Table 4.

Study characteristics (N = 21).

Authors	Study Aim	EMG Variable and Lumbar Paraspinal Muscles Recorded (LMU = Lumbar Multifidus, LES = Lumbar Erector Spinae, TES = Thoracic Erector Spinae)	Lumbar Kinematic Measurements	Study Findings	Participants	Analysis
Arjmand et al. (2010) [41]	To compare a single joint model to kinematic driven model during trunk flexion.	Normalised EMG activity. Muscles Longissimus (3 cm lateral to L1) Iliocostalis (3 cm lateral to L3) Multifidus (2 cm lateral to L5).	Optotrak 4 camera system (regional) Lumbar region LED’s placed on pelvis and T12.	In both models, global extensor activity peaked around 30° of flexion, due to the increase in contribution of passive structures at this point. Extensors became silent between 50–70°.	N = 1 A male participant with no recent history of LBP.	Quantitative comparison was not performed.
Burnett et al. (2004) [42]	To determine whether differences exist in spinal kinematics and trunk muscle activity in cyclists with and without NSCLBP.	EMG activity was quantified by obtaining the mean activation, during a 5 crank revolution period. Muscles TES (5 cm lateral to T9) LMU (2–3 cm lateral to L4–L5).	3-Space Fastrak (regional) Lower lumbar L3 relative to S2 Upper lumbar T12 relative to L3.	The LBP group demonstrated greater lower lumbar flexion than controls associated with a loss of multifidus co-contraction.	N = 18 mean age 37.6 years 9 non low back pain 9 NSCLBP.	Independent sample t-tests.
Callaghan and Dunk 2002 [43]	To determine if FRP occurs in seated and slumped postures.	Ensemble average normalised EMG activity. Muscles TES (5 cm lateral to T9) LES (3 cm lateral to L3).	3-Space ISOTRAK (regional) Lumbar region Sacrum relative to L1.	FRP was shown in the TES, but not the LES during Slumped sitting. TES silence during sitting also happened at earlier angle of lumbar flexion than during standing.	N = 22 low back pain free participants 11 males mean age 21.3 years 11 females mean age 21.9 years.	Three way ANOVA, and Tukey’s post hoc multiple comparisons.
Cholewicki et al. (1997) [44]	To test the hypothesis that the flexors and extensors of the trunk are co-activated around a neutral spine posture.	Normalised EMG activity. Muscles TES (5 cm lateral to T9) LES (3 cm lateral to L3) LMU (2 cm lateral to L5–L5).	The use of 2 pieces of string attached to a chest harness and two potentiometers (regional).	Co-activation of trunk flexors and extensors was shown in healthy participants around a neutral posture.	N = 10 low back pain free participants 8 males and 2 females mean age 27 years.	A two factor repeated measures ANOVA.
Dankaerts et al. (2009) [7]	To test the ability of a model to distinguish between flexion pattern (FP) and active extension pattern (AEP) subgroups and healthy controls using lumbar kinematics and trunk muscle activity.	Normalised EMG activity. Superficial LMU (at the level of L5 orientated by a line between the PSIS and the L1–L2 interspace. Iliocostalis lumborum pars thoracis (lateral to L1).	3-Space Fastrak (regional) Upper lumbars T12 relative to L3 Lower lumbars L3 relative to S2.	Differences in muscle activity and spinal kinematics during flexion suggest that 2 distinct motor control patterns can exist in CNSLBP patients.	N = 67 participants 34 low back pain free controls, mean age 32 20 Flexion pattern NSLBP patients, mean age 36 13 Extension pattern NSLBP patients, mean age 40.	ANOVA and post hoc Bonferroni.
Hashemirad et al. (2009) [45]	To investigate the relationship between lumbar spine flexibility and LES activity during sagittal flexion and return.	Normalised EMG amplitude and signal onset/offset. Muscle LES (4 cm lateral to L3–L4).	Estimated using a camera and markers placed at the spinous processes of T12, L3 and S2 (regional).	During bending the ES of participants with high toe touch score deactivated at greater trunk and hip angles. Those with high modified Schober scores deactivated later and reactivated sooner in accordance with lumbar angle.	N = 30 low back pain free participants.	Pearson correlations and multiple linear regression analysis.
Hay et al. (2016) [46]	To show that wavelet coherence and phase plots can be used to provide insight into how muscle activation relates to kinematics.	EMG amplitude (linear envelope). Muscle Lumbar erector spinae (no details of positioning).	Oqus 400 motion capture system (regional) Reflective markers placed over T12 and S1.	The study showed good agreement between lumbar kinematics and linear enveloped sEMG. Validating the use of the wavelet coherence technique.	N = 14 low back pain free male participants.	The coefficient of determination (R²).
Kaigle et al. (1998) [20]	To concurrently quantify muscle activation of LES with the kinematics of lumbar motion segments, in low back patients and controls.	Root mean square (RMS) sEMG amplitude. Muscle LES (3 cm lateral to L3–L4).	A linkage transducer system secured by interosseous pins to L2-L3, L3-4 and L4-L5 motion segments (inter-vertebral).	ROM was less in low back pain patients and FRP occurred in participants when IV-ROM was complete before full trunk flexion	N = 13 6 low back pain free participants, mean age 40. 7 low back pain patients with suspected lumbar instability, mean age 51.	Wilcoxon rank-sum test and Wilcoxon matched-pairs signed rank test.
Kienbacher et al. (2016) [47]	To determine whether lumbar extensor activity and flexion relaxation ratios could differentiate low back pain patients (of various age groups) during flexion-extension task.	Normalised RMS sEMG amplitudes. Muscle LMU (lateral to L5) a line joining the iliac crests, and 2–3 cm bilateral and distal from their middle).	3-D accelerometers placed at the levels of T4 and L5. Used to calculate hip, lumbothoracic and gross trunk regions. (regional).	The sEMG activation was highest in over 60′s and female groups during standing. This possibly relates to why this group showed minimal changes during flexion. This group also demonstrated the highest hip, and lowest lumbothoracic angle changes.	N = 216 low back pain patients. 62 (60–90 year olds) 84 (40–59 year olds) 70 (18–39 year olds).	ANOVA and bootstrap confidence intervals.
Lariviere et al. (2000) [6]	To evaluate the sensitivity of trunk muscle EMG waveforms to trunk ROM and low back pain status during flexion-extension tasks.	Mean normalised EMG activity. Muscles LES and TES (exact locations not specified).	Video cameras and reflective markers. Trunk angles relative to the vertical plane were used to determine trunk flexion (A line between the hips and the centre of C7-T1) (regional).	Principal component analysis (PCA) distance measures were sensitive to trunk ROM but not low back status. The usefulness of PCA as an effective clinical tool was not established.	N = 33 15 low back pain patients, mean age 40 18 low back pain free participants, mean age 39.	ANOVA and ICC’s.
Liu et al. (2011) [48]	To develop a new test based on lumbar sEMG activity (the sEMG coordination network analysis approach) during flexion-extension, to distinguish between healthy control and low back pain groups.	Normalised RMS sEMG activity. Muscles An sEMG electrode array placed over the lumbar region (16 electrodes, target muscles not specified).	30° of trunk flexion, measured by a protractor (no further details) (regional).	Group network analysis shows a loss of global symmetric patterns in the low back pain group.	N = 21 11 low back pain patients, mean age 40. 10 low back pain free participants, mean age 28.	Did not specify. (However, groups comparison statistics and symmetry scores were used).
Luhring et al. (2015) [16]	To determine a kinematic measurement that best determines the onset and offset of the FRP.	Normalised sEMG onset and cessation. Muscle LES (4 cm lateral to L3).	Vicon MX motion capture camera system. Reflective markers placed at various locations throughout the spine including T12, L5 and pelvis (regional).	Lumbar kinematic measurements are preferential when the FRP is considered clinically.	N = 20 low back pain free participants, mean age 24.	Coefficients of Variation (CV) and ICC’s.
Mayer et al. (2009) [49]	To determine when FRP occurs in patients and to correlate the findings with lumbar ROM.	Mean RMS sEMG with pre-determined cut-off values. Muscles Not identified within paper.	Gross lumbar, hip/pelvic ROM using an inclinometer (no further details provided) (regional).	After a functional restoration program, both normal FRP and normal lumbar ROM were restored in the majority of patients.	N = 134 30 low back pain free participants, mean age 38. 104 low back pain patients (mean age not provided).	Descriptive statistics including mean and SD. Sensitivity and specificity. P-values and Odds ratios (not specified).
McGill and Kippers 1994 [15]	To examine the tissue loading during the period of transition between active and passive tissues during flexion.	Normalised sEMG activity. Muscles TES (5 cm lateral to T9) LES (3 cm lateral to L3).	3-Space Isotrak (regional) with sensors placed over the sacrum and T10.	The deactivation of lumbar extensor muscles during FRP occurs only in an electrical sense as they still provide force elastically.	N = 8 low back pain free participants, mean age 26.	Dynamic modelling.
Nairn et al. (2013) [50]	To quantify slumped sitting both in terms of spinal kinematics and sEMG.	Mean normalised sEMG activity. Muscles Lower TES (5 cm lateral to T9) LES (4 cm lateral to L3) LMU (Adjacent to L5 orientated along a line between the PSIS and the L1–L2 interspinous space.	Vicon motion capture camera system. Reflective markers placed at various locations throughout the spine including T12, L1 and bilateral PSIS’s (regional).	During slumped sitting lower sEMG activity was found in the thoracic and lumbar erector spinae compared to upright sitting. Patterns varied depending on the degree of bending at each area of the spine. Thoracic kinematic and EMG information is therefore useful in these type of studies	N = 12 low back pain free participants, mean age 23.	ANOVA and Bonferroni correction.
Neblett et al. (2003) [51]	To assess EMG activity in terms of the FRP during dynamic flexion and to determine whether abnormal FRP patterns in NSLBP patients can be normalised.	RMS sEMG cut-off values. Muscles LES (2 cm lateral to L3).	Inclinometers at T12 and the sacrum (regional).	In asymptomatic participants, the flexion relaxation (FR) angle was always less than the maximal voluntary flexion (MVF) angle. Of the patients that completed a functional restoration program, 94% achieved FR compared to 30% pre-treatment.	N = 66 12 low back pain free participants, mean age 34. 54 chronically disabled work-related spinal disorder (CDWRSD) patients	Descriptive statistics for ROM and FRP
Ning et al. (2012) [52]	To determine a boundary at which the passive tissues begin to take a significant role in trunk extensor moment (and therefore at what point EMG assisted modelling is no longer valid).	Normalised EMG activity. Muscles LES at two levels (3 cm lateral to L3 and 4 cm lateral to L4).	A magnetic-field based motion tracking system with sensors placed at T12 and S1. Lumbar flexion calculated as the pitch of T12 relative to S1 (regional).	EMG-assisted models should consider the action of the passive tissues at lower flexion angles than previously thought.	N = 11 low back pain free participants, mean age 26.	ANOVA and Tukey–Kramer post-hoc testing
O’Sullivan et al. (2006) [17]	To investigate the FRP of spinal muscles in healthy participants during slumped sitting from an upright position.	Normalised EMG activity offset. Muscles TES (5 cm lateral to T9) LMU (Adjacent to L5 orientated along a line between the PSIS and the L1–L2 interspinous space.	3- Space Fastrak with sensors placed over T6, T12 and S2. (regional).	LMU is active during neutral sitting and demonstrates FRP when moving from upright to slumped sitting. FRP of these muscles is also different to when standing. More variation was found in EMG patterns of the TES.	N = 24 low back pain free participants, mean age 32.	ANOVA and ICC’s
Paquet et al. (1994) [53]	To compare healthy controls and low back pain patients in terms of hip-spine movement interaction and EMG, and to verify the relationships between kinematics and EMG in these groups.	Raw EMG envelope. Area under the curve and ratio of activity at different parts of the flexion-extension cycle (not-specified). Muscles LES (at the level of L3, distance not-specified).	Electro goniometers measured angular displacements at the hip and lumbar spine using landmarks of T8 and S1 (regional).	LES activation patterns were found to be significantly different between groups when flexion was performed at the same rate and range. Abnormal hip-spine movement related to an absence of the FRP at full flexion.	N = 20 10 low back pain free participants, mean age 34. 10 low back pain patients, mean age 38.	Mann-Whitney U test and Kruskal-Wallis test
Peach et al. (1998) [35]	To document the lumbar kinematics and trunk EMG activation patterns of healthy controls during tasks including sagittal flexion	Mean normalised EMG. Muscles TES (5 cm lateral to T9) LES (3 cm lateral to L3) LMU (1–2 cm lateral to L5).	3-Space Isotrak with sensors placed over T12 and Sacrum. (regional).	A database of normal lumbar spinal kinematics and EMG patterns was created for future reference against LBP groups.	N = 24 low back pain free participants, mean age 22.	Descriptive statistics, ANOVA and Tukey’s honestly significant difference (HSD) post-hoc testing
Sanchez-Zuriaga et al. (2015) [26]	To compare healthy controls and LBP patients in terms of lumbopelvic kinematics and erector spinae activity	Mean normalised EMG activity, and start and end of FRP. Muscle LES (3 cm lateral to L3).	A 3-dimensional videophotogrammetric system, with markers placed at T12, L3, L5 and the sacrum (regional).	During pain free periods, recurrent LBP patients showed significantly greater LES activity during flexion and extension. Lumbar ROM and FRP were not found to be useful to distinguish between groups.	N = 30 15 low back pain free participants, mean age 41. 15 patients with recurring low back pain (currently in a pain free stage), mean age 45.	Mann-Whitney U test