- Email: [email protected]
Assessment of the activation modalities of gastrocnemius lateralis and tibialis anterior during gait: A statistical analysis
Journal of Electromyography and Kinesiology 23 (2013) 1428–1433
Contents lists available at SciVerse ScienceDirect
Journal of Electromyography and Kinesiology journal homepage: www.elsevier.com/locate/jelekin
Assessment of the activation modalities of gastrocnemius lateralis and tibialis anterior during gait: A statistical analysis Francesco Di Nardo a,⇑, Giacomo Ghetti b, Sandro Fioretti a a b
Department of Information Engineering, Università Politecnica delle Marche, 60131 Ancona, Italy Posture and Movement Analysis Laboratory, Italian National Institute of Health and Science on Aging (INRCA), 60131 Ancona, Italy
a r t i c l e
i n f o
Article history: Received 12 December 2012 Received in revised form 28 May 2013 Accepted 29 May 2013
Keywords: EMG Ankle flexor muscles Gait analysis
a b s t r a c t Aim of the study was to identify the different modalities of activation of gastrocnemius lateralis (GL) and tibialis anterior (TA) during gait at self-selected speed, by a statistical analysis of surface electromyographic signal from a large number (hundreds) of strides per subject. The analysis on fourteen healthy adults showed a large variability in the number of activation intervals, in their occurrence rate, and in the on-off instants, within different strides of the same walk. For each muscle, the assessment of the different modalities of activation (five for muscle) allowed to identify a single pattern, common for all the modalities and able to characterize the behavior of muscles during normal gait. The pattern of GL activity centered in two regions of the gait cycle: the transition between flat foot contact and push-off (observed in 100% of total strides) and the final swing (67.1 ± 15.9%). Two regions characterized also the pattern of TA activity: from pre-swing to following loading response (100%), and the mid-stance (30.5 ± 15.0%). This ‘‘normality’’ pattern represents the first attempt for the development in healthy young adults of a reference for dynamic EMG activity of GL and TA, in terms of variability of on-off muscular activity and occurrence rate during gait. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Ankle flexor muscles play an important role in everyday activities, especially walking. During gait, ankle plantar flexors act to restrain the forward rotation of the tibia on the talus during stance phase, provide ankle stability, contribute to knee stability, and conserve energy by minimizing vertical oscillation of the whole-body center of mass (Sutherland et al., 1980a,b). The main role of the ankle dorsi-flexors during gait is to prevent slapping of the foot on the ground in initial stance, to permit the forefoot to clear the ground in initial swing, and to hold the ankle in position for initial contact (Perry, 1992). Surface electromyography (sEMG) has been largely used for the assessment of the activation patterns of the ankle flexor muscles during normal and pathological gait (Petersen et al., 2013; Stewart et al., 2007; Sutherland et al., 1980a,b). The general consensus is that ankle plantar flexors are active during the stance phase, while the ankle dorsi-flexors are active mostly during the swing phase, with a continuation of their activity till the next loading response (Perry, 1992; Sutherland, 2001). In particular, normal EMG activity
⇑ Corresponding author. Address: Department of Information Engineering, Università Politecnica delle Marche, Via Brecce Bianche, 60131 Ancona, Italy. Fax: +39 0712204224. E-mail address: [email protected] (F. Di Nardo). 1050-6411/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jelekin.2013.05.011
for the gastrocnemius has been reported to occur from the loading response to the terminal stance. A short activation at the beginning of gait cycle and a further activation from pre-swing to the end of gait cycle have been identified as the normal EMG activity for the tibialis anterior (Perry, 1992; Shiavi et al., 1987; Sutherland, 2001). However, both in old (Hirschberg and Nathanson, 1952; Richter, 1966) and in more recent (Agostini et al., 2010) studies, activities outside the typical activation intervals for gastrocnemius and tibialis anterior and, more generally, a large stride-to-stride variability in the EMG profiles have been reported (Arsenault et al., 1986; Winter and Yack, 1987; Yang and Winter, 1984). It is, therefore, important to study the natural variability associated with muscle activity during free walking, in order to improve the interpretation of EMG signals in both physiological and pathological conditions. This can be achieved only by recording and analyzing the electromyographic signal over a large number of steps per subject, which allows to detect the different modalities of muscle activation. From this point of view, data on ankle flexor modalities of activation, reported by cited studies, are limited by the small number of gait cycles considered during an assessment session. A study over a larger number of steps should be considered. This goal appears to be reachable by means of surface EMG, because it is a non-invasive and completely painless technique. Thus, the aim of the present study was to identify the different modalities of activation of gastrocnemius lateralis (GL) and tibialis
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 23 (2013) 1428–1433
anterior (TA) in healthy adults during gait at self-selected speed, by analyzing surface EMG signal from a large number (hundreds) of strides per subject. The study is based on the recent availability of robust techniques for the detection of muscle activation intervals, and specific tools for statistical gait analysis (Bonato et al., 1998; Balestra et al., 2002; Staude et al., 2001). 2. Materials and methods 2.1. Subjects Fourteen healthy adult volunteers were recruited (mean age ± SD: 23.9 ± 2.3 years; height 174 ± 10 cm; weight 63.4 ± 14.1 kg; body mass index (BMI) 20.7 ± 2.2 kg m 2, male-female ratio: 7/7). Exclusion criteria included history of neurological pathology, orthopedic surgery within the previous year, acute or chronic knee pain or pathology, BMI P 25, or abnormal gait. Before the beginning of the test, all participants signed informed consent. 2.2. Recording system : signal acquisition and processing Signals were acquired by means of a multichannel recording system for statistical gait analysis (Step32, DemItalia, Italy). Each subject was instrumented with foot-switches, goniometers and sEMG probes on both right and left lower limb. Three foot-switches (size: 11 11 0.5 mm; activation force: 0.2 N) were attached beneath the heel, the first and the fifth metatarsal heads of each foot. A goniometer (accuracy: 0.5°) was attached to the lateral side of each lower limb for measuring the knee joint angles in the sagittal plane. Single differential (SD) sEMG probes with fixed geometry constituted by Ag-disks (manufacturer: DemItalia, size: 7 27 19 mm; interelectrode distance: 12 mm, gain: 1000, high-pass filter: 20 Hz) were attached over Gastrocnemius Lateralis (GL) and Tibialis Anterior (TA), following the SENIAM recommendations (Freriks et al., 1999). Subjects were asked to walk barefoot on level ground for around 5 min at their natural pace, following the path schematized in Fig. 1. Time events were identified from the footswitch signals corresponding to Heel contact (H), Flat foot contact (F), Push off (P), Swing (S) and processed to segment and classify the different gait cycles. During acceleration, deceleration, and changes in direction the strides are different from those of steady state walking. Therefore, knee angles in the sagittal plane (lowpass filtered with cut-off frequency of 15 Hz) along with sequences and durations of gait phases derived by the basographic signal, were used by a multivariate statistical filter embedded in the Step32 system, to detect outlier cycles like those relative to deceleration, reversing, and acceleration. Thus, cycles with improper sequences of gait phases (i.e. different from H–F–P–S), with abnormal timing and knee angles, with respect to a mean value computed on each single subject, were discarded. (Agostini and Knaflitz, 2012b). Natural pace was chosen because walking at a
Fig. 1. Schematic representation of the path walked by the recruited subjects during the experiment; subjects walked barefoot over the floor for 5 min at their natural pace.
1429
self-selected speed improves the repeatability of EMG data (Kadaba et al., 1989), while variability increases when subjects are required to walk abnormally slow (Powers et al., 1996; Winter and Yack, 1987). EMG signals were high-pass filtered (cut-off frequency of 20 Hz) and then processed by a double-threshold statistical detector, embedded in the Step32 system, that provides the onset and offset time instants of muscle activity in a completely user-independent way. This technique (Bonato et al., 1998) consists of selecting a first threshold f and observing m successive samples: if at least r0 out of successive m samples are above the first threshold f, the presence of the signal is acknowledged. In this approach, the second threshold is represented by r0. Thus, the behavior of the double-threshold detector is fixed by three parameters: the first threshold f, the second threshold r0, and the length of the observation window m. Their values are selected to jointly minimize the value of false-alarm probability and maximize probability of detection for each specific signal-to-noise ratio. The setting of the first threshold, f, is based on the assessment of the background noise level, as a necessary input parameter. Furthermore, the double-threshold detector requires to estimate the signal-to-noise ratio in order to fine tune the second threshold, r0. The values of the background noise level and the signal-to-noise ratio, necessary to run doublethreshold algorithm, is estimated for each signal by Step32 system, using the statistical approach proposed by Agostini and Knaflitz (2012a). The length duration of the observation window, m, of 30 ms is considered a suitable value for the study of muscle activation in gait analysis (Bonato et al., 1998). During gait, a muscle activates a number of times which is usually variable from cycle to cycle (Agostini and Knaflitz, 2012b). Thus, muscle on/off instants should be averaged considering each single modality of activation by itself. With modality of activation we mean the number of times when muscle activates during a single gait cycle, i.e. n-activation modality consists of n activation intervals for the considered muscle, during a single gait cycle. In the present study, mean activation intervals (normalized with respect to the gait cycle) for each modality of activation were achieved by means of the Step32 system, according to the following steps. First, muscle activations relative to each gait cycle were identified. Then, for all the gait cycles corresponding to straight line walking, muscle activations were grouped according to the number of activations detected, i.e. relatively to the modalities of activations detected. Finally the on/off time instants were averaged, for each specific modality of activation observed, and relative standard deviation and standard error were computed. In the present study only gait cycles consisting of the sequence of H–F–P–S were considered.
3. Results For each subject, a mean (±standard deviation, SD) of 461 ± 128 strides has been considered. From the total of 6448 strides considered, 146 strides (2.3% of total strides) have been removed from the analysis because they did not follow the H–F–P–S foot-switch pattern. Thus, the mean value (±SD) of 450 ± 126 strides was obtained for each subject, and the mean (±SD) duration of H, F, P and S phases over our population was computed. The H-phase lasts 5.6 ± 1.7% of the gait cycle (mean and standard deviation), the Fphase 28.9 ± 6.0%, the P-phase 23.4 ± 4.8% and the S-phase 44.1 ± 2.8%. The mean results are reported with data from right and left lower limb considered all together. The analysis of the myoelectric signal confirms that muscles show a different number of activation intervals in different strides of the same walk. As example, EMG signals from the TA muscle of a subject have been depicted in Fig. 2.
1430
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 23 (2013) 1428–1433
Fig. 2. An example of EMG signals of TA muscle showing 2 activations (panel A), 3 activations (panel B) and 4 activations (panel C), respectively. The signals have been extracted from different strides of the same subject, during the same walk. The heel contact, flat foot contact, push off and swing phases are delimited by dashed vertical lines.
The most recurrent modality of activation for GL (Fig. 3) consists of two activations (2-activation modality), observed in 39.5 ± 9.8% of total strides; the first occurs in the transition between flat foot contact and push-off phases (6.5–36.3% of gait cycle), and the second during the mid-swing. The second most recurrent modality of activation is similar to the 2-activation modality but with no activation during the swing. This 1-activation modality was observed in 33.0 ± 20.1% of total strides. In a further 21.8 ± 11.9% of total strides, three activations were observed for GL: in the flat foot contact phase, in the push-off phase and in terminal swing (3-activation modality). The remaining 5.2 ± 3.9% of total strides was characterized by two activations during the stance and two activations during the swing (4-activation modality). For TA (Fig. 4), the most recurrent modality of activation during gait cycle consists of three activations (3-activation modality), observed in 37.4 ± 9.4% of total strides: the first occurs at the beginning of gait cycle, the second around stance-to swing-transition and the third in the terminal swing. In a 23.1 ± 13.2% of total strides a 4-activation modality was observed, similar to the 3-activation modality but with a further activation in the transition between flat foot contact and push-off phases. In a 32.3 ± 23.4% of total strides, only two activations were observed: from the beginning up to 10.9 ± 6.7% of the gait cycle, and during the whole swing (2-activation modality). The remaining 7.4 ± 7.2% of total strides was characterized by five activations, three during the stance and two during the swing (5-activation modality).
Fig. 3. Gastrocnemius lateralis: mean (+SE) percentage frequency of each of the five different modalities of activation patterns (panel A) and mean (+SE) activation intervals vs. percentage of gait cycle for the modalities with 1, 2, 3 and 4 activations, respectively (panel B), detected during walk. Mean and standard error (SE) refer to the whole population. Standard deviations are reported in the text and in Table 1. The heel contact, flat foot contact, push off and swing phases are delimited by dashed vertical lines.
The 5-activation modality for GL and the 1-activation modality for TA were observed only in 1% of the total stride and showed a large variability ; thus, they are not considered in the present analysis. Fig. 5 showed a pictorial representation of GL and TA activation onset and offset instants in function of the number of subjects where muscular activity is observed; this representation has been achieved for each muscle, considering the four main modalities of activation all together. No significant difference was found between mean muscle activation intervals of the male subgroup (7 subjects) with those of the female subgroup (7 subjects), in terms of number and modalities of GL and TA activation, as shown in Tables 1 and 2, respectively.
4. Discussion In the present study, EMG signals were recorded in fourteen healthy adult subjects, in order to determine the picture of the activation modalities of gastrocnemius lateralis (GL) and tibialis anterior (TA) during normal gait. A statistical analysis of a large number (approximately 475) of strides for each subject was performed to this aim. The muscle activation intervals, detected in each subject at self-selected walking speed, followed roughly the typical pattern reported for each muscle during gait (Basmajian, 1962; Perry, 1992; Sutherland, 2001). However, the statistical analysis put into evidence that the considered muscles showed different modalities in the number of activations and in the timing of signal onset and
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 23 (2013) 1428–1433
Fig. 4. Tibialis anterior: mean (+SE) percentage frequency of each of the five different modalities of activation patterns (panel A) and mean (+SE) activation intervals vs. percentage of gait cycle for the modalities with 2, 3, 4 and 5 activations, respectively (panel B), detected during the walk. Mean and standard error (SE) refer to the whole population. Standard deviations are reported in the text and in Table 1. The heel contact, flat foot contact, push off and swing phases are delimited by dashed vertical lines.
Fig. 5. Muscle activation onset and offset instants over the population for GL and TA, as percentage of gait cycle, considering, for each muscle, the four main modalities of activation all together. Horizontal bars are grey-level coded, according to the number of subjects where a certain condition is observed; black: condition observed for all subjects in every modality of activation, white: condition never met. The phases of Heel contact (H), Flat foot contact (F), Push off (P), Swing (S) are shown superimposed, delimited by dashed vertical lines.
offset (Figs. 3 and 4), as previously shown for different muscles (Di Nardo and Fioretti, 2013). This suggests that it is worth considering not only the activation patterns of each muscle, but also how frequently they are observed, i.e. their occurrence rate. The most recurrent modality of activation for GL, (observed in 39.5 ± 9.8% of total strides, Fig. 3/A) consists of two activations. The first is located during the stance, starting after the end of heel rocker, and ending at the heel off, and the second during the midswing. (Fig. 3/B). In a further percentage (27.0 ± 12.5%) of strides, the activation during the stance splits into two smaller activations;
1431
in these strides, single (21.8 ± 11.9%) or double (5.2 ± 3.9%) activations were detected during the swing. In the remaining (33.0 ± 20.1%) of strides, the activity during the swing lacks, though activity in the stance is still present. The analysis of the five modalities of activation showed that the activity of GL is centered mainly in two regions of the gait cycle. The first region is located around the transition between flat foot contact and push-off, and the second one is in the final swing phase. This observation allowed to identify a single pattern, common for all the five modalities of activation and able to characterize completely the behavior of the gastrocnemius lateralis during normal gait. This common pattern consists of the activity detected in the two regions defined above. The first region, observed in the totality of the strides, can be characterized by a single or a double activity, detected in the 72.5 ± 22.4% and in the 27.0 ± 12.5% of total strides, respectively. As reported in literature (Perry, 1992; Sutherland et al., 1980a,b), the activity during this region has been interpreted as the active participation of the gastrocnemius in restraining the forward progression of the tibia over the talus during the second rocker, hence controlling dorsiflexion. Most of this activity occurs during the mid-stance and terminal stance phases. The 3-activation modality, however, showed GL activity during pre-swing. A controversy has been reported as to whether the plantar flexor muscles produce ankle plantar flexion and knee flexion in pre-swing. Sutherland asserted that these muscles are silent during this period (Sutherland, 2001). Winter, on the contrary, believed that the knee flexion during this period resulted from plantar flexor muscle action (Winter, 1990). The large number of strides, considered in the present study, enables to show that in 77.7 ± 22.7% of the total strides GL is silent during pre-swing (1-, 2- and 4-activation modalities in Fig. 3B) as asserted by Sutherland (Sutherland, 2001, 1966), but also that in 21.8 ± 11.9% of strides (3activation modality in Fig. 3B) GL is active in pre-swing, as asserted by Winter (Winter, 1990). The second region of activity is located in swing phase, with large variability. This activity occurs in 67.1 ± 15.9% of the total strides. The presence of GL activity during this part of the gait cycle is not of general consensus. However, short activities of GL during mid-swing (as in our 2-activation modality) have been reported by Basmajian (1962) and Perry (1992). Recently, Agostini et al. (2010), in their recent study on school-age children, showed GL activities during swing, similar to those reported here. An explanation of the recruitment of GL in this phase of gait cycle could be related to the activity of the GL as a foot invertor muscle, acting in synergy with the TA for the correct positioning of the foot, in preparation of the following heel strike. A pictorial representation of the pattern of GL activation is showed in Fig. 5. For TA (Fig. 4), the most recurrent modality of activation during gait cycle consists of three activations (observed in 37.4 ± 9.4% of total strides): at the beginning of gait cycle, around stanceto-swing transition, and in the terminal swing. The 2-activation modality (32.2 ± 23.4%) differs from the most common modality for the prolonged initial activity up to 10.9 ± 6.7% of gait cycle, and for the continuous activation during swing. The 4- (23.1 ± 13.2%) and 5-activation (7.4 ± 7.2%) modalities are characterized by a double or a triple activity from pre-swing to the end of gait cycle; during stance, they present the typical activation at the beginning of gait cycle and introduce a further activity in the late mid-stance. In all the modalities of activation analyzed in the present study, the tibialis anterior is active during the beginning and the end of the gait cycle. Taking into account that the gait is periodic, it is clear that the activation in the beginning of the gait cycle is just a continuation of the activation that occurs at the end of the gait cycle, in all the modalities of activation (Fig. 2). However, following the literature (Perry, 1992), we choose to represent the modalities of activation in function of the typical gait
1432
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 23 (2013) 1428–1433
Table 1 Activation intervals of gastrocnemius lateralis in its main modalities of activation, expressed as the timing, in the percentage of gait cycle, of signal onset and offset. Values are expressed for female and male population as mean ± standard deviation (SD). Gastrocnemius lateralis
First activation (% gait cycle)
Second activation (% gait cycle)
Third activation (% gait cycle)
Fourth activation (% gait cycle)
ON
OFF
ON
OFF
ON
OFF
ON
OFF
Female (n = 7) 1-Activation mod 2-Activation mod 3-Activation mod 4-Activation mod
11.5 ± 6.2 6.2 ± 3.5 2.3 ± 1.5 0.5 ± 0.5
48.3 ± 2.1 36.0 ± 7.7 17.9 ± 8.2 12.3 ± 9.4
67.6 ± 11.9 29.3 ± 7.5 20.4 ± 9.0
80.8 ± 9.1 52.7 ± 5.6 42.6 ± 11.0
87.7 ± 6.1 55.0 ± 11.9
94.8 ± 5.0 66.4 ± 7.9
91.8 ± 4.1
98.2 ± 1.8
Male (n = 7) 1-Activation 2-Activation 3-Activation 4-Activation
10.8 ± 8.4 6.8 ± 4.9 3.6 ± 2.8 1.8 ± 1.6
47.2 ± 2.7 36.6 ± 5.7 21.1 ± 7.4 11.8 ± 9.1
69.0 ± 10.9 33.4 ± 8.4 17.5 ± 9.5
80.6 ± 9.4 55.4 ± 6.1 41.0 ± 9.5
86.5 ± 7.0 55.3 ± 14.9
92.5 ± 5.8 67.0 ± 11.5
93.2 ± 4.1
97.8 ± 2.1
mod mod mod mod
Table 2 Activation intervals of tibialis anterior in its main modalities of activation, expressed as the timing, in the percentage of gait cycle, of signal onset and offset. Values are expressed for female and male population as mean ± standard deviation (SD). Anterior
First activation (% gait cycle)
Second activation (% gait cycle)
Third activation (% gait cycle)
Fourth activation (% gait cycle)
Fifth activation (% gait cycle) ON
ON
OFF
ON
OFF
ON
OFF
ON
Female (n = 7) 2-Activation mod 3-Activation mod 4-Activation mod 5-Activation mod
3.0 ± 2.9 0.8 ± 0.7 0.1 ± 0.1 0.0 ± 0.0
11.8 ± 7.1 7.8 ± 3.4 5.7 ± 2.4 4.4 ± 2.2
63.0 ± 13.8 49.1 ± 6.5 30.1 ± 8.9 20.7 ± 10.4
97.6 ± 8.2 65.9 ± 7.0 38.5 ± 10.0 27.6 ± 10.8
79.9 ± 6.7 57.4 ± 5.9 43.7 ± 10.9
99.8 ± 0.2 72.7 ± 5.6 51.9 ± 11.9
83.3 ± 4.9 63.1 ± 8.8
99.9 ± 0.1 77.3 ± 4.9
85.7 ± 4.4
99.9 ± 0.1
Male (n = 7) 2-Activation 3-Activation 4-Activation 5-Activation
0.3 ± 0.3 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0
10.2 ± 6.5 6.9 ± 3.9 5.0 ± 2.5 4.2 ± 1.2
54.2 ± 3.6 34.6 ± 9.0 22.6 ± 9.9 14.7 ± 10.1
99.9 ± 0.1 48.2 ± 12.3 29.4 ± 11.1 20.6 ± 10.6
65.3 ± 8.5 45.2 ± 10.9 36.2 ± 9.3
99.9 ± 0.1 60.2 ± 15.2 43.5 ± 13.0
70.2 ± 11.5 59.5 ± 7.2
100.0 ± 0.0 76.0 ± 4.4
82.6 ± 4.4
100.0 ± 0.0
mod mod mod mod
cycle (Fig. 4), which starts with the initial heel contact and ends at the successive heel contact. Conforming to this convention, the activity of the tibialis anterior during the beginning and end of the gait cycle has been categorized as two separate activations. The analysis of the five modalities of activation showed that the activity of TA centered mainly in two regions of the gait cycle. The first region occurs from pre-swing to the following loading response and the second in the mid-stance. As for GL, this observation allowed to identify a single pattern common for all modalities of activation, and able to characterize completely the behavior of the tibialis anterior during normal gait. This pattern consists of the activity detected in the two regions defined above. The activity in the first region is recognized as the typical TA activation for ankle dorsiflexion (Perry, 1992; Sutherland, 2001). This region is characterized mostly by a double activity (60.5 ± 16.2% of total strides) as reported by Perry (1992), but can show also a single (32.3 ± 23.4%), or a triple (7.4 ± 7.2%) activity. The activity in the second region was detected only in the 4- and 5-activation modalities. It is not frequent (30.5 ± 15.0% of total strides) and short, lasting approximately 10% of gait cycle. Similar TA activity during midstance, usually not reported in healthy adults, has been observed in children (Agostini et al., 2010; Sutherland et al., 1980a,b). An interpretation of the recruitment of TA in this phase could be related to the activity of the TA as a foot invertor muscle for controlling balance during single support and contralateral limb swing. A pictorial representation of the pattern of TA activation is reported in Fig. 5. To evaluate the gender difference as a potential confound to the data (Chiu and Wang, 2007), the mean muscle activation intervals of the male subgroup were compared with those of the female subgroup. As shown in Tables 1 and 2, no significant difference was found in terms of number and modalities of GL and TA activation.
OFF
OFF
A more detailed analysis of gender-related differences in sEMG signal during walking could be valuable but is beyond the goal of the present study. The criteria for the recruitment of the subjects excluded overweight and obese adults based on BMI. As obesity continues to increase worldwide, a further aim of our research could be the generalization of the results achieved in the present study to a larger population, which includes obese subjects. Despite the accuracy of the methodology and the reliability of the results achieved, the small sample of participants in this study could be acknowledged as a limitation for the generality of the findings. 5. Conclusions The study was designed to describe the distribution of activation intervals for gastrocnemius lateralis and tibialis anterior during normal gait. The large number of strides considered for each participant allowed to stress that: (1) a single muscle can show a different number of activation intervals in different strides of the same walk, indicating a large variability associated with muscle activity during normal gait; (2) GL and TA show activity also in phases of the gait cycle usually not reported in healthy adults. For each muscle, the assessment of the different modalities of activation (five for each muscle) and of their statistics allowed to identify a single pattern common for all the activation modalities, and able to characterize completely the behavior of the muscles during normal gait. To our knowledge, the ‘‘normality’’ pattern, identified in the present study, represents the first attempt for the development in healthy young adults of a reference frame for dynamic EMG activity of GL and TA, in terms of variability of on-off muscular activity and frequency of occurrence during normal gait. Thus, it
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 23 (2013) 1428–1433
can be useful for comparison between normal and pathological walking in the clinical context and for designing future gait studies. References Agostini V, Knaflitz M. An algorithm for the estimation of the signal-to-noise ratio in surface myoelectric signals generated during cyclic movements. IEEE Trans Biomed Eng 2012a;59:219–25. Agostini V, Knaflitz M. Statistical gait analysis in distributed diagnosis and home healthcare. In: Rajendra Acharya U, Filippo M, Toshiyo T, Subbaram Naidu D, Jasjit SS, Stevenson Ranch, editors, vol. 2. California (USA):American Scientific Publishers; 2012b:99–121. Agostini V, Nascimbeni A, Gaffuri A, Imazio P, Benedetti MG, Knaflitz M. Normative EMG activation patterns of school-age children during gait. Gait Posture 2010;32(3):285–9. Arsenault AB, Winter DA, Marteniuk RG. Is there a ‘normal’ profile of EMG activity in gait? Med Biol Eng Comput 1986;24:337–43. Balestra G, Knaflitz M, Molinari F. Principles of statistical gait analysis. In: Proceedings of XIV ISEK conference; 2002. Basmajian JV. Muscles alive, their functions revealed by electromyography. Baltimore: Williams & Wilkins Company; 1962. Bonato P, D’Alessio T, Knaflitz M. A statistical method for the measurement of muscle activation intervals from surface myoelectric signal during gait. IEEE Trans Biomed Eng 1998;45:287–99. Chiu MC, Wang MJ. The effect of gait speed and gender on perceived exertion, muscle activity, joint motion of lower extremity, ground reaction force and heart rate during normal walking. Gait Posture 2007;25:385–92. Di Nardo F, Fioretti S. Statistical analysis of surface electromyographic signal for the assessment of rectus femoris modalities of activation during gait. J Electromyogr Kinesiol 2013;23:56–61. Freriks B, Hermens HJ, Disselhorst-Klug C, Rau G. The recommendations for sensors and sensor placement procedures for surface electromyography. In: Hermens HJ, Freriks B, Merletti R, Stegeman D, Bok J, Rau G, et al., editors. European recommendations for surface electromyography. Results of the SENIAM project. Enschede: Roessingh Research and, Development; 1999. Hirschberg GG, Nathanson M. Electromyographic recording of muscular activity in normal and spastic gaits. Arch Phys Med 1952;1952(33):217–25. Kadaba MP, Ramakrishnan HK, Wootten ME, Gainey J, Gorton G, Cochran GVB. Repeatability of kinematic, kinetic, and electromyographic data in normal adult gait. J Orthop Res 1989;6:849–60. Perry J. Gait analysis –normal and pathological function. USA: Slack Inc.; 1992. Petersen TH, Farmer SF, Kliim-Due M, Nielsen JB. Failure of normal development of central drive to ankle dorsiflexors relates to gait deficits in children with cerebral palsy. J Neurophysiol 2013;109:625–39. Powers CM, Landel RF, Perry J. Timing and intensity of vastus muscle activity during functional activities in subjects with and without patellofemoral pain. Phys Ther 1996;76:946–55. Richter J. Vorläufige Mitteilung über elektromyographische Untersuchungen am unbelasteten und belasteten Fuß. Arch Orthop Unfall-Chir 1966;59:168–76. Shiavi R, Bugle HJ, Limbird T. Electromyographic gait assessment, part 1: adult EMG profiles and walking speed. J Rehabil Res 1987;24(2):13–23. Staude G, Flachenecker C, Daumer M, Wolf W. Onset detection in surface electromyographic signals: a systematic comparison of methods. EURASIP J Appl Signal Process 2001;2:67–81. Stewart C, Postans N, Schwartz MH, Rozumalski A, Roberts A. An exploration of the function of the triceps surae during normal gait using functional electrical stimulation. Gait Posture 2007;26:482–8. Sutherland DH. An electromyographic study of the plantar flexors of the ankle in normal walking on the level. J Bone Joint Surg Am 1966;48:66–71. Sutherland DH. The evolution of clinical gait analysis part l: kinesiological EMG. Gait Posture 2001;14:61–70. Sutherland DH, Cooper L, Daniel D. The role of the ankle plantar flexors in normal walking. J Bone Joint Surg Am 1980a;62:354–63.
1433
Sutherland DH, Olshen R, Cooper L, Woo SL. The development of mature gait. J Bone Joint Surg Am 1980b;62:336–53. Winter DA. Biomechanics and motor control of human movement. 2nd ed. New York: Wiley; 1990. Winter DA, Yack HJ. EMG profiles during normal human walking: stride to stride and inter-subject variability. Electroencephalogr Clin Neurophysiol 1987;67:402–11. Yang JF, Winter DA. Electromyographic amplitude normalization methods: improving their sensitivity as diagnostic tools in gait analysis. Arch Phys Med Rehabil 1984;65:517–21.
Francesco Di Nardo received his Ph.D. in Artificial intelligence systems from Università Politecnica delle Marche, Ancona, Italy in 2005. His main research activity involves the development and the clinical application (type-2 diabetes, insulin resistance and hypertension) of mathematical models of metabolic and endocrine systems. In this field, he published several papers in peer review journals and participated in National research projects. Recently, he is addressing his research in the field of movement analysis for motor rehabilitation, with particular involvement in the acquisition and the processing of surface electromyography signal to assess the muscular function during gait task.
Giacomo Ghetti graduated in Health and Rehabilitation Profession Sciences at University of L’Aquila and currently is the coordinator of the Physiotherapy Unit at the Posture and Movement Analysis Laboratory, Italian National Institute of Health and Science on Aging (INRCA), Ancona, Italy. His clinical activity at the Functional Recovery and Rehabilitation Unit at INRCA concerns the prevention, treatment, and rehabilitation of neurological, surgical, rheumatologic, orthopaedic and traumatologic disorders. His main research activity involves the field of movement analysis for motor rehabilitation. In this field, he is author of numerous scientific publications in national and international journals and congress proceedings.
Sandro Fioretti graduated in 1979 in Electronic Engineering at Ancona University and presently is Associate Professor in Bioengineering at the Department of Information Engineering – Università Politecnica delle Marche - Ancona. He teaches Movement Biomechanics and Bioengineering of Motor Rehabilitation at the Biomedical Engineering course of the same University. His main research interests are in the field of human movement analysis and its related fields such as: stereophotogrammetry, linear and nonlinear filtering, joint kinematics, analysis and identification of postural control, static and perturbed posturography, gait analysis, dynamic electromyography. He participated in various European and National research projects in the field of movement analysis for motor rehabilitation. He is author of numerous scientific publications in international journals, books and congress proceedings.
Contents lists available at SciVerse ScienceDirect
Journal of Electromyography and Kinesiology journal homepage: www.elsevier.com/locate/jelekin
Assessment of the activation modalities of gastrocnemius lateralis and tibialis anterior during gait: A statistical analysis Francesco Di Nardo a,⇑, Giacomo Ghetti b, Sandro Fioretti a a b
Department of Information Engineering, Università Politecnica delle Marche, 60131 Ancona, Italy Posture and Movement Analysis Laboratory, Italian National Institute of Health and Science on Aging (INRCA), 60131 Ancona, Italy
a r t i c l e
i n f o
Article history: Received 12 December 2012 Received in revised form 28 May 2013 Accepted 29 May 2013
Keywords: EMG Ankle flexor muscles Gait analysis
a b s t r a c t Aim of the study was to identify the different modalities of activation of gastrocnemius lateralis (GL) and tibialis anterior (TA) during gait at self-selected speed, by a statistical analysis of surface electromyographic signal from a large number (hundreds) of strides per subject. The analysis on fourteen healthy adults showed a large variability in the number of activation intervals, in their occurrence rate, and in the on-off instants, within different strides of the same walk. For each muscle, the assessment of the different modalities of activation (five for muscle) allowed to identify a single pattern, common for all the modalities and able to characterize the behavior of muscles during normal gait. The pattern of GL activity centered in two regions of the gait cycle: the transition between flat foot contact and push-off (observed in 100% of total strides) and the final swing (67.1 ± 15.9%). Two regions characterized also the pattern of TA activity: from pre-swing to following loading response (100%), and the mid-stance (30.5 ± 15.0%). This ‘‘normality’’ pattern represents the first attempt for the development in healthy young adults of a reference for dynamic EMG activity of GL and TA, in terms of variability of on-off muscular activity and occurrence rate during gait. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Ankle flexor muscles play an important role in everyday activities, especially walking. During gait, ankle plantar flexors act to restrain the forward rotation of the tibia on the talus during stance phase, provide ankle stability, contribute to knee stability, and conserve energy by minimizing vertical oscillation of the whole-body center of mass (Sutherland et al., 1980a,b). The main role of the ankle dorsi-flexors during gait is to prevent slapping of the foot on the ground in initial stance, to permit the forefoot to clear the ground in initial swing, and to hold the ankle in position for initial contact (Perry, 1992). Surface electromyography (sEMG) has been largely used for the assessment of the activation patterns of the ankle flexor muscles during normal and pathological gait (Petersen et al., 2013; Stewart et al., 2007; Sutherland et al., 1980a,b). The general consensus is that ankle plantar flexors are active during the stance phase, while the ankle dorsi-flexors are active mostly during the swing phase, with a continuation of their activity till the next loading response (Perry, 1992; Sutherland, 2001). In particular, normal EMG activity
⇑ Corresponding author. Address: Department of Information Engineering, Università Politecnica delle Marche, Via Brecce Bianche, 60131 Ancona, Italy. Fax: +39 0712204224. E-mail address: [email protected] (F. Di Nardo). 1050-6411/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jelekin.2013.05.011
for the gastrocnemius has been reported to occur from the loading response to the terminal stance. A short activation at the beginning of gait cycle and a further activation from pre-swing to the end of gait cycle have been identified as the normal EMG activity for the tibialis anterior (Perry, 1992; Shiavi et al., 1987; Sutherland, 2001). However, both in old (Hirschberg and Nathanson, 1952; Richter, 1966) and in more recent (Agostini et al., 2010) studies, activities outside the typical activation intervals for gastrocnemius and tibialis anterior and, more generally, a large stride-to-stride variability in the EMG profiles have been reported (Arsenault et al., 1986; Winter and Yack, 1987; Yang and Winter, 1984). It is, therefore, important to study the natural variability associated with muscle activity during free walking, in order to improve the interpretation of EMG signals in both physiological and pathological conditions. This can be achieved only by recording and analyzing the electromyographic signal over a large number of steps per subject, which allows to detect the different modalities of muscle activation. From this point of view, data on ankle flexor modalities of activation, reported by cited studies, are limited by the small number of gait cycles considered during an assessment session. A study over a larger number of steps should be considered. This goal appears to be reachable by means of surface EMG, because it is a non-invasive and completely painless technique. Thus, the aim of the present study was to identify the different modalities of activation of gastrocnemius lateralis (GL) and tibialis
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 23 (2013) 1428–1433
anterior (TA) in healthy adults during gait at self-selected speed, by analyzing surface EMG signal from a large number (hundreds) of strides per subject. The study is based on the recent availability of robust techniques for the detection of muscle activation intervals, and specific tools for statistical gait analysis (Bonato et al., 1998; Balestra et al., 2002; Staude et al., 2001). 2. Materials and methods 2.1. Subjects Fourteen healthy adult volunteers were recruited (mean age ± SD: 23.9 ± 2.3 years; height 174 ± 10 cm; weight 63.4 ± 14.1 kg; body mass index (BMI) 20.7 ± 2.2 kg m 2, male-female ratio: 7/7). Exclusion criteria included history of neurological pathology, orthopedic surgery within the previous year, acute or chronic knee pain or pathology, BMI P 25, or abnormal gait. Before the beginning of the test, all participants signed informed consent. 2.2. Recording system : signal acquisition and processing Signals were acquired by means of a multichannel recording system for statistical gait analysis (Step32, DemItalia, Italy). Each subject was instrumented with foot-switches, goniometers and sEMG probes on both right and left lower limb. Three foot-switches (size: 11 11 0.5 mm; activation force: 0.2 N) were attached beneath the heel, the first and the fifth metatarsal heads of each foot. A goniometer (accuracy: 0.5°) was attached to the lateral side of each lower limb for measuring the knee joint angles in the sagittal plane. Single differential (SD) sEMG probes with fixed geometry constituted by Ag-disks (manufacturer: DemItalia, size: 7 27 19 mm; interelectrode distance: 12 mm, gain: 1000, high-pass filter: 20 Hz) were attached over Gastrocnemius Lateralis (GL) and Tibialis Anterior (TA), following the SENIAM recommendations (Freriks et al., 1999). Subjects were asked to walk barefoot on level ground for around 5 min at their natural pace, following the path schematized in Fig. 1. Time events were identified from the footswitch signals corresponding to Heel contact (H), Flat foot contact (F), Push off (P), Swing (S) and processed to segment and classify the different gait cycles. During acceleration, deceleration, and changes in direction the strides are different from those of steady state walking. Therefore, knee angles in the sagittal plane (lowpass filtered with cut-off frequency of 15 Hz) along with sequences and durations of gait phases derived by the basographic signal, were used by a multivariate statistical filter embedded in the Step32 system, to detect outlier cycles like those relative to deceleration, reversing, and acceleration. Thus, cycles with improper sequences of gait phases (i.e. different from H–F–P–S), with abnormal timing and knee angles, with respect to a mean value computed on each single subject, were discarded. (Agostini and Knaflitz, 2012b). Natural pace was chosen because walking at a
Fig. 1. Schematic representation of the path walked by the recruited subjects during the experiment; subjects walked barefoot over the floor for 5 min at their natural pace.
1429
self-selected speed improves the repeatability of EMG data (Kadaba et al., 1989), while variability increases when subjects are required to walk abnormally slow (Powers et al., 1996; Winter and Yack, 1987). EMG signals were high-pass filtered (cut-off frequency of 20 Hz) and then processed by a double-threshold statistical detector, embedded in the Step32 system, that provides the onset and offset time instants of muscle activity in a completely user-independent way. This technique (Bonato et al., 1998) consists of selecting a first threshold f and observing m successive samples: if at least r0 out of successive m samples are above the first threshold f, the presence of the signal is acknowledged. In this approach, the second threshold is represented by r0. Thus, the behavior of the double-threshold detector is fixed by three parameters: the first threshold f, the second threshold r0, and the length of the observation window m. Their values are selected to jointly minimize the value of false-alarm probability and maximize probability of detection for each specific signal-to-noise ratio. The setting of the first threshold, f, is based on the assessment of the background noise level, as a necessary input parameter. Furthermore, the double-threshold detector requires to estimate the signal-to-noise ratio in order to fine tune the second threshold, r0. The values of the background noise level and the signal-to-noise ratio, necessary to run doublethreshold algorithm, is estimated for each signal by Step32 system, using the statistical approach proposed by Agostini and Knaflitz (2012a). The length duration of the observation window, m, of 30 ms is considered a suitable value for the study of muscle activation in gait analysis (Bonato et al., 1998). During gait, a muscle activates a number of times which is usually variable from cycle to cycle (Agostini and Knaflitz, 2012b). Thus, muscle on/off instants should be averaged considering each single modality of activation by itself. With modality of activation we mean the number of times when muscle activates during a single gait cycle, i.e. n-activation modality consists of n activation intervals for the considered muscle, during a single gait cycle. In the present study, mean activation intervals (normalized with respect to the gait cycle) for each modality of activation were achieved by means of the Step32 system, according to the following steps. First, muscle activations relative to each gait cycle were identified. Then, for all the gait cycles corresponding to straight line walking, muscle activations were grouped according to the number of activations detected, i.e. relatively to the modalities of activations detected. Finally the on/off time instants were averaged, for each specific modality of activation observed, and relative standard deviation and standard error were computed. In the present study only gait cycles consisting of the sequence of H–F–P–S were considered.
3. Results For each subject, a mean (±standard deviation, SD) of 461 ± 128 strides has been considered. From the total of 6448 strides considered, 146 strides (2.3% of total strides) have been removed from the analysis because they did not follow the H–F–P–S foot-switch pattern. Thus, the mean value (±SD) of 450 ± 126 strides was obtained for each subject, and the mean (±SD) duration of H, F, P and S phases over our population was computed. The H-phase lasts 5.6 ± 1.7% of the gait cycle (mean and standard deviation), the Fphase 28.9 ± 6.0%, the P-phase 23.4 ± 4.8% and the S-phase 44.1 ± 2.8%. The mean results are reported with data from right and left lower limb considered all together. The analysis of the myoelectric signal confirms that muscles show a different number of activation intervals in different strides of the same walk. As example, EMG signals from the TA muscle of a subject have been depicted in Fig. 2.
1430
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 23 (2013) 1428–1433
Fig. 2. An example of EMG signals of TA muscle showing 2 activations (panel A), 3 activations (panel B) and 4 activations (panel C), respectively. The signals have been extracted from different strides of the same subject, during the same walk. The heel contact, flat foot contact, push off and swing phases are delimited by dashed vertical lines.
The most recurrent modality of activation for GL (Fig. 3) consists of two activations (2-activation modality), observed in 39.5 ± 9.8% of total strides; the first occurs in the transition between flat foot contact and push-off phases (6.5–36.3% of gait cycle), and the second during the mid-swing. The second most recurrent modality of activation is similar to the 2-activation modality but with no activation during the swing. This 1-activation modality was observed in 33.0 ± 20.1% of total strides. In a further 21.8 ± 11.9% of total strides, three activations were observed for GL: in the flat foot contact phase, in the push-off phase and in terminal swing (3-activation modality). The remaining 5.2 ± 3.9% of total strides was characterized by two activations during the stance and two activations during the swing (4-activation modality). For TA (Fig. 4), the most recurrent modality of activation during gait cycle consists of three activations (3-activation modality), observed in 37.4 ± 9.4% of total strides: the first occurs at the beginning of gait cycle, the second around stance-to swing-transition and the third in the terminal swing. In a 23.1 ± 13.2% of total strides a 4-activation modality was observed, similar to the 3-activation modality but with a further activation in the transition between flat foot contact and push-off phases. In a 32.3 ± 23.4% of total strides, only two activations were observed: from the beginning up to 10.9 ± 6.7% of the gait cycle, and during the whole swing (2-activation modality). The remaining 7.4 ± 7.2% of total strides was characterized by five activations, three during the stance and two during the swing (5-activation modality).
Fig. 3. Gastrocnemius lateralis: mean (+SE) percentage frequency of each of the five different modalities of activation patterns (panel A) and mean (+SE) activation intervals vs. percentage of gait cycle for the modalities with 1, 2, 3 and 4 activations, respectively (panel B), detected during walk. Mean and standard error (SE) refer to the whole population. Standard deviations are reported in the text and in Table 1. The heel contact, flat foot contact, push off and swing phases are delimited by dashed vertical lines.
The 5-activation modality for GL and the 1-activation modality for TA were observed only in 1% of the total stride and showed a large variability ; thus, they are not considered in the present analysis. Fig. 5 showed a pictorial representation of GL and TA activation onset and offset instants in function of the number of subjects where muscular activity is observed; this representation has been achieved for each muscle, considering the four main modalities of activation all together. No significant difference was found between mean muscle activation intervals of the male subgroup (7 subjects) with those of the female subgroup (7 subjects), in terms of number and modalities of GL and TA activation, as shown in Tables 1 and 2, respectively.
4. Discussion In the present study, EMG signals were recorded in fourteen healthy adult subjects, in order to determine the picture of the activation modalities of gastrocnemius lateralis (GL) and tibialis anterior (TA) during normal gait. A statistical analysis of a large number (approximately 475) of strides for each subject was performed to this aim. The muscle activation intervals, detected in each subject at self-selected walking speed, followed roughly the typical pattern reported for each muscle during gait (Basmajian, 1962; Perry, 1992; Sutherland, 2001). However, the statistical analysis put into evidence that the considered muscles showed different modalities in the number of activations and in the timing of signal onset and
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 23 (2013) 1428–1433
Fig. 4. Tibialis anterior: mean (+SE) percentage frequency of each of the five different modalities of activation patterns (panel A) and mean (+SE) activation intervals vs. percentage of gait cycle for the modalities with 2, 3, 4 and 5 activations, respectively (panel B), detected during the walk. Mean and standard error (SE) refer to the whole population. Standard deviations are reported in the text and in Table 1. The heel contact, flat foot contact, push off and swing phases are delimited by dashed vertical lines.
Fig. 5. Muscle activation onset and offset instants over the population for GL and TA, as percentage of gait cycle, considering, for each muscle, the four main modalities of activation all together. Horizontal bars are grey-level coded, according to the number of subjects where a certain condition is observed; black: condition observed for all subjects in every modality of activation, white: condition never met. The phases of Heel contact (H), Flat foot contact (F), Push off (P), Swing (S) are shown superimposed, delimited by dashed vertical lines.
offset (Figs. 3 and 4), as previously shown for different muscles (Di Nardo and Fioretti, 2013). This suggests that it is worth considering not only the activation patterns of each muscle, but also how frequently they are observed, i.e. their occurrence rate. The most recurrent modality of activation for GL, (observed in 39.5 ± 9.8% of total strides, Fig. 3/A) consists of two activations. The first is located during the stance, starting after the end of heel rocker, and ending at the heel off, and the second during the midswing. (Fig. 3/B). In a further percentage (27.0 ± 12.5%) of strides, the activation during the stance splits into two smaller activations;
1431
in these strides, single (21.8 ± 11.9%) or double (5.2 ± 3.9%) activations were detected during the swing. In the remaining (33.0 ± 20.1%) of strides, the activity during the swing lacks, though activity in the stance is still present. The analysis of the five modalities of activation showed that the activity of GL is centered mainly in two regions of the gait cycle. The first region is located around the transition between flat foot contact and push-off, and the second one is in the final swing phase. This observation allowed to identify a single pattern, common for all the five modalities of activation and able to characterize completely the behavior of the gastrocnemius lateralis during normal gait. This common pattern consists of the activity detected in the two regions defined above. The first region, observed in the totality of the strides, can be characterized by a single or a double activity, detected in the 72.5 ± 22.4% and in the 27.0 ± 12.5% of total strides, respectively. As reported in literature (Perry, 1992; Sutherland et al., 1980a,b), the activity during this region has been interpreted as the active participation of the gastrocnemius in restraining the forward progression of the tibia over the talus during the second rocker, hence controlling dorsiflexion. Most of this activity occurs during the mid-stance and terminal stance phases. The 3-activation modality, however, showed GL activity during pre-swing. A controversy has been reported as to whether the plantar flexor muscles produce ankle plantar flexion and knee flexion in pre-swing. Sutherland asserted that these muscles are silent during this period (Sutherland, 2001). Winter, on the contrary, believed that the knee flexion during this period resulted from plantar flexor muscle action (Winter, 1990). The large number of strides, considered in the present study, enables to show that in 77.7 ± 22.7% of the total strides GL is silent during pre-swing (1-, 2- and 4-activation modalities in Fig. 3B) as asserted by Sutherland (Sutherland, 2001, 1966), but also that in 21.8 ± 11.9% of strides (3activation modality in Fig. 3B) GL is active in pre-swing, as asserted by Winter (Winter, 1990). The second region of activity is located in swing phase, with large variability. This activity occurs in 67.1 ± 15.9% of the total strides. The presence of GL activity during this part of the gait cycle is not of general consensus. However, short activities of GL during mid-swing (as in our 2-activation modality) have been reported by Basmajian (1962) and Perry (1992). Recently, Agostini et al. (2010), in their recent study on school-age children, showed GL activities during swing, similar to those reported here. An explanation of the recruitment of GL in this phase of gait cycle could be related to the activity of the GL as a foot invertor muscle, acting in synergy with the TA for the correct positioning of the foot, in preparation of the following heel strike. A pictorial representation of the pattern of GL activation is showed in Fig. 5. For TA (Fig. 4), the most recurrent modality of activation during gait cycle consists of three activations (observed in 37.4 ± 9.4% of total strides): at the beginning of gait cycle, around stanceto-swing transition, and in the terminal swing. The 2-activation modality (32.2 ± 23.4%) differs from the most common modality for the prolonged initial activity up to 10.9 ± 6.7% of gait cycle, and for the continuous activation during swing. The 4- (23.1 ± 13.2%) and 5-activation (7.4 ± 7.2%) modalities are characterized by a double or a triple activity from pre-swing to the end of gait cycle; during stance, they present the typical activation at the beginning of gait cycle and introduce a further activity in the late mid-stance. In all the modalities of activation analyzed in the present study, the tibialis anterior is active during the beginning and the end of the gait cycle. Taking into account that the gait is periodic, it is clear that the activation in the beginning of the gait cycle is just a continuation of the activation that occurs at the end of the gait cycle, in all the modalities of activation (Fig. 2). However, following the literature (Perry, 1992), we choose to represent the modalities of activation in function of the typical gait
1432
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 23 (2013) 1428–1433
Table 1 Activation intervals of gastrocnemius lateralis in its main modalities of activation, expressed as the timing, in the percentage of gait cycle, of signal onset and offset. Values are expressed for female and male population as mean ± standard deviation (SD). Gastrocnemius lateralis
First activation (% gait cycle)
Second activation (% gait cycle)
Third activation (% gait cycle)
Fourth activation (% gait cycle)
ON
OFF
ON
OFF
ON
OFF
ON
OFF
Female (n = 7) 1-Activation mod 2-Activation mod 3-Activation mod 4-Activation mod
11.5 ± 6.2 6.2 ± 3.5 2.3 ± 1.5 0.5 ± 0.5
48.3 ± 2.1 36.0 ± 7.7 17.9 ± 8.2 12.3 ± 9.4
67.6 ± 11.9 29.3 ± 7.5 20.4 ± 9.0
80.8 ± 9.1 52.7 ± 5.6 42.6 ± 11.0
87.7 ± 6.1 55.0 ± 11.9
94.8 ± 5.0 66.4 ± 7.9
91.8 ± 4.1
98.2 ± 1.8
Male (n = 7) 1-Activation 2-Activation 3-Activation 4-Activation
10.8 ± 8.4 6.8 ± 4.9 3.6 ± 2.8 1.8 ± 1.6
47.2 ± 2.7 36.6 ± 5.7 21.1 ± 7.4 11.8 ± 9.1
69.0 ± 10.9 33.4 ± 8.4 17.5 ± 9.5
80.6 ± 9.4 55.4 ± 6.1 41.0 ± 9.5
86.5 ± 7.0 55.3 ± 14.9
92.5 ± 5.8 67.0 ± 11.5
93.2 ± 4.1
97.8 ± 2.1
mod mod mod mod
Table 2 Activation intervals of tibialis anterior in its main modalities of activation, expressed as the timing, in the percentage of gait cycle, of signal onset and offset. Values are expressed for female and male population as mean ± standard deviation (SD). Anterior
First activation (% gait cycle)
Second activation (% gait cycle)
Third activation (% gait cycle)
Fourth activation (% gait cycle)
Fifth activation (% gait cycle) ON
ON
OFF
ON
OFF
ON
OFF
ON
Female (n = 7) 2-Activation mod 3-Activation mod 4-Activation mod 5-Activation mod
3.0 ± 2.9 0.8 ± 0.7 0.1 ± 0.1 0.0 ± 0.0
11.8 ± 7.1 7.8 ± 3.4 5.7 ± 2.4 4.4 ± 2.2
63.0 ± 13.8 49.1 ± 6.5 30.1 ± 8.9 20.7 ± 10.4
97.6 ± 8.2 65.9 ± 7.0 38.5 ± 10.0 27.6 ± 10.8
79.9 ± 6.7 57.4 ± 5.9 43.7 ± 10.9
99.8 ± 0.2 72.7 ± 5.6 51.9 ± 11.9
83.3 ± 4.9 63.1 ± 8.8
99.9 ± 0.1 77.3 ± 4.9
85.7 ± 4.4
99.9 ± 0.1
Male (n = 7) 2-Activation 3-Activation 4-Activation 5-Activation
0.3 ± 0.3 0.1 ± 0.1 0.0 ± 0.0 0.0 ± 0.0
10.2 ± 6.5 6.9 ± 3.9 5.0 ± 2.5 4.2 ± 1.2
54.2 ± 3.6 34.6 ± 9.0 22.6 ± 9.9 14.7 ± 10.1
99.9 ± 0.1 48.2 ± 12.3 29.4 ± 11.1 20.6 ± 10.6
65.3 ± 8.5 45.2 ± 10.9 36.2 ± 9.3
99.9 ± 0.1 60.2 ± 15.2 43.5 ± 13.0
70.2 ± 11.5 59.5 ± 7.2
100.0 ± 0.0 76.0 ± 4.4
82.6 ± 4.4
100.0 ± 0.0
mod mod mod mod
cycle (Fig. 4), which starts with the initial heel contact and ends at the successive heel contact. Conforming to this convention, the activity of the tibialis anterior during the beginning and end of the gait cycle has been categorized as two separate activations. The analysis of the five modalities of activation showed that the activity of TA centered mainly in two regions of the gait cycle. The first region occurs from pre-swing to the following loading response and the second in the mid-stance. As for GL, this observation allowed to identify a single pattern common for all modalities of activation, and able to characterize completely the behavior of the tibialis anterior during normal gait. This pattern consists of the activity detected in the two regions defined above. The activity in the first region is recognized as the typical TA activation for ankle dorsiflexion (Perry, 1992; Sutherland, 2001). This region is characterized mostly by a double activity (60.5 ± 16.2% of total strides) as reported by Perry (1992), but can show also a single (32.3 ± 23.4%), or a triple (7.4 ± 7.2%) activity. The activity in the second region was detected only in the 4- and 5-activation modalities. It is not frequent (30.5 ± 15.0% of total strides) and short, lasting approximately 10% of gait cycle. Similar TA activity during midstance, usually not reported in healthy adults, has been observed in children (Agostini et al., 2010; Sutherland et al., 1980a,b). An interpretation of the recruitment of TA in this phase could be related to the activity of the TA as a foot invertor muscle for controlling balance during single support and contralateral limb swing. A pictorial representation of the pattern of TA activation is reported in Fig. 5. To evaluate the gender difference as a potential confound to the data (Chiu and Wang, 2007), the mean muscle activation intervals of the male subgroup were compared with those of the female subgroup. As shown in Tables 1 and 2, no significant difference was found in terms of number and modalities of GL and TA activation.
OFF
OFF
A more detailed analysis of gender-related differences in sEMG signal during walking could be valuable but is beyond the goal of the present study. The criteria for the recruitment of the subjects excluded overweight and obese adults based on BMI. As obesity continues to increase worldwide, a further aim of our research could be the generalization of the results achieved in the present study to a larger population, which includes obese subjects. Despite the accuracy of the methodology and the reliability of the results achieved, the small sample of participants in this study could be acknowledged as a limitation for the generality of the findings. 5. Conclusions The study was designed to describe the distribution of activation intervals for gastrocnemius lateralis and tibialis anterior during normal gait. The large number of strides considered for each participant allowed to stress that: (1) a single muscle can show a different number of activation intervals in different strides of the same walk, indicating a large variability associated with muscle activity during normal gait; (2) GL and TA show activity also in phases of the gait cycle usually not reported in healthy adults. For each muscle, the assessment of the different modalities of activation (five for each muscle) and of their statistics allowed to identify a single pattern common for all the activation modalities, and able to characterize completely the behavior of the muscles during normal gait. To our knowledge, the ‘‘normality’’ pattern, identified in the present study, represents the first attempt for the development in healthy young adults of a reference frame for dynamic EMG activity of GL and TA, in terms of variability of on-off muscular activity and frequency of occurrence during normal gait. Thus, it
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 23 (2013) 1428–1433
can be useful for comparison between normal and pathological walking in the clinical context and for designing future gait studies. References Agostini V, Knaflitz M. An algorithm for the estimation of the signal-to-noise ratio in surface myoelectric signals generated during cyclic movements. IEEE Trans Biomed Eng 2012a;59:219–25. Agostini V, Knaflitz M. Statistical gait analysis in distributed diagnosis and home healthcare. In: Rajendra Acharya U, Filippo M, Toshiyo T, Subbaram Naidu D, Jasjit SS, Stevenson Ranch, editors, vol. 2. California (USA):American Scientific Publishers; 2012b:99–121. Agostini V, Nascimbeni A, Gaffuri A, Imazio P, Benedetti MG, Knaflitz M. Normative EMG activation patterns of school-age children during gait. Gait Posture 2010;32(3):285–9. Arsenault AB, Winter DA, Marteniuk RG. Is there a ‘normal’ profile of EMG activity in gait? Med Biol Eng Comput 1986;24:337–43. Balestra G, Knaflitz M, Molinari F. Principles of statistical gait analysis. In: Proceedings of XIV ISEK conference; 2002. Basmajian JV. Muscles alive, their functions revealed by electromyography. Baltimore: Williams & Wilkins Company; 1962. Bonato P, D’Alessio T, Knaflitz M. A statistical method for the measurement of muscle activation intervals from surface myoelectric signal during gait. IEEE Trans Biomed Eng 1998;45:287–99. Chiu MC, Wang MJ. The effect of gait speed and gender on perceived exertion, muscle activity, joint motion of lower extremity, ground reaction force and heart rate during normal walking. Gait Posture 2007;25:385–92. Di Nardo F, Fioretti S. Statistical analysis of surface electromyographic signal for the assessment of rectus femoris modalities of activation during gait. J Electromyogr Kinesiol 2013;23:56–61. Freriks B, Hermens HJ, Disselhorst-Klug C, Rau G. The recommendations for sensors and sensor placement procedures for surface electromyography. In: Hermens HJ, Freriks B, Merletti R, Stegeman D, Bok J, Rau G, et al., editors. European recommendations for surface electromyography. Results of the SENIAM project. Enschede: Roessingh Research and, Development; 1999. Hirschberg GG, Nathanson M. Electromyographic recording of muscular activity in normal and spastic gaits. Arch Phys Med 1952;1952(33):217–25. Kadaba MP, Ramakrishnan HK, Wootten ME, Gainey J, Gorton G, Cochran GVB. Repeatability of kinematic, kinetic, and electromyographic data in normal adult gait. J Orthop Res 1989;6:849–60. Perry J. Gait analysis –normal and pathological function. USA: Slack Inc.; 1992. Petersen TH, Farmer SF, Kliim-Due M, Nielsen JB. Failure of normal development of central drive to ankle dorsiflexors relates to gait deficits in children with cerebral palsy. J Neurophysiol 2013;109:625–39. Powers CM, Landel RF, Perry J. Timing and intensity of vastus muscle activity during functional activities in subjects with and without patellofemoral pain. Phys Ther 1996;76:946–55. Richter J. Vorläufige Mitteilung über elektromyographische Untersuchungen am unbelasteten und belasteten Fuß. Arch Orthop Unfall-Chir 1966;59:168–76. Shiavi R, Bugle HJ, Limbird T. Electromyographic gait assessment, part 1: adult EMG profiles and walking speed. J Rehabil Res 1987;24(2):13–23. Staude G, Flachenecker C, Daumer M, Wolf W. Onset detection in surface electromyographic signals: a systematic comparison of methods. EURASIP J Appl Signal Process 2001;2:67–81. Stewart C, Postans N, Schwartz MH, Rozumalski A, Roberts A. An exploration of the function of the triceps surae during normal gait using functional electrical stimulation. Gait Posture 2007;26:482–8. Sutherland DH. An electromyographic study of the plantar flexors of the ankle in normal walking on the level. J Bone Joint Surg Am 1966;48:66–71. Sutherland DH. The evolution of clinical gait analysis part l: kinesiological EMG. Gait Posture 2001;14:61–70. Sutherland DH, Cooper L, Daniel D. The role of the ankle plantar flexors in normal walking. J Bone Joint Surg Am 1980a;62:354–63.
1433
Sutherland DH, Olshen R, Cooper L, Woo SL. The development of mature gait. J Bone Joint Surg Am 1980b;62:336–53. Winter DA. Biomechanics and motor control of human movement. 2nd ed. New York: Wiley; 1990. Winter DA, Yack HJ. EMG profiles during normal human walking: stride to stride and inter-subject variability. Electroencephalogr Clin Neurophysiol 1987;67:402–11. Yang JF, Winter DA. Electromyographic amplitude normalization methods: improving their sensitivity as diagnostic tools in gait analysis. Arch Phys Med Rehabil 1984;65:517–21.
Francesco Di Nardo received his Ph.D. in Artificial intelligence systems from Università Politecnica delle Marche, Ancona, Italy in 2005. His main research activity involves the development and the clinical application (type-2 diabetes, insulin resistance and hypertension) of mathematical models of metabolic and endocrine systems. In this field, he published several papers in peer review journals and participated in National research projects. Recently, he is addressing his research in the field of movement analysis for motor rehabilitation, with particular involvement in the acquisition and the processing of surface electromyography signal to assess the muscular function during gait task.
Giacomo Ghetti graduated in Health and Rehabilitation Profession Sciences at University of L’Aquila and currently is the coordinator of the Physiotherapy Unit at the Posture and Movement Analysis Laboratory, Italian National Institute of Health and Science on Aging (INRCA), Ancona, Italy. His clinical activity at the Functional Recovery and Rehabilitation Unit at INRCA concerns the prevention, treatment, and rehabilitation of neurological, surgical, rheumatologic, orthopaedic and traumatologic disorders. His main research activity involves the field of movement analysis for motor rehabilitation. In this field, he is author of numerous scientific publications in national and international journals and congress proceedings.
Sandro Fioretti graduated in 1979 in Electronic Engineering at Ancona University and presently is Associate Professor in Bioengineering at the Department of Information Engineering – Università Politecnica delle Marche - Ancona. He teaches Movement Biomechanics and Bioengineering of Motor Rehabilitation at the Biomedical Engineering course of the same University. His main research interests are in the field of human movement analysis and its related fields such as: stereophotogrammetry, linear and nonlinear filtering, joint kinematics, analysis and identification of postural control, static and perturbed posturography, gait analysis, dynamic electromyography. He participated in various European and National research projects in the field of movement analysis for motor rehabilitation. He is author of numerous scientific publications in international journals, books and congress proceedings.