Clinical Biomechanics 32 (2016) 124130

Contents lists available at ScienceDirect

Clinical Biomechanics

journal homepage: www.elsevier.com/locate/clinbiomech

Muscle activation timing and balance response in chronic lower back

pain patients with associated radiculopathy
Lydia R. Frost, Stephen H.M. Brown
Department of Human Health and Nutritional Sciences, University of Guelph, Guelph, ON, Canada

a r t i c l e i n f o a b s t r a c t

Article history: Background: Patients with chronic low back pain and associated radiculopathy present with neuromuscular
Received 9 September 2015 symptoms both in their lower back and down their leg; however, investigations of muscle activation have so
Accepted 2 December 2015 far been isolated to the lower back. During balance perturbations, it is necessary that lower limb muscles activate
with proper timing and sequencing along with the lower back musculature to efciently regain balance control.
Keywords: Methods: Patients with chronic low back pain and radiculopathy and matched controls completed a series of
Low back pain
balance perturbations (rapid bilateral arm raise, unanticipated and anticipated sudden loading, and rapid rise
Sciatica
Balance
to toe). Muscle activation timing and sequencing as well as kinetic response to the perturbations were analyzed.
Perturbation Findings: Patients had signicantly delayed lower limb muscle activation in rapid arm raise trials as compared to
Muscle activation timing controls. In sudden loading trials, muscle activation timing was not delayed in patients; however, some differ-
Kinetics ences in posterior chain muscle activation sequencing were present. Patients demonstrated less anteriorposte-
rior movement in unanticipated sudden loading trials, and greater mediallateral movement in rise to toe trials.
Interpretation: Patients with low back pain and radiculopathy demonstrated some signicant differences from
control participants in terms of muscle activation timing, sequencing, and overall balance control. The presence
of differences between patients and controls, specically in the lower limb, indicates that radiculopathy may play
a role in altering balance control in these patients.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
Chronic lower back pain (LBP) is a highly prevalent musculoskeletal
disorder, affecting nearly 80% of individuals at some point in their life-
time (Andersson, 1999; Cassidy, 1998). Chronic LBP is a heterogeneous
pathology with a variety of associated pathophysiological conditions.
Due to the wide scope of LBP, it can be difcult to target specic treat-
ments to chronic patients; much of the research to date has focused
on chronic non-specic LBP. In order to fully understand specic subsets
of chronic LBP, it is imperative that more focussed research is
conducted.
One of the most common conditions associated with chronic LBP is
radiculopathy (LBP-R), also known as sciatica. The lifetime prevalence
of sciatica ranges from 1.2% to 43% (Konstantinou and Dunn, 2008). In
patients under the age of 50, LBP-R is most frequently the result of an
intervertebral disc herniation or bulge in the lumbar spine that com-
presses on the nerve root; in later years, lumbar spinal stenosis is the
most likely cause (Tarulli and Raynor, 2007). Overall, approximately
90% of sciatica cases are the result of a disc herniation or bulge (Valat
Corresponding author at: Department of Human Health and Nutritional Sciences,
University of Guelph, 50 Stone Rd East, Guelph, ON N1G2W1, Canada. Tel.: +1 519 824
4120x53651; fax: +1 519 763 5902.
E-mail address: shmbrown@uoguelph.ca (S.H.M. Brown).
et al., 2010). Compression of the nerve, and the associated inammatory
process, contributes to symptoms of radiating pain, tingling or numb-
ness down the course of the sciatic nerve and its branches. Symptoms
are reported in the gluteal, hamstring, posterior leg, and foot regions
(Van Boxem et al., 2010). It is unknown how these lower limb symp-
toms of chronic nerve compression are related to the functional control
of lower limb muscle activation.
Electromyography (EMG) has been used to demonstrate altered
motor control of trunk muscles in LBP patient groups. Leinonen et al.
(2001) tested participants with disc herniation-related LBP in response
to unexpected and expected upper limb perturbations and found that
feed-forward activation of erector spinae (ES) and multidus were
slower compared to controls in expected perturbations, but not in
unexpected perturbations. In contrast, Magnusson et al. (1996) found
delayed ES muscle activation in unilaterally affected LBP patients as
compared to controls in response to an unexpected perturbation, as
well as a further delay in activation on the painful side relative to the
contralateral side. This research to date has been isolated to the trunk
musculature.
Following a balance perturbation, the goal of the neuromuscular
system is to effectively regain balance equilibrium. One widely accepted
way to assess the effectiveness of the whole-body balance control is to
record and analyze force plate data acquired during experimentally in-
duced perturbations. Mok et al. conducted two studies that investigated

http://dx.doi.org/10.1016/j.clinbiomech.2015.12.001
0268-0033/ 2015 Elsevier Ltd. All rights reserved.
 L.R. Frost, S.H.M. Brown / Clinical Biomechanics 32 (2016) 124130
balance control strategies in non-specic LBP patients using voluntary
arm raise trials (Mok et al., 2011a) and sudden loading trials (Mok
et al., 2011b). In both studies, there were no differences in COP
excursion along the anteriorposterior (AP) axis between patients and
controls; however, COP stabilization metrics demonstrated that pa-
tients took longer to stabilize following a perturbation and required
more ne-tuning adjustments during the stabilization process. In partial
contrast to this, Frost et al. (2015) found no difference between LBP-R
patients and healthy controls in balance stabilization measures during
rapid arm raise trials, but like Mok et al., they also found no difference
in COP excursion measures.
Possible detriments in whole-body balance control in LBP-R patients
are not isolated to deciencies in trunk muscle coordination; proper
lower limb muscle activation and sequencing is also imperative for bal-
ance control. There has been no prior research investigating lower limb
muscle activation in any balance perturbations of LBP patients. Thus, the
purpose of this work was to investigate functional muscle activation
timing in the lower back and lower limb and determine if and how
this affects the whole-body kinetic response during balance perturba-
tions in a chronic LBP-R population. It was hypothesized that LBP-R pa-
tients would demonstrate delays in muscle activation timing of the
lower back and lower limb musculature during balance perturbations
as compared to control participants. Second, LBP-R patients would
have altered whole-body balance control during balance perturbations,
as evidenced by kinetic (force plate) analysis. Specically, LBP-R
patients would demonstrate no difference in COP excursions and veloc-
ities along the anteriorposterior (AP) and mediallateral (ML) axes of
motion; however, it would require more time and a greater number of
direction changes to stabilize their AP COP velocity back to baseline
levels following perturbations in the AP plane compared to controls.
2. Methods
2.1. Participant characteristics
Patients with LBP-R (n = 17) and matched healthy control partici-
pants (n = 17) were recruited. LBP-R patients had clinically diagnosed
lumbar intervertebral disc herniation or bulge resulting in lower back
pain and unilateral radiculopathy symptoms (at least one of pain, tin-
gling or numbness radiating down the leg and/or into the foot) for a
minimum of 3 months. Control participants (n = 17) had no history
of chronic LBP, musculoskeletal disorder, or neurological decit and
were matched to LBP-R patients for age, sex, mass, height, foot domi-
nance, and physical activity status (Table 1). Both LBP-R and control
participants completed a medical questionnaire pertaining to musculo-
skeletal and neurological disorders, an Oswestry Disability Index (ODI),
visual analog scales (VAS) of pain for the lower back and legs, a Water-
loo Footedness Questionnaire (all participants were right foot
dominant), and a Baecke score of physical activity. Institutional ethics
board approval was acquired, and all participants signed informed
consent.
2.2. Experimental protocol
Four balance perturbations were conducted; three successful trials
of each balance perturbation were recorded. Participants stood in a
Table 1
Participant characteristics of LBP-R patients (n = 17) and matched healthy controls (n = 17) (mean SD).
Age Sex (M:F) Height (m) Mass (kg)
LBP-R 44.2 14.9 6:11 1.69 7.7 69.5 9.2
Control 44.2 15.7 6:11 1.72 7.9 69.1 10.8
1
Oswestry Disability Index, on a percentage scale; scores 020% indicate minimal disability, 2040% indicate moderate disability, 4060% severe disability.
2
Visual Analog Scale of pain recorded at the back and the affected leg on separate scales, on a scale of 010. A score of 00.4 indicates no pain, 0.54.4 indicates mild pain, 4.57.4
indicates moderate pain, 7.510 indicates severe pain.
3
Baecke score of physical activity for work, sport and leisure activity.
125
self-selected neutral stance on a force plate (AccuSway, AMTI, Water-
town MA). During all trials, one kinematic infrared marker (Optotrak
3D Investigator, Northern Digital, Waterloo ON, Canada) was adhered
to a body landmark (wrist or ankle) using double sided tape, to allow
for detection of onset and velocity of movement. Kinematic data were
sampled at 512 Hz. Additionally, EMG data were recorded during all bal-
ance perturbation trials. Participants were instrumented with 8 pairs of
disposable Ag/AgCl surface EMG electrodes, placed bilaterally on the
muscle bellies of erector spinae (ES) at L4, biceps femoris (BF), medial
gastrocnemius (MG), and tibialis anterior (TA). Raw EMG signals were
band-pass ltered at 101000 Hz, amplied (AMT-16, Bortec, Calgary
AB, Canada), and sampled at 2048 Hz.
Participants rst completed the voluntary rise to toe perturbation
(Frank et al., 2000). For these trials, the kinematic infrared marker was
adhered to the right ankle. Participants began standing on the force
plate in a normal-width stance with eyes open, then on a countdown
verbal cue, they were instructed to rise up onto their toes as rapidly as
possible, maintain this position for a count of 3 s, and then slowly
lower back down to neutral stance. A successful trial required that
participants did not take a step to recover balance and maintained the
position up on their toes for 3 s.
The second balance perturbation was a rapid bilateral arm raise
(Hodges and Richardson, 1996), in which the kinematic infrared marker
was placed on the right wrist. Participants began standing in a normal-
width stance with their eyes open and arms at their sides, then on a ver-
bal countdown, they exed their arms at the shoulder to 90 as rapidly
as possible, held that position for a count of 3 s, then slowly lowered
their arms to the starting position.
The nal two balance perturbations were applied by external sud-
den loads. Participants held a box while maintaining a 90 elbow
angle; the kinematic marker was adhered to the wrist. A 2.3 kg mass
was dropped into the box by an experimenter, from a height of 5 cm
(Leinonen et al., 2001; Mok et al., 2011b). The participant was instructed
to respond to the perturbation by returning the box to its original posi-
tion as soon as possible. The participant's view of the box was obscured
so that he/she could not see when the mass was dropped. This proce-
dure was conducted under two conditions: unanticipated (10 s window
of time in which the mass could be dropped) and anticipated (experi-
menter verbal countdown to mass drop).
2.3. Data analysis
A custom program was used determine bilateral muscle activation
timing within 100 ms prior to, and 1000 ms following movement
onset, as determined by the kinematic marker. EMG was processed as
a linear envelope by rectication followed by a dual low-pass 4th-
order Butterworth lter with a cutoff frequency of 50 Hz. The threshold
for activation onset was dened as the processed EMG exceeding 3 stan-
dard deviations above baseline for at least 20 ms; baseline activation
was averaged from 250 ms to 150 ms prior to kinematic movement
onset. All muscle activation onsets were visually inspected and
conrmed.
Force plate data were processed by ltering using a dual low-pass
4th-order Butterworth lter with a cutoff frequency of 6 Hz. The centre
of pressure (COP) in the AP and ML directions was calculated from the
recorded forces and moments. For all trials, the following parameters
ODI (%)1 VAS back2 VAS leg2 Baecke score3 Duration (months)
19.9 12.8 2.6 2.3 1.8 1.2 8.18 1.4 126 143
0.69 1.3 0.3 0.4 0.3 0.4 8.35 1.4 0
126 L.R. Frost, S.H.M. Brown / Clinical Biomechanics 32 (2016) 124130
were computed along both AP and ML axes of motion: maximum COP
excursion (cm), maximum COP velocity (cm/s), and mean absolute
value of the COP velocity (cm/s). Excursions in the ML plane were quan-
tied in terms of going toward the unaffected leg or toward the affected
leg; control participants were assigned unaffected and affected legs to
mimic their LBP-R match. Additional COP stabilization parameters
were analyzed for the balance recovery phase, dened as the 2 s follow-
ing movement initiation (as measured by the kinematic marker). Time
to recover stability was dened as the time until the rectied COP AP ve-
locity (cm/s) returned back to a baseline value (average of 1000 ms to
500 ms prior to the perturbation) and remained there for 50 ms. The
number of corrections (direction changes) was calculated as the num-
ber of times that non-rectied COP AP velocity (cm/s) crossed the
zero-point from the time of perturbation until the previously dened
return to baseline.
2.4. Statistical analyses
All statistical analyses were completed using SAS 9.2. Normality of
data was veried using the ShapiroWilk W statistic. For force plate
parameters, one-way analyses of variance (ANOVA) were completed
to compare between groups (LBP-R and control) for each outcome mea-
sure. For EMG muscle activation timing, 2-way ANOVAs were complet-
ed for each type of balance perturbation, to compare between groups
(control, LBP-R unaffected side, LBP-R affected side) and muscles (ES,
BF, MG, TA). Further, a 2-way ANOVA was completed to compare be-
tween sudden loading type (unanticipated vs anticipated) and muscle
(ES, BF, MG, TA). Main effects and interactions were examined. For all,
Tukey-adjusted post hoc tests of signicant main effects were used. Ad-
ditionally, Pearson correlations were computed between measures of
LBP-R severity (ODI, VAS at the back, VAS at the affected leg, and dura-
tion of symptoms) and each outcome measure. All statistical analyses
were completed at a signicance level of = 0.05.
3. Results
3.1. Muscle activation
For the sudden loading trials, no differences in activation onset times
of individual muscles were found between LBP-R patients on their unaf-
fected versus affected sides, or versus matched controls (Table 2, Fig. 1).
However, when muscle activation sequencing was compared between
LBP-R patients and control participants, there were some interesting
relationships. In the anticipated condition, participants in all groups
demonstrated a relatively sequenced activation pattern by activating
rst the ES (4248 ms), then activating from superior to inferior in the
order of BF (5661 ms) followed by MG (5865 ms) (Fig. 1). Interesting-
ly, BF and MG activation was signicantly later (p b 0.05) than ES activa-
tion in control participants and on the unaffected side of LBP-R patients,
but on the affected side of LBP-R patients these muscles were not
Table 2
Mean SE EMG muscle activation timing (ms) of the control participants, LBP-R patients on their unaffected leg, and LBP-R patients on their affected leg. A negative value indicates muscle
activation prior to the onset of movement. Data from all four balance perturbations are presented.
ES
Rise to toe Control 23.8 11.3
LBP-R unaffected 57.5 16.3
LBP-R affected 36.2 15.6
Arm raise Control 96.4 10.1
LBP-R unaffected 90.9 13.3
LBP-R affected 89.3 13.2
Sudden loading Control 62.4 9.7
Unanticipated LBP-R unaffected 64.9 13.2
LBP-R affected 73.3 13.4
Sudden loading Control 42.4 10.4
Anticipated LBP-R unaffected 43.8 13.3
LBP-R affected 48.3 13.6
activated at signicantly different times (p N 0.1). This pattern was not
apparent in the unanticipated trials, where participants in all groups ac-
tivated all three muscles at similar times. In both conditions, the TA
muscle was activated signicantly later than other muscles (p b
0.0001; unanticipated 8899 ms, anticipated 8393 ms), and was not
presented in Fig. 1 to enable direct comparison of posterior chain mus-
cle responses to the perturbation.
Bilateral arm raise trials represent an internally generated perturba-
tion, and therefore involve voluntary activation of muscles prior to the
onset of movement. Angular velocity averaged over the 90 of arm
raise (degrees/second) was not signicantly different between LBP-R
participants (mean SE = 256.2 13.2/s) and control participants
(mean SE = 250.0 10.9 /s). MG muscle activation was signicantly
delayed on the affected side of LBP-R patients (13.4 ms) as compared
to controls ( 56.3 ms) (p b 0.01); however, there were no between-
group signicant differences in activation timing of the ES or BF muscles
(Table 2, Fig. 2). To perform the voluntary arm raise, both patients and
controls used a similar sequence strategy of activating the ES and BF
muscles 85 to 96 ms prior to arm movement, followed by signicantly
later activation of the MG muscle (p b 0.01), and then by the initiation
of arm movement. The activation of the TA muscle was signicantly
later than all other muscles (p b 0.01) and was used as stabilization ei-
ther before or after the initiation of arm movement (Table 2). TA activa-
tion was signicantly delayed on the affected side of LBP-R patients, as
compared to controls (p b 0.01).
Muscle activation during the rise to toe perturbation was less consis-
tent, reecting the varied ability of participants to complete this task
(Table 2). The average rise to toe velocity (mm/s) was not signicantly
different between LBP-R participants (mean SE = 165.4 14.6 mm/s)
and control participants (mean SE = 156.7 10.4 mm/s). The MG
muscle was activated 80 to 92 ms prior to onset of ankle movement,
which was signicantly earlier than the activation of all other muscles
(p b 0.01), and was not signicantly different between groups. Similarly,
the BF muscle did not demonstrate between-group differences and was
activated slightly before (13 ms) or after (2 ms) onset of movement.
In order to stabilize during the perturbation, the ES and the TA muscles
turned on following movement onset: ES at 23 to 56 ms, TA at 12 to
71 ms. The activation of the TA muscle was slightly, although not statisti-
cally signicantly (p = 0.08), delayed on the affected leg of the LBP-R
patients (71 ms) as compared to controls (12 ms).
3.2. Balance
In both sudden loading conditions (unanticipated and anticipated),
LBP-R patients had signicantly lower magnitude COP measures for
some parameters as compared to matched controls (Table 3). Specical-
ly, in the unanticipated condition, LBP-R patients had lower maximum
anterior COP excursion (p = 0.05) and velocity (p = 0.05) and lower
maximum posterior COP velocity (p b 0.01). In the anticipated condi-
tion, LBP-R patients had smaller maximum anterior (p = 0.01) and
BF MG TA
2.9 10.1 82.1 6.1 12.7 14.9
13.5 12.3 80.4 8.3 49.7 20.0
8.2 12.0 92.5 8.4 71.4 20.7
86.9 9.7 56.3 6.8 4.1 25.9
96.1 11.5 36.8 9.2 7.8 44.5
85.6 12.0 13.4 9.6 49.8 29.2
66.2 9.2 67.0 6.0 88.9 13.5
68.1 12.0 59.3 8.4 93.2 18.8
71.3 12.2 60.9 8.5 99.8 18.7
56.6 9.4 58.6 6.2 83.5 13.5
61.5 12.0 65.8 8.3 90.1 19.3
57.8 12.4 60.0 9.0 93.8 21.4
 L.R. Frost, S.H.M. Brown / Clinical Biomechanics 32 (2016) 124130 127

Fig. 1. Mean SE EMG muscle activation timing (mean SE) for ES, BF, and MG during the sudden loading perturbations in unanticipated [A] and anticipated [B] conditions. The per-
turbation occurred at 0 ms, as indicated by the black arrow. Data are represented for control participants, LBP-R patients on their unaffected leg, and LBP-R patients on their affected leg.

posterior (p b 0.01) COP velocity than controls. Across all outcome mea-
sures, COP measures were of larger magnitude in the unanticipated con-
dition, as compared to anticipated.
The rise to toe perturbations also demonstrated some signicant
between-group differences (Table 3). Specically, LBP-R patients had
larger maximum COP excursions in the anterior direction (p b 0.01)
and toward the unaffected side (p b 0.05), larger maximum COP velocity
toward both the unaffected (p b 0.05) and affected sides (p b 0.01), and
greater mean absolute ML velocity (p b 0.05), as compared to controls.
COP measures during the arm raise perturbations did not demon-
strate any signicant differences between LBP-R patients and controls
(Table 3). Further, there were no between-group differences in COP sta-
bilization latency or number of crossings during arm raise or sudden
loading trials (Fig. 3).

Fig. 2. Mean SE EMG muscle activation timing (mean SE) for ES, BF, and MG during the bilateral arm raise perturbation. The initiation of arm raise movement was at 0 ms, as indicated
by the black arrow. Data are represented for control participants, LBP-R patients on their unaffected leg, and LBP-R patients on their affected leg. * p b 0.05.
128 L.R. Frost, S.H.M. Brown / Clinical Biomechanics 32 (2016) 124130

Table 3
Mean SE anteriorposterior [A] and mediallateral [B] force plate COP parameters for LBP-R and control participants in each of the balance perturbations. Measures that are signicantly
different (p b 0.05) between LBP-R and control groups are shown in bold.

Maximum Maximum posterior Maximum anterior Maximum Mean absolute

anterior excursion (cm) velocity posterior velocity AP
excursion (cm/s) (cm/s) velocity
(cm) (cm/s)

Rise to toe LBP-R 12.1 0.31 1.2 0.11 78.1 4.1 17.9 1.8 8.8 0.33
Control 10.7 0.34 1.5 0.12 74.6 4.5 17.3 2.0 8.7 0.36
Arm raise LBP-R 1.9 0.29 0.29 0.12 12.4 4.0 6.1 1.8 2.3 0.32
Control 1.9 0.31 0.39 0.11 12.9 4.4 6.8 1.8 2.6 0.36
Sudden loading LBP-R 3.9 0.30 0.61 0.09 36.9 3.9 18.2 1.8 5.5 0.37
unanticipated Control 4.9 0.33 0.82 0.12 49.9 4.4 34.3 1.9 7.2 0.35
Sudden loading LBP-R 3.2 0.29 0.36 0.17 25.0 3.6 9.6 1.6 3.9 0.32
anticipated Control 3.8 0.33 0.49 0.11 36.1 4.4 19.3 1.5 5.3 0.36

Maximum Maximum excursion Maximum velocity to Maximum Mean absolute

excursion to to affected side unaffected side velocity to ML velocity
unaffected side (cm) (cm/s) affected side (cm/s)
(cm) (cm/s)

Rise to toe LBP-R 1.7 0.10 1.4 0.12 16.9 1.0 18.9 1.2 4.7 0.19
Control 1.3 0.11 1.5 0.13 13.5 1.1 14.3 1.3 3.9 0.20
Arm raise LBP-R 0.7 0.10 0.5 0.10 6.6 0.94 5.8 1.0 1.5 0.18
Control 0.5 0.09 0.5 0.11 4.5 1.0 5.2 1.1 1.4 0.20
Sudden loading LBP-R 0.6 0.09 0.6 0.11 7.1 0.93 7.9 1.0 1.7 0.18
unanticipated Control 0.8 0.10 0.7 0.12 8.0 1.0 8.7 1.2 1.9 0.19
Sudden loading LBP-R 0.6 0.10 0.4 0.12 5.3 0.97 4.5 1.0 1.3 0.18
anticipated Control 0.6 0.08 0.6 0.13 5.3 1.1 5.3 1.2 1.4 0.20

3.3. Correlations with LBP-R severity and |r| N 0.20) with ODI, VAS at the lower back, VAS at the affected
leg, or symptom duration.
Across all muscle activation and balance outcome measures, there
were no statistically signicant Pearson correlations (p-value b 0.05 4. Discussion

The primary nding of the current study is that, in response to

whole-body balance perturbations, LBP-R patients demonstrated some
A delays in MG and TA activation as well as evidence of altered muscle se-
quencing in comparison to matched healthy controls. In terms of the ki-
netic response, LBP-R patients demonstrated greater COP movement
during the rise to toe perturbation, yet lesser movement in response
to sudden loading perturbations.

4.1. Sudden loading perturbations

The activation timing of individual muscles during the sudden load-

B muscle activation in LBP patients has been identied in anticipated
Fig. 3. Mean SE stabilization latency [A] and number of crossings [B] of the AP COP ve-
locity in the 2 s following movement initiation for control and LBP-R participants.
ing trials was not different between LBP-R patients on either their
affected or unaffected sides, nor with control participants. This is con-
trary to ndings in the literature that have investigated muscle activa-
tion timing at the lower back musculature, where delayed lower back
(e.g., Leinonen et al., 2001), or both anticipated and unanticipated sud-
den trunk loading perturbations (Magnusson et al., 1996). Our results
may not correspond to the ndings of Leinonen et al. (2001), as they
studied a more severely affected LBP group, with an average ODI of
38.4% and VAS of 6.8 (on a scale of 010), compared to average values
of 19.9% and 2.6, respectively, in our population; Magnusson et al.
(1996) did not specify LBP patient severity. However, our ndings are
in agreement with Lariviere et al. (2010), who found no difference
between LBP patients (VAS of approximately 3.6) and controls in ES
muscle activation timing during unanticipated sudden trunk loading.
Relating differences in outcomes to patient severity is difcult, as
patients rate severity subjectively, which possibly reects the lack of
signicant correlations between severity and outcomes in the current
study. However, this lack of correlation could also be inuenced by
the relatively low variance in severity in the studied patient population.
 L.R. Frost, S.H.M. Brown / Clinical Biomechanics 32 (2016) 124130
No previous studies have investigated muscle activation timing in
the lower limb musculature of LBP patients in response to sudden
trunk loading perturbations. We did not see any signicant differences
in LBP-R patients, between affected and unaffected legs or compared
to controls, in the timing of lower limb muscles in these sudden loading
conditions; however, this may be due to the relatively moderate level of
pain and disability in our LBP-R population.
In comparing anticipated to unanticipated sudden loading perturba-
tions, we hypothesized more between-group differences in the antici-
pated condition due to potential decits in feed-forward planning in
the LBP group (Leinonen et al., 2001). However, we did not observe
this relationship. LBP-R patients were not signicantly different from
control participants in the timing of individual muscles in either unan-
ticipated or anticipated trials. What we did observe, however, was
that control participants as well LBP-R patients on their unaffected
side demonstrated a more clear top-down sequence of muscle activa-
tion in response to anticipated sudden loading, by activating rst ES,
then signicantly later BF and MG. Alternatively, LBP-R patients on
their affected side co-activated all muscles at the same time. This co-
activation pattern was similar to how participants in all groups
responded to the unanticipated sudden loading perturbations.
Linking this to the kinetic balance response, LBP-R patients demon-
strated lower COP excursions and velocities when compared to controls,
which indicates a stiffer balance control strategy. This could result from
greater muscle co-activation in anticipation of the perturbation, which
would agree with the pain-adaptation model (van Dien et al., 2003).
4.2. Arm raise perturbations
In preparation for a rapid bilateral arm raise, ES and BF muscles did
not demonstrate any differences in activation timing between the
affected and unaffected sides of LBP-R patients, nor between LBP-R pa-
tients and controls. Delays in ES activation, and potentially lower limb
muscle activation, were expected based on previous research that
showed delays in multidus, as estimated by surface EMG, in non-
specic chronic LBP patients during voluntary arm movement perturba-
tions (Mehta et al., 2010). However, the LBP patients studied by Mehta
et al. (2010) were slightly more severely affected than our cohort, and
their surface EMG placement was designed to target multidus as
opposed to ES, which might explain the differences between their nd-
ings and ours.
No previous research has examined muscle activation of lower limb
muscles in LBP-R patients in response to an arm raise perturbation;
therefore, this is the rst report of delays in MG and TA activation. Dur-
ing a rapid bilateral arm raise, MG activates in a feed-forward manner,
following ES and BF activation, to assist with pulling the centre of
mass of the body posteriorly to prepare for the anterior shift in centre
of mass that occurs during an arm raise. A later activation of MG may
impair the feed-forward preparation for this movement; however,
there are no between-group differences in the kinetic balance outcome
measures for the arm raise trials, so it appears that the effect of this MG
activation delay in LBP-R patients is minimal. TA activates slightly prior
to the onset of arm movement in controls and LBP-R patients on their
unaffected side; however, activation is delayed on the affected side of
LBP-R patients. The activation of TA functions to stabilize the lower
limb by providing a counter balanced response to the feed-forward ac-
tivation of posterior chain muscles.
Despite the delays in MG and TA activation, there were no signicant
differences between LBP-R patients and control participants in any of
the measured COP balance parameters in the bilateral arm raise pertur-
bation. LBP-R patients execute their arm raise at the same angular veloc-
ity as controls; therefore, they were capable of kinetically preparing and
responding to this voluntary balance perturbation as well as controls.
This nding corresponds to the AP COP excursion and velocity results
of Mok et al. (2011a) in a group of LBP patients of similar levels of dis-
ability and pain. However, Mok et al. (2011a) reported that LBP patients
129
required more time and more corrections to stabilize their AP COP ve-
locity after arm movement, a nding that was not reproduced in the
present study.
4.3. Rise to toe perturbations
During the rise to toe perturbations, MG was activated rst to pre-
pare for the movement, followed by BF activation at the approximate
time when the movement of the heel was initiated, and then by ES
and TA muscles activating after movement initiation. TA is an antagonist
to MG; therefore, the activation of the TA functions to balance the MG
activation that propels the body up onto the toes. However, due to the
self-timed nature of the task, the activation timing of TA was quite var-
iable in all participants, both LBP-R and control. The clinical and func-
tional relevance of the non-statistically signicant delay in TA
activation in the affected leg of LBP-R patients may be questionable;
however, kinetic differences between groups were found, as detailed
next.
Rise to toe perturbations have not previously been investigated in
LBP or LBP-R patients; therefore, the nding of greater COP excursions
and velocities in the ML plane of the patient group is novel. Larger
COP excursion toward the unaffected side indicates that LBP-R patients
were favouring their radiculopathy-unaffected side when rising up onto
their toes and stabilizing in that position. Further, the increased ML COP
velocity (both toward the unaffected and affected limbs) provides in-
sight into the control capability; larger COP velocity may be interpreted
as poorer balance control in the LBP-R patients.
4.4. Mechanisms of delay
Across all investigations, it was hypothesized that LBP-R patients
would have delayed muscle activation timing relative to controls; this
was hypothesized for two reasons. First, nerve conduction velocity de-
creases in chronic nerve compression pathologies. Kim et al. (2013)
found a positive correlation between swelling (increased CSA) of the
median nerve and decreased median nerve conduction velocity in car-
pal tunnel syndrome. In a concurrent investigation on the same patient
population as presented here, greater sciatic nerve CSA was found in the
affected leg of the LBP-R patient population (Frost and Brown, 2015),
which could be expected to therefore correspond to decreased nerve
conduction velocity. Further, in the pain-adaptation model, it is pro-
posed that in a pain state motor neuron excitability of agonist muscles
is reduced, thereby reducing the magnitude and increasing the latency
of neuromotor commands and responses (Lund et al., 1991; van Dien
et al., 2003). In the current study, muscle activation timing was delayed
in only a limited number of conditions; therefore, it appears that this
population of LBP-R patients is likely experiencing only relatively
minor decreases in sciatic nerve conduction velocity or motor system
timing-based adaptations to pain.
4.5. Limitations
To ensure that all participants were able to safely and comfortably
complete data collection, we chose to not record maximum voluntary
contractions of the muscles studied; instead, we focused on the muscle
activation timing in response to the balance perturbations. However,
this limited our ability to interpret muscle activation magnitudes during
the balance responses. If we had access to measures of muscle activation
magnitude, we would have been better able to test theories related to
muscle co-contraction and stiffening responses related to pain-
adaptation.
5. Conclusion
The purpose of this research was to expand previous LBP neuromus-
cular research that focussed on the lower back and trunk, into the lower
130 L.R. Frost, S.H.M. Brown / Clinical Biomechanics 32 (2016) 124130
limb of LBP-R patients to determine the neuromuscular effect of
radiculopathy. Muscles in the lower back and lower limb of LBP-R pa-
tients were activated with a different sequencing in anticipated sudden
loading conditions, and lower limb muscles (MG and TA) demonstrated
delayed activation in certain instances. Further, evidence of differing
balance control strategies was apparent when kinetic data were ana-
lyzed. Specically, LBP-R patients had less AP COP movement during un-
anticipated sudden loading, potentially due to a stiffer control strategy,
and more ML COP movement in voluntary rise to toe perturbations,
demonstrating poorer balance control as well as an unloading strategy.
Overall, despite these differences, neuromuscular changes were not
consistently present in the lower limb; however, as they were not pres-
ent in the lower back region either, it is possible that the moderately
affected LBP-R patient group studied here functioned at a comparable
neuromuscular level to controls, and more substantial differences
could be detected in a more severely affected group.
Acknowledgments
The authors wish to acknowledge the assistance of Grace
Glofcheskie in data collection.
References
Andersson, G.B., 1999. Epidemiological features of chronic low-back pain. Lancet 354,
581585.
Cassidy, J.D., 1998. Saskatchewan health and back pain survey. Spine 23, 18601867.
Frank, J.S., Horack, F.B., Nutt, J., 2000. Centrally initiated postural adjustments in
partinsonian patients on and off levodopa. J. Neurophysiol. 84 (5), 24402448.
Frost, L.R., Bijman, M., Strzalkowski, N.D.J., Bent, L.R., Brown, S.H.M., 2015. Decits in foot
skin sensation are related to alterations in balance control in chronic low back
patients experiencing clinical signs of lumbar nerve root impingement. Gait Posture
41 (4), 923928.
Hodges, P.W., Richardson, C.A., 1996. Inefcient muscular stabilization of the lumbar
spine associated with low back pain: a motor control evaluation of transverses
abdominis. Spine 21, 26402650.
Kim, J.M., Kim, M.W., Ko, Y.J., 2013. Correlating ultrasound ndings of carpal tunnel
syndrome with nerve conduction studies. Muscle Nerve 48, 905910.
Konstantinou, K., Dunn, K.M., 2008. Sciatica review of epidemiological studies and preva-
lence estimates. Spine 33, 24642472.
Lariviere, C., Forget, R., Vadeboncoeur, R., Bilodeau, M., Mecheri, H., 2010. The effect of sex
and chronic low back pain on back muscle reex responses. Eur. J. Appl. Physiol. 109,
577590.
Leinonen, V., et al., 2001. Disc herniation-related back pain impairs feed-forward control
of paraspinal muscles. Spine 26, 367372.
Lund, J.P., Donga, R., Widmer, C.G., Stohler, C.S., 1991. The pain-adaption model: a
discussion of the relationship between chronic musculoskeletal pain and motor
activity. Can. J. Physiol. Pharmacol. 69, 683694.
Magnusson, M.L., et al., 1996. European spine societythe AcroMed Prize for Spinal
Research 1995. Unexpected load and asymmetric posture as etiologic factors in low
back pain. Eur. Spine J. 5, 2335.
Mehta, R., Cannella, M., Smith, S.S., Siles, S.P., 2010. Altered trunk motor planning in pa-
tients with nonspecic low back pain. J. Mot. Behav. 42, 135144.
Mok, N.W., Brauer, S.G., Hodges, P.W., 2011a. Postural recovery following voluntary arm
movement is impaired in people with chronic low back pain. Gait Posture 34,
97102.
Mok, N.W., Brauer, S.G., Hodges, P.W., 2011b. Changes in lumbar movement in people
with low back pain are related to compromised balance. Spine 36, 4552.
Tarulli, A.W., Raynor, E.M., 2007. Lumbosacral radiculopathy. Neurol. Clin. 25, 387405.
Valat, J.-P., Genevay, S., Marty, M., Rozenberg, S., Koes, B., 2010. Sciatica. Best Pract. Res.
Clin. Rheumatol. 24, 241252.
Van Boxem, K., Cheng, J., Patjin, J., Van Kleef, M., Lataster, A., Mekhail, N., Van Zundert, J.V.,
2010. Lumosacral radicular pain. Pain Pract. 10 (4), 339358.
van Dien, J.H., Selen, L.P.J., Cholewicki, J., 2003. Trunk muscle activation in low-back pain
patients, an analysis of the literature. J. Electromyogr. Kinesiol. 13, 333351.