Comparison of trunk muscle reflex activation patterns between active and passive trunk flexion–extension loading conditions

Human Movement Science 34 (2014) 12–27

Contents lists available at ScienceDirect

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

Comparison of trunk muscle reflex activation patterns between active and passive trunk flexion–extension loading conditions Michael W. Olson ⇑ Southern Illinois University, Department of Kinesiology, 1075 S. Normal Avenue, Carbondale, IL 62901, United States

a r t i c l e

i n f o

Article history: Available online 29 March 2014 Keywords: Reflex response Tension–relaxation Trunk Electrophysiology

a b s t r a c t The aim of the present study was to determine the effects of trunk flexion–extension loading on the neuromuscular reflexive latencies and amplitude responses of the trunk musculature. Eighteen male and female subjects (18–27 yrs) participated in active and passive trunk flexion extension, performed 7 days apart. Subjects performed 60 trunk flexion–extension repetitions. Surface electromyography (EMG) was collected bilaterally from paraspinal and abdominal muscles. In the active condition, subjects volitionally moved their trunks, while in the passive condition the dynamometer controlled the movements. The trunk was perturbed before and immediately after 30 repetitions. Latency of muscle onset, latency of first peak, latency of maximum peak, and peak EMG amplitude were evaluated. No differences between conditions, sides, or perturbation session were apparent. Overall latencies were shorter in females (p < .05) and abdominal muscles compared to paraspinals (p < .05). Thoracic paraspinal muscle amplitudes were greater than all other muscles (p < .05). Based upon the present results, the neuromuscular system engages trunk flexor muscles prior to the paraspinals in order to provide possible stabilization of the trunk when flexor moments are generated. Overall, the results indicate no difference in response of the neuromuscular system to active or passive repetitive loading. Ó 2014 Elsevier B.V. All rights reserved.

⇑ Tel.: +1 618 536 2244. E-mail address: [email protected] http://dx.doi.org/10.1016/j.humov.2014.03.004 0167-9457/Ó 2014 Elsevier B.V. All rights reserved.

M.W. Olson / Human Movement Science 34 (2014) 12–27

13

1. Introduction The ability of the neuromuscular system to effectively respond to external perturbations is of great concern, especially when considering the ramifications loading schemes present to the health of individuals. Work related physical exertions, due to manual material handling, influence the ability of the musculoskeletal system to coordinate movements (Granata & Sanford, 2000; Marras, 2000) and may significantly contribute to low back pain (Gallagher, Marras, Litsky, & Burr, 2005; Marras, Lavender, Ferguson, Splittstoesser, & Yang, 2010). Many factors have been presented to explain the etiology of low back pain and injury, such as muscle fatigue (Granata, Slota, & Wilson, 2004), muscle fiber type distribution (Mannion et al., 2000), and repetitive loading (de Looze et al., 1996). Neuromuscular fatigue of the low back muscles does influence the ability of the system to control movement and respond accordingly to applied forces. Postural sense (Madigan, Davidson, & Nussbaum, 2006; Taimela, Kankaanpää, & Luoto, 1999; Wilson, Madigan, Davidson, & Nussbaum, 2006), neuromuscular coordination (Gorelick, Brown, & Groeller, 2003; Potvin & O’Brien, 1998), and reflexive responses (Hermann, Madigan, Davidson, & Granata, 2006) are modified due to the inability of the neuromuscular system to effectively receive, interpret and send information to the corresponding effectors (Taylor, Butler, & Gandevia, 2000; Taylor, Todd, & Gandevia, 2006). However, there is evidence to suggest that neuromuscular fatigue may not significantly alter the latency of the reflexive responses to perturbations (Mawston, McNair, & Boocock, 2007; Sanchez-Zuriaga, Adams, & Dolan, 2010). Additionally, neuromuscular fatigue of trunk flexor and extensor muscles is observed to increase the stiffness of the spine, possibly to compensate for the reduced ability of the system to respond when a load is introduced (Grondin & Potvin, 2009). In addition, when there is a greater activation of the muscle prior to introduction of a perturbation, the reflexive response amplitude gain decreases (Stokes, Gardner-Morse, Henry, & Badger, 2000; Vera-Garcia, Brown, Gray, & McGill, 2006). Creep and tension–relaxation behaviors in human in vivo models are documented in the low back tissues (Granata, Rogers, & Moorehouse, 2005; McGill & Brown, 1992; Olson, Li, & Solomonow, 2009; Parkinson, Beach, & Callaghan, 2004; Rogers & Granata, 2006; Shin & Mirka, 2007). Granata et al. and Rogers and Granata report increased response gains from the paraspinal muscles after mechanical creep loading schemes, as well as reduction in the reflexive response of the muscles. Similarly, others have reported the modification of the paraspinal muscle response to either prolonged (Shin, D’Souza, & Liu, 2009; Solomonow, Baratta, Banks, Freudenberger, & Zhou, 2003) or repeated trunk flexion (Dickey, McNorton, & Potvin, 2003; Olson, Li, & Solomonow, 2004; Olson et al., 2009), with most results indicating an extended activation of the paraspinal muscles. Brief periods of static loading are observed to induce passive tissue creep in the lumbar region which is believed to significantly influence the response of the mechanoreceptors within the ligamentous tissues (Rogers & Granata, 2006; Shin, D’Souza, & Liu, 2009). In feline models, prolonged mechanical loading of the spinal ligaments is observed to significantly desensitize the embedded mechanoreceptors resulting in reduced neuromuscular responses (Sbriccoli et al., 2004; Solomonow, Zhou, Baratta, Lu, & Harris, 1999; Solomonow et al., 2000). Sanchez-Zuriaga et al. (2010) report significant reflex latency increases after a prolonged paraspinal creep, but no reflex latency changes when neuromuscular fatigue was induced. Although, reflex gain was reported to increase in some paraspinal muscles post-fatigue. There is also evidence to suggest the abdominal muscles provide additional support to the trunk to assist in increased trunk stability when flexor moments are applied (Cresswell, Oddsson, & Thorstensson, 1994; Hodges, 2001). Based upon previous experiments, the relationship between neuromuscular response of the paraspinal muscles and creep/tension–relaxation loading has been inconclusive. Likewise, fatiguing of the paraspinal muscles provides a range of information when paraspinal muscle reflexes are induced. Data are currently scarce regarding the influence of repetitive passive movement (tension– relaxation) on neuromuscular reflex response, as compared to creep protocols in humans. Therefore, the purpose of this experiment is to compare passive trunk loading and active muscle contraction conditions while subjects perform similar trunk movements. It is hypothesized that the reflexive responses from the trunk muscles will be different between the loading schemes as differences between passive and active loading of the spine have been observed previously. As a second hypothesis, it was believed the activation pattern of the abdominal muscles would compensate for the re-

14

M.W. Olson / Human Movement Science 34 (2014) 12–27

duced compliance of the soft viscoelastic paraspinal tissues, as well as the possible fatigue in the trunk extensors. 2. Methods 2.1. Subjects Eighteen healthy individuals (9 male, 9 females; age 21.7 ± 2.3 yrs, height 1.75 ± 0.08 m, mass 72.5 ± 12.0 kg; males: 22.9 ± 2.6 yrs, height 1.80 ± 0.07 m, mass 80.0 ± 10.1 kg, females: 20.6 ± 1.3 yrs, height 1.70 ± 0.04 m, mass 65.0 ± 8.7 kg) with no history of low back or lower extremity pain/injury were recruited for this study. Subjects were excluded if they had any current or previous incidents of upper extremity, back, or lower extremity pain/dysfunction within the past 12 months. The procedures were approved by the local Institutional Review Board on human subjects. All subjects agreed to perform in the procedures and signed a written informed consent form prior to participation. 2.2. Instrumentation 2.2.1. Isokinetic dynamometer A Biodex System 3 dynamometer (Biodex Medical Systems, Inc., Shirley, NY, USA) was used to collect reaction moment data and control the movement velocity of the trunk while a Biodex back attachment unit was secured to the axis. The dynamometer was calibrated before and after data collection and was within the manufacture’s specifications. Data were collected at a rate of 100 Hz and saved for future processing. A dial located on the left side of the back attachment and encompassing the chair axis was used to position the back attachment and ensure minimal variability within and between days of data collection. During the active condition, the isometric mode was used, while the passive mode was utilized during the passive condition. 2.2.2. Kinematics An infrared motion capture system (Qualisys AB, Gothenburg, Sweden) was used to collect kinematics data (100 Hz) from retro-reflective markers adhered to the specific boney landmarks on each subject. For the purposes of this study, only the angular displacement and velocity of the trunk will be reported. Markers of 14 mm diameter were affixed on the skin over the spinous processes at C7, T6, T10, L1, L3, L5 and S1, bilaterally on the upper extremities at the acromion processes, lateral humoral epicondyles, and the styloid process of the distal ulna. Residuals were calculated to be less than 0.97 mm during calibration trials. These kinematics data were synchronize with electromyography (EMG) recordings during the reflexive response trials using the Qualisys Track Manager (QTM) software interfaced with a USB 2533 A/D board (Measurement Computing, Inc., Norton, MA, USA). Markers positioned along the spine were used to ensure the subjects were positioned in a consistent erect stance between trials. 2.2.3. Electromyography Surface EMG was collected from four regions of the trunk, bilaterally, using a MA-300 system (Motion Lab Systems, Baton Rouge, LA, USA). The skin was abraded and cleaned with alcohol pads prior to electrode placement. Pre-gelled Ag–AgCl electrodes (Biopac Systems, Inc., Goleta, CA, USA) were positioned at a distance of 2.0 cm center to center from the 1.0 cm2 collection area of each electrode and aligned parallel along the length of the respective muscle. Each electrode pair was positioned with a bipolar configuration over the thoracic paraspinals (TP) (4.0 cm lateral from the body midline at T11), lumbar paraspinals (LP) (3.0 cm lateral from the body midline at L3), rectus abdominis (RA) (3.0 cm lateral from the umbilicus), and external oblique (EO) (15 cm lateral from the umbilicus, midway between the 12th rib and iliac crest at a 45° angle from vertical). A ground electrode was positioned on the skin over the left iliac crest. Surface EMG signals were band-pass filtered 20–500 Hz with a common mode rejection ratio of >100 dB at a frequency of 60 Hz, an input impedance of >100 MX, and

M.W. Olson / Human Movement Science 34 (2014) 12–27

15

amplified up to 1000 times. Data were collected at a rate of 1200 Hz using a 12 bit A/D board and saved for future processing. 2.2.4. Accelerometer A triaxial accelerometer (Kister Corporation, Type 8692B50, Buffalo, NY, USA) was securely fixed to a box which was held by the subjects during perturbation trials. The accelerometer was synchronized with the same USB 2533 A/D board as the EMG system and was used in determining the onset of the perturbation during reflexive response trials. 2.3. Procedures Subjects performed two conditions separated by a minimum of 7 days. The two conditions consisted of active (isokinetic mode) and passive (passive mode) trunk flexion–extension, and the order of presentation was balanced between subjects. Each condition was performed while sitting in a chair attachment of the isokinetic dynamometer. This chair was used to reduce potential movement of the pelvis and lower extremities during each loading protocol. Velcro shoulder straps were used to secure each subject’s trunk to the chair back to minimize extraneous movement, while a Velcro strap located along the proximal third of the thigh was used to limit lower extremity movement. The arms of each subject were positioned across the chest so that the hands were resting on the shoulder straps. Subjects performed three trials of maximal voluntary isometric contractions (MVIC) in both flexion and extension while seated in an erect trunk position. Flexion and extension trials were performed in alternating order and 60 s rest was provided between maximal efforts. A 10 min rest period was provided between the end of the last MVIC and the testing procedures. For each condition, range of motion began at erect sitting position (0°) and finished once the individual’s maximum trunk flexion position was attained. The angular velocity was set at 10°/s for each protocol. The range of motion settings were locked from 0° to the angle of full trunk flexion for each subject. In the active condition, subjects were required to attain between 20% and 25% of their maximal isometric effort force during extension, while the flexion phase was performed more passively as the inertial load of the trunk assisted in anterior movement of the chair back. The 20–25% MVIC load during active extension efforts was chosen to ensure that subjects could perform the required number of repetitions for both conditions (Place, Maffiuletti, Bailay, & Lepers, 2005). In the passive condition, subjects were instructed to relax while the machine moved the trunk. Each protocol consisted of 60 flexion–extension repetitions which were divided into two 30 repetition (5 min) periods per protocol. Perturbation sessions (PS) were performed before (PSpre) and immediately after (PSpost1, PSpost2) each 30 repetition period in both active and passive flexion–extension conditions such that three perturbation sessions were recorded. The duration between the end of one 30 repetition period and the ensuing perturbation session (PSpost1 or PSpost2) was approximately 90 s. During the PS, subjects stood erect on a platform while holding a box (0.33 m  0.33 m  0.28 m, 1.2 kg) which was affixed to a pulley system at its base (Fig. 1). The shoulders were maintained at 0° (anatomically) in the sagittal plane while the forearms were flexed  90° while holding the box. Foot placement was ensured by aligning the feet with the end of the platform on which each subject stood. The accelerometer was affixed on the right side of the box (subject reference) and positioned towards the bottom of the box approximately 10 cm from the pulley insertion. A mass of 9.07 kg was attached to the other end of the pulley system and was used to supply the perturbation. Briefly, pilot testing with male and female subjects assisted in determining the load amount that would be sufficient to impose a determinable muscle response. This 9.07 kg mass was dropped from a height of 1.0 m and displaced 0.7 m vertically to initiate the perturbation and impose a flexion moment about the trunk. Tension was maintained in the pulley line to minimize slack and to reduce possible anticipation by the subjects to changes in the pulley tension. Once the subjects were positioned and ready the load was released randomly within an 8 s window. Subjects were blinded to the introduction of the perturbation by a curtain between them and the load (Fig. 1). Each PS consisted of three trials separated by 30 s. Subjects were instructed to maintain an upright posture while remaining relaxed during all trials. Additionally, surface EMG were used to determine the delay in muscle activation initiation between anterior deltoid and paraspinal muscles during pilot testing in order to exam the effect of either maintaining or deviating elbow angle

16

M.W. Olson / Human Movement Science 34 (2014) 12–27

A

B

C7

T6 T10 L1 L3 L5 S1 Fig. 1. (A) Schematic representation of the perturbation trials. The black rectangle represents the 9.07 kg load, while the square with upward diagonal lines represents the box that each subject held. The subjects were blind to the presentation of the perturbation by a curtain (squiggle line). (B) The stick figure on the right provides exemplar data depicting the position of the spine during the perturbation trials.

from 90° at loading. Independent of initial arm movement at the onset of loading, there was 20 ms delay in the initiation of paraspinal activity indicating minimal influence of movement about the elbow in determining trunk muscle activation. 2.4. Data processing Surface EMG signals collected during the MVIC trials were full-wave rectified and smoothed with a fourth-order low pass Butterworth filter at 25 Hz. The maximum EMG value for each muscle attained during the MVIC trials of flexion and extension efforts was used for normalization. The EMG signals from the PS were full wave rectified and low pass filtered with a fourth-order zero lag Butterworth filter at 25 Hz (Granata et al., 2005). The EMG signals from the PS were then normalized to the maximum EMG values attained during MVIC efforts for the respective muscles. Similarly, the signal from the accelerometer was low pass filtered with a zero-lag Butterworth filter at 25 Hz. Kinematics data were low pass filtered with a zero-lag Butterworth filter at 10 Hz. Markers positioned at C7 and L5 were used to evaluate displacements and velocities of the trunk due to the perturbation. Four variables were of interest during PS: EMG onset latency, latency to first peak EMG, latency to maximum EMG amplitude, and the maximum EMG amplitude (Dupeyron, Perrey, Micallef, & Pélissier, 2010; Sanchez-Zuriaga et al., 2010) (Fig. 2). The temporal parameters were determined from the accelerometer data. The perturbation time was determined as the deviation of the resultant acceleration by more than two standard deviations from the mean, based upon a 200 ms window prior to perturbation, while subjects were relaxed (Hermann et al., 2006). The temporal peak reflexive response was determined within 150 ms after introduction of the perturbation, as any response beyond this time is considered voluntary (Dupeyron et al., 2010; Larivière, Forget, Vandeboncoeur, Bilodeau, & Mecheri, 2010). Onset latencies were indicated when the EMG signal attained two standard deviations above a mean threshold level based upon the EMG signal levels 200 ms prior to introduction of the perturbation (Hermann et al., 2006). The activation level of the muscles 200 ms prior to the onset of the perturbation was also assessed to determine any significant changes between perturbations session and conditions. During both active and passive conditions, each 30 repetition period lasted approximately 5 min. Electromyography data from each condition were recorded every other cycle to ensure complete cycles of trunk flexion–extension were collected. The EMG data from the MVIC efforts were reprocessed at 5 Hz with a low-pass Butterworth filter to normalize the EMG signals collected during the passive and active conditions. The EMG signals were full-wave rectified, and smoothed with a fourth-order zero-lag low-pass Butterworth filter at 5 Hz (Olson, 2011). The EMG data were then normalized to the respective reprocessed EMG from MVIC performed before each condition. The average normalized

17

M.W. Olson / Human Movement Science 34 (2014) 12–27

4 1.2

1.2

12

3

1

1

0.4 0.2 0

0.8 0.6

8

1

0.4

6 0.2 0

4 3.1

3.15

3.2

3.25

3.3

-0.2

2

3.35

Trunk Posion (°)

0.6

Accelerometer (mV)

Normalized EMG

10 0.8

3.4 2

-0.4 -0.6

0

5 Time (s) Fig. 2. Schematic of the accelerometer (gray line), normalized EMG recording from the left lumber paraspinal (thick black line), and trunk position (thin black line) during a single perturbation trial. The numbers and corresponding arrows denote: (1) threshold of accelerometer in determining perturbation initiation, (2) onset of EMG response, (3) first EMG peak response, (4) maximum EMG amplitude response, and (5) initial kinematics response of the trunk.

EMG signal from each flexion and extension movement phase were analyzed to determine the change, if any, in EMG during each condition. Moment data collected by the isokinetic dynamometer were low-pass filtered at 5 Hz with a fourth-order Butterworth filter (Olson, 2011). The peak moments from each cycle in each condition were initially normalized to the MVIC moment for each protocol, analyzed, and used to determine the trends of the data. Only data from the extension phase during the active condition (active loading) and flexion phase (passive loading) from the passive condition were used for analyses since these provided the loading schemes of interest. 2.5. Statistical analyses Analysis of variance (ANOVA) with repeated measures (condition  PS (sessions)  gender  side (left versus right)  muscle group) was used to evaluate the dependent variables onset latency, latency to first peak, latency to maximum peak amplitude, and maximum peak amplitude. Three-way ANOVAs (condition  repetition  gender) were performed to evaluate the effects of the 30 repetitions and the active and passive conditions on the dependent variables. Additionally, the normalized EMG amplitude of the muscles during the two 30 repetition periods from both conditions were assessed in separate flexion and extension phases using a repeated measures ANOVA (condition  repetition  gender). A repeated measures ANOVA (condition  sessions  muscle group  gender) was used to assess the activation level of the muscle groups 200 ms prior to the onset of the perturbation. No differences between sides were indicated in initial statistical evaluations of pre-perturbation EMG activity, therefore the pooled values from each muscle group were used for analysis. Trunk angular displacements and velocities were evaluated using a three-way ANOVA (condition  session  gender). When significant interaction effects were observed a Tukey’s HSD was performed. The Statistical Package for Social Science (IBM SPSS Statistics, Chicago, IL, USA) version 20 was used for statistical analyses. The level of significance was set at alpha < .05. 3. Results The muscle activity 200 ms prior to perturbation onset was not significantly different between conditions for any muscle group (all p > .05). Similarly, there were no significant differences between

18

M.W. Olson / Human Movement Science 34 (2014) 12–27

sessions for any muscle (all p > .12), nor any differences between gender (p > .05). No statistically significant differences were present between the activation levels of the muscle groups (all p > .05). Mean pre-perturbation muscle activation were 0.056 (±0.092) for TP, 0.042 (±0.048) for LP, 0.033 (±0.063) for EO, and 0.060 (±0.055) for the RA, respectively. 3.1. EMG onset latency The mean EMG onset latency was not significantly different between sessions (p > .598), or side (p > .969). There were significant differences between muscle groups (F3,2464 = 8.35, p < .0001) as the RA muscle group onset latency was significantly shorter than the other muscle groups (Fig. 3). A significant condition  gender interaction (F1,2464 = 8.55, p < .03) was observed (Fig. 3), while no other interaction effects were noted. 3.2. Latency to first EMG peak There were no significant main effects for condition (p > .374), session (p > .23), or side (p > .624), but significant differences were observed between gender (F1,2464 = 6.68, p < .011) and muscle groups (F3,2464 = 4.69, p < .005) (Fig. 4). The latency of first EMG peak was, on average, 4.9% (3.7 ms) shorter for females than for males over all muscles. The RA muscle group attained an initial peak response 8.9%

A

Pooled Acve Passive

120 100

* 80 60 40

Time (ms)

20 0

B

TP

LP

EO

RA

120

Male Female

100 80 60 40 20 0

Acve

Passive

Fig. 3. (A) Mean (+sd) onset latency of muscle activity for thoracic paraspinal (TP), lumbar paraspinal (LP), external oblique (EO), and rectus abdominis (RA) muscles, respectively, represented as pooled data and individual conditions. (B) Gender specific differences were observed in the data as females had shorter onset latencies compared to males independent of condition (active versus passive). Asterisks (⁄) denote significant differences (p < .001) between RA and the other muscle groups.

19

M.W. Olson / Human Movement Science 34 (2014) 12–27

120

A *

Pooled Acve Passive

100 80 60 40

Time (ms)

20 0 120

TP

B

LP

EO

RA

† Male

100

Female 80 60 40 20 0

Pooled

Acve

Passive

Fig. 4. (A) Mean (+sd) latency to first electromyography peak muscle activity for thoracic paraspinal (TP), lumbar paraspinal (LP), external oblique (EO), and rectus abdominis (RA) muscles, respectively, represented as pooled data and individual conditions. The asterisk (⁄) indicates a significant difference between LP and RA muscles (p < .01). (B) Gender specific differences were observed in the data (pooled data) as females had shorter first peak latencies compared to males overall, but this difference was significant during the active condition ( p < .01).

(6.8 ms) sooner, on average, than the LP muscle group, but not significantly before the TP and EO muscle groups (Fig. 4). There were no significant interaction effects observed. 3.3. Latency to maximum EMG amplitude There were no significant differences between conditions (p > .075), sessions (p > .126), or side (p > .156) in determining the latencies of the maximum EMG amplitude response. A significant main effect for muscle groups was present (F3,2464 = 12.62, p < .0001) as the RA muscle group attained a peak amplitude significantly earlier than the other muscle groups (Fig. 5). A significant main effect for gender (F1,2464 = 91.28, p < .0001) was present as the latency of the peak amplitude for the females (106.6 ± 29.6 ms) was overall statistically less than that of the males (117.0 ± 23.7 ms, Fig. 5). There were no significant interactions observed. 3.4. Maximum EMG amplitude There was no main effect for sessions (p > .284) or side (p > .209) observed. There were significant muscle groups  gender (F3,2464 = 7.154, p < .01) and condition  muscle groups (F3,2464 = 7.981, p < .01) interactions (Tables 1 and 2). Post hoc analysis for the muscle groups  gender interaction observed greater TP activity for both genders compared to the other muscle groups (p < .01). Post hoc analysis for the condition  muscle groups interaction observed significantly greater TP activity be-

20

M.W. Olson / Human Movement Science 34 (2014) 12–27

A 160

*

140 120

Pooled Acve

100

Passive

80 60 40

Time (ms)

20 0

B

TP

LP

EO

RA

160 140

†

120 100

Male

80

Female

60 40 20 0

Pooled

Acve

Passive

Fig. 5. (A) Mean (+sd) onset latency of maximal amplitude (peak) muscle activity for thoracic paraspinal (TP), lumbar paraspinal (LP), external oblique (EO), and rectus abdominis (RA) muscles, respectively. The asterisk (⁄) denotes significant differences (p < .01) between RA and the other muscle groups. (B) Gender specific differences were observed in the data as females had shorter onset latencies compared to males in both conditions ( p < .01). Pooled data and data from each condition are given.

tween muscle groups in the active condition (p < .01), while a significantly lower EO muscle activity was observed between muscles in both conditions (p < .01). No other interaction effects were observed.

3.5. Active and passive forces During active extension efforts, there was no significant decline in the normalized force attained over time during either 5-min period (p > .1). Similarly, there were no significant differences between gender (p > .4). Subjects were able to attain a level of consistent effort between 20% and 30% of their maximal efforts, but a trend in the data was present (p = .081) indicating a reduction in force output with increasing number of repetitions (repetitions 1–30: 26.0 ± 14.0% versus repetitions 31–60: 22.3 ± 13.6%). There were significant increases in the reaction moment measured from the dynamometer during the passive condition (F29,1135 = 6.55, p < .01). Similarly, there were significant differences between the two periods (F1,1135 = 4.65, p < .01) during the passive condition, indicating a recovery between periods with further reduction in the tension within the posterior passive tissues with increasing repetition (Fig. 6). Significant differences between gender were present (F1,1135 = 239.7, p < .01; men 5.6 ± 4.5% versus women 9.9 ± 4.4%, respectively).

Active condition Session 1

TP LP EO RA

Passive condition Session 2

Session 3

Session 1

Session 2

Session 3

M

F

M

F

M

F

M

F

M

F

M

F

0.582 (0.42) 0.345 (0.30) 0.157 (0.13) 0.245 (0.29)

0.571 (0.40) 0.365 (0.26) 0.405 (0.43) 0.480 (0.84)

0.523 (0.31) 0.279 (0.29) 0.209 (0.33) 0.317 (0.41)

0.587 (0.36) 0.356 (0.24) 0.354 (0.44) 0.260 (0.32)

0.495 (0.32) 0.337 (0.34) 0.144 (0.13) 0.480 (0.56)

0.473 (0.35) 0.393 (0.28) 0.344 (0.30) 0.358 (0.57)

0.385 (0.24) 0.372 (0.29) 0.165 (0.15) 0.189 (0.25)

0.427 (0.26) 0.353 (0.25) 0.399 (0.60) 0.427 (0.47)

0.390 (0.25) 0.381 (0.35) 0.152 (0.13) 0.246 (0.37)

0.370 (0.30) 0.424 (0.33) 0.341 (0.41) 0.303 (0.35)

0.322 (0.20) 0.331 (0.26) 0.191 (0.24) 0.384 (0.29)

0.470 (0.31) 0.444 (0.34) 0.360 (0.35) 0.404 (0.37)

TP = thoracic paraspinal, LP = lumbar paraspinal, EO = external obliques, RA = rectus abdominis, M = males, F = females.

M.W. Olson / Human Movement Science 34 (2014) 12–27

Table 1 Mean (±sd) normalized muscle EMG amplitude between genders across conditions and sessions.

21

22

M.W. Olson / Human Movement Science 34 (2014) 12–27

Table 2 Mean (± sd) muscle x gender and condition x muscle interactions for maximum EMG amplitude. Muscle

Gender

Condition

Male TP LP EO RA

0.449 0.341 0.170 0.310

(0.29) (0.31) (0.19) (0.36)

Female

Active

0.483 0.389 0.367 0.372

0.539 0.346 0.269 0.357

(0.33) (0.28) (0.42) (0.49)

Passive (0.36) (0.29) (0.29) (0.50)

0.394 0.384 0.268 0.326

(0.26) (0.30) (0.31) (0.35)

TP = thoracic paraspinal, LP = lumbar paraspinal, EO = external obliques, RA = rectus abdominis.

Normalized Moment

0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 0

10

20

30

40

50

60

Repeons Fig. 6. Mean (+sd) non-normalized peak reaction moment during trunk flexion in the passive condition. The data are grouped based upon the two 30 repetition periods of loading.

3.6. Electromyography from the active and passive sessions During trunk flexion, significant differences between conditions were observed for the average EMG for the left TP (F1,1030 = 71.63, p < .01), right TP (F1,1030 = 71.63, p < .001), left LP (F1,1030 = 76.75, p < .001), right LP (F1,1030 = 56.18, p < .001), left RA (F1,1030 = 17.03, p < .001) and right RA (F1,1030 = 37.97, p < .001) muscle groups (Table 3). The right EO group was not significantly different between conditions, but the left EO was significantly different (F1,1030 = 37.47, p < .001) between conditions. There were no significant differences between the repetitions (p > .6) or gender (p > .05). Overall, significantly greater muscle activity was present in the active condition compared to the passive condition.

Table 3 Mean (±sd) average EMG per muscle for each condition during flexion and extension movement phases. Muscle

Active

Passive

Flexion Left TP LP EO RA

0.031 (0.41) 0.099 (0.21) 0.176 (0.51) 0.285 (1.24)

Extension Right 0.048 (0.08) 0.097 (0.21) 0.037 (0.087) 0.247 (0.73)

Left 0.043 (0.045) 0.102 (0.15) 0.160 (0.47) 0.268 (1.18)

Flexion Right 0.057 (0.088) 0.124 (0.30) 0.026 (0.023) 0.223 (0.75)

Left

Extension Right

*

0.017 (0.016) 0.016* (0.03) 0.032* (0.051) 0.031* (0.047)

Left *

0.017 (0.018) 0.024* (0.034) 0.029 (0.059) 0.025* (0.036)

TP = thoracic paraspinal, LP = lumbar paraspinal, EO = external obliques, RA = rectus abdominis. Denote significant differences between conditions (p < .001).

*

Right *

0.018 (0.022) 0.015* (0.025) 0.033* (0.057) 0.032* (0.067)

0.016* (0.017) 0.024* (0.034) 0.027 (0.053) 0.025* (0.038)

M.W. Olson / Human Movement Science 34 (2014) 12–27

23

During the extension phase, significant differences between conditions were observed for the average EMG for the left TP (F1,1030 = 130.7, p < .001), right TP (F1,1030 = 104.6, p < .001), left LP (F1,1030 = 167.9, p < .001), right LP (F1,1030 = 56.35, p < .01), left RA (F1,1030 = 15.88, p < .001), and right RA (F1,1030 = 29.85, p < .001) muscle groups (Table 3). The left EO was significantly different between conditions (F1,1030 = 36.16, p < .001), while the right EO was similar between conditions (p > .05). There were no significant differences observed for gender (p > .05) or repetition (p > .5). There was no significant trend present in the data to suggest increasing paraspinal activity over repetitions during the active condition. 3.7. Trunk kinematics Ranges of motion were not significantly different between active (67.6 ± 7.8°) and passive (68.4 ± 3.8°) conditions during the loading schemes (p > .7). Trunk angular displacements were not different between sessions or conditions (10.5 ± 5.2°, 9.1 ± 5.3°, and 8.9 ± 6.0° for the active condition, and 10.2 ± 5.0°, 8.6 ± 5.5°, and 8.8 ± 5.5° for the passive condition, all p > .101). No significant differences were apparent for gender in either trunk angular displacement (men: 9.47 ± 5.4°; women: 9.20 ± 5.2°) or peak angular velocity (men: 57.8 ± 37.5°/s; women: 60.2 ± 36.1°/s). 4. Discussion The purpose of this study was to examine the latency and amplitude response of the trunk musculature between active and passive loading schemes when each was performed in trunk flexion–extension. From the data collected it appears that no significant temporal latencies were present between the conditions, but other variables, such as gender and muscle group, were observed to present significant differences independent of the conditions. In addition, amplitude response of the muscles was greatest in the TP during active conditions and engaging at higher level in both genders while EO muscles in both conditions was the least active among muscle groups, indicating little compensation by the abdominal muscles when the paraspinal tissues are actively loaded for relatively short intervals. The second hypothesis stated compensation by the abdominal muscles would be present when the paraspinal tissues were either fatigued or mechanically compromised. In all subjects, the RA muscle latencies were shorter than those of the paraspinal muscles, independent of the conditions. This may indicate a mechanism used by the neuromuscular system to enhance trunk stability while the system is in erect stance. Although the main hypothesis was not supported in the data, there were other variables that were of significance which will be discussed further. Tension–relaxation was induced in the posterior lumbar soft tissues, while very little muscle activity was detected during these passive trunk flexion–extension repetitions. The increased reaction moment measured is indicative of increased tissue mechanical compliance. Although the relative reaction moment was greater for women than men, this is indicative of the larger moments generated by the men during MVIC efforts, as both showed evidence of increased tissue compliance during the loading periods. It is assumed that the lumbar paraspinal tissues were principally affected by these loading schemes since these tissues are primarily strained during trunk flexion, particularly when the pelvis is in a fixed position. These strains would result in significant mechanical behavior modifications to localized tissues in the lumbar region and influence the structural stability of the local spinal tissues. Tissue stiffness increases when sufficient rest is provided to the lumbar tissues between loading schemes (Sbriccoli et al., 2004). A reduction in force from the end of the first loading period to the beginning of the second loading period (Fig. 6) was apparent. It is possible that the time between the loading periods, although minimized during the PS, allowed for this reduction. Secondary to these mechanical behavior changes, a modified response of the embedded mechanoreceptors is believed to occur, based upon work by Solomonow et al. (1999, 2000). It was believed that the latency of the reflexive response to the perturbation would increase once tension–relaxation was induced due to desensitization of the mechanoreceptors embedded within the soft passive viscoelastic tissues. In feline models, reduced stiffness of the soft connective tissues influences the initiation of tension dependent mechanoreceptors embedded within these tissues (Sbriccoli et al., 2004; Solomonow et al., 1999,

24

M.W. Olson / Human Movement Science 34 (2014) 12–27

2000). Thus, direct empirical evidence provides significant support for the desensitization of the mechanoreceptors. Human in vivo studies indicate prolonged trunk flexion adversely influences the neuromuscular response of the paraspinal muscles in humans (Granata et al., 2005), while short duration creep applied repeatedly also reduces reflex responses (Rogers & Granata, 2006). Sanchez-Zuriaga et al. (2010) report a significant delay in muscle reflex activation after tissue creep. However, their subjects were creep induced for at least 1 h, compared to the 2–5 min bouts of tension–relaxation induced in the present study. Further, temporal modifications of voluntary muscle activity are shown after exposure to trunk flexion–extension in fatiguing and non-fatiguing conditions (Dickey et al., 2003; Shin, D’Souza, & Liu, 2009) and hypothesized to result from modifications to the mechanical behavior of the soft viscoelastic passive connective tissues. Tension–relaxation and mechanical creep do influence the ability of the muscles to transfer muscle force (Olson, 2011), and it is believed that these loading schemes reduce the mechanical stability of the spine and influence modifications in neuromuscular response (Adams & Dolan, 1996; Rogers & Granata, 2006). The active trunk flexion–extension movements resulted in greater activation of the paraspinal muscles compared to the passive condition, and a trend of reduced moment was present to indicate a possible reduction in force generating capacity of the trunk extensor muscles. Since greater levels of muscular effort reduce the ability of the muscle to sustain force and maintain the task, it was believed that muscular efforts that matched 20–25% of MVIC would allow for subject to continue performing flexion–extension for the two 5-min (30 repetition) periods. A trend was present in the moment data to indicate a reduction in force generation over the repetitions. It is possible that the motor unit pools of the paraspinal muscles were differentially recruited to ensure a steady-state force level was maintained in proceeding cycles to reduce the influence of neuromuscular fatigue (van Dieën, Böke, Oosterhuis, & Toussaint, 1996). Although the trunk extensors were engaged in the active condition, more than in the passive condition, it is difficult to determine if a level of fatigue was attained that would influence the latency of muscle response to perturbations. It is also possible that the flexion phase of the movement provide a brief window of recovery during each flexion–extension cycle as the inertial load of the trunk assisted the flexion movement. However, the amplitudes of the paraspinal muscles were higher than in the passive condition during the flexion movement phases, indicating that the muscles were engaged during the entire movement. Amplitudes of the neuromuscular signals from any muscle in either condition did not vary significantly. Sanchez-Zuriaga et al. (2010) report increased amplitude of the paraspinal activity as a result of neuromuscular fatigue, but also state that no changes in the latency of the neuromuscular response were apparent. Others have reported increased baseline levels of EMG signals after inducing muscle fatigue which may increase spinal stiffness (Grondin & Potvin, 2009) and reduce the detection of reflex responses (Stokes et al., 2000; Vera-Garcia et al., 2006). Although neuromuscular fatigue does influence postural sense (Madigan et al., 2006; Taimela et al., 1999) and neuromuscular coordination (Potvin & O’Brien, 1998), its impact on reflex timing is variable in both healthy individuals (Hermann et al., 2006; Sanchez-Zuriaga et al., 2010) and those with low back pain (Leinonen et al., 2001; Wilder et al., 1996). Therefore, importance was placed on determining whether passively straining the soft passive viscoelastic tissues alone, or the additional neuromuscular activation of the paraspinal muscles (active loading), was influential in modifying the neuromuscular reflexive response. Based upon the data of this current study, no significant changes in neuromuscular response of the trunk musculature occurred when brief periods of repeated loading were induced. Gender differences are apparent in the current data, specifically in regards to muscle response amplitude and reflex latency measures. Consistent across both conditions and all PS, it was observed that the muscle latencies of onset, first EMG peak, and maximum EMG peak are shorter in women. The reflex latency results in this current work are in agreement with some authors (Miller, Slota, Agnew, & Madigan, 2010). However, these latency results from the paraspinal muscles contradict the works of others (Granata et al., 2005; Larivière et al., 2010; Wilder et al., 1996). Additionally, higher EMG amplitudes were observed in women compared to men during PS. Since the load used for the perturbations was the same for both genders (9.07 kg), it is understandable that paraspinal activity was greater in amplitude for women during the PS (Granata et al., 2005). This standard load imposed upon all subjects without factoring trunk and head moments of inertia could substantially explain the differences in muscle response latencies between men and women. The pre-load muscle activation was consistent

M.W. Olson / Human Movement Science 34 (2014) 12–27

25

between sessions for both genders, thus pre-activation of the trunk muscles prior to the introduction of the perturbation cannot be factors in determining these latency differences. Regardless of the condition, the latencies were unaffected, and the overall amplitude response did not change. Further information provided in the data suggests differential recruitment of the trunk musculature when external loads are presented. Independent of the condition or PS, the primary trunk flexors (RA) were observed to initiate prior to the paraspinal and EO muscles. The activation of the RA muscles prior to the paraspinal muscles is unclear in this current study. Since a flexion moment was presented to the trunk, it was assumed that either the paraspinal or EO muscles would initially engage, since the paraspinals must resist the flexion moment, while EO muscles are shown to respond earlier than RA (Hodges, 2001). Interestingly, Cresswell et al. (1994) also observed shorter activation latencies for deep and superficial abdominal muscles compared with paraspinal activity during unexpected ventral loading. It is possible the trunk flexors were engaged prior to the other muscle groups to initiate an increased pressurization of the abdominal cavity that would allow for increased stability of the trunk (Cholewicki, Julura, & McGill, 1999; Cresswell et al., 1994), however, this idea has been doubted recently (Stokes, Gardner-Morse, & Henry, 2011). Similarly, the motor activation of the abdominal muscles is dependent upon the postural task and demands imposed upon the system in providing sufficient support to the trunk (Carpenter, Tokuno, Thorstensson, & Cresswell, 2008; Eriksson Crommert & Thorstensson, 2009; Tokuno, Cresswell, Thorstensson, & Carpenter, 2013). Based upon the data presented in this current experiment, it is possible that the neuromuscular system does indeed use the trunk musculature to increase stiffness of the trunk prior to engaging the trunk extensors. The results of this current study indicate that short duration passive loading of the lumbar tissues is insufficient in modifying the neuromuscular reflexive response of the paraspinal muscles. Similarly, low level neuromuscular activity during active trunk movements did not change the reflexive response of the trunk muscles. It is possible that longer duration loading schemes may provide the required time for tissue compliance to significantly influence the mechanoreceptor threshold level attainments that would alter neuromuscular response during repetitive movement tasks. Previous studies report significant neuromuscular adaptations to prolonged passive loading which influence the latency of the neuromuscular response to perturbations (Granata et al., 2005; Sanchez-Zuriaga et al., 2010). The previous studies used perturbations that were either harnessed to the trunk or employed a quick release from the trunk to induce neuromuscular responses. In the present study, a crate held in the hands was loaded unexpectedly (for the subject) to induce muscle responses while subjects were standing erect and this may have been a limitation of the study. It is possible that the movement of the arms may have attenuated some of the force exerted on the system and influenced the reflexive response. Pilot testing indicated arm displacements, independent of magnitude, did very little to modify the activation of trunk muscles. Further, onset of deltoid and paraspinal muscles activations were examined in pilot tests and observed to be non-significant (unreported data). Additionally, MVICs were performed from an erect seated position for each muscle group. This position may not elicit a true ‘‘maximum’’ based upon the position of the trunk, but it did provide for a reference value in determining changes to the EMG signal amplitudes.

5. Conclusion In conclusion, the paraspinal tissues were exposed to passive strain loading and active muscle loading during comparable trunk flexion–extension tasks. Mechanically, there is evidence to indicate that these loading schemes influenced the behaviors of the soft passive viscoelastic tissues, as well as the muscle force generation in respective conditions. Although no significant reflex differences were identified during the conditions, there were significant gender differences between muscle onset latencies, as well as between muscles independent of condition. The neuromuscular system may provide a mechanism to support the trunk–spine system during loading in upright posture which incorporates early activation of abdominal muscles in order to buttress the abdominal cavity. Additional studies are warranted to further examine the trunk muscle activation patterns and how these responses are influence by external loading of the trunk.

26

M.W. Olson / Human Movement Science 34 (2014) 12–27

References Adams, M. A., & Dolan, P. (1996). Time dependent changes in the lumbar spine’s resistance to bending. Clinical Biomechanics, 11, 194–200. Carpenter, M. G., Tokuno, C. D., Thorstensson, A., & Cresswell, A. G. (2008). Differential control of abdominal muscles during multi-directional support-surface translations in man. Experimental Brain Research, 188, 445–455. Cholewicki, J., Julura, K., & McGill, S. M. (1999). Intra-abdominal pressure mechanism for stabilizing the lumbar spine. Journal of Biomechanics, 32, 13–17. Cresswell, A. G., Oddsson, L., & Thorstensson, A. (1994). The influence of sudden perturbation on trunk muscle activity and intraabdominal pressure while standing. Experimental Brain Research, 98, 336–341. de Looze, M. P., Visser, B., Houting, I., van Roy, M. A. G., van Dieën, J. H., & Toussaint, H. M. (1996). Weight and frequency effects on spinal loading in a bricklaying task. Journal of Biomechanics, 29, 1425–1433. Dickey, J. P., McNorton, S., & Potvin, J. R. (2003). Repeated spinal flexion modulates the flexion–relaxation phenomenon. Clinical Biomechanics, 18, 783–789. Dupeyron, A., Perrey, S., Micallef, J. P., & Pélissier, J. (2010). Influence of back muscle fatigue on lumbar reflex adaptation during sudden external force perturbations. Journal of Electromyography & Kinesiology, 20, 426–432. Eriksson Crommert, A. E., & Thorstensson, A. (2009). Trunk muscle reactions to sudden unexpected and expected perturbations in the absence of upright postural demand. Experimental Brain Research, 196, 385–392. Gallagher, S., Marras, W. S., Litsky, A. L., & Burr, D. (2005). Torso flexion loads and the fatigue failure of human lumbosacral motion segments. Spine, 30, 2265–2273. Gorelick, M., Brown, J. M. M., & Groeller, H. (2003). Short-duration fatigue alters neuromuscular coordination of trunk musculature: Implications for injury. Applied Ergonomics, 34, 317–325. Granata, K. P., Rogers, E., & Moorehouse, K. (2005). Effects of static flexion-relaxation on paraspinal reflex behavior. Clinical Biomechanics, 20, 16–24. Granata, K. P., & Sanford, A. H. (2000). Lumbar–pelvic coordination is influenced by lifting task parameters. Spine, 25, 1413–1418. Granata, K. P., Slota, G. P., & Wilson, S. E. (2004). Influence of fatigue in neuromuscular control of spinal stability. Human Factors, 46, 81–91. Grondin, D. E., & Potvin, J. R. (2009). Effects of trunk muscle fatigue and load timing on spinal responses during sudden hand loading. Journal of Electromyography & Kinesiology, 19, e237–e245. Hermann, K. M., Madigan, M. L., Davidson, B. C., & Granata, K. P. (2006). Effects of lumbar extensor fatigue on paraspinal muscle reflexes. Journal of Electromyography & Kinesiology, 16, 637–641. Hodges, P. W. (2001). Changes in motor planning of feedforward postural responses of the trunk muscles in low back pain. Experimental Brain Research, 141, 261–266. Larivière, C., Forget, R., Vandeboncoeur, R., Bilodeau, M., & Mecheri, H. (2010). The effect of sex and chronic low back pain on back muscle reflex response. European Journal of Applied Physiology, 109, 577–590. Leinonen, V., Kankaanpää, M., Luukkonen, M., Hänninen, O., Airaksinen, O., & Taimela, S. (2001). Disc herniation-related back pain impairs feed-forward control of paraspinal muscles. Spine, 26, E367–E372. Madigan, M. L., Davidson, B. C., & Nussbaum, M. A. (2006). Postural sway and joint kinematics during quiet standing are affected by lumbar extensor fatigue. Human Movement Science, 25, 788–799. Mannion, A. F., Käser, L., Weber, E., Rhyner, A., Dvorak, J., & Müntener, M. (2000). Influence of age and duration of symptoms on fibre type distribution and size of the back muscles in chronic low back patients. European Spine Journal, 9, 273–281. Marras, W. S. (2000). Occupational low back disorder causation and control. Ergonomics, 43, 880–902. Marras, W. S., Lavender, S. A., Ferguson, S. A., Splittstoesser, R. E., & Yang, G. (2010). Quantitative dynamic measures of physical exposure predict low back functional impairments. Spine, 35, 914–923. Mawston, G. A., McNair, P. J., & Boocock, M. G. (2007). The effects of prior warning and lift-induced fatigue on trunk muscle and postural responses to sudden loading during manual handling. Ergonomics, 50, 2157–2170. McGill, S. M., & Brown, S. (1992). Creep response of the lumbar spine to prolonged full flexion. Clinical Biomechanics, 17, 43–46. Miller, E. M., Slota, G. P., Agnew, M. J., & Madigan, M. L. (2010). Females exhibit shorter paraspinal reflex latencies than males in response to sudden trunk flexion perturbations. Clinical Biomechanics, 25, 541–545. Olson, M. W. (2011). Passive repetitive loading of the lumbar tissues influences force output and EMG during maximal efforts. European Journal of Applied Physiology, 111, 1269–1278. Olson, M. W., Li, L., & Solomonow, M. (2004). Flexion–relaxation response to cyclic lumbar flexion. Clinical Biomechanics, 19, 769–776. Olson, M. W., Li, L., & Solomonow, M. (2009). Interaction of viscoelastic tissue compliance with lumbar muscles during passive cyclic flexion-extension. Journal of Electromyography and Kinesiology, 19, 30–38. Parkinson, R. J., Beach, T. A. C., & Callaghan, J. P. (2004). The time-varying response of the in vivo lumbar spine to dynamic repetitive flexion. Clinical Biomechanics, 19, 330–336. Place, N., Maffiuletti, N. A., Bailay, Y., & Lepers, R. (2005). Twitch potentiation is greater after a fatiguing submaximal isometric contraction performed at short vs. long quadriceps muscle length. Journal of Applied Physiology, 98, 429–436. Potvin, J. R., & O’Brien, P. R. (1998). Trunk muscle co-contraction increases during fatiguing, isometric, lateral bend exertions: Possible implications for spine stability. Spine, 23, 774–780. Rogers, E. L., & Granata, K. P. (2006). Disturbed paraspinal reflex following prolonged flexion-relaxation and recovery. Spine, 31, 839–845. Sanchez-Zuriaga, D., Adams, M. A., & Dolan, P. (2010). Is activation of the back muscles impaired by creep or muscle fatigue? Spine, 35, 517–525. Sbriccoli, P., Youssef, K., Kiperstein, I., Solomonow, M., Zhou, B., Shu, M., et al (2004). Static load repetition is a risk factor in the development of lumbar cumulative musculoskeletal disorder. Spine, 29, 2643–2653. Shin, G., D’Souza, C., & Liu, Y. H. (2009). Creep and fatigue development in the low back in static flexion. Spine, 34, 1873–1878.

M.W. Olson / Human Movement Science 34 (2014) 12–27

27

Shin, G., & Mirka, G. (2007). An in vivo assessment of the low back response to prolonged flexion; interplay between active and passive tissues. Clinical Biomechanics, 22, 965–971. Solomonow, M., Baratta, R. V., Banks, A., Freudenberger, C., & Zhou, B. H. (2003). Flexion-relaxation response to static lumbar flexion in males and females. Clinical Biomechanics, 18, 273–279. Solomonow, M., Zhou, B. H., Baratta, R. V., Lu, Y., & Harris, M. (1999). Biomechanics of increased exposure to lumbar injury due to cyclic loading: I. Loss of reflexive muscular stabilization. Spine, 24, 2426–2434. Solomonow, M., Zhou, B. H., Baratta, R. V., Lu, Y., Zhu, M., & Harris, M. (2000). Bi-exponential recovery model of lumbar viscoelastic laxity and reflexive muscular activity after prolonged cycling loading. Clinical Biomechanics, 15, 167–175. Stokes, I. A. F., Gardner-Morse, M., & Henry, S. M. (2011). Abdominal muscle activation increases lumbar spinal stability: Analysis of contributions of different muscle groups. Clinical Biomechanics, 26, 797–803. Stokes, I. A. F., Gardner-Morse, M., Henry, S. M., & Badger, G. J. (2000). Decrease in trunk muscular response to perturbation with preactivation of lumbar spinal musculature. Spine, 25, 1957–1964. Taimela, S., Kankaanpää, M., & Luoto, S. (1999). The effects of lumbar fatigue on the ability to sense a change in lumbar position. Spine, 24, 1322–1331. Taylor, J. L., Butler, J. E., & Gandevia, S. C. (2000). Changes in muscle afferents, motoneurons and motor drive during fatigue. European Journal of Applied Physiology, 83, 106–115. Taylor, J. L., Todd, G., & Gandevia, S. C. (2006). Evidence for a supraspinal contribution to human muscle fatigue. Clinical and Experimental Pharmacology and Physiology, 33, 400–405. Tokuno, C. D., Cresswell, A. G., Thorstensson, A., & Carpenter, M. G. (2013). Recruitment order of the abdominal muscles varies with postural task. Scandinavian Journal of Medicine & Science in Sports, 23, 349–354. van Dieën, J. H., Böke, B., Oosterhuis, W., & Toussaint, H. M. (1996). The influence of torque and velocity on erector spinae muscle fatigue and its relationship to changes of electromyogram spectrum density. European Journal of Applied Physiology, 72, 310–315. Vera-Garcia, F. J., Brown, S. H. M., 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, 443–455. Wilder, D. G., Aleksiev, A. R., Magnusson, M. L., Pope, M. H., Spratt, K. F., & Goel, V. K. (1996). Muscular response to sudden load. A tool to evaluate fatigue and rehabilitation. Spine, 21, 2628–2639. Wilson, E. L., Madigan, M. L., Davidson, B. C., & Nussbaum, M. A. (2006). Postural strategy changes with fatigue of the lumbar extensor muscles. Gait & Posture, 23, 348–354.