Do surface electromyograms provide physiological estimates of conduction velocity from the medial gastrocnemius muscle?

Journal of Electromyography and Kinesiology 23 (2013) 319–325

Contents lists available at SciVerse ScienceDirect

Journal of Electromyography and Kinesiology journal homepage: www.elsevier.com/locate/jelekin

Do surface electromyograms provide physiological estimates of conduction velocity from the medial gastrocnemius muscle? Alessio Gallina a, Cintia H. Ritzel b, Roberto Merletti a, Taian M.M. Vieira a,c,⇑ a

Laboratorio di Ingegneria del Sistema Neuromuscolare (LISiN), Politecnico di Torino, Italy Faculdade de Medicina, Serviço de Ortopedia e Traumatologia, Hospital de Clínicas de Porto Alegre, Universidade Federal do Rio Grande do Sul, Brazil c Escola de Educação Física e Desportos, Universidade Federal do Rio de Janeiro, Brazil b

a r t i c l e

i n f o

Article history: Received 17 September 2012 Received in revised form 9 November 2012 Accepted 11 November 2012

Keywords: Medial gastrocnemius Surface electromyogram Conduction velocity

a b s t r a c t Muscle fiber conduction velocity (CV) is commonly estimated from surface electromyograms (EMGs) collected with electrodes parallel to muscle fibers. If electrodes and muscle fibers are not located in parallel planes, CV estimates are biased towards values far over the physiological range. In virtue of their pinnate architecture, the fibers of muscles such as the gastrocnemius are hardly aligned in planes parallel to surface electrodes. Therefore, in this study we investigate whether physiological CV estimates can be obtained from the gastrocnemius muscle. Specifically, with a large grid of 16  8 electrodes we map CV estimates over the whole gastrocnemius muscle while eleven subjects exerted isometric plantar flexions at three different force levels. CV was estimated for couples of single differential EMGs and estimate locations (i.e., channels) were classified as physiological and non-physiological, depending on whether CV estimates were within the physiological range (3–6 ms 1) or not. Physiological CV values could be estimated from a markedly small muscle region for eight participants; channels providing physiological CV estimates corresponded to about 5% of the total number of channels. As expected, physiological and non-physiological channels were clustered in distinct regions. CV estimates within the physiological range were obtained for the most distal gastrocnemius portion (ANOVA, P < 0.001), where occurrences of propagating potentials were often verified through visual analysis. For the first time, this study shows that CV might be reliably assessed from surface EMGs collected from the most distal gastrocnemius region. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Unique physiological and anatomical information on skeletal muscles might be obtained from high-density surface electromyography. While a conventional system of bipolar electrodes has been successfully used to identify the timing and degree of muscle activation (Ferrari et al., 2008), the use of a grid of surface electrodes allows for: (i) the evaluation of localized muscle activity (Vieira et al., 2011; Falla and Farina, 2008); (ii) identifying the location of innervation zones and tendon regions (Barbero et al., 2011); (iii) the estimation of muscle fibers length (Merletti et al., 2003); (iv) the decomposition of surface electromyograms (EMGs) into the action potentials of individual motor units (Holobar et al., 2009); and (v) the estimation of muscle fiber conduction velocity (CV; Merletti et al., 1990). Each of these features likely posits marked relevance in specific clinical settings, in sports and in rehabilitation. The possibility of estimating CV from EMGs is of particular physiological interest. Peripheral changes in the neuromuscular ⇑ Corresponding author. Address: LISiN, Politecnico di Torino, Via Cavalli 22/h, 10138 Torino (TO), Italy. E-mail address: [email protected] (T.M.M. Vieira). 1050-6411/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jelekin.2012.11.007

system, such as the structural changes in the muscle sarcolemma, might speed either up or down the propagation of motor unit action potentials. For instance, CV slowing is typically reported from EMGs recorded during prolonged or forceful muscle effort (i.e., fatiguing contractions at constant force; Merletti et al., 1990; Rainoldi et al., 2008a; Cescon and Gazzoni, 2010). Moreover, considering that action potentials propagate at higher speeds along muscle fibers of greater diameters, relative differences between CV estimates could possibly be associated to motor units of different sizes. It is not surprising that sprinters and long distance runners exhibit different CV values (Rainoldi et al., 2008b). Moreover, CV evaluation proved useful for the diagnosis of neuromuscular disorders (Falla and Farina, 2005; Gerdle et al., 2008). Notwithstanding its potentialities in different settings, the attractive possibility of estimating CV non-invasively demands caution. Most importantly, consecutive surface electrodes must be positioned on skin regions parallel to muscle fibers (Merletti et al., 2003). While such a condition is easily met for some muscles, for others, the parallelism between fibers and electrodes is hardly observed. Obtaining physiological CV estimates from EMGs collected from the gastrocnemius muscle is somewhat challenging. Because of their pinnate architecture, gastrocnemius fibers are mostly not

A. Gallina et al. / Journal of Electromyography and Kinesiology 23 (2013) 319–325

parallel to electrodes positioned on the skin surface. Consequently, the delay estimated between surface potentials recorded by successive electrodes along the gastrocnemius length is often not physiological (Vieira et al., 2010b). Previous studies, indeed, reported markedly high (higher than 6 ms 1) CV values estimated for the gastrocnemius muscle (Rainoldi et al., 2003). By triggering EMGs collected with an array of electrodes with the firing pattern of motor units in the medial gastrocnemius (MG) muscle, however, we observed a consistently frequent occurrence of propagating potentials in the very distal muscle region (Vieira et al., 2011). It seems then worth investigating whether and where physiological CV estimates might be obtained from propagating potentials in the pinnate gastrocnemius. In this study we systematically investigate if and where physiological CV estimates in gastrocnemius fibers might be obtained from EMGs. Specifically, with a large matrix of surface electrodes, we test whether CV in the MG muscle might be estimated within the physiological range of 3–6 ms 1 (Andreassen and Arendt-Nielsen, 1987; Hedayatpour et al., 2007). If so, then, we further investigate whether the likelihood of obtaining physiological CV estimates depends on where EMGs are recorded from the MG muscle. Following our previous evidence showing propagating potentials in the most distal MG region (Vieira et al., 2011), we expect these physiological estimates to be obtained from the most distal electrodes. 2. Methods 2.1. Subjects After providing written informed consent, eleven male subjects (mean ± standard deviation; age: 27 ± 4 years; body mass: 76 ± 6 kg; height: 182 ± 6 cm) volunteered to participate in this experiment. None of the participants reported neurological disorders or muscle-skeletal dysfunction in their limbs. All experimental procedures conformed to the latest amendment to the Declaration of Helsinki and were approved by the Institutional Ethical Committee. 2.2. Experiment protocol Ankle torque was recorded with subjects in seating position, with their knee fully extended and their hip flexed at angles smaller than 90° to avoid discomfort due to excessive lengthening of the hamstring muscles. Participants had their foot positioned over a piezoelectric force-plate (9286AA Kistler, Zurich, Switzerland), mounted vertically on a rigid structure in front of them. The distance between the force-plate and the seat was adjusted separately for each individual, so that the foot was passively pushing against the force-plate (see Fig. 1 in Gallina et al., 2011 for detailed description on the seat setup and on the calculation of ankle torque). The torque of plantar flexion was calculated as the torque applied about the force-plate transverse axis. For this reason, the axis of ankle rotation in the sagittal plane, roughly defined as crossing lateral and medial malleolus (Wu et al., 2002), was aligned as parallel as possible to the force-plate transverse axis. While being provided with visual feedback of their ankle torque, subjects were asked to exert isometric plantar flexion contractions at 10%, 30% and 60% of their maximal effort. A custom interface developed in Matlab (Version 7.0.4, The Mathworks, Natick, USA) was used to provide subject with visual feedback of their ankle torque profile (i.e., the computer screen showing 5 s of ankle torque was updated every 100 ms). Isometric contractions lasted 15 s and were applied in random order. A brief period of training (from 5 to 10 min) was provided before starting measurements.

Right Leg Junction between medial and lateral Gastrocnemius

16

Rows

320

1

1 8 Columns

Junction between medial gastrocnemius and Achilles tendon

Fig. 1. Matrix positioning on the calf. The positioning of the electrode grid over the MG muscle is schematically shown in the left panel. Achilles tendon and the junction between the two gastrocnemii were marked on the skin after ultrasound assessment of the right calf of each subject. The adhesive electrode grid was then placed according to these anatomical landmarks, covering as much as possible the MG muscle. Circles appearing on the skin are residues of the conductive paste left over after removal of the grid.

Sub-maximal intensities were calculated with respect to the torque of ankle plantar flexion averaged across three maximal voluntary contractions. Each maximal attempt lasted 5 s and a 2 min resting interval in-between was provided. Subjects were verbally encouraged to reach strongest torque levels during these maximal contractions. While not contracting their plantar flexors, participants were allowed to place their leg in a comfortable position close to the force-plate. The foot was then repositioned in a standardized position for the following contraction. 2.3. Electromyographic recordings Single-differential EMGs were detected using a large matrix of 128 surface electrodes, disposed into 8 columns and 16 rows with 10 mm interelectrode distance. EMGs were amplified by a variable factor, ranging from 1000 to 5000, to ensure the highest signal to noise ratio without resulting in saturation (10–750 Hz bandwidth amplifier, EMG-USB amplifier, LISiN and OTBioelettronica, Turin, Italy). After that, EMGs and reaction forces measured by the force-plate were digitized synchronously at 2048 samples/s with a 12-bit A/D converter (±2.5 V dynamic range). As the analog signals provided by the force-plate were fed-in the EMG-USB amplifier, EMGs detected by the last column of electrodes were discarded (i.e., a total of 105 single-differential EMGs were recorded). 2.4. Matrix positioning Given our interest in mapping CV estimates across the whole MG muscle, we used a specific procedure to position the matrix of electrodes on the calf. With the use of an ultrasound device (Fukuda Denshi, UF 4000, 7.5 MHz linear probe), the junction between both gastrocnemius muscle was identified and marked on the skin. Similarly, the medial MG border was identified and

321

A. Gallina et al. / Journal of Electromyography and Kinesiology 23 (2013) 319–325

from 1.6 to 3.3 ms (CV range: 3–6 ms 1, (Andreassen and ArendtNielsen, 1987; Hedayatpour et al., 2007)) were termed, respectively, as non-physiological and physiological. Finally, the baricenter coordinates for each of these two categories of pairs of channels were calculated to test for whether the estimation of physiological delays depended on the location where EMGs were detected from the MG muscles. All EMGs were visually inspected for the occurrence of low quality signals. EMGs of low quality were conceived in terms of short-circuits (identified as flat channels), noisy channels (highamplitude EMGs with no clear occurrences of action potentials), channels with movement artifacts (sporadic occurrences of lowfrequency variations of the signal) or channels affected by power line interference (marked 50 Hz and harmonics interference). Whenever any of these features was verified, the corresponding channel was noted and excluded from analysis.

marked on the skin. With the ultrasound probe oriented longitudinally along the leg, we further scanned the muscle–tendon junction and marked it on the skin surface. For a detailed description on the scanning procedure, the reader is referred to the supplemental material in Vieira et al. (2010a). After obtaining a coarse representation of the MG contour on the skin (Fig. 1), we carefully positioned the grid of electrodes. Specifically, the bottom row of electrodes was positioned immediately proximal to the most distal muscle–tendon region, while the most right column was located about 1 cm medial with respect to the junction between the two gastrocnemii (Fig. 1). Before positioning the matrix, the whole skin covering the calf region was shaved, cleansed with abrasive paste and then wet with a soaked cloth. The electrode grid was positioned on the skin of each subject with a bi-adhesive pad of foam. In correspondence of electrodes, there were 128 cavities in the pad. Electrical contact between electrodes and skin was ensured by filling cavities with a conductive paste.

2.6. Statistical analysis 2.5. Data analysis Descriptive statistics were used to quantify the number of physiological and non-physiological CV estimates, within and between subjects. Differences in the longitudinal MG location (distal and proximal) providing physiological and non-physiological estimates were tested with a two-way analysis of variance (ANOVA; three contraction intensities  2 delay categories). Parametric statistics were used after ensuring our data distribution differed from a Gaussian distribution with a probability smaller than 5% (Shapiro-wilk test, P < 0.05).

CV values in the MG muscle were estimated across the whole matrix of electrodes. Initially, EMGs were offline band-pass filtered with a second order, zero-phase shift Butterworth filter (15– 400 Hz cut-off frequency). Then, using the algorithm proposed by Farina et al. (2001), we estimated the delay minimizing the least square error between the power spectrum of pairs of EMGs detected by consecutive channels along the MG longitudinal axis. Delays were calculated for 250 ms epochs (i.e., 512 samples), thus providing a total of 60 estimates (15 s  4 epochs/s) per couple of channels. CV was calculated as the interelectrode distance (10 mm) divided by estimated delays and then averaged across epochs. Therefore, from the 105 single-differential EMGs, we obtained 98 CV estimates across the MG muscle, for each contraction intensity and for each participant. To distinguish between physiological and non-physiological CV estimates, we categorized each pair of channels according to the delay estimate it provided. Considering that the center-to-center distance of our electrodes was 1 cm, pairs of channels detecting EMGs leading to delays in- and out-side the range defined

3. Results 3.1. Some observations on propagating potentials Anatomical features typically present in EMGs collected from fusiform muscles were observed for the MG muscle of all participants tested. Consider, for example, the representation of a single motor unit action potential across all channels in the matrix, shown in Fig. 2 for a representative subject (subject 11). Surface potentials of this motor unit were clearly evident in the most distal

15

Rows of single differential channels

14 13 12 11 10 9 8 7 6 5 4

0.5 mV

Ankle

3 2 1

Medial 7

6

5

4

Columns (cm)

3

2

1

15 ms

Fig. 2. Propagating potentials in the gastrocnemius muscle. An example of a short epoch (15 ms) of raw, single-differential EMGs, is shown for subject 11. The Achilles tendon is located towards the bottom of the figure (i.e., distally) whereas the lateral gastrocnemius is to the right of the matrix. Action potentials can be observed in the very distal portion of the muscle (rows 1–5), with an innervation zone located nearby the third channel.

322

A. Gallina et al. / Journal of Electromyography and Kinesiology 23 (2013) 319–325

Table 1 Number of channels providing physiological estimates of conduction velocity. For each subject, there were 98 pairs of channels. The median and the range were calculated across three contraction levels and two conditions, before and after fatigue. Subject number

Number of pairs of channels

1 2 3 4 5 6 7 8 9 10 11

Median

Range

0 7 11 0 14 5 0 5 6 16 11

0–2 5–7 9–11 0–2 8–18 4–6 0–1 5–6 4–7 16–18 9–11

3.3. Location of channels providing physiological estimates of conduction velocity

channels, from row 1 to row 7, and at the most medial columns. Potentials with small amplitude (third row) and potentials with phase inversion (e.g., potentials in row 2 and 4) indicate the occurrence of an innervation zone in the proximity of the third row of electrodes. At consecutive channels located either proximally or distally from the third row, consecutive surface potentials appear with a consistent time delay (see the temporal difference between peaks of potentials in the column 6). The delay between consecutive potentials, however, was not consistently similar across all channels in the matrix. 3.2. Occurrences of physiological and non-physiological conduction velocity estimates Notwithstanding the consistent representation of propagating potentials in the EMGs of all participants, not all of them provided physiological CV estimates. Table 1 reports the number of pairs of single-differential EMGs, for each subject, whose CV was estimated within the physiological range (i.e., from 3 to 6 ms 1). Across the three different contraction levels and the two testing sessions, none of the 98 channels provided physiological CV estimates for subjects 1, 4 and 7. The other eight participants provided, at least, four channels with physiological CV estimates. On average, CV in the gastrocnemius muscle was overestimated from EMGs. The distribution of CV estimates, across participants and conditions, is shown in Fig. 3. The majority of non-physiolog-

3578 140 120

Occurrences

ical estimates were exceeding the figure of six meters per second (94%; N = 3234 occurrences; 98 channels  3 contraction levels  11 subjects). The markedly infrequent occurrences of physiological estimates (5%), interestingly, was not represented at random skin locations; they were rather represented at the more distal muscle region.

The average coordinates of pairs of channels providing physiological and non-physiological estimates of CV were centered at different regions in the matrix of electrodes. A short epoch of EMGs collected from subject 9 (Fig. 4A), for example, indicates the occurrence of propagating potentials exclusively at the most distal channels, as observed for the subject 11 in Fig. 2. Indeed, only the most distal channels provided physiological estimates of CV (Fig. 4B). Specifically, the small group of channels (grey circles in Fig. 4B) providing physiological estimates was centered at 0.9 cm proximal and 3.1 cm medial to the bottom-right corner of the matrix of electrodes (see the large dark grey circle in Fig. 4B). Non-physiological estimates, instead, were obtained for a large group of channels centered at a somewhat proximal location (large white circle in Fig. 4B). The proximo-distal difference between physiological and nonphysiological estimates was observed consistently across all participants (Fig. 5). Pairs of channels providing physiological CV values were centered at a significantly more distal region (interquartile interval: 2.5–4.5 cm from the bottom row of electrodes) than those providing non-physiological values (inter-quartile interval 8.5–10.5 cm; Fig. 5A; ANOVA F = 176.9, P < 0.001, N = 24 cases; 3 contraction levels  8 subjects providing physiological estimates of CV). Differences in the lateral baricenter coordinate, on the other hand, did not reach statistical significance (Fig. 5B; ANOVA F = 2.31, P = 0.14). 4. Discussion In this study we investigated whether physiological CV estimates might be obtained from the MG muscle. We further assessed the MG location most likely providing physiological CV estimates. From eleven participants, our results suggest that CV values within the physiological range 3–6 ms 1 were obtained from a small skin region (0–18 channels out of 98 for each contraction). These electrodes were significantly clustered at the most distal MG region. 4.1. Physiological conduction velocity is exclusively estimable from a small and distal gastrocnemius region

100 80 60 40 20 0 3

6

10

20

30

40

50+

Conduction Velocity (m/s) Fig. 3. Estimates of conduction velocity in the gastrocnemius muscle. Histogram of conduction velocity values are shown for all subjects and contraction intensities. The range of values of conduction velocity is represented along the abscissa. Light grey bars represent occurrences of physiological values of conduction velocity. Occurrences of values higher than 50 ms 1 are represented in the last bin.

Estimations of muscle fiber CV are based on the use of systems of multiple electrodes. Consecutive surface electrodes are placed over skin regions in such a way that each electrode samples from a different transversal section of the same muscle fibers. If electrodes and muscle fibers are not aligned parallel with respect to each other, CV estimates reach markedly high and non-physiological values (Farina et al., 2002). Misalignment between electrodes and muscle fibers might occur either in the depth or in the transversal direction. These are the likely reasons explaining the small number and the distal location of physiological CV estimates observed in this study. Depending on where EMGs are collected from the gastrocnemius muscle, the alignment between electrodes and fibers changes. Parallel alignment between electrodes and gastrocnemius fibers, for example, is not viable for electrodes located on skin

323

A. Gallina et al. / Journal of Electromyography and Kinesiology 23 (2013) 319–325

A

B Barycenter denoting physiological estimates

Surface EMGs from the medial gastrocnemius in the right calf

Barycenter denoting non-physiological estimates

15 13 12

13

Distance from the bottom row (cm)

Rows of single differential channels

14

12 11 10 9 8 7 6 5 4

Ankle

3

11 10 9 8 7 6 5 4 3 2 1

0.8 mV

2 1 Medial

0 6

7

6

5

4

Columns (cm)

5

4

3

2

1

0

Distance from the right most column (cm)

20 ms

Fig. 4. Propagating potentials and their location in the matrix of electrodes. (A) shows an example of raw EMGs collected for subject 9. The distribution of conduction velocity values for this subject across skin regions is shown in (B). Propagation of action potentials is observable in the distal portion of the muscle, where conduction velocity estimates obtained from couples of single-differential EMGs was typically within the physiological range (3–6 ms 1). Physiological values in the centro-proximal muscle portion, instead, were not observed. The baricenters of the two areas, i.e. physiological and not physiological values, are represented as grey and white crosses respectively.

A

B *

6

Distance of Y baricenter from the bottom row of electrodes (cm)

4 2

Non-physiological Occurrences

0

6 4 2

12

10

10

0 5

Medial

12

4

3

2

8

8

6

6

4

4

2

2

0

0 6 5 4 3 2 1 0

1

Distance of X baricenter from the most right column (cm)

Ankle

Physiological Occurrences

8

Physiological Occurrences

0

2

4

6

8

Non-physiological Occurrences

Fig. 5. Location of physiological and non-physiological estimates in the gastrocnemius muscle. The median, the interquartile interval (boxes) and the range (whiskers) for the baricenter coordinates along the columns (A) and rows (B) in the grid of electrodes are shown for all subjects and contraction intensities. Distribution of baricenter values are shown separately for conduction velocity estimates classified as physiological and non-physiological. Statistical significance was observed for data shown in panel B.

regions covering the muscle superficial aponeurosis (i.e., fibers are inclined in the muscle depth direction). Consecutive electrodes over the superficial aponeurosis sample the motor units action potentials when in the proximity of the distal tendons (Mesin et al., 2011), not when propagating along the gastrocnemius fibers. Conversely, in the most distal muscle region, from the distal extremity of the superficial aponeurosis to the MG-Achilles tendon junction, MG fibers run parallel to the skin surface. In this distal

location, consecutive electrodes may sample from different transverse sections of the same MG fibers. Because of these relative proximo-distal differences in gastrocnemius geometry with respect to the skin (see Fig. 1 in Hodson-Tole et al., 2012), propagating potentials are predominantly observed at the very distal MG region (see potentials in Figs. 2 and 4). Presumably for this reason, physiological CV estimates were obtained exclusively from the most distal MG region (Fig. 5).

324

A. Gallina et al. / Journal of Electromyography and Kinesiology 23 (2013) 319–325

The small length of gastrocnemius fiber in relation to the muscle length likely accounted for the small number of channels providing physiological CV estimates. Reported values for the length of MG fibers are typically 5 times smaller than those for the total MG length at rest (25 cm; Narici et al., 1996; Kawakami et al., 1998). Thus, because of the 1 cm inter-electrode distance, and considering that we positioned our matrix of electrodes with its bottom row corresponding to the MG-Achilles tendon junction, about five consecutive rows of electrodes should be covering the MG region where fibers were aligned parallel to the skin. The average number of expected channels detecting propagating potentials (35 channels; 5 rows  7 columns) amounts to 33% of the total number of channels in the matrix. This figure is markedly smaller than the relative number of physiological CV estimates we obtained per subject (Fig. 3; see Table 1). Not all propagating potentials in the distal MG region, then, provide CV estimates within the physiological range. 4.2. Why not all propagating potentials provided physiological estimates of conduction velocity? Misalignment between surface electrodes and distal MG fibers in the transverse direction was a key factor possibly leading to non-physiological CV estimates in the muscle distal region. At the most distal MG region, fascicles run from the deep to the superficial aponeurosis in a fan-like and curvilinear fashion (i.e., the longitudinal axes of distal MG fibers are not parallel to each other). Consequently, although distal MG fibers reside in a plane parallel to the skin, within such plane, these fibers are not parallel to each other; pinnation occurs also in a plane parallel to the skin. While some distal MG fibers were parallel to our distal electrodes, others were not. Theoretical and experimental evidence has shown that small deviations of an array of surface electrodes from the longitudinal axis of muscle fibers lead to CV overestimation (Cescon et al., 2008; Farina et al., 2002; Merletti et al., 1999; Mesin et al., 2007). Indeed, close inspection of Figs. 2 and 4 reveals smaller delays between consecutive potentials detected proximally with respect to those detected distally from the innervation zone of distal MG fibers (compare pairs of traces immediately above and below the third row in Figs. 2 and 4). Distal, propagating potentials providing CV overestimation were thus likely recorded by channels not parallel to MG fibers. Besides the transverse inclination of distal MG fibers with respect to the skin, other anatomical factors could have accounted for non-physiological CV estimations in distal channels. Specifically, the presence of myotendinuous junctions and innervation zones under the electrodes could result in higher estimates of conduction velocity (Farina et al., 2001; Roy et al., 1986). In the present study, the position of the distal myotendinuous junction was carefully assessed using ultrasound before the placement of the electrode grid (Fig. 1). Therefore, the presence of the MG-Achilles tendon junction under our detection system was unlikely. Surface potentials in Figs. 2 and 4 clearly show phase opposition (row 3 of columns 3–7 in both figures), which is consistent with innervation zones being located closely beneath the third row of electrodes. Surface electromyograms detected at the neighborhood of the innervation zone typically lead to a biased CV estimation, towards either markedly low or high CV values (Farina et al., 2001; Roy et al., 1986). It is therefore possible that distal CV estimates outside the physiological range were due also to innervation zone or endof-fiber effects. The relevance of physiological CV estimates obtained from the distal MG portion, however, remains to be elucidated. Evidence on the regional activation of skeletal muscles has grown markedly in the last years, in particular for the MG muscle. Local variations in MG activation are typically observed for specific ankle force direc-

tions (Staudenmann et al., 2009), motor tasks (Hodson-Tole et al., 2012), during quiet standing (Vieira et al., 2010b, 2011) and during fatiguing contractions (Gallina et al., 2011), among other circumstances. It is possible then that only a very limited number of the total number of MG motor units has contributed to our CV estimates taken distally from the MG muscle. The cautious user of surface electromyography should bear in mind that, although CV estimates might be reliably obtained from the most distal MG region, they might not be representative of, or sensitive to, physiological events (i.e., muscle fatigue) affecting the muscle globally. 4.3. Limitation of the study The distance between consecutive electrodes was the main limitation of our study. Previous studies have shown that CV estimates are increasingly more robust as the number of electrodes used for its estimation increases and when the double-differential spatial filter is applied (Merletti et al., 2003; Farina and Mesin, 2005). In this study, a relatively small number of channels for each contraction detected propagating potentials. A presumably greater number of propagating potentials could have been detected with a denser system of electrodes (i.e., a matrix with inter-electrode distance smaller than 1 cm). As decreasing the distance between electrodes and increasing the number of electrodes are currently technically difficult and under development, we could not use a denser system of electrodes to map CV estimates over the whole MG muscle. Considering the inter-electrode distance of our system and the average length of MG fibers (5 cm; Narici et al., 1996; Kawakami et al., 1998), the use of double-differential spatial filters would provide a small and possibly unrepresentative spatial resolution for reliable, CV estimation. Specifically, EMGs detected over the innervation zone would likely be included in the computation of double-differential EMGs. Acknowledgements This work was supported by Compagnia di San Paolo, Fondazione C.R.T., Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (INST – 110.842/2012), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). References Andreassen S, Arendt-Nielsen L. Muscle fibre conduction velocity in motor units of the human anterior tibial muscle: a new size principle parameter. J. Physiol. 1987;391:561–71. Barbero M, Gatti R, Lo Conte L, Macmillan F, Coutts F, Merletti R. Reliability of surface EMG matrix in locating the innervation zone of the upper trapezius. J. Electromyogr. Kinesiol. 2011;21(5):827–33. Cescon C, Gazzoni M. Short term bed-rest reduces conduction velocity of individual motor units in leg muscles. J. Electromyogr. Kinesiol. 2010;20(5):860–7. Cescon C, Rebecchi P, Merletti R. Effect of array position and subcutaneous tissue thickness on conduction velocity estimation in upper trapezius muscle. J. Electromyogr. Kinesiol. 2008;18:628–36. Falla D, Farina D. Muscle fiber conduction velocity of the upper trapezius muscle during dynamic contraction of the upper limb in patients with chronic neck pain. Pain 2005;116(1–2):138–45. Falla D, Farina D. Non-uniform adaptation of motor unit discharge rates during sustained static contraction of the upper trapezius muscle. Exp. Brain Res. 2008;191(3):363–70. Farina D, Mesin L. Sensitivity of surface EMG-based conduction velocity estimates to local tissue in-homogeneities – influence of the number of channels and inter-channel distance. J. Neurosci. Methods 2005;142:83–9. Farina D, Muhammad W, Fortunato E, Meste O, Merletti R, Rix IH. Estimation of single motor unit conduction velocity from surface electromyogram signals detected with linear electrode arrays. Med. Biol. Eng. Comput. 2001;39:225–36. Farina D, Cescon C, Merletti R. Influence of anatomical, physical, and detectionsystem parameters on surface EMG. Biol. Cybern. 2002;86(6):445–56. Ferrari A, Benedetti MG, Pavan E, Frigo C, Bettinelli D, Rabuffetti M, et al. Quantitative comparison of five current protocols in gait analysis. Gait Posture 2008;28(2):207–16.

A. Gallina et al. / Journal of Electromyography and Kinesiology 23 (2013) 319–325 Gallina A, Merletti R, Vieira TMM. Are the myoelectric manifestations of fatigue distributed regionally in the human medial gastrocnemius muscle? J. Electromyogr. Kinesiol. 2011;21:929–38. Gerdle B, Ostlund N, Gronlund C, Roeleveld K, Karlsson S. Firing rate and conduction velocity of single motor units in the trapezius muscle in fibromyalgia patients and healthy controls. J. Electromyogr. Kinesiol. 2008;18:707–16. Hedayatpour N, Arendt-Nielsen L, Farina D. Motor unit conduction velocity during sustained contraction of the vastus medialis muscle. Exp. Brain Res. 2007;180:509–16. Hodson-Tole EF, Loram ID, Vieira TMM. Myoelectric activity along human gastrocnemius medialis: different spatial distributions of postural and electrically elicited surface potentials. J Electromyogr Kinesiol; 2012. September 8. [Epub ahead of print]. Holobar A, Farina D, Gazzoni M, Merletti R, Zazula D. Estimating motor unit discharge patterns from high-density surface electromyogram. Clin. Neurophysiol. 2009;120:551–62. Kawakami Y, Ichinose Y, Fukunaga T. Architectural and functional features of human triceps surae muscles during contraction. J. Appl. Physiol. 1998;85:398–404. Merletti R, Knaflitz M, De Luca CJ. Myoelectric manifestations of fatigue in voluntary and electrically elicited contractions. J. Appl. Physiol. 1990;69(5):1810–20. Merletti R, Roy SH, Kupa E, Roatta S, Granata A. Modeling of surface myoelectric signals – Part II: model-based signal interpretation. IEEE Trans. Biomed. Eng. 1999;46(7):821–9. Merletti R, Farina D, Gazzoni M. The linear electrode array: a useful tool with many applications. J. Electromyogr. Kinesiol. 2003;13:37–47. Mesin L, Damiano L, Farina D. Estimation of average muscle fiber conduction velocity from simulated surface EMG in pinnate muscles. J. Neurosci. Methods 2007;160(2):327–34. Mesin L, Merletti R, Vieira TMM. Insights gained into the interpretation of surface electromyograms from the gastrocnemius muscles: a simulation study. J. Biomech. 2011;44(6):1096–103. Narici MV, Binzoni T, Hiltbrand E, Fasel J, Terrier F, Cerretelli P. In vivo human gastrocnemius architecture with changing joint angle at rest and during graded isometric contraction. J. Physiol. 1996;496:287–97. Rainoldi A, Melchiorri G, Caruso I. A method of positioning electrodes during surface EMG recordings in lower limb muscles. J. Neurosci. Methods 2003;134(1):37–43. Rainoldi A, Falla D, Mellor R, Bennel K, Hodges P. Myoelectric manifestations of muscle fatigue in vastus lateralis, medialis obliquus and medialis longus muscles. J. Electromyogr. Kinesiol. 2008a;18:1032–7. Rainoldi A, Gazzoni M, Melchiorri G. Differences in myoelectric manifestations of fatigue in sprinters and long distance runners. Physiol. Meas. 2008b;29(3):331–40. Roy SH, De Luca CJ, Schneider J. Effects of electrode location on myoelectric conduction velocity and median frequency estimates. J. Appl. Physiol. 1986;61(4):1510–7. Staudenmann D, Kingma I, Daffertshofer A, Stegeman DF, Van Dieen JH. Heterogeneity of muscle activation in relation to force direction: a multichannel surface electromyography study on the triceps surae muscle. J. Electromyogr. Kinesiol. 2009;19:882–95. Vieira TMM, Merletti R, Mesin L. Automatic segmentation of surface EMG images: improving the estimation of neuromuscular activity. J. Biomech. 2010a;43(11):2149–58. Vieira TMM, Windhorst U, Merletti R. Is the stabilization of quiet upright stance in humans driven by synchronized modulations of the activity of medial and lateral gastrocnemius muscles? J. Appl. Physiol. 2010b;108:85–97. Vieira TMM, Loram I, Muceli S, Merletti R, Farina D. Postural activation of the human medialis gastrocnemius muscle: are the muscle units spatially localised? J. Physiol. 2011;589(2):431–43. Wu G, Siegler S, Allard P, Kirtley C, Leardini A, Rosenbaum D, et al. ISB recommendation on definitions of joint coordinate system of various joints for the reporting of human joint motion – part I: ankle, hip, and spine. J. Biomech. 2002;35:543–68.

Alessio Gallina obtained, in November 2009, the Bachelor degree in Physiotherapy from the University of Turin, Italy. In February 2012 he completed a master in the field of rehabilitation, hosted by the Neuroscience Department of the University of Pisa. Alessio Gallina has been doing research at the Laboratory for Neuromuscular System Engineering in Torino since January 2010. His research mainly focused on the analysis of the within-muscle distribution of surface electroyographic activity in healthy subjects using high-density systems. Further interests concern the applications of surface electromyography techniques in physiotherapy and in the rehabilitation of musculoskeletal disorders.

325

Cintia Helena Ritzel obtained a degree in Physical Therapy from the Feevale University, Brazil. She is specialist in Kinesiology from Federal University of Rio Grande do Sul, Brazil and specialist in Sports Physiotherapy - SONAFE, Brazil. She obtained a Master degree in Human Movement Science from the Federal University of Rio Grande do Sul, Brazil and a PhD degree in Medicine, Surgery Sciences, from the same university. During her PhD, she spent one year in the Laboratory for Engineering of the Neuromuscular System, Politecnico of Torino, Italy. Currently she is Professor and Research Advisor in Physiotherapy, Postgraduate of Research and Education Institute, Hospital Moinhos de Vento, Brazil. She is also Professor in Physiotherapy, Luterana University of Brazil.

Roberto Merletti is graduated in Electronics Engineering from Politecnico di Torino, Italy, and obtained his M.Sc. and Ph.D. in Biomedical Engineering from the Ohio State University. He has been Associate Professor of Biomedical Engineering at Boston University were he was also Research Associate at the NeuroMuscular Research Center. He is now Full Professor of Biomedical Engineering at Politecnico di Torino where he established, in 1996, the Laboratory for Engineering of the Neuromuscular System (LISiN) of which he is currently Director. He is Senior Member of IEEE, Fellow of ISEK, and member of the Editorial Board of three international journals.

Taian de Mello Martins Vieira is graduated in Physical Education and, in January 2005, obtained his M.Sc. in Biomedical Engineering, from the Federal University of Rio de Janeiro, Brazil. With a doctoral scholarship provided by the Brazilian Research Council (CNPq), at January 2011, he obtained the PhD degree in Biomedical Engineering from the Politecnico di Torino, Italy. Throughout his doctoral studies, he received two student presentation awards by international, scientific societies. Recently, in July 2011, he was the winner of the first edition of the Emerging Scientist Award, sponsored by Prof. Carlo De Luca. Currently, Taian Vieira is reviewer in three peerreviewed international journals and is Assistant Professor within the School of Sports Science, hosted in the Federal University of Rio de Janeiro. His research interest is chiefly focused on the use of electromyography to gain insights into the control of human posture and balance.