Analysis of the activation modalities of the lower limb muscles during walking

Abstract

BACKGROUND:

Walking is a basic human activity and many orthopedic diseases can manifest with gait abnormalities. However, the muscle activation intervals of lower limbs are not clear.

OBJECTIVE:

The aim of this study was to explore the contraction patterns of lower limb muscles by analyzing activation intervals using surface electromyography (SEMG) during walking.

METHODS:

Four muscles including the tibialis anterior (TA), lateral gastrocnemius (LG), medial gastrocnemius (MG), and rectus femoris (RF) of bilateral lower extremity of 92 healthy subjects were selected for SEMG measurements. The number of activations (activation intervals) and the point of the highest root mean square (RMS) EMG signal in the percentage of the gait cycle (GC) were used to analyze muscle activities.

RESULTS:

The majority of TA and RF showed two activation intervals and both gastrocnemius parts three activation intervals during walking. The point of the highest RMS EMG signal in the percentage of the GC for TA, LG, MG and RF are 5%, 41%, 40%, and 8%, respectively. The activation intervals were mostly affected by age, height, different genders and bilateral limbs.

CONCLUSION:

This study identified the different activation intervals (four for each muscle) and the proportion of healthy adults in which they occurred during the normal gait cycle. These different activation intervals provided a new insight to evaluate the function of nerves and muscles. In addition, the activation interval and RMS peak time proposed in this study can be used as new parameters for gait analysis.

Keywords

Surface electrography activation intervals peak time gait cycle walking

1. Introduction

Walking is a basic human activity and many orthopedic diseases can manifest with gait abnormalities [1]. During walking, the muscles around the ankle play an important role. Ankle plantar flexors act to restrain the forward rotation of the tibia on the talus during the stance phase, provide ankle stability, contribute to knee stability, and conserve energy by minimizing vertical oscillation of the whole-body center of mass [2]. The main role of the ankle dorsiflexors during walking is to prevent the foot from slapping on the ground during the initial stance, to permit the forefoot to clear the ground during the initial swing, and to hold the ankle in position during initial contact [3]. Therefore, the dynamic characteristics of muscles may better reflect the degree of disease and rehabilitation [4].

In most studies, it is considered useful to have a profile of normal surface electromyography (SEMG) activity that can be used as a reference. Clinical investigators need these references to diagnose individual patients, and basic researchers interested in motor patterns use it to discuss the roles of various muscles during the gait cycle [5]. The normal distribution and motor patterns of the lower limb muscles have already been described [3]. However, those patterns are not sufficient enough to explain complicated muscle activities, and variations in SEMG signals have been found between people and even among different strides in the same person [5]. There are various attempts to identify the SEMG patterns in healthy people, and the results have shown that a variation of myoelectric interval exists in children of different ages [6], across genders [6, 7, 8, 9], and across different strides in the same person [10]. However, few studies have focused on the variation of activation intervals in healthy adults with statistical gait analysis method and the variation of activation intervals may be an important factor that affects the walking ability [11].

Therefore, the aim of this study was to use SGA to quantitatively assess the natural variability of activation intervals in healthy adults through measuring four important gait-related muscles: tibialis anterior (TA), lateral gastrocnemius (GL), medial gastrocnemius (MG), and rectus femoris (RF).

2. Methods

2.1 Participants

Ninety-two healthy adults between the ages of 20–87 (average $\pm$ standard deviation [SD]: 44.03 $\pm$ 18.61 years), with heights ranging from 150–184 cm (average $\pm$ SD: 165.40 $\pm$ 8.56 cm) and weights ranging from 43–97 kg (average $\pm$ SD: 63.80 $\pm$ 11.28 kg) were recruited (Table 1). SEMG signals were collected from 736 muscle recordings from healthy volunteers.

Table 1
The characteristics of the healthy subjects

Number	32		14		10		13		14		10
	20–29 years		30–39 years		40–49 years		50–59 years		60–69 years		70–90 years
Male:Female	20:11		6:8		2:8		4:9		4:10		2:8
Age (years)	25.13	$\pm$ 1.81	33.86	$\pm$ 2.93	45.6	$\pm$ 2.98	55.92	$\pm$ 3.33	62.15	$\pm$ 1.63	79.22	$\pm$ 6.14
Height (cm)	171.63	$\pm$ 6.19	167.21	$\pm$ 8.77	159.4	$\pm$ 7.09	161.62	$\pm$ 8.08	161.36	$\pm$ 6.64	158.88	$\pm$ 5.61
Weight (kg)	63.31	$\pm$ 11.18	69.21	$\pm$ 15.24	66	$\pm$ 8.49	63	$\pm$ 13.62	63.79	$\pm$ 6.42	55.56	$\pm$ 5.38

The information is shown as average $\pm$ standard deviation [SD].

2.2 Testing procedures

SEMG was measured using a DELSYS wireless dynamic EMG tester (Trigno™ Wireless Systems, Delsys Inc., USA). The bilateral TA, LG, MG, and RF were collected and analyzed as they are widely used to evaluate lower limb functions [8, 12]. The participant’s skin was scraped, sanded, and cleaned prior to pasting electrodes (size: 37 $\times$ 26 $\times$ 15 mm) on the designated areas (Fig. 1), as per the SENIAM recommendations [13]. Subjects walked at a natural speed and cadence on a 10-meter flat floor back and forth while the SEMG signals were collected and transmitted via Bluetooth wireless connection. Each parameter was averaged from 20 gait cycles to obtain a representative EMG profile for each muscle [14]. A small, portable, high-speed camera (GoPro Hero3, GoPro Inc., USA) was used to identify the start and end points of each gait cycle. The signal from the GoPro was synchronous with SEMG and the GoPro shooting rate was set to 120 frames per second.

Figure 1.

The position of the electrodes.

2.3 Signal processing

The EMG signals were processed and analyzed using EMGworks Analysis and MATLAB software. The sampling frequency was 1200 Hz. Four parameters were analyzed. Root mean square (RMS) and RMS peak time (the position of the dominant wave’s peak, recorded as a percentage of a gait cycle) were used to observe muscular activity [15, 16]. We combined EMG data from 20 gait cycles using the EMG analysis software. The EMG synchronization points were subjected to analysis using MATLAB software after eliminating noise. For the time parameter, the RMS value of the EMG signal in each gait cycle was calculated using a 30 ms window and a 20 ms step size [17, 18]. The time was normalized, and the average value of the 20 gait cycles was calculated [14]. The RMS was plotted to get the maximum RMS value and the peak time. Next, the intervals were analyzed. Intervals are defined as the number of times a muscle is activated and is expressed as the waves present during a single gait cycle [10]. Each interval refers to a different interval pattern, characterized by a specific number of intervals over a gait cycle, i.e. n-interval modality consists of n interval waves for a particular muscle during a single gait cycle [10]. Perry’s [3] theory of the gait cycle was used to classify the RMS peak time into one of eight phases: initial contact (0%–2%), loading response (2%–12%), mid stance (12%–31%), terminal stance (31%–50%), pre-swing (50%–62%), initial swing (62%–75%), mid swing (75%–87%) and terminal swing (87%–100%).

2.4 Statistics

Data are reported as mean $\pm$ SD. The Lilliefors test was used to evaluate the hypothesis that each data vector had a normal distribution. Comparisons between two normally distributed samples were performed with two-tailed, unpaired Student’s t-test or paired Student’s t-test. Comparisons between two non-normally distributed samples were performed with a Wilcoxon rank sum test. A Bonferroni-correction was used for the multiple repeated measurements. Spearman rank correlation coefficient was used to evaluate the relationship between interval modality and possible influencing factors. Statistical significance was set at $p<$ 0.05.

Table 2
The distribution of RMS value and peak time with each interval modality in all muscles

TA	1-intervals	First peak	LR	100	51.55		6
			Phases	Occurrence %	RMS values (Median (IQR))		Peak time % (Median (IQR))
	2-intervals	First peak	LR	74.51	47.72	(48.69)	5	(2)
		Second peak	ISw	83.33	26.96	(12.54)	69	(6)
	3-intervals	First peak	LR	68.25	28.9	(33.9)	5	(4)
		Second peak	ISw	57.14	18.67	(26.95)	69.5	(10)
		Third peak	TSt	41.27	12.5	(12.69)	40	(13)
	4-intervals	First peak	LR	44.44	29.47	(20.03)	5.5	(2.25)
		Second peak	ISw	38.89	20.53	(10.44)	69	(5.5)
		Third peak	MSt	33.33	18.11	(11.59)	18	(3)
		Fourth peak	TSt	33.33	14.08	(13.4)	38	(3.75)
	All subjects	Highest peak	–	–	38.93	(38.01)	5	(2.5)
LG	1-intervals	First peak	TSt	95.65	29.41	(15.30)	42	(2)
	2-intervals	First peak	TSt	68.63	25.89	(19.93)	42	(3)
		Second peak	LR	43.14	11.04	(4.67)	8	(3.5)
	3-intervals	First peak	TSt	51.56	23.6	(12.90)	43	(5)
		Second peak	LR	37.50	14.06	(9.94)	7	(4)
		Third peak	IS	43.75	9.04	(4.44)	70	(6)
	4-intervals	First peak	TSt	52.17	17.78	(11.16)	42	(3.25)
		Second peak	LR	30.43	13.77	(7.43)	7.5	(5)
		Third peak	IS	43.48	8.3	(4.18)	67	(2.25)
		Fourth peak	IS	39.13	8.13	(4.26)	65	(10)
	All subjects	Highest peak		–	27.34	(19.35)	41	(13)
MG	1-intervals	First peak	TSt	88.68	41.52	(23.95)	40	(4.5)
	2-intervals	First peak	TSt	78.95	32.80	(17.92)	39	(4)
		Second peak	MSw	56.14	10.29	(6.38)	82	(4.25)
	3-intervals	First peak	TSt	74.14	26.85	(16.39)	39	(4)
		Second peak	LR	31.03	14.05	(15.36)	7.5	(5)
		Third peak	MSw	39.66	10.43	(5.88)	80	(6.5)
	4-intervals	First peak	TSt	37.50	31.69	(35.65)	38	(0.75)
		Second peak	LR	43.75	33.73	(21.59)	9	(3)
		Third peak	TSw	37.50	22.51	(14.27)	93.5	(4)
		Fourth peak	IS	31.25	16.81	(1.7)	66	(7)
	All subjects	Highest peak		–	36.28	(33.08)	40	(7)
RF	1-intervals	First peak	LR	61.54	9.68	(2.07)	7.5	(2.5)
	2-intervalss	First peak	LR	66.67	9.09	(3.26)	7.5	(2)
		Second peak	ISw	57.89	5.99	(0.66)	68	(8)
	3-intervals	First peak	LR	62.50	8.82	(5.4)	7	(2)
		Second peak	ISw	21.42	6.46	(4.64)	68.5	(3.75)
		Third peak	ISw	50	4.52	(1.88)	66.5	(7.25)
	4-intervals	First peak	LR	53.33	8.31	(8.11)	7	(3.5)
		Second peak	MSt	33.33	6.69	(2.03)	19	(9)
		Third peak	ISw	35.56	5.64	(2.37)	66.5	(4.75)
		Fourth peak	TSt	33.33	5.02	(4.1)	44	(7.5)
	All subjects	Highest peak	LR	–	9.03	(7.77)	8	(9)

3. Results

The values of 736 muscles from 92 healthy volunteers were used for analysis. Four activation intervals were found, including one, two, three, and four muscle activation intervals. The analysis of the myoelectric signals suggested that muscles show a different number of activation intervals in different subjects and even in different limbs from the same individual.

Table 3
The p value and correlation between interval modality and gender, bilateral limbs, age, weight, height and BMI. The results show that there was a significant difference between bilateral limbs in LG. The age and height show some correlation to the variation of interval modality

	Interval modality	Gender	Bilateral limbs	Age (years)	Weight (kg)		Height (cm)		BMI (kg/m ${}^{2}$ )
TA	$P$ values (2-tailed)	0.094	0.064	0.041*	0.	25	0.	005*	0.67
	Correlation coefficient	–	–	0.15	$-$ 0.	08	$-$ 0.	21	0.03
LG	$P$ values (2-tailed)	0.084	$<$ 0.001*	0.02*	0.	49	0.	01*	0.52
	Correlation coefficient	–	–	0.17	$-$ 0.	50	$-$ 0.	18	0.48
MG	$P$ values (2-tailed)	0.17	0.091	0.005*	0.	311	0.	108	0.043*
	Correlation coefficient	–	–	0.21	0.	08	$-$ 0.	12	0.15
RF	$P$ values (2-tailed)	0.169	0.128	0.174	0.	665	0.	392	0.735
	Correlation coefficient	–	–	–0.11	0.	03	$-$ 0.	06	0.25

*The statistical significant relationship was set at $p\leqslant$ 0.05.

Figure 2.

The RMS of TA shows four basic activation intervals: 1-intervals (panel a), 2-intervals (panel b), 3-intervals (panel c) and 4-intervals (panel d).

Four activation intervals were found in the TA (Fig. 2). The most common interval modality in the TA was a 2-interval modality (55.43% of all participants) (Fig. 3). The peak time of TA with 2-interval modality mainly occurred in the loading response phase (74.51% of all participants), and the second burst’s peak time mainly occurred in the initial swing phase (83.33% of all participants). The other modalities are described in Table 2. Of all activation intervals, the RMS peak time mainly occurred in the loading response phase (70.65% of participants) (Fig. 4). The interval modality was not significantly different between genders ( $p=$ 0.09) or between bilateral limbs ( $p=$ 0.06). There was no correlation between interval modality and weight ( $p=$ 0.25) or body mass index (BMI) ( $p=$ 0.67). The activation intervals were correlated with age ( $p=$ 0.04) and height ( $p=$ 0.01); however, the correlation coefficients were very low ( $r=$ 0.15 and $r=$ $-$ 0.21, respectively) (Tables 3 and 4).

For the LG, there were also four activation intervals (Fig. 5). The most common interval modality was the 3-interval modality (34.78% of participants) (Fig. 3). The peak time of LG with 3-interval modality mainly occurred in the terminal stance phase (51.56% of participants), the second peak time mainly occurred in the loading response phase (37.5% of participants), and the third peak time occurred in the initial swing phase (43.75% of all participants). The other modalities are described in Table 2. Of all activation intervals, the RMS peak time focused on the terminal stance phase (68.48% of participants) (Fig. 4). The interval modality of the LG was not significantly different in terms of gender ( $p=$ 0.84), but there was a significant difference between bilateral limbs ( $p<$ 0.001). There was no correlation with weight ( $p=$ 0.49) or BMI ( $p=$ 0.52). Although the interval modality was correlated with age ( $p=$ 0.02) and height ( $p=$ 0.01), the correlation coefficients were very low ( $r=$ 0.17 and $r=$ $-$ 0.18, respectively) (Tables 3 and 4).

Table 4

The mean and median value of the peak time and interval models of each muscle in healthy subjects

		LTA	RTA	LLG	RLG	LMG	RMG	LRF	RRF
RMS peak	Mean $\pm$ SD	11.71 $\pm$ 19.69	10.71 $\pm$ 19.42	32.71 $\pm$ 21.42	40.82 $\pm$ 18.40	40.63 $\pm$ 15.29	43.27 $\pm$ 18.94	15.57 $\pm$ 19.33	18.45 $\pm$ 26.14
time	Median (IQR)	5 (3)	5 (3)	40 (34.5)	42 (5)	39.50 (5.75)	40 (8.75)	8 (9)	7.5 (10.75)
Intervals	Mean $\pm$ SD	2.62 $\pm$ 0.68	2.46 $\pm$ 0.67	2.73 $\pm$ 0.83	1.98 $\pm$ 1.01	2.11 $\pm$ 0.88	2.30 $\pm$ 1.01	2.77 $\pm$ 1.03	2.57 $\pm$ 0.99
	Median (IQR)	3 (1)	2 (1)	3 (1)	2 (2)	2 (2)	2 (2)	3 (2)	3 (1)

SEMG, surface electromyography; TA, tibialis anterior; LG, lateral gastrocnemius; RMS, root mean square; MPF, mean power frequency; MF, median frequency; SD, standard deviation; IQR, inter-quartile range.

Figure 3.

55.43% of all participants’ interval modalities in the TA were 2-interval modalities (a); 34.78% of all participants’ interval modalities in the LG were 3-interval modalities (b); 31.52% of all participants’ interval modalities in the MG were 3-interval modalities (c); 30.98% of all participants’ interval modalities in the RF were 3-interval modalities (d).

Figure 4.

70.65% of all participants’ RMS peak time in the TA mainly occurred in the loading response phase (a); 68.48% of all participants’ RMS peak time in the LG mainly occurred in the terminal stance phase (b); 77.71% of all participants’ RMS peak time in the MG mainly occurred in the terminal stance phase (c); 61.96% of all participants’ RMS peak time in the RF mainly occurred in the loading response phase (d).

Figure 5.

The RMS of LG shows four basic activation intervals: 1-intervals (panel a), 2-intervals (panel b), 3-intervals (panel c) and 4-intervals (panel d).

Concerning the MG, there were also four activation intervals (Fig. 6). The most common interval modality was also the 3-interval modality (31.52% of participants) (Fig. 3). The peak time of MG with 3-interval modality mainly occurred in the terminal stance phase (74.14% of participants), the second peak time occurred in the loading response phase (31.03% of all participants), and the third peak time occurred in the mid swing phase (39.66% of all participants). The other modalities are described in Table 2. Of all the activation intervals, the RMS peak time mainly focused on the terminal stance phase (77.71% of participants) (Fig. 4). The results also showed that the interval modality was not significantly different between genders ( $p=$ 0.17) or between bilateral limbs ( $p=$ 0.09). There was no correlation with weight ( $p=$ 0.311) or height ( $p=$ 0.108). Although the interval modality was correlated with age ( $p=$ 0.005) and BMI ( $p=$ 0.04), the correlation coefficients were very low ( $r=$ 0.21 and $r=$ 0.15, respectively) (Tables 3 and 4).

Figure 6.

The RMS of MG shows four basic intervals: 1-intervals (panel a), 2-intervals (panel b), 3-intervals (panel c) and 4-intervals (panel d).

Concerning the RF, there were also four activation intervals (Fig. 7). The most common interval modality was the 2-interval modality (30.98% of participants) (Fig. 3). The peak time of RF with 2-interval modality mainly occurred in the loading response phase (66.67% of participants), and the second peak time occurred in the initial swing phase (57.89% of all participants). The other modalities are described in Table 2. Of all activation intervals, the RMS peak time mainly focused on the terminal stance phase (61.96% of participants) (Fig. 4). The interval modality was not significantly different between genders ( $p=$ 0.17) or bilateral limbs ( $p=$ 0.13). There was no correlation with weight ( $p=$ 0.67), height ( $p=$ 0.39), age ( $p=$ 0.17), or BMI ( $p=$ 0.74) (Tables 3 and 4).

Figure 7.

The RMS of RF shows four basic interval intervals: 1-intervals (panel a), 2-intervals (panel b), 3-intervals (panel c) and 4-intervals (panel d).

4. Discussion

The present study was designed to analyze the activation intervals of lower-limb muscles during walking at a self-selected speed. The results were provided to gain a novel insight into the muscle function during GC.

In a gait cycle, four muscle activation intervals were found in different muscles from healthy subjects (Fig. 2, Figs 5–7). However, the dominant activation modality and the RMS peak times varied per muscle (Figs 3 and 4). The different activation intervals and the difference in RMS peak time in healthy adults may be attributable to segmental motor overlaps and anatomical variations. In the lower extremity, most muscles have 2 or 3 levels of nerve roots innervation. Therefore, a single muscle may express different activation intervals in different people. Of all the nerves that innervate the muscle, one nerve will be dominant [19, 20, 21, 22]. For example, the TA and the extensor hallucis longus are innervated by the L5 nerve root in 90% of the people, and the remaining 10% of people are innervated by the S1 nerve root [19]. Additionally, the nerve root’s variations are exited in a considerable percentage of healthy subjects [23]. The differences in dominant nerves may be a reason for the differences in RMS peak times.

The dominant intervals for the four different muscles also varied. The TA and RF had a 2-interval modality, whereas the LG and MG had a 3-interval modality. Additionally, the proportions of people exhibiting the dominant interval modality was different for the difference muscles (TA: 55.43%; LG: 34.78%; MG: 31.52%; LG: 30.98%). This finding supports the assumption of the overlap theory that multiple segmental roots innervate the same muscle together [24, 25]. The muscle activation pattern and RMS peak time may be decided by the neuromechanical function of the individual muscle [26]. The weight-bearing muscles (TA, RF) have their major peak in the first 15% of the stride for the purpose to bear weight when the walking is initiated. The plantar flexors (MG, LG) have peaks at push-off terminal stance phase (50% of the stride) so as to produce the pushing power and propel the body [3, 25]. The other interval waves may be caused by co-contraction of the antagonist muscle, and the co-contraction between agonist muscle and antagonist muscle would stabilize the joint movement.

The amplitude parameters to quantitatively evaluate SEMG signal (peak amplitude, area under curve, mean value) are widely used in clinical evaluation [27]. However, these parameters may be affected by electrode location and the volume conductor [28]. Thus, only temporal parameters are generally accepted as suitable indices to quantify SEMG signals. The present study identified two additional parameters from SEMG signal: the RMS peak time and the interval modality. These two parameters are able to provide additional information to that which is supplied by classical temporal parameters, and these parameters present significant differences between healthy subjects and patients with lumbar disc herniation [29]. In addition, RMS peak times have small individual variation and can be used to compare different subjects. Despite the robustness of the results achieved, the participants in this study were limited to those of Chinese origin, which could limit the generalizability of the findings.

5. Conclusion

In conclusion, the present study identified the different activation modalities of healthy subjects during walking. These different activation modalities intervals provided a new insight to evaluate the degree of rehabilitation of patients with nerve or muscle diseases. In addition, the interval modality and RMS peak time proposed in this study can be used as a new parameter to complement the classic temporal parameters when evaluating the function of nerves or muscle diseases.

Footnotes

Acknowledgments

This work was partly supported by a grant from the Beijing Rehabilitation Hospital (No. 2018-004, 2018-006).

Conflict of interest

None to report.

Abbreviations

References

Huang

Bruijn

Lin

Meijer

Abbasibafghi

Lin

Dieën

JHV

. Gait adaptations in low back pain patients with lumbar disc herniation: trunk coordination and arm swing. European Spine Journal. 2011; 20(3): 491.

Sutherland

Cooper

Daniel

. The role of the ankle plantar flexors in normal walking. Journal of Bone & Joint Surgery American Volume. 1980; 62(3): 354-363.

Perry

Burnfield

Cabico

. Gait analysis: normal and pathological function. Slack, 2010.

Sutherland

. The evolution of clinical gait analysis. Part II kinematics. Gait & Posture. 2002; 16(2): 159-179.

Winter

Yack

. EMG profiles during normal human walking: stride-to-stride and inter-subject variability. Electroencephalography & Clinical Neurophysiology. 1987; 67(5): 402-411.

Laureati

Strazza

Mengarelli

Burattini

Agostini

Nascimbeni

Knaflitz

Fioretti

. Is child walking conditioned by gender? Surface EMG patterns in female and male children. Gait & Posture. 2017; 53: 254.

Chong

Vojnic

Quanbury

Letts

. The assessment of the internal rotation gait in cerebral palsy: an electromyographic gait analysis. Clinical Orthopaedics & Related Research. 1978; NA(132): 145.

Thorstensson

Carlson

Zomlefer

Nilsson

. Lumbar back muscle activity in relation to trunk movements during locomotion in man. Acta Physiologica. 1982; 116(1): 13-20.

Tolj

Luetić

Schwarz

Bilić

Jurcić

Gabrić

. The impact of age, sex, body mass index and menstrual cycle phase on gastric myoelectrical activity characteristics in a healthy Croatian population. Coll Antropol. 2007; 31(4): 955-962.

10.

Maranesi

Mengarelli

Ghetti

Burattini

Fioretti

. Assessment of the variability of vastii myoelectric activity in young healthy females during walking: A statistical gait analysis. Journal of Electromyography & Kinesiology. 2015; 25(5): 800-807.

11.

Harriss

Macsween

Atkinson

. Standards for Ethics in Sport and Exercise Science Research: 2018 Update. International Journal of Sports Medicine. 2017; 38(14): 1126-1131.

12.

Hortobágyi

Solnik

Gruber

Rider

Steinweg

Helseth

Devita

. Interaction between age and gait velocity in the amplitude and timing of antagonist muscle cointerval. Gait & Posture. 2009; 29(4): 558-564.

13.

Hermens

Freriks

Disselhorstklug

Rau

. Development of recommendations for SEMG sensors and sensor placement procedures. Journal of Electromyography & Kinesiology. 2000; 10(5): 361.

14.

Gabel

Brand

. The effects of signal conditioning on the statistical analyses of gait EMG. Electroencephalography & Clinical Neurophysiology. 1994; 93(3): 188.

15.

Berni

Dibai-Filho

Pires

Rodrigues-Bigaton

. Accuracy of the surface electromyography RMS processing for the diagnosis of myogenous temporomandibular disorder. Journal of Electromyography & Kinesiology. 2015; 25(4): 596-602.

16.

Molinari

Knaflitz

Bonato

Actis

. Electrical manifestations of muscle fatigue during concentric and eccentric isokinetic knee flexion-extension movements. IEEE Transactions on Bio-Medical Engineering. 2006; 53(7): 1309.

17.

Merletti

Botter

Troiano

Merlo

Minetto

. Technology and instrumentation for detection and conditioning of the surface electromyographic signal: state of the art. Clinical Biomechanics. 2009; 24(2): 122-134.

18.

Malone

Meldrum

Gleeson

Bolger

. Reliability of surface electromyography timing parameters in gait in cervical spondylotic myelopathy. Journal of Electromyography & Kinesiology. 2011; 21(6): 1004.

19.

Young

Getty

Jackson

Kirwan

Sullivan

Parry

. Variations in the pattern of muscle innervation by the L5 and S1 nerve roots. Spine. 1983; 8(6): 616.

20.

Tsao

Levin

Bodner

. Comparison of surgical and electrodiagnostic findings in single root lumbosacral radiculopathies. Muscle & Nerve. 2003; 27(1): 60-64.

21.

Yue

Andrew

. Foot drop resulting from degenerative lumbar spinal diseases: clinical characteristics and prognosis. Clinical Neurology & Neurosurgery. 2014; 117(2): 33-39.

22.

Daube

. Electrodiagnostic studies in amyotrophic lateral sclerosis and other motor neuron disorders. Muscle & Nerve. 2000; 23(10): 1488.

23.

Chotigavanich

Sawangnatra

. Anomalies of the lumbosacral nerve roots. An anatomic investigation. Clinical Orthopaedics & Related Research. 1992; 278(278): 46.

24.

Chang

Lee

Kim

. Effects of the ankle angle of an ankle foot orthosis on foot pressure during the gait in healthy adults. Journal of Physical Therapy Science. 2015; 27(4): 1033-1035.

25.

Gianakos

Yasui

Murawski

Kennedy

. Effects of gastrocnemius recession on ankle motion, strength, and functional outcomes: a systematic review and national healthcare database analysis. Knee Surgery Sports Traumatology Arthroscopy. 2015; 42(3): 1-10.

26.

Hagen

Schwiertz

Landorf

. Selective activation of lower leg muscles during maximum voluntary isometric contractions[J]. Human Movement Science. 2016; 50: 30-37.

27.

Sutherl

. The evolution of clinical gait analysis part l: kinesiological EMG. Gait and Posture. 2001; 14: 61-70.

28.

Reaz

MBI

Hussain

Mohd-Yasin

. Techniques of EMG signal analysis: detection, processing, classification and applications (Correction). Biological Procedures Online. 2006; 8(1): 163-163.

29.

Liu

Zheng

Miao

Chen

Quan

Yan

Zhang

. Diagnosis of Compressed Nerve Root in Lumbar Disc Herniation Patients by Surface Electromyography. Orthop Surg. 2018; 10(1).