill, S. M. (2006). Effects of abdominal muscle coactivation on the externally preloaded trunk: Variations in motor control and its effect on spine stability. Spine, 31(13), E387–E393. Caron, O., Fontanari, P., Cremieux, J., & Joulia, F. (2004). Effects of ventilation on body sway during human standing. Neuroscience Letters, 366(1), 6–9. Ekstrom, R. A., Donatelli, R. A., & Carp, K. C. (2007). Electromyographic analysis of core trunk, hip, and thigh muscles during 9 rehabilitation exercises. Journal of Orthopedic & Sports Physical Therapy, 37(12), 754–762. Godfrey, K. E., Kindig, L. E., & Windell, E. J. (1977). Electromyographic study of duration of muscle-activity in sit-up variations. Archives of Physical Medicine and Rehabilitation, 58(3), 132–135. Gray, H. (1942). Anatomy of the human body, Vol. 1970. Philadelphia: Lea & Febiger. Guimaraes, A., Vaz, M. A., De Campos, M., & Marantes, R. (1991). The contribution of the rectus abdominis and rectus femoris in twelve selected abdominal exercises. An electromyographic study. The Journal of Sports Medicine and Physical Fitness, 31(2), 222–230. Hermens, H. J., Freriks, B., Disselhorst-Klug, C., & Rau, G. (2000). Development of recommendations for SEMG sensors and sensor placement procedures. Journal of Electromyography and Kinesiology, 10(5), 361–374. Hermens, H. J., Freriks, B., Merletti, R., Stegeman, D. F., Blok, J., Rau, G., et al. (1999). European recommendations for surface ElectroMyoGraphy, results of the SENIAM project, Vol. 8. Enschede: Roessingh Research and Development b.v. Juker, D., McGill, S., Kropf, P., & Steffen, T. (1998). Quantitative intramuscular myoelectric activity of lumbar portions of psoas and the abdominal wall during a wide variety of tasks. Medicine & Science in Sports & Exercise, 30(2), 301–310. Kendall, F. P., Kendall Mc Geary, E., Provance, P. G., Rodgers, M. M., & Romani, W. A. (2005). Muscles testing and function with posture and pain (4th ed.). Philadelphia: Lippincott Williams & Wilkins. Lavender, S. A., Tsuang, Y. H., Andersson, G. B., Hafezi, A., & Shin, C. C. (1992). Trunk muscle cocontraction: The effects of moment direction and moment magnitude. Journal of Orthopedic Research, 10(5), 691–700. Lavender, S. A., Tsuang, Y. H., Hafezi, A., Andersson, G. B., Chaffin, D. B., & Hughes, R. E. (1992). Coactivation of the trunk muscles during asymmetric loading of the torso. Human Factors, 34(2), 239–247. Lipetz, S., & Gutin, B. (1970). An electromyographic study of four abdominal exercises. Medicine and Science in Sports, 2(1), 35–38. Ng, J. K., Kippers, V., & Richardson, C. A. (1998). Muscle fibre orientation of abdominal muscles and suggested surface EMG electrode positions. Electromyography and Clinical Neurophysiology, 38(1), 51–58. Noble, L. (1981). Effects of various types of situps on iEMG of the abdominal musculature. Journal of Human Movement Studies, 7(2), 124–130. Perez, M. A., & Nussbaum, M. A. (2002). Lower torso muscle activation patterns for high-magnitude static exertions: Gender differences and the effects of twisting. Spine, 27(12), 1326–1335. SENIAM. (2015), from www.seniam.org. Shamsi, M., Sarrafzadeh, J., Jamshidi, A., Zarabi, V., & Pourahmadi, M. R. (2016). The effect of core stability and general exercise on abdominal muscle thickness in non-specific chronic low back pain using ultrasound imaging. Physiotherapy Theory and Practice, 32(4), 277–283. Tokuno, C. D., Carpenter, M. G., Thorstensson, A., & Cresswell, A. G. (2006). The influence of natural body sway on neuromuscular responses to an unpredictable surface translation. Experimental Brain Research, 174(1), 19–28. Vera-Garcia, F. J., Brown, S. H., Gray, J. R., & McGill, S. M. (2006). Effects of different levels of torso coactivation on trunk muscular and kinematic responses to posteriorly applied sudden loads. Clinical Biomechanics, 21(5), 443–455. Walters, C. E. P. D., & Partridge, M. J. B. S. (1957). Electromyographic study of the differential action of the abdominal muscles during exercise. American Journal of Physical Medicine, 36(5), 259–268. Willett, G. M., Hyde, J. E., Uhrlaub, M. B., Wendel, C. L., & Karst, G. M. (2001). Relative activity of abdominal muscles during commonly prescribed strengthening exercises. Journal of Strength and Conditioning Research, 15(4), 480–485.

224