- Email: [email protected]
Assessment of the variability of vastii myoelectric activity in young healthy females during walking: A statistical gait analysis
Journal of Electromyography and Kinesiology 25 (2015) 800–807
Contents lists available at ScienceDirect
Journal of Electromyography and Kinesiology journal homepage: www.elsevier.com/locate/jelekin
Assessment of the variability of vastii myoelectric activity in young healthy females during walking: A statistical gait analysis Francesco Di Nardo a,⇑, Elvira Maranesi a,b, Alessandro Mengarelli a, Giacomo Ghetti b, Laura Burattini a, 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 10 November 2014 Received in revised form 17 June 2015 Accepted 3 July 2015
Keywords: Surface electromyography Knee-flexor muscles Statistical gait analysis Variability Quadriceps femoris
a b s t r a c t The study was designed to assess the natural variability of the activation modalities of vastus medialis (VM) and vastus lateralis (VL) during walking at a self-selected speed and cadence of 30 young, healthy, females. This was achieved by conducting statistical gait analysis on the surface electromyographic signals from hundreds of strides for each subject. Results revealed variability in the number of activations, occurrence frequency, and onset-offset instants across the thousands of strides analyzed. However, despite the variability, there was one activation occurrence which remained consistent across subjects for both VM and VL. This occurred from terminal swing to the following loading response (observed in 100% of strides). A second, less frequent, activation occurred between mid-stance up to pre-swing (observed in 39.3 ± 22.4% of strides for VM and in 35.1 ± 20.6% for VL). No significant differences (p > 0.05) were observed in the onset–offset instants or in the occurrence frequency, which suggest a simultaneous recruitment of VM and VL. This ‘‘normality’’ pattern represents the first attempt at developing a reference frame for vastii sEMG activity during walking, that is able to include the physiological variability of the phenomenon and control the confounding effects of age and gender. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction The knee is the key to stance stability, and the muscles of the quadriceps femoris group are the most direct source of extensor control (Perry, 1992). During walking, however, the knee extensor muscles are mainly used to restrain the shock-absorbing flexion during loading response; the quadriceps femoris eccentric contraction in the touching phase of the gait cycle (GC), indeed, represents the primary absorbing mechanism of impact during weight acceptance. In stance, the extensors act to decelerate knee flexion, while in swing they contribute to limb progression (Perry, 1992). Moreover, vastus medialis (VM) and vastus lateralis (VL) play the important role of stabilizing the patella and the knee joint during walking (Grelsamer and McConnell, 1988). Thus, the vastii muscles play a fundamental role in normal walking physiology and in the etiology of common knee pathologies. A VM–VL activity imbalance, observed in thirty-three patients (vs. thirty-three controls) with ⇑ Corresponding author at: 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). http://dx.doi.org/10.1016/j.jelekin.2015.07.004 1050-6411/Ó 2015 Elsevier Ltd. All rights reserved.
patellofemoral pain syndrome, was reported as a possible mechanism for abnormal patellar tracking (Cowan et al., 2001). Gender-related differences in both knee kinematic and vastii myoelectric activity were also reported (Chumanov et al., 2008; Di Nardo et al., 2014; Kerrigan et al., 1998; Malinzak et al., 2001). In a recent electromyographic study of ours, involving twenty-two healthy adults (Di Nardo et al., 2015b), females showed a higher VL-recruitment frequency during walking than males. In a study on seventeen young females and seventeen age-matched males (Chumanov et al., 2008), the former group showed higher VL myoelectric activity during the terminal swing and initial running loading transition than the latter. Eventually, in a study on nine females and eleven males (Malinzak et al., 2001), females showed a higher quadriceps activation during different motor tasks than males. It was also observed that advancing age modifies both the activation timing of lower limb muscles (including the vastii), and the activity duration of agonist and antagonist muscles (Hortobagyi et al., 2009; Tirosh and Sparrow, 2005). Thus, there is much evidence that vastii activations change with both age and gender, so that availability of electromyographic reference frame, stratified for age and gender, is desirable.
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807
Surface electromyography (sEMG) is regularly used to assess the activation patterns of knee extensor muscles during normal and pathological gait (Agostini et al., 2010; Karst and Willett, 1995). Vastii activity begins in terminal swing (around 90% of GC). Early in loading response of the following stride (5% of GC), vastii intensity rapidly increases to a peak of 25% of maximum manual muscle test value, a level of effort that is maintained throughout the remainder of the loading response period. With the beginning of mid-stance, the vastii rapidly reduce their effort, which ceases by the 15% GC point (Perry, 1992). However, activities outside typical activation intervals for VM and VL and, more generally, a large stride-to-stride variability in EMG profiles were also reported (Agostini et al., 2010; Winter and Yack, 1987). Thus, besides the effect of age and gender, a reliable reference frame of vastii recruitment in healthy adults should also consider the natural muscle-activation variability during free walking. The availability of such a tool, simultaneously including natural variability of the phenomenon and controlling for the confounding effect of age and gender, would indeed be useful for discriminating physiological and pathological conditions. A possible way to create a frame of reference is to use statistical gait analysis (SGA) (Agostini and Knaflitz, 2012) to analyze electromyographic recordings containing several steps from several patients. SGA is a recently developed methodology, which performs a statistical characterization of gait by averaging spatial–temporal and sEMG-based parameters over hundreds of strides, during the same episode of walking. Thus, the aim of the study was to use SGA for the quantitative assessment of natural variability of VL and VM activation modalities. To this aim, electromyographic recordings of thirty healthy, young, females during self-selected walking were analyzed. 2. Materials and methods 2.1. Subjects Thirty healthy young female adult volunteers were recruited (age 25.1 ± 2.1 years; height 167.9 ± 0.1 cm; weight 57.7 ± 5.0 kg; body mass index (BMI) 19.8 ± 4.1 kg m 2). Exclusion criteria included history of neurological disorders, orthopedic surgery, acute/chronic knee pain or pathology, BMI > 25, or abnormal gait. None of the recruited subjects were involved in competitive sports activities. All participants signed informed consent. 2.2. Signal acquisition Signals were acquired (sampling rate: 2000 Hz; resolution: 12 bit) and processed by the multichannel recording system Step32 (Version PCI-32 ch2.0.1. DV), DemItalia, Italy. Each subject was instrumented with foot-switches, knee electro-goniometers and sEMG probes on left lower limb. Three foot-switches (DemItalia, Italy; size: 11 11 0.5 mm; activation force: 3 N) were attached beneath the heel, the first, and the fifth metatarsal heads. An electro-goniometer (Step32, DemItalia, Italy; accuracy: 0.5°) was attached to the lateral side of the lower limb for measuring knee joint angles in sagittal plane. sEMG signals were detected with single-differential sEMG probes with fixed geometry constituted by Ag/Ag–Cl disks (manufacturer: DemItalia, size: 7 27 19 mm; interelectrode distance: 12 mm, gain: 1000, high-pass filter: 10 Hz, input impedance >1.5 GX, CMRR > 126 dB, input referred noise 61 lVrms). sEMG signals were further amplified and low-pass filtered (450 Hz) by the recording system. The skin was shaved, cleansed with abrasive paste and moistened. To assure proper electrode–skin contact, each electrode
801
was dressed with highly-conductive gel. sEMG probes were applied over VM and VL, following the SENIAM recommendations for electrode location and orientation over muscle with respect to tendons, motor point position, and fiber direction (Freriks et al., 2000). Then, after being accurately instructed (Di Nardo and Fioretti, 2013), subjects were asked to walk barefoot over the floor for 6 min at their natural speed and cadence, following the path described in Fig. 1. A time length of 6 min was chosen in order to have an appropriate number of consecutive strides, since Agostini and Knaflitz (2012) indicate that the characteristic gait cannot be quantified in a reliable way by SGA unless at least 100–200 gait cycles are analyzed. This path was chosen to allow the subjects to walk uninterruptedly, without perturbing their natural pace, in order to improve the repeatability of sEMG data (Kadaba et al., 1989). 2.3. Signal processing Footswitch signals were debounced and converted to four levels, Heel contact (H), Foot Flat contact (F), Push off (P), and Swing (S), then processed to segment and classify the different GCs (Agostini et al., 2014). Electro-goniometric signals were low-pass filtered (FIR filter, 100 taps, cut-off frequency 15 Hz). Knee angles in the sagittal plane, along with sequences and durations of gait phases derived by the basographic signal, were used by a multivariate statistical filter to detect outlier cycles like those relative to deceleration, reversing, and acceleration. Cycles with improper sequences of gait phases (i.e. different from H–F–P–S sequence) not corresponding to straight walking, and with abnormal timing and knee angles, with respect to a mean value computed on each single subject, were discarded (Agostini and Knaflitz, 2012). sEMG signals were high-pass filtered (FIR filter, 100 taps, cut-off frequency of 20 Hz) and processed by a double-threshold statistical detector that allows a user-independent assessment of the muscle activation intervals (Bonato et al., 1998). This technique (Bonato et al., 1998) consists of selecting a first threshold f and observing m successive samples. If at least r0 (second threshold) out of successive m samples are above f, the presence of the signal is acknowledged. The time instant when the presence of the signal is acknowledged is defined as the onset of muscle activation. The time instant when the presence of the signal is no longer acknowledged is defined as the offset of muscle activation. Values of the three parameters f, r0, and m are selected to jointly minimize the false-alarm probability value and maximize the detection probability for each specific signal-to-noise ratio. The setting of f is based on the assessment of background noise level, as a necessary input parameter. Furthermore, the double-threshold detector requires estimating the signal-to-noise ratio in order to fine tune r0. Background noise level and signal-to-noise ratio, necessary to run double-threshold algorithm, are estimated for each signal by
Fig. 1. Schematic representation of the path walked by the recruited subjects during the experiment. From a total of 9787 strides acquired, 38.1% were discarded, due to not following the H–F–P–S foot-switch pattern and/or being outlier cycles relative to deceleration, reversing, and acceleration.
802
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807
Step32 system, using the statistical approach proposed by Agostini and Knaflitz (2012). Eventually, m = 30 ms is considered a suitable value for evaluating muscle activation in gait analysis (Bonato et al., 1998). 2.4. Statistical gait analysis (SGA) Statistical gait analysis (Agostini and Knaflitz, 2012) is a recently developed methodology defined as the statistical characterization of gait, arrived at by averaging spatial–temporal and sEMG-based parameters. The aim of SGA is the acquisition and statistical analysis of sEMG signals over hundreds of strides during the same walking trial, without the limitation of three-four strides of classic sEMG acquisitions in a gait analysis laboratory (Agostini et al., 2010; Agostini and Knaflitz, 2012; Benedetti et al., 2012). SGA relies on the fact that the number of muscle activations is cycle dependent, so that averaging should be performed only over onset/offset instants of cycles including the same number of activations, i.e. belonging to the same activation modality. Activation modality is defined as the number of times a muscle activates during a single GC; each modality refers to a different activation pattern, characterized by a specific number of activations over GC, i.e. n-activation modality consists of n activation intervals for the considered muscle during a single GC. Mean activation intervals (normalized with respect to GC duration) for each modality of activation were achieved, according to the following steps. First, muscle activations relative to each GC were identified, computing muscle onset/offset instants in temporal space (Bonato et al., 1998), as previously described. Then, muscle activations were grouped according to their activation modalities. Eventually, the onset/offset time instants of each activation modality were averaged over the thirty population subjects. Averaged onset/offset percentage time instants were normalized with respect to GC to provide mean activation intervals in percentage of GC. SGA was performed by Step32 system. 2.5. Statistics The following variables were considered in the statistical analysis, for each muscle and for each activation modality: onset and offset instants of activation, length of activation intervals, and occurrence frequency. Data are reported as mean ± standard deviation (SD). The Lilliefors test was used to evaluate the hypothesis that each data vector had a normal distribution. Comparisons among normally distributed samples were performed with two-tailed, non-paired Student’s t-test (two groups) or one-way analysis of variance (ANOVA) (more than two groups) followed by multiple comparison test. The Kruskal–Wallis test, followed by a multiple comparison test, were used to compare samples not normally distributed. Statistical significance was set at p < 0.05.
3. Results Overall, 3729 strides out of 9787 were discarded for not following the H–F–P–S foot-switch pattern, and/or for being outlier cycles relative to deceleration, reversing, and acceleration. H-phase lasted 5.7 ± 1.6% (percent of GC), F-phase 29.0 ± 6.5%, P-phase 23.7 ± 6.1% and S-phase 41.5 ± 6.2%. The SGA of myoelectric signals indicated that muscles show different numbers of activation intervals in different strides of the same walking. An example of sEMG signals from the VL muscle of a subject is depicted in Fig. 2. Details of activation intervals of VM and VL, expressed as the timing, in the percentage of GC, of signal onset and offset are reported in Figs. 3 and 4 (panel B).
Fig. 2. An example of raw sEMG signals of VL muscle showing 2 activations (panel A), 3 activations (panel B) and 4 activations (panel C), respectively. The signals were extracted from different strides of the same subject, during the same walk. H, F, P S phases are delimited by dashed light-gray vertical lines.
The most recurrent (p = 8.2 10 12) activation modality for VM (Fig. 3, panel A) consists of two activations (2-activation modality), observed in 53.4 ± 23.2% of strides: the first occurs at the beginning of GC and the second in terminal swing. The second most recurrent activation modality is similar to the 2-activation modality but with a further activation in the beginning of the push-off phase. This 3-activation modality was observed in 34.1 ± 14.3% of strides. In a further 10.3 ± 9.3% of strides, four activations were observed: in the beginning of GC, in the Foot Flat contact phase, in the push-off phase and in terminal swing (4-activation modality). The remaining 1.8 ± 2.0% of strides were characterized by five activations during the GC (5-activation modality). This modality showed a rare occurrence and a large variability; thus, it is not considered in the present analysis. Participant-specific data are reported in Table 1 for VM and in Table 2 for VL. For VL, the most recurrent (p = 8.3 10 11) activation modality (Fig. 4, panel A) consists of two activations (2-activation modality) observed in 59.4 ± 27.8% of strides: the first occurs at the beginning of GC, the second in terminal swing. The second most recurrent activation modality is similar to the 2-activation modality but with a further activation in the beginning of the push-off phase. This 3-activation modality was observed in 28.9 ± 16.6% of strides. In a further 9.5 ± 10.6% of strides, four activations were observed: in the beginning of GC, in the Foot Flat contact phase, in the
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807
Fig. 3. Vastus medialis: mean (+SD) percentage occurrence frequency of each of the five different activation modalities (panel A) and mean (+SD) activation intervals vs. percentage of gait cycle for the modalities with 2, 3 and 4 activations, respectively (panel B), detected during walking. Mean and SD refer to the present female population. H, F, P S phases are delimited by dashed light-gray vertical lines. ⁄ Significantly different, compared to 1-, 3-, 4- and 5-activation modality.
803
2008; Di Nardo et al., 2015b; Kerrigan et al., 1998; Malinzak et al., 2001), indicating the suitability of stratifying biomechanical and electromyographic reference frames for males and females. Moreover, it was observed that advancing age modifies the timing of onset/offset activation of lower limb muscles, including the vastii (Hortobagyi et al., 2009). Thus, to control for age and gender, only young (20 years < age < 30 years) females were recruited. Further recruitment exclusion criteria (see Section 2) were used to control for possible additional confounding effects. The application of all the above-mentioned criteria limited the number of participants, but also reduced the variability, and therefore the number of subjects necessary, for the study. The observed vastii activation intervals roughly followed the typical pattern reported for each muscle during walking (Perry, 1992; Sutherland, 2001). The present analysis put into evidence that the considered muscles showed different modalities in the number of activations and in the timing of signal onset/offset (Figs. 3 and 4). In previous studies (Di Nardo et al., 2013, 2015a), we observed the same behavior in tibialis anterior and gastrocnemius lateralis. This suggests that, besides the muscular activation instants, it is worth considering how frequently are observed, i.e. their occurrence frequency. Reporting such a parameter is quite innovative, because it is seldom considered due to the low number of strides usually analyzed in classic EMG studies. Statistically, the most recurrent activation modality during GC consists of two activations for both VM (observed in 53.4 ± 23.2% of total strides, Fig. 3) and VL (observed in 59.4 ± 27.8% of total strides, Fig. 4), at the beginning of GC and in terminal swing, respectively. The 3-activation modality (34.1 ± 14.3% for VM and 28.9 ± 16.6% for VL) differs from the most common modality, because of a further
push-off phase, and in terminal swing (4-activation modality). The remaining 1.5 ± 2.3% of strides were characterized by five activations during the GC (5-activation modality); as with VM, it is not considered in the present analysis. No significant differences (p > 0.05) were observed in onset/offset instants of activation, or in the length of activation intervals between VM and VL, for all the activation modalities. This study also compared VM and VL activity in terms of the frequency each modality of muscle activation occurs, quantified by the percentage number of strides over the total number of strides of the whole subject population, where the muscle is recruited with the specific activation modality: no significant differences (p > 0.05) were detected between VM and VL. Fig. 5 shows a pictorial representation of VM and VL activation intervals in a function of the number of subjects where muscular activity is observed; this representation was achieved for each muscle, considering the three main activation modalities together.
4. Discussion In this study, sEMG signals were recorded in thirty healthy young females, in order to assess the natural variability of the activation modalities of VM and VL during walking at a self-selected pace. SGA was used to analyze hundreds of strides for each subject. The SENIAM recommendations for electrode location and orientation over muscle with respect to tendons, motor point location, and fiber direction (Freriks et al., 2000) were strictly followed in order to soften the possible impact of differences in muscle fiber amount and orientation between VM and VL. Recent studies suggested gender-related differences in knee kinematic and sEMG activity of the vastii (Chumanov et al.,
Fig. 4. Vastus lateralis: mean (+SD) percentage occurrence frequency of each of the five different activation modalities (panel A) and mean (+SD) activation intervals vs. percentage of gait cycle for the modalities with 2, 3 and 4 activations, respectively (panel B), detected during walk. Mean and SD refer to the present female population. H, F, P S phases are delimited by dashed light-gray vertical lines. ⁄ Significantly different, compared to 1-, 3-, 4- and 5-activation modality.
804
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807
Table 1 Occurrence frequencies of all five VM activation modalities for each subject. Subject
1-activation modality (%)
2-activation modality (%)
3-activation modality (%)
4-activation modality (%)
5-activation modality (%)
#1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15 #16 #17 #18 #19 #20 #21 #22 #23 #24 #25 #26 #27 #28 #29 #30
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 3.1 0.0 0.5 0.0 0.0 0.9 0.0 0.6 0.0 0.2 0.0 1.1 0.0 0.0 0.0 0.0 1.0 0.0 0.0
63.0 47.8 61.1 59.9 54.7 54.8 79.9 79.5 92.1 88.2 28.4 16.9 56.2 20.1 42.2 51.1 68.7 63.2 20.9 46.5 89.2 9.5 33.0 75.6 20.0 68.0 23.5 58.8 65.5 63.3
34.4 41.4 36.3 37.0 39.2 31.9 17.7 17.0 7.1 11.4 37.6 48.4 37.0 51.0 42.2 33.7 22.5 27.7 42.4 45.2 10.1 72.0 42.1 20.1 49.2 26.0 53.6 37.8 23.7 26.1
2.6 9.7 2.6 3.1 4.9 13.3 2.4 3.2 0.4 0.4 33.0 23.6 6.4 23.0 12.2 10.9 5.7 6.8 32.0 7.7 0.2 14.9 18.2 4.3 23.1 5.9 20.3 1.5 7.9 8.3
0.0 0.8 0.0 0.0 1.2 0.0 0.0 0.3 0.0 0.0 0.9 6.2 0.5 5.4 3.3 3.3 2.2 1.4 3.5 0.7 0 2.9 5.7 0.0 6.1 0.1 2.6 1.0 2.8 2.2
Values are reported as percentage of total strides. The occurrence frequency for the most recurrent activation modality is reported in bold.
Table 2 Occurrence frequencies of all five VL activation modalities for each subject. Subject
1-activation modality (%)
2-activation modality (%)
3-activation modality (%)
4-activation modality (%)
5-activation modality (%)
#1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15 #16 #17 #18 #19 #20 #21 #22 #23 #24 #25 #26 #27 #28 #29 #30
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 6.3 0.0 0.0 0.0 0.0 3.5 0.0 0.0 0.0 0.0 1.8 0.0 1.8 0.8 0.0 0.0 1.5 0.0 0.0
55.8 78.5 47.4 74.1 20.3 5.3 61.5 73.1 81.9 93.1 18.4 30.6 75.8 77.9 20.6 65.2 86.3 55.9 14.5 81.3 83.1 23.8 60.2 90.2 19.2 88.9 66.0 94.6 87.6 49.4
35.7 17.6 44.1 24.8 45.1 68.5 30.1 24.0 15.0 6.6 47.7 36.0 21.5 21.1 54.4 29.4 9.3 30.0 51.2 18.1 16.2 51.8 22.7 5.5 43.1 11.1 31.4 3.9 11.3 40.6
7.5 3.9 7.8 1.0 29.7 22.8 7.6 2.3 2.6 0.4 28.4 19.8 2.3 1.0 23.3 3.3 0.9 11.8 33.1 0.7 0.7 19.0 14.8 2.4 26.9 0.0 2.6 0.0 0.6 9.4
1.0 0.0 0.7 0.0 4.9 2.9 0.8 0.6 0.4 0.0 5.5 6.3 0.5 0.0 1.11 2.2 0.0 2.3 1.2 0.0 0.0 2.9 2.3 0.0 9.2 0.0 0.0 0.0 0.6 0.6
Values are reported as percentage of total strides. The occurrence frequency for the most recurrent activation modality is reported in bold.
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807
Fig. 5. VM and VL activations over the population, as percentage of gait cycle, considering, for each muscle, the three main modalities of activation all together. Horizontal bars are gray-level coded, according to the number of subjects where a certain condition is observed; black: condition observed for all subjects in every activation modality, white: condition never met. H, F, P S phases are delimited by dashed light-gray vertical lines.
activity around heel off. In the 4-activation modality (10.3 ± 9.3% for VM and 9.5 ± 10.6% for VL), this activation splits into two distinct activations. Considering the gait periodicity, it is clear that, in all activation modalities, the activation at the beginning of the GC is just a continuation of the activation at the end of the previous GC (Figs. 3 and 4, panel B). Still, to conform to the literature (Perry, 1992), the activation modalities were represented as functions of the typical GC, which conventionally starts with the initial heel contact and ends at the successive heel contact. Consequently, the single activity of the vastii during the beginning and the end of the GC was split in two separate activations. Results revealed variability in the number of activations, occurrence frequency, and onset–offset instants across the thousands of strides analyzed (Figs. 3 and 4), despite the narrowly focused population (young, healthy, non-obese females) recruited for the study. The apparently large amount of variability in the occurrence frequency across participants (Figs. 3 and 4, Panel A) is due to the fact that not every participant presents the same preferred (i.e. most recurrent) activation modality. Indeed, as reported in Tables 1 and 2, around 70% of the participants prefer walking with 2-activation modality and around 30% of the participants prefer walking with 3-activation modality. Thus, two sub-groups of people are observed; those who favor a 2-activation modality and those who favor a 3-activation modality. This difference is due to the presence/absence of muscle activity during mid/terminal stance. No further differences in the gait patterns of the people that appear to belong to each of these groups were detected. Variability was also detected in onset–offset instants (Figs. 3 and 4, Panel B), in particular for the activity around heel off. Despite the variability, there was one activation occurrence which remained consistent across subjects for both VM and VL. This activation occurred from terminal swing to the following loading response. It is recognized as the typical activation for both VM and VL during normal walking, and is required to generate tension in terminal swing in preparation for weight bearing at initial contact, and to control knee flexion during weight acceptance (Perry, 1992; Sutherland, 2001). This activation is observed in the totality of the strides for both muscles. A second, less frequent, activation occurred between mid-stance and pre-swing (observed in 44.4 ± 17.1% of strides for VM and in 38.4 ± 19.7% for VL), in monophasic (3-activation modalities, Figs. 3 and 4, panel B) or biphasic (4-activation modalities) shape, and presented activations significantly shorter than those detected between terminal swing and subsequent loading response. This activation is characterized by a large variability, making it difficult, at this stage, to indicate a single and precise explanation for the variability. Agostini et al. (2010) suggested that
805
the recruitment in this phase may be related to the activity of the muscle to modulate rapid knee flexion in the transition between mid and terminal stance, and to stabilize the patella before entering the pre-swing phase. However, since knee flexion modulation and patellar stabilization occur on every gait cycle, one could wonder why this pattern is not observed more consistently. A further possible interpretation could be the mutual crosstalk between vastii and rectus femoris. Although we are not able to completely exclude this hypothesis, the presence of this vastii activity, in each subject and in a limited number of strides (44.4 ± 17.1%), suggests that it is related to the need to fulfill a specific, less frequent task, rather than to crosstalk. Possible future studies aimed at investigating cross-talk could be performed by triggering sEMGs in the vastii muscles with the firing patterns of motor units identified, with intramuscular electrodes, in the rectus femoris (Merletti and Farina, 2009). Results also indicate that the meaning of the mid-stance activation is still not completely clear. Despite this, the present study has the merit to point out and quantify, for the first time to our knowledge, this vastii activity in healthy females. Moreover, the study evokes further analysis on this sEMG activity, which seems critical to understand the possible cause for the different activation patterns detected here. No significant differences were observed in onset/offset instants of activation, in the activation length, and in the occurrence frequency, suggesting a coincident VM and VL recruitment in the present female population (Fig. 5). This finding is consistent with previous results (Cowan et al., 2001; Karst and Willett, 1995; Powers et al., 1996) and concurs with the hypothesis that, in healthy people, the onset and cessation of activity of the vastii muscles are balanced. Indeed, although the VM and VL may have antagonistic actions for the mediolateral control of the patella, ultimately, the recruitment of VM and VL should be appropriately timed for efficient biomechanical function of the knee (Grelsamer and McConnell, 1988). To our knowledge, the present analysis represents the first attempt to quantify the VM and VL dynamic sEMG activity during walking in young, healthy females. Such quantification was performed in terms of variability of onset–offset muscular activation and occurrence frequency, and provides a first contribution to the process of building a ‘‘normality’’ reference frame for females. By including the physiological variability of the phenomenon and controlling the confounding effects due to age and gender, the reference frame proposed here, even in its early stages, could already support discrimination of pathological from physiological behaviors. Moreover, due to its selectivity, the model could be suitable for designing more focused gait studies on the effect of age and gender on the variability of physiological activity of the vastii. 4.1. Limitations of the study A first limitation of the study lies in the limited number of enrolled females. Thus, further studies on larger populations, possibly including male and/or older populations, are needed to confirm or contrast the present findings. Another limitation consists of the lack of information about the Q angle, which was not measurable in the present study. The Q angle represents an estimate of the resultant force of the quadriceps on the patella and is a predictor of the patella lateral movement under dynamic conditions. Convincing evidence indicates that young women have greater mean Q angles than their male counterparts, with differences in magnitude ranging from 2.7 to 5.8° (Livingston, 1998), and that these Q-angle differences could be associated with gender differences in VL and VM EMG parameters (Malinzak et al., 2001; Sigward and Powers, 2006). These findings support not only the suitability of developing a specific
806
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807
reference frame for females as indicated by the present study, but also the need for future studies to complete such a reference frame with Q-angle measurements.
5. Conclusions Despite the large variability detected in VM and VL activity, one activation occurrence remains consistent across subjects for both muscles. This occurs from terminal swing to the following loading response and is observed in 100% of strides). A second, less frequent, activation occurs between mid-stance up to pre-swing. This pattern of activation represents the first contribution toward developing a reference frame able to describe the physiological variability of vastii sEMG activity during walking.
Conflict of interest
Kerrigan DC, Todd MK, Della Croce U. Gender differences in joint biomechanics during walking: normative study in young adults. Am J Phys Med Rehabil 1998;77:2–7. Livingston LA. The quadriceps angle: a review of the literature. J Orthop Sports Phys Ther 1998;28(2):105–9. Malinzak RA, Colby SM, Kirkendall DT, et al. A comparison of knee joint motion patterns between men and women in selected athletic tasks. Clin Biomech 2001;16(5):438–45. Merletti R, Farina D. Analysis of intramuscular electromyogram signals. Philos Transact A Math Phys Eng Sci 2009;367:357–68. Perry J. Gait analysis – normal and pathological function. USA: Slack Inc.; 1992. 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. Sigward SM, Powers CM. The influence of gender on knee kinematics, kinetics and muscle activation patterns during side-step cutting. Clin Biomech (Bristol, Avon) 2006;21:41–8. Sutherland DH. The evolution of clinical gait analysis part l: kinesiological EMG. Gait Posture 2001;14:61–70. Tirosh O, Sparrow WA. Age and walking speed effects on muscle recruitment in gait termination. Gait Posture 2005;21:279–88. Winter DA, Yack HJ. EMG profiles during normal human walking: stride to stride and inter-subject variability. Electroencephalogr Clin Neurophysiol 1987;67:402–11.
The authors declare that there are no conflicts of interest.
References Agostini V, Balestra G, Knaflitz M. Segmentation and classification of gait cycles. IEEE Trans Neural Syst Rehabil Eng 2014;22(5):946–52. 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, editors. Stevenson Ranch, California, (USA): American Scientific Publishers, vol. 2; 2012. p. 99–121. Agostini V, Nascimbeni A, Gaffuri A, et al. Normative EMG activation patterns of school-age children during gait. Gait Posture 2010;32(3):285–9. Benedetti MG, Agostini V, Knaflitz M, et al. Self-reported gait unsteadiness in mildly impaired neurological patients: an objective assessment through statistical gait analysis. J Neuroeng Rehabil 2012;29(9):64. 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. Chumanov ES, Wall-Scheffler C, Heiderscheit BC. Gender differences in walking and running on level and inclined surfaces. Clin Biomech (Bristol, Avon) 2008;23:1260–8. Cowan SM, Bennell KL, Hodges PW, et al. Delayed onset of electromyographic activity of vastus medialis obliquus relative to vastus lateralis in subjects with patellofemoral pain syndrome. Arch Phys Med Rehabil 2001;82(2):183–9. 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. Di Nardo F, Ghetti G, Fioretti S. Assessment of the activation modalities of gastrocnemius lateralis and tibialis anterior during gait: a statistical analysis. J Electromyogr Kinesiol 2013;23(6):1428–33. Di Nardo F, Mengarelli A, Maranesi E, et al. Influence of gender on the myoelectric signal of thigh muscles. In: MESA 2014 – 10th IEEE/ASME international conference on mechatronic and embedded systems and applications conference proceedings, Senigallia; 2014. http://dx.doi.org/10.1109/MESA. 2014.6935538. Di Nardo F, Mengarelli A, Maranesi E, Burattini L, Fioretti S. Assessment of the ankle muscle co-contraction during normal gait: a surface electromyography study. J Electromyogr Kinesiol 2015a;25(2):347–54. Di Nardo F, Mengarelli A, Maranesi E, Burattini L, Fioretti S. Gender differences in the myoelectric activity of lower limb muscles in young healthy subjects during walking. Biomed Signal Process Control 2015b;18:14–22. Freriks B, Hermens HJ, Disselhorst-Klug C, Rau G. Development of recommendations for SEMG sensors and sensor placement procedures. J Electromyogr Kinesiol 2000;10(5):361–74. Grelsamer RP, McConnell J. The patella: a team approach. Gaithersburg (MD): Aspen; 1988. Hortobagyi T, Solnik S, Gruber A, et al. Interaction between age and gait velocity in the amplitude and timing of antagonist muscle coactivation. Gait Posture 2009;29:558–64. Kadaba MP, Ramakrishnan HK, Wootten ME, et al. Repeatability of kinematic, kinetic, and electromyographic data in normal adult gait. J Orthop Res 1989;6:849–60. Karst GM, Willett GM. Onset timing of electromyographic activity in the vastus medialis oblique and vastus lateralis muscles in subjects with and without patellofemoral pain syndrome. Phys Ther 1995;75:813–23.
Francesco Di Nardo currently teaches the classes of Models and Control of Biological Systems in Università Politecnica delle Marche, Ancona, Italy, where he received his Ph.D., in 2005. His main research activity involved the development and the clinical application (type-2 diabetes, insulin resistance and hypertension) of mathematical models of metabolic and endocrine systems. 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. He published several papers in peer review international journals, books and conference proceedings.
Elvira Maranesi received her Ph.D. in Biomedical Engineering from Università Politecnica delle Marche, Ancona, Italy in 2013. Currently, she is Postdoctoral Researcher with the Department of Information Engineering, Università Politecnica delle Marche and collaborates with the Posture and Movement Analysis Laboratory at the Italian National Institute of Health and Science on Aging. Furthermore she is a charter member of B.M.E.D. Srl, an academic spin-off of Università Politecnica delle Marche. Her main research interests include the electromyographic signal processing and the movement analysis during static and dynamic postural tests to evaluate the risk of falls in the elderly. In this field, she is author of scientific publications in international journals and congress proceedings.
Alessandro Mengarelli received his Master Degree in Biomedical Engineering in 2012 from Università Politecnica delle Marche, Ancona, Italy, where he is actually attending Ph.D. course in Information Engineering. His research field involves dynamic posturography and gait analysis, focused mainly on the assessing the muscles’ behavior and functions by means of surface electromyography signal. In this field he is the co-author of some national and international conference proceedings.
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807 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.
Laura Burattini is currently Assistant Professor of Biomedical Engineering at the Polytechnic University of Marche, where she teaches the classes of Biomedical Engineering and Biomedical Signal and Data Processing, and President and CEO of B.M.E.D. SRL, an academic spinoff she founded herself. She graduated in Electrical/Biomedical Engineering at the Politecnico di Milano, Italy, in 1993 and took the PhD at the University of Rochester, USA, in 1998. Then, returned in Italy, she matured extensive working experience in both academic and corporate. Her main research interest includes the biomedical signal processing. She is author of numerous scientific publications in international journals, books and congress proceedings.
807
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 ScienceDirect
Journal of Electromyography and Kinesiology journal homepage: www.elsevier.com/locate/jelekin
Assessment of the variability of vastii myoelectric activity in young healthy females during walking: A statistical gait analysis Francesco Di Nardo a,⇑, Elvira Maranesi a,b, Alessandro Mengarelli a, Giacomo Ghetti b, Laura Burattini a, 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 10 November 2014 Received in revised form 17 June 2015 Accepted 3 July 2015
Keywords: Surface electromyography Knee-flexor muscles Statistical gait analysis Variability Quadriceps femoris
a b s t r a c t The study was designed to assess the natural variability of the activation modalities of vastus medialis (VM) and vastus lateralis (VL) during walking at a self-selected speed and cadence of 30 young, healthy, females. This was achieved by conducting statistical gait analysis on the surface electromyographic signals from hundreds of strides for each subject. Results revealed variability in the number of activations, occurrence frequency, and onset-offset instants across the thousands of strides analyzed. However, despite the variability, there was one activation occurrence which remained consistent across subjects for both VM and VL. This occurred from terminal swing to the following loading response (observed in 100% of strides). A second, less frequent, activation occurred between mid-stance up to pre-swing (observed in 39.3 ± 22.4% of strides for VM and in 35.1 ± 20.6% for VL). No significant differences (p > 0.05) were observed in the onset–offset instants or in the occurrence frequency, which suggest a simultaneous recruitment of VM and VL. This ‘‘normality’’ pattern represents the first attempt at developing a reference frame for vastii sEMG activity during walking, that is able to include the physiological variability of the phenomenon and control the confounding effects of age and gender. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction The knee is the key to stance stability, and the muscles of the quadriceps femoris group are the most direct source of extensor control (Perry, 1992). During walking, however, the knee extensor muscles are mainly used to restrain the shock-absorbing flexion during loading response; the quadriceps femoris eccentric contraction in the touching phase of the gait cycle (GC), indeed, represents the primary absorbing mechanism of impact during weight acceptance. In stance, the extensors act to decelerate knee flexion, while in swing they contribute to limb progression (Perry, 1992). Moreover, vastus medialis (VM) and vastus lateralis (VL) play the important role of stabilizing the patella and the knee joint during walking (Grelsamer and McConnell, 1988). Thus, the vastii muscles play a fundamental role in normal walking physiology and in the etiology of common knee pathologies. A VM–VL activity imbalance, observed in thirty-three patients (vs. thirty-three controls) with ⇑ Corresponding author at: 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). http://dx.doi.org/10.1016/j.jelekin.2015.07.004 1050-6411/Ó 2015 Elsevier Ltd. All rights reserved.
patellofemoral pain syndrome, was reported as a possible mechanism for abnormal patellar tracking (Cowan et al., 2001). Gender-related differences in both knee kinematic and vastii myoelectric activity were also reported (Chumanov et al., 2008; Di Nardo et al., 2014; Kerrigan et al., 1998; Malinzak et al., 2001). In a recent electromyographic study of ours, involving twenty-two healthy adults (Di Nardo et al., 2015b), females showed a higher VL-recruitment frequency during walking than males. In a study on seventeen young females and seventeen age-matched males (Chumanov et al., 2008), the former group showed higher VL myoelectric activity during the terminal swing and initial running loading transition than the latter. Eventually, in a study on nine females and eleven males (Malinzak et al., 2001), females showed a higher quadriceps activation during different motor tasks than males. It was also observed that advancing age modifies both the activation timing of lower limb muscles (including the vastii), and the activity duration of agonist and antagonist muscles (Hortobagyi et al., 2009; Tirosh and Sparrow, 2005). Thus, there is much evidence that vastii activations change with both age and gender, so that availability of electromyographic reference frame, stratified for age and gender, is desirable.
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807
Surface electromyography (sEMG) is regularly used to assess the activation patterns of knee extensor muscles during normal and pathological gait (Agostini et al., 2010; Karst and Willett, 1995). Vastii activity begins in terminal swing (around 90% of GC). Early in loading response of the following stride (5% of GC), vastii intensity rapidly increases to a peak of 25% of maximum manual muscle test value, a level of effort that is maintained throughout the remainder of the loading response period. With the beginning of mid-stance, the vastii rapidly reduce their effort, which ceases by the 15% GC point (Perry, 1992). However, activities outside typical activation intervals for VM and VL and, more generally, a large stride-to-stride variability in EMG profiles were also reported (Agostini et al., 2010; Winter and Yack, 1987). Thus, besides the effect of age and gender, a reliable reference frame of vastii recruitment in healthy adults should also consider the natural muscle-activation variability during free walking. The availability of such a tool, simultaneously including natural variability of the phenomenon and controlling for the confounding effect of age and gender, would indeed be useful for discriminating physiological and pathological conditions. A possible way to create a frame of reference is to use statistical gait analysis (SGA) (Agostini and Knaflitz, 2012) to analyze electromyographic recordings containing several steps from several patients. SGA is a recently developed methodology, which performs a statistical characterization of gait by averaging spatial–temporal and sEMG-based parameters over hundreds of strides, during the same episode of walking. Thus, the aim of the study was to use SGA for the quantitative assessment of natural variability of VL and VM activation modalities. To this aim, electromyographic recordings of thirty healthy, young, females during self-selected walking were analyzed. 2. Materials and methods 2.1. Subjects Thirty healthy young female adult volunteers were recruited (age 25.1 ± 2.1 years; height 167.9 ± 0.1 cm; weight 57.7 ± 5.0 kg; body mass index (BMI) 19.8 ± 4.1 kg m 2). Exclusion criteria included history of neurological disorders, orthopedic surgery, acute/chronic knee pain or pathology, BMI > 25, or abnormal gait. None of the recruited subjects were involved in competitive sports activities. All participants signed informed consent. 2.2. Signal acquisition Signals were acquired (sampling rate: 2000 Hz; resolution: 12 bit) and processed by the multichannel recording system Step32 (Version PCI-32 ch2.0.1. DV), DemItalia, Italy. Each subject was instrumented with foot-switches, knee electro-goniometers and sEMG probes on left lower limb. Three foot-switches (DemItalia, Italy; size: 11 11 0.5 mm; activation force: 3 N) were attached beneath the heel, the first, and the fifth metatarsal heads. An electro-goniometer (Step32, DemItalia, Italy; accuracy: 0.5°) was attached to the lateral side of the lower limb for measuring knee joint angles in sagittal plane. sEMG signals were detected with single-differential sEMG probes with fixed geometry constituted by Ag/Ag–Cl disks (manufacturer: DemItalia, size: 7 27 19 mm; interelectrode distance: 12 mm, gain: 1000, high-pass filter: 10 Hz, input impedance >1.5 GX, CMRR > 126 dB, input referred noise 61 lVrms). sEMG signals were further amplified and low-pass filtered (450 Hz) by the recording system. The skin was shaved, cleansed with abrasive paste and moistened. To assure proper electrode–skin contact, each electrode
801
was dressed with highly-conductive gel. sEMG probes were applied over VM and VL, following the SENIAM recommendations for electrode location and orientation over muscle with respect to tendons, motor point position, and fiber direction (Freriks et al., 2000). Then, after being accurately instructed (Di Nardo and Fioretti, 2013), subjects were asked to walk barefoot over the floor for 6 min at their natural speed and cadence, following the path described in Fig. 1. A time length of 6 min was chosen in order to have an appropriate number of consecutive strides, since Agostini and Knaflitz (2012) indicate that the characteristic gait cannot be quantified in a reliable way by SGA unless at least 100–200 gait cycles are analyzed. This path was chosen to allow the subjects to walk uninterruptedly, without perturbing their natural pace, in order to improve the repeatability of sEMG data (Kadaba et al., 1989). 2.3. Signal processing Footswitch signals were debounced and converted to four levels, Heel contact (H), Foot Flat contact (F), Push off (P), and Swing (S), then processed to segment and classify the different GCs (Agostini et al., 2014). Electro-goniometric signals were low-pass filtered (FIR filter, 100 taps, cut-off frequency 15 Hz). Knee angles in the sagittal plane, along with sequences and durations of gait phases derived by the basographic signal, were used by a multivariate statistical filter to detect outlier cycles like those relative to deceleration, reversing, and acceleration. Cycles with improper sequences of gait phases (i.e. different from H–F–P–S sequence) not corresponding to straight walking, and with abnormal timing and knee angles, with respect to a mean value computed on each single subject, were discarded (Agostini and Knaflitz, 2012). sEMG signals were high-pass filtered (FIR filter, 100 taps, cut-off frequency of 20 Hz) and processed by a double-threshold statistical detector that allows a user-independent assessment of the muscle activation intervals (Bonato et al., 1998). This technique (Bonato et al., 1998) consists of selecting a first threshold f and observing m successive samples. If at least r0 (second threshold) out of successive m samples are above f, the presence of the signal is acknowledged. The time instant when the presence of the signal is acknowledged is defined as the onset of muscle activation. The time instant when the presence of the signal is no longer acknowledged is defined as the offset of muscle activation. Values of the three parameters f, r0, and m are selected to jointly minimize the false-alarm probability value and maximize the detection probability for each specific signal-to-noise ratio. The setting of f is based on the assessment of background noise level, as a necessary input parameter. Furthermore, the double-threshold detector requires estimating the signal-to-noise ratio in order to fine tune r0. Background noise level and signal-to-noise ratio, necessary to run double-threshold algorithm, are estimated for each signal by
Fig. 1. Schematic representation of the path walked by the recruited subjects during the experiment. From a total of 9787 strides acquired, 38.1% were discarded, due to not following the H–F–P–S foot-switch pattern and/or being outlier cycles relative to deceleration, reversing, and acceleration.
802
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807
Step32 system, using the statistical approach proposed by Agostini and Knaflitz (2012). Eventually, m = 30 ms is considered a suitable value for evaluating muscle activation in gait analysis (Bonato et al., 1998). 2.4. Statistical gait analysis (SGA) Statistical gait analysis (Agostini and Knaflitz, 2012) is a recently developed methodology defined as the statistical characterization of gait, arrived at by averaging spatial–temporal and sEMG-based parameters. The aim of SGA is the acquisition and statistical analysis of sEMG signals over hundreds of strides during the same walking trial, without the limitation of three-four strides of classic sEMG acquisitions in a gait analysis laboratory (Agostini et al., 2010; Agostini and Knaflitz, 2012; Benedetti et al., 2012). SGA relies on the fact that the number of muscle activations is cycle dependent, so that averaging should be performed only over onset/offset instants of cycles including the same number of activations, i.e. belonging to the same activation modality. Activation modality is defined as the number of times a muscle activates during a single GC; each modality refers to a different activation pattern, characterized by a specific number of activations over GC, i.e. n-activation modality consists of n activation intervals for the considered muscle during a single GC. Mean activation intervals (normalized with respect to GC duration) for each modality of activation were achieved, according to the following steps. First, muscle activations relative to each GC were identified, computing muscle onset/offset instants in temporal space (Bonato et al., 1998), as previously described. Then, muscle activations were grouped according to their activation modalities. Eventually, the onset/offset time instants of each activation modality were averaged over the thirty population subjects. Averaged onset/offset percentage time instants were normalized with respect to GC to provide mean activation intervals in percentage of GC. SGA was performed by Step32 system. 2.5. Statistics The following variables were considered in the statistical analysis, for each muscle and for each activation modality: onset and offset instants of activation, length of activation intervals, and occurrence frequency. Data are reported as mean ± standard deviation (SD). The Lilliefors test was used to evaluate the hypothesis that each data vector had a normal distribution. Comparisons among normally distributed samples were performed with two-tailed, non-paired Student’s t-test (two groups) or one-way analysis of variance (ANOVA) (more than two groups) followed by multiple comparison test. The Kruskal–Wallis test, followed by a multiple comparison test, were used to compare samples not normally distributed. Statistical significance was set at p < 0.05.
3. Results Overall, 3729 strides out of 9787 were discarded for not following the H–F–P–S foot-switch pattern, and/or for being outlier cycles relative to deceleration, reversing, and acceleration. H-phase lasted 5.7 ± 1.6% (percent of GC), F-phase 29.0 ± 6.5%, P-phase 23.7 ± 6.1% and S-phase 41.5 ± 6.2%. The SGA of myoelectric signals indicated that muscles show different numbers of activation intervals in different strides of the same walking. An example of sEMG signals from the VL muscle of a subject is depicted in Fig. 2. Details of activation intervals of VM and VL, expressed as the timing, in the percentage of GC, of signal onset and offset are reported in Figs. 3 and 4 (panel B).
Fig. 2. An example of raw sEMG signals of VL muscle showing 2 activations (panel A), 3 activations (panel B) and 4 activations (panel C), respectively. The signals were extracted from different strides of the same subject, during the same walk. H, F, P S phases are delimited by dashed light-gray vertical lines.
The most recurrent (p = 8.2 10 12) activation modality for VM (Fig. 3, panel A) consists of two activations (2-activation modality), observed in 53.4 ± 23.2% of strides: the first occurs at the beginning of GC and the second in terminal swing. The second most recurrent activation modality is similar to the 2-activation modality but with a further activation in the beginning of the push-off phase. This 3-activation modality was observed in 34.1 ± 14.3% of strides. In a further 10.3 ± 9.3% of strides, four activations were observed: in the beginning of GC, in the Foot Flat contact phase, in the push-off phase and in terminal swing (4-activation modality). The remaining 1.8 ± 2.0% of strides were characterized by five activations during the GC (5-activation modality). This modality showed a rare occurrence and a large variability; thus, it is not considered in the present analysis. Participant-specific data are reported in Table 1 for VM and in Table 2 for VL. For VL, the most recurrent (p = 8.3 10 11) activation modality (Fig. 4, panel A) consists of two activations (2-activation modality) observed in 59.4 ± 27.8% of strides: the first occurs at the beginning of GC, the second in terminal swing. The second most recurrent activation modality is similar to the 2-activation modality but with a further activation in the beginning of the push-off phase. This 3-activation modality was observed in 28.9 ± 16.6% of strides. In a further 9.5 ± 10.6% of strides, four activations were observed: in the beginning of GC, in the Foot Flat contact phase, in the
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807
Fig. 3. Vastus medialis: mean (+SD) percentage occurrence frequency of each of the five different activation modalities (panel A) and mean (+SD) activation intervals vs. percentage of gait cycle for the modalities with 2, 3 and 4 activations, respectively (panel B), detected during walking. Mean and SD refer to the present female population. H, F, P S phases are delimited by dashed light-gray vertical lines. ⁄ Significantly different, compared to 1-, 3-, 4- and 5-activation modality.
803
2008; Di Nardo et al., 2015b; Kerrigan et al., 1998; Malinzak et al., 2001), indicating the suitability of stratifying biomechanical and electromyographic reference frames for males and females. Moreover, it was observed that advancing age modifies the timing of onset/offset activation of lower limb muscles, including the vastii (Hortobagyi et al., 2009). Thus, to control for age and gender, only young (20 years < age < 30 years) females were recruited. Further recruitment exclusion criteria (see Section 2) were used to control for possible additional confounding effects. The application of all the above-mentioned criteria limited the number of participants, but also reduced the variability, and therefore the number of subjects necessary, for the study. The observed vastii activation intervals roughly followed the typical pattern reported for each muscle during walking (Perry, 1992; Sutherland, 2001). The present analysis put into evidence that the considered muscles showed different modalities in the number of activations and in the timing of signal onset/offset (Figs. 3 and 4). In previous studies (Di Nardo et al., 2013, 2015a), we observed the same behavior in tibialis anterior and gastrocnemius lateralis. This suggests that, besides the muscular activation instants, it is worth considering how frequently are observed, i.e. their occurrence frequency. Reporting such a parameter is quite innovative, because it is seldom considered due to the low number of strides usually analyzed in classic EMG studies. Statistically, the most recurrent activation modality during GC consists of two activations for both VM (observed in 53.4 ± 23.2% of total strides, Fig. 3) and VL (observed in 59.4 ± 27.8% of total strides, Fig. 4), at the beginning of GC and in terminal swing, respectively. The 3-activation modality (34.1 ± 14.3% for VM and 28.9 ± 16.6% for VL) differs from the most common modality, because of a further
push-off phase, and in terminal swing (4-activation modality). The remaining 1.5 ± 2.3% of strides were characterized by five activations during the GC (5-activation modality); as with VM, it is not considered in the present analysis. No significant differences (p > 0.05) were observed in onset/offset instants of activation, or in the length of activation intervals between VM and VL, for all the activation modalities. This study also compared VM and VL activity in terms of the frequency each modality of muscle activation occurs, quantified by the percentage number of strides over the total number of strides of the whole subject population, where the muscle is recruited with the specific activation modality: no significant differences (p > 0.05) were detected between VM and VL. Fig. 5 shows a pictorial representation of VM and VL activation intervals in a function of the number of subjects where muscular activity is observed; this representation was achieved for each muscle, considering the three main activation modalities together.
4. Discussion In this study, sEMG signals were recorded in thirty healthy young females, in order to assess the natural variability of the activation modalities of VM and VL during walking at a self-selected pace. SGA was used to analyze hundreds of strides for each subject. The SENIAM recommendations for electrode location and orientation over muscle with respect to tendons, motor point location, and fiber direction (Freriks et al., 2000) were strictly followed in order to soften the possible impact of differences in muscle fiber amount and orientation between VM and VL. Recent studies suggested gender-related differences in knee kinematic and sEMG activity of the vastii (Chumanov et al.,
Fig. 4. Vastus lateralis: mean (+SD) percentage occurrence frequency of each of the five different activation modalities (panel A) and mean (+SD) activation intervals vs. percentage of gait cycle for the modalities with 2, 3 and 4 activations, respectively (panel B), detected during walk. Mean and SD refer to the present female population. H, F, P S phases are delimited by dashed light-gray vertical lines. ⁄ Significantly different, compared to 1-, 3-, 4- and 5-activation modality.
804
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807
Table 1 Occurrence frequencies of all five VM activation modalities for each subject. Subject
1-activation modality (%)
2-activation modality (%)
3-activation modality (%)
4-activation modality (%)
5-activation modality (%)
#1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15 #16 #17 #18 #19 #20 #21 #22 #23 #24 #25 #26 #27 #28 #29 #30
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 3.1 0.0 0.5 0.0 0.0 0.9 0.0 0.6 0.0 0.2 0.0 1.1 0.0 0.0 0.0 0.0 1.0 0.0 0.0
63.0 47.8 61.1 59.9 54.7 54.8 79.9 79.5 92.1 88.2 28.4 16.9 56.2 20.1 42.2 51.1 68.7 63.2 20.9 46.5 89.2 9.5 33.0 75.6 20.0 68.0 23.5 58.8 65.5 63.3
34.4 41.4 36.3 37.0 39.2 31.9 17.7 17.0 7.1 11.4 37.6 48.4 37.0 51.0 42.2 33.7 22.5 27.7 42.4 45.2 10.1 72.0 42.1 20.1 49.2 26.0 53.6 37.8 23.7 26.1
2.6 9.7 2.6 3.1 4.9 13.3 2.4 3.2 0.4 0.4 33.0 23.6 6.4 23.0 12.2 10.9 5.7 6.8 32.0 7.7 0.2 14.9 18.2 4.3 23.1 5.9 20.3 1.5 7.9 8.3
0.0 0.8 0.0 0.0 1.2 0.0 0.0 0.3 0.0 0.0 0.9 6.2 0.5 5.4 3.3 3.3 2.2 1.4 3.5 0.7 0 2.9 5.7 0.0 6.1 0.1 2.6 1.0 2.8 2.2
Values are reported as percentage of total strides. The occurrence frequency for the most recurrent activation modality is reported in bold.
Table 2 Occurrence frequencies of all five VL activation modalities for each subject. Subject
1-activation modality (%)
2-activation modality (%)
3-activation modality (%)
4-activation modality (%)
5-activation modality (%)
#1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15 #16 #17 #18 #19 #20 #21 #22 #23 #24 #25 #26 #27 #28 #29 #30
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 6.3 0.0 0.0 0.0 0.0 3.5 0.0 0.0 0.0 0.0 1.8 0.0 1.8 0.8 0.0 0.0 1.5 0.0 0.0
55.8 78.5 47.4 74.1 20.3 5.3 61.5 73.1 81.9 93.1 18.4 30.6 75.8 77.9 20.6 65.2 86.3 55.9 14.5 81.3 83.1 23.8 60.2 90.2 19.2 88.9 66.0 94.6 87.6 49.4
35.7 17.6 44.1 24.8 45.1 68.5 30.1 24.0 15.0 6.6 47.7 36.0 21.5 21.1 54.4 29.4 9.3 30.0 51.2 18.1 16.2 51.8 22.7 5.5 43.1 11.1 31.4 3.9 11.3 40.6
7.5 3.9 7.8 1.0 29.7 22.8 7.6 2.3 2.6 0.4 28.4 19.8 2.3 1.0 23.3 3.3 0.9 11.8 33.1 0.7 0.7 19.0 14.8 2.4 26.9 0.0 2.6 0.0 0.6 9.4
1.0 0.0 0.7 0.0 4.9 2.9 0.8 0.6 0.4 0.0 5.5 6.3 0.5 0.0 1.11 2.2 0.0 2.3 1.2 0.0 0.0 2.9 2.3 0.0 9.2 0.0 0.0 0.0 0.6 0.6
Values are reported as percentage of total strides. The occurrence frequency for the most recurrent activation modality is reported in bold.
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807
Fig. 5. VM and VL activations over the population, as percentage of gait cycle, considering, for each muscle, the three main modalities of activation all together. Horizontal bars are gray-level coded, according to the number of subjects where a certain condition is observed; black: condition observed for all subjects in every activation modality, white: condition never met. H, F, P S phases are delimited by dashed light-gray vertical lines.
activity around heel off. In the 4-activation modality (10.3 ± 9.3% for VM and 9.5 ± 10.6% for VL), this activation splits into two distinct activations. Considering the gait periodicity, it is clear that, in all activation modalities, the activation at the beginning of the GC is just a continuation of the activation at the end of the previous GC (Figs. 3 and 4, panel B). Still, to conform to the literature (Perry, 1992), the activation modalities were represented as functions of the typical GC, which conventionally starts with the initial heel contact and ends at the successive heel contact. Consequently, the single activity of the vastii during the beginning and the end of the GC was split in two separate activations. Results revealed variability in the number of activations, occurrence frequency, and onset–offset instants across the thousands of strides analyzed (Figs. 3 and 4), despite the narrowly focused population (young, healthy, non-obese females) recruited for the study. The apparently large amount of variability in the occurrence frequency across participants (Figs. 3 and 4, Panel A) is due to the fact that not every participant presents the same preferred (i.e. most recurrent) activation modality. Indeed, as reported in Tables 1 and 2, around 70% of the participants prefer walking with 2-activation modality and around 30% of the participants prefer walking with 3-activation modality. Thus, two sub-groups of people are observed; those who favor a 2-activation modality and those who favor a 3-activation modality. This difference is due to the presence/absence of muscle activity during mid/terminal stance. No further differences in the gait patterns of the people that appear to belong to each of these groups were detected. Variability was also detected in onset–offset instants (Figs. 3 and 4, Panel B), in particular for the activity around heel off. Despite the variability, there was one activation occurrence which remained consistent across subjects for both VM and VL. This activation occurred from terminal swing to the following loading response. It is recognized as the typical activation for both VM and VL during normal walking, and is required to generate tension in terminal swing in preparation for weight bearing at initial contact, and to control knee flexion during weight acceptance (Perry, 1992; Sutherland, 2001). This activation is observed in the totality of the strides for both muscles. A second, less frequent, activation occurred between mid-stance and pre-swing (observed in 44.4 ± 17.1% of strides for VM and in 38.4 ± 19.7% for VL), in monophasic (3-activation modalities, Figs. 3 and 4, panel B) or biphasic (4-activation modalities) shape, and presented activations significantly shorter than those detected between terminal swing and subsequent loading response. This activation is characterized by a large variability, making it difficult, at this stage, to indicate a single and precise explanation for the variability. Agostini et al. (2010) suggested that
805
the recruitment in this phase may be related to the activity of the muscle to modulate rapid knee flexion in the transition between mid and terminal stance, and to stabilize the patella before entering the pre-swing phase. However, since knee flexion modulation and patellar stabilization occur on every gait cycle, one could wonder why this pattern is not observed more consistently. A further possible interpretation could be the mutual crosstalk between vastii and rectus femoris. Although we are not able to completely exclude this hypothesis, the presence of this vastii activity, in each subject and in a limited number of strides (44.4 ± 17.1%), suggests that it is related to the need to fulfill a specific, less frequent task, rather than to crosstalk. Possible future studies aimed at investigating cross-talk could be performed by triggering sEMGs in the vastii muscles with the firing patterns of motor units identified, with intramuscular electrodes, in the rectus femoris (Merletti and Farina, 2009). Results also indicate that the meaning of the mid-stance activation is still not completely clear. Despite this, the present study has the merit to point out and quantify, for the first time to our knowledge, this vastii activity in healthy females. Moreover, the study evokes further analysis on this sEMG activity, which seems critical to understand the possible cause for the different activation patterns detected here. No significant differences were observed in onset/offset instants of activation, in the activation length, and in the occurrence frequency, suggesting a coincident VM and VL recruitment in the present female population (Fig. 5). This finding is consistent with previous results (Cowan et al., 2001; Karst and Willett, 1995; Powers et al., 1996) and concurs with the hypothesis that, in healthy people, the onset and cessation of activity of the vastii muscles are balanced. Indeed, although the VM and VL may have antagonistic actions for the mediolateral control of the patella, ultimately, the recruitment of VM and VL should be appropriately timed for efficient biomechanical function of the knee (Grelsamer and McConnell, 1988). To our knowledge, the present analysis represents the first attempt to quantify the VM and VL dynamic sEMG activity during walking in young, healthy females. Such quantification was performed in terms of variability of onset–offset muscular activation and occurrence frequency, and provides a first contribution to the process of building a ‘‘normality’’ reference frame for females. By including the physiological variability of the phenomenon and controlling the confounding effects due to age and gender, the reference frame proposed here, even in its early stages, could already support discrimination of pathological from physiological behaviors. Moreover, due to its selectivity, the model could be suitable for designing more focused gait studies on the effect of age and gender on the variability of physiological activity of the vastii. 4.1. Limitations of the study A first limitation of the study lies in the limited number of enrolled females. Thus, further studies on larger populations, possibly including male and/or older populations, are needed to confirm or contrast the present findings. Another limitation consists of the lack of information about the Q angle, which was not measurable in the present study. The Q angle represents an estimate of the resultant force of the quadriceps on the patella and is a predictor of the patella lateral movement under dynamic conditions. Convincing evidence indicates that young women have greater mean Q angles than their male counterparts, with differences in magnitude ranging from 2.7 to 5.8° (Livingston, 1998), and that these Q-angle differences could be associated with gender differences in VL and VM EMG parameters (Malinzak et al., 2001; Sigward and Powers, 2006). These findings support not only the suitability of developing a specific
806
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807
reference frame for females as indicated by the present study, but also the need for future studies to complete such a reference frame with Q-angle measurements.
5. Conclusions Despite the large variability detected in VM and VL activity, one activation occurrence remains consistent across subjects for both muscles. This occurs from terminal swing to the following loading response and is observed in 100% of strides). A second, less frequent, activation occurs between mid-stance up to pre-swing. This pattern of activation represents the first contribution toward developing a reference frame able to describe the physiological variability of vastii sEMG activity during walking.
Conflict of interest
Kerrigan DC, Todd MK, Della Croce U. Gender differences in joint biomechanics during walking: normative study in young adults. Am J Phys Med Rehabil 1998;77:2–7. Livingston LA. The quadriceps angle: a review of the literature. J Orthop Sports Phys Ther 1998;28(2):105–9. Malinzak RA, Colby SM, Kirkendall DT, et al. A comparison of knee joint motion patterns between men and women in selected athletic tasks. Clin Biomech 2001;16(5):438–45. Merletti R, Farina D. Analysis of intramuscular electromyogram signals. Philos Transact A Math Phys Eng Sci 2009;367:357–68. Perry J. Gait analysis – normal and pathological function. USA: Slack Inc.; 1992. 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. Sigward SM, Powers CM. The influence of gender on knee kinematics, kinetics and muscle activation patterns during side-step cutting. Clin Biomech (Bristol, Avon) 2006;21:41–8. Sutherland DH. The evolution of clinical gait analysis part l: kinesiological EMG. Gait Posture 2001;14:61–70. Tirosh O, Sparrow WA. Age and walking speed effects on muscle recruitment in gait termination. Gait Posture 2005;21:279–88. Winter DA, Yack HJ. EMG profiles during normal human walking: stride to stride and inter-subject variability. Electroencephalogr Clin Neurophysiol 1987;67:402–11.
The authors declare that there are no conflicts of interest.
References Agostini V, Balestra G, Knaflitz M. Segmentation and classification of gait cycles. IEEE Trans Neural Syst Rehabil Eng 2014;22(5):946–52. 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, editors. Stevenson Ranch, California, (USA): American Scientific Publishers, vol. 2; 2012. p. 99–121. Agostini V, Nascimbeni A, Gaffuri A, et al. Normative EMG activation patterns of school-age children during gait. Gait Posture 2010;32(3):285–9. Benedetti MG, Agostini V, Knaflitz M, et al. Self-reported gait unsteadiness in mildly impaired neurological patients: an objective assessment through statistical gait analysis. J Neuroeng Rehabil 2012;29(9):64. 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. Chumanov ES, Wall-Scheffler C, Heiderscheit BC. Gender differences in walking and running on level and inclined surfaces. Clin Biomech (Bristol, Avon) 2008;23:1260–8. Cowan SM, Bennell KL, Hodges PW, et al. Delayed onset of electromyographic activity of vastus medialis obliquus relative to vastus lateralis in subjects with patellofemoral pain syndrome. Arch Phys Med Rehabil 2001;82(2):183–9. 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. Di Nardo F, Ghetti G, Fioretti S. Assessment of the activation modalities of gastrocnemius lateralis and tibialis anterior during gait: a statistical analysis. J Electromyogr Kinesiol 2013;23(6):1428–33. Di Nardo F, Mengarelli A, Maranesi E, et al. Influence of gender on the myoelectric signal of thigh muscles. In: MESA 2014 – 10th IEEE/ASME international conference on mechatronic and embedded systems and applications conference proceedings, Senigallia; 2014. http://dx.doi.org/10.1109/MESA. 2014.6935538. Di Nardo F, Mengarelli A, Maranesi E, Burattini L, Fioretti S. Assessment of the ankle muscle co-contraction during normal gait: a surface electromyography study. J Electromyogr Kinesiol 2015a;25(2):347–54. Di Nardo F, Mengarelli A, Maranesi E, Burattini L, Fioretti S. Gender differences in the myoelectric activity of lower limb muscles in young healthy subjects during walking. Biomed Signal Process Control 2015b;18:14–22. Freriks B, Hermens HJ, Disselhorst-Klug C, Rau G. Development of recommendations for SEMG sensors and sensor placement procedures. J Electromyogr Kinesiol 2000;10(5):361–74. Grelsamer RP, McConnell J. The patella: a team approach. Gaithersburg (MD): Aspen; 1988. Hortobagyi T, Solnik S, Gruber A, et al. Interaction between age and gait velocity in the amplitude and timing of antagonist muscle coactivation. Gait Posture 2009;29:558–64. Kadaba MP, Ramakrishnan HK, Wootten ME, et al. Repeatability of kinematic, kinetic, and electromyographic data in normal adult gait. J Orthop Res 1989;6:849–60. Karst GM, Willett GM. Onset timing of electromyographic activity in the vastus medialis oblique and vastus lateralis muscles in subjects with and without patellofemoral pain syndrome. Phys Ther 1995;75:813–23.
Francesco Di Nardo currently teaches the classes of Models and Control of Biological Systems in Università Politecnica delle Marche, Ancona, Italy, where he received his Ph.D., in 2005. His main research activity involved the development and the clinical application (type-2 diabetes, insulin resistance and hypertension) of mathematical models of metabolic and endocrine systems. 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. He published several papers in peer review international journals, books and conference proceedings.
Elvira Maranesi received her Ph.D. in Biomedical Engineering from Università Politecnica delle Marche, Ancona, Italy in 2013. Currently, she is Postdoctoral Researcher with the Department of Information Engineering, Università Politecnica delle Marche and collaborates with the Posture and Movement Analysis Laboratory at the Italian National Institute of Health and Science on Aging. Furthermore she is a charter member of B.M.E.D. Srl, an academic spin-off of Università Politecnica delle Marche. Her main research interests include the electromyographic signal processing and the movement analysis during static and dynamic postural tests to evaluate the risk of falls in the elderly. In this field, she is author of scientific publications in international journals and congress proceedings.
Alessandro Mengarelli received his Master Degree in Biomedical Engineering in 2012 from Università Politecnica delle Marche, Ancona, Italy, where he is actually attending Ph.D. course in Information Engineering. His research field involves dynamic posturography and gait analysis, focused mainly on the assessing the muscles’ behavior and functions by means of surface electromyography signal. In this field he is the co-author of some national and international conference proceedings.
F. Di Nardo et al. / Journal of Electromyography and Kinesiology 25 (2015) 800–807 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.
Laura Burattini is currently Assistant Professor of Biomedical Engineering at the Polytechnic University of Marche, where she teaches the classes of Biomedical Engineering and Biomedical Signal and Data Processing, and President and CEO of B.M.E.D. SRL, an academic spinoff she founded herself. She graduated in Electrical/Biomedical Engineering at the Politecnico di Milano, Italy, in 1993 and took the PhD at the University of Rochester, USA, in 1998. Then, returned in Italy, she matured extensive working experience in both academic and corporate. Her main research interest includes the biomedical signal processing. She is author of numerous scientific publications in international journals, books and congress proceedings.
807
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.