ANNALES KINESIOLOGIAE • 1 • 2010 • 1
61
TWITCH PARAMETERS IN TRANSVERSAL AND LONGITUDINAL 
BICEPS BRACHII RESPONSE
Boštjan ŠIMUNIČ 1, Dejan KRIŽAJ 2, Marco NARICI 3, Rado PIŠOT 1
1 University of Primorska, Science and Research Centre of Koper, Institute for Kinesiology Research, 
Slovenia
2 University of Ljubljana, Faculty of Electrical Engineering, Laboratory for Muscle Biomechanics, Slovenia
3 Manchester Metropolitan University, Institute for Biomedical Research into Human Movement and 
Health, United Kingdom
Corresponding Author:
Boštjan Šimunič
University of Primorska, Science and Research Centre of Koper, Institute for Kinesi-
ology Research, Garibaldijeva 1, 6000 Koper, Slovenia
e-mail: bostjan.simunic@zrs.upr.si
ABSTRACT
Assessment of the contractile properties of skeletal muscles is continuing to be an 
important issue and a diffi cult task methodologically. Longitudinal direction of skeletal 
muscle contraction blurs intrinsic muscle belly contractile properties with many fac-
tors. This study evaluates and explains contractile properties such as: delay time (Td), 
contraction time (Tc), half relaxation time (Tr) and maximal amplitude (Dm) extracted 
from twitch transversal response and compare them with torque response. In fi fteen 
healthy males (age 23.7 ± 3.4 years) isometric twitch transversal and torque responses 
were simultaneously recorded during graded electrically elicited contractions in the 
biceps brachii muscle. The amplitude of electrical stimulation was increased in 5 mA 
steps from a threshold up to a maximal response. The muscles’ belly transversal re-
sponse was measured by a high precision mechanical displacement sensor while elbow 
joint torque was calculated from force readings. Results indicate a parabolic relation 
between the transversal displacement and the torque Dm. A signifi cantly shorter Tc 
was found in transversal response without being correlated to torque Tc (r = -0.12; 
p > 0.05). A signifi cant correlation was found between torque Tc and the time occur-
rence of the second peak in the transversal response (r = 0.83; p < 0.001). Electrical 
stimulation amplitude dependant variation of the Tc was notably different in trans-
versal than in torque response. Td was similar at submaximal and maximal responses 
but larger in transversal at just above threshold contractions. Tr has a similar linear 
original scientifi c article UDC: 616.74-073
received: 2010-09-07
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
62
trend in both responses, however, the magnitude and the slope are much larger in the 
transversal response. We could conclude that different mechanisms affect longitudinal 
and transversal twitch skeletal muscle deformations. Contractile properties extracted 
from the transversal response enable alternative insights into skeletal muscle contrac-
tion mechanics.
Keywords: skeletal muscle, contraction time, tensiomyography, torque
NAVOR IN ODEBELITEV TREBUHA MIŠICE BICEPS BRACHII NA 
ELEKTRIČNI DRAŽLJAJ
IZVLEČEK
Merjenje kontraktilnih lastnosti skeletne mišice je pomembno a metodološko zelo 
zahtevno področje. Običajen zajem informacije v vzdolžni smeri mišične akcije zamegli 
intrinzične kontraktilne lastnosti mišičnega trebuha. Študija obravnava kontraktilne 
lastnosti skeletne mišice, kot so: čas zakasnitve (Td), čas krčenja (Tc), polovični čas 
sproščanja (Tr) in največjo doseženo amplitudo (Dm), izračunane iz odziva mišice na 
en električni dražljaj - ‘twitch’ dražljaj. V študiji je sodelovalo petnajst zdravih moških 
preiskovancev (povprečna starost 23.7 ± 3.4 let). Mehanski odzivi mišice dvo-glave 
upogibalke komolca so bili zajeti iz odzivov sile (navora) in prečne odebelitve trebuha 
mišice tekom povečevanja amplitude električne stimulacije od praga krčenja, po 5 mA 
do največjega doseženega odziva. Prečna odebelitev trebuha mišice je bila merjena z 
linearnim senzorjem odmika, medtem, ko je bil navor v komolcu izračunan iz odziva sile 
v vzdolžni smeri. Rezultati pričajo o parabolni odvisnosti med Dm prečne odebelitve in 
vzdolžne sile mišice. Pomembno krajši je bil Tc prečne odebelitve, a ni bil povezan z Tc 
vzdolžnega navora (r = -0.12; p > 0.05). Medtem, ko smo ugotovili pomembno pove-
zavo med Tc vzdolžnega navora in trenutkom pojava drugega vrha prečnega odziva (r 
= 0.83; p < 0.001). Variacija Tc, glede na amplitudo stimulacije, je bila drugačna kot 
pri vzdolžnem navoru. Ugotovili smo da je imel Td obeh metod podobne vrednosti pri 
submaksimalni in maksimalni električni stimulaciji, vendar je bil Td prečnega odziva 
pomembno višji pri nadpražni električni stimulaciji. Ugotovili smo, da je bil parameter 
Tr, izračunan iz obeh metod, linearno odvisen od amplitude stimulacije, vendar je bil 
naklon pri prečnem odzivu 2-krat višji. Zaključimo lahko, da različni mehanizmi vpli-
vajo na deformacijo mišičnih struktur v vzdolžni in prečni smeri. Kontraktilne lastnosti 
izmerjene iz prečnega odziva nam nudijo nove vpoglede v mehaniko mišičnega krčenja.
Ključne besede: skeletna mišica, čas krčenja, tenziomiografi ja, navor
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
63
INTRODUCTION
Skeletal muscle contractile properties are studied in order to make conclusions 
about muscle composition (McComas & Thomas, 1968; Dahmane, Valenčič, Knez 
& Eržen, 2000), exercise velocity (Maffi uletti & Martin, 2001), muscle resistance to 
fatigue (Lindström, Magnusson & Petersén, 1970; Burke, Levine, Tsairis & Engel, 
1971; Merletti, LoConte & Orizio, 1991), motor unit spatial distribution (Dahmane, 
Djordjević, Šimunič & Valenčič, 2005) and fi ring rate (Botterman, Iwamoto & Gon-
yea, 1986; Bernardi, Solomonow, Nguyen & Baratto, 1996). Most exact results would 
be obtained by in vitro experiments using invasive techniques. However, several dif-
fi culties and obstacles surrounding the regular use of invasive techniques result in the 
continuing search for alternative – non-invasive techniques that would improve our 
understanding of the contractile properties of muscles.
Distally detected muscle force/torque response invoked by an electrical twitch 
is frequently used to analyse muscle contractile properties (Hill, 1953; Close, 1972; 
Bülow, Norregaard, Danneskoild & Mehlsen, 1993; Hamada, Sale, MacDougall & 
Tarnopolsky, 2000). Allen, Lee & Westerblad (1989) suggest that such an approach 
should be considered with caution as single twitch stimulation does not release suf-
fi cient intracellular Ca2+ to uncover enough actin-myosin binding sites for development 
and transmission of maximum twitch force. Hoyle (1983) and Kawakami & Lieber 
(2000) further demonstrated that twitch force exerted by a contracted muscle belly has 
to transmit through connective tissue to be measured by an external force transducer. 
Usually there is already a slack of exerted muscle force present before the connective 
tissue is completely stretched and therefore representative muscle belly twitch force 
could not be measured unless longer tetanic stimulation is applied. Beside the elastic 
component of the connective tissue the damping component of the passive surrounding 
tissue also affects longitudinal twitch response making interpretation of the results to-
ward extraction of intrinsic contractile properties of a muscle diffi cult if not impossible.
To reduce these effects, an alternative approach for the evaluation of muscle con-
tractile properties has been introduced through mechanomyographic methods based on 
measurements of lateral vibrations in resonant frequency and thickening of the mus-
cle fi bres. Several different sensors (transducers) have been evaluated, developed and 
tested: phonomyography (Maton, Petitjean & Cnockaert, 1990) and soundmyography 
(Barry, Geiringer & Ball, 1985; Orizio & Veicsteinas, 1992) use microphones to detect 
muscle sound as a consequence of muscle fi bre mechanical oscillations; vibromyog-
raphy (Zhang, Frank, Rangayyan & Bell, 1992) uses accelerometers and laser beams 
(Orizio, Baratta, Zhou, Solomonow & Veicsteinas, 1999 and 2000) to detect thickening 
and vibration of a whole muscle belly. Promising results have been obtained using the 
above methods, however, several diffi culties – mostly of a technological origin – have 
been stated (Orizio, 2000; Wong, 2001): low signal to noise ratio and consecutively 
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
64
high variability, complex measuring set-up, expensive hardware adjustments and nec-
essary post-processing of detected signals, etc.
Valenčič & Knez (1997) introduced another mechanomyographic method named 
Tensiomyography (TMG). TMG assesses muscle mechanical contractile properties with 
a linear displacement sensor applied directly on the skin surface above the muscle belly. 
The method has certain similarities with the intra-muscular pressure measurements (Hill, 
1948; Parker, Körner & Kadefors, 1984; Sejersted et al., 1984) as well as other mechano-
myographic measurements (Barry et al., 1985; Zhang et al., 1992; Orizio et al., 1999). 
The basic difference between the TMG method and other MMG type methods and simi-
larity to the intramuscular pressure method is the slight pre-tension the displacement 
sensor exerts on the muscle belly before the measurement is performed. This pre-tension 
increases the muscle response amplitude when stimulated with an electrical pulse and 
increases the repeatability rate and the signal to noise ratio (Valenčič & Knez, 1997). 
The TMG method was used in several studies (Burger, Valenčič, Marinček & Kogovšek, 
1996; Valenčič & Knez, 1997; Dahmane et al., 2000 and 2005; Kerševan, Valenčič, 
Djordjević & Šimunič, 2002; Grabljevec et al., 2004; Križaj, Šimunič & Žagar, 2008; 
Pišot et al., 2008). First evaluation of the TMG method was performed by comparing 
contraction time extracted from the transversal displacement of the muscle belly to the 
histological fi bre typing analysis obtained from cadaver muscles (Dahmane et al., 2000 
and 2005). A high correlation between the contraction time and the percentage of the type 
I muscle fi bres was found (r = 0.93, Dahmane et al., 2000). The usefulness of the TMG 
method was further demonstrated in monitoring muscle atrophy of above-knee amputees 
(Burger et al., 1996), after 35 days of horizontal bed rest (Pišot et al., 2008), patients 
with spastic muscles (Grabljevec et al., 2004) and muscle adaptation on specifi c training 
process (Kerševan et al., 2002; Dahmane, Djordjević & Smerdu, 2007). The short-term 
repeatability rate of the TMG method has been found very high, making it a reliable es-
timator of muscle contraction properties (Križaj et al., 2008; Tous-Fajardo et al., 2010).
Despite several specifi c investigations and uses of the method has been demon-
strated, so far no detailed explanation of the twitch mechanisms involved in the TMG 
response and its direct relation to the force/torque response of a muscle has been given. 
A goal of this study was to correlate the transversal response measured by the TMG 
method and the torque twitch response and to assess whether TMG may be used as an 
alternative method of the measurement of muscle contraction amplitude and its kinetics.
METHODS AND MATERIALS
Subjects
Fifteen male subjects (Table 1) with no history of neuromuscular disorders volun-
teered for this investigation. The study was approved by the Slovenian National Medi-
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
65
cal Ethics Committee and has therefore been performed in accordance with the ethical 
standards laid down in the 1964 Helsinki Declaration. Each subject was fully informed 
about possible risks and the nature of the experiments and signed the informed consent. 
Each subject was seated in a relaxed position on a straight back chair with elastic straps 
crossing their shoulders to fi x their upper body. Their left arms were isometrically hung 
on a force transducer at 80º of elbow fl exion (0º represents totally extended arm) in a 
pronated wrist position (palm rotated to face down direction).
Torque measurements
An analogue force transducer (FORT 5000, WPI Inc.) was mounted 26.7 ± 1.5 cm 
of the moment arm distally from the elbow joint axis. The forearm was hung on a force 
transducer, which was secured on the other side with a magnetic hand on a heavy metal 
stand. The torque signal was amplifi ed using a DC amplifi er (TGM4M, WPI Inc.).
Tensiomyographic measurements
The principle of the TMG method is presented in Figure 1. A custom-made digital 
displacement sensor (G40, RLS Inc.) was applied perpendicularly to the skin above the 
muscle belly to acquire the transversal response of the biceps brachii muscle. The sen-
sor’s sensitivity was 4 mm and the spring constant was 77 Nm-1.
The measuring point was chosen carefully at the thickest part of the muscle belly 
from where also skin fold data was collected. The thickest part of the muscle belly 
was identifi ed visually and through few measured transversal responses at different 
surrounding locations fi xed at a position where the largest transversal response was 
Figure 1: Principle of the Tensyomyographic method: linear displacement sensor (a) is 
pressed against resting muscle (b) and during twitch muscle contraction muscle belly 
thickens and vertically presses displacement rod (c) to measure a mechanical twitch 
response (d).
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
66
obtained. Two electrodes were placed 5 cm distally (cathode) and proximally (anode) 
from the measuring point. Passive (initial) muscle stiffness (Di) was evaluated in re-
laxed muscle as shown in Figure 1b and as it was previously proposed by Charles et 
al. (2001).
Signal recordings
The transversal and torque signals were simultaneously and continuously stored on 
a portable computer. Torque signal acquisition was performed through the interface and 
analogue/digital converter (DAQCard 1200, National Instruments, Inc.) and sampled 
with a Matlab Compiler Toolbox (The MathWorks, Inc.) with a sampling frequency of 
1 kHz from the 16-bit up/down hardware counter storage. Torque signals were further 
smoothed using a digital low-pass Butterworth fi lter with a cut-off frequency of 25 Hz.
Both recorded signals were further analysed on the basis of seven extracted param-
eters presented in Figure 2.
Figure 2: Explanation of the contractile parameters extracted from the twitch transver-
sal displacement (TD) or torque response.
Contractile parameters were defi ned from the response as follows: Dm – maximal amplitude; Dpp – differ-
ence in the amplitude between both peaks; Tpp – time between both peaks; Td – delay time between stimula-
tion and 10% of the Dm; Tc1-100 – contraction time from 1% and 100% of the Dm; Tc10-90 – contraction 
time from 10% to 90% of the Dm; Tr – half relaxation time from 90% to 50% of the Dm.
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
67
Experiment
Two rounded (5 cm diameter) self-adhesive electrodes (Axelgaard, Pulse) were 
placed bipolarly on the skin above the muscle belly. A cathode was placed distally 
while anode proximally to measuring point. Single monopolar rectangular pulses of 1 
ms duration from the electro-stimulator (TMG-S1, Furlan & Co. Ltd.) were delivered 
to the electrodes. A single pulse (twitch) amplitude was increased every 15 s from the 
threshold to the maximal amplitude in steps of 5 mA. The threshold stimulation ampli-
tude was determined as the amplitude necessary for evoking minimal detectable muscle 
response. Maximal stimulation amplitude was determined as minimum required stimu-
lation pulse amplitude which elicited the largest transversal response. Further increase 
of the stimulation amplitude pulse would not result in the further increase of the trans-
versal response. Typical threshold and maximal stimulation amplitude were around 10 
mA and 65 mA, respectively.
Statistics
The torque peak amplitude values and the transversal peak amplitude values were 
correlated by the equation: y = b + ax2, where y represents the transversal peak ampli-
tude value, x represents the torque peak amplitude value, and a and b are the model 
constants (a is the regression coeffi cient expressing the change in y for a given change 
in x, b is the y-intercept).
Furthermore, a one-way ANOVA was used for comparing the means of two groups 
of data. A Pearson correlation coeffi cient (r) was used to test two series of data for its 
relations. An alpha of p < 0.05 was considered statistically signifi cant for all compari-
sons.
RESULTS
Table 1 presents average data for relative displacement differences between peaks 
(relative Dpp), compression depth (Di) and contraction time (Tc10-90) extracted from 
the transversal response together with anthropometrical (skin fold over biceps brachii 
muscle).
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
68
A typical example of the transversal and the torque responses is presented in Figure 
3. Several differences in the responses can be qualitatively identifi ed: 
- The beginning of the muscle contraction appears sooner in the torque response.
- The peak of the response appears sooner in transversal response
- The relaxation phase of the torque response is shorter.
- Two peaks were identifi ed only in the transversal responses. They were identifi ed 
close to and at maximal responses in 14 subjects.
No statistically signifi cant correlation was found between relative Dpp to the skin 
fold above the biceps brachii muscle belly (r = 0.056) nor to muscle stiffness Di (r = 
-0.043). The highest, but still not signifi cant, correlation was found between the relative 
Dpp and Tc10-90 (r = -0.424).
Table 1: Summary of subjects’ personal, morphological, anthropometrical and trans-
versal displacement characteristics.
Subject Age / years Height / 
cm
Weight 
/ kg
Skin fold / 
mm
Relative Dpp / % Di / mm Tc10-90 
/ ms
1 22 181 80 2.5 44.5 8.7 26.9
2 27 184 85 2.6 24.3 9.5 30.9
3 30 183 95 5 17.7 9.3 28.6
4 26 176 86 6.5 31.3 7.7 25.5
5 23 187 83 3.7 27.0 8.7 29.5
6 22 175 72 3 26.0 5.8 28.1
7 20 184 90 3.5 25.3 8.9 30.4
8 19 188 76 4 21.3 8.7 32.6
9 20 186 75 4.2 45.4 8.3 28.7
10 23 181 72 3.5 21.0 5.2 28.0
11 27 171 78 5 27.9 7.2 28.4
12 25 179 73 3.1 0.0 8.7 34.3
13 24 188 75 4.6 23.8 9.5 25.0
14 28 193 110 5 19.5 8.0 25.3
15 20 189 76 2.7 20.7 8.4 25.9
Average 23.7 183.0 81.7 3.9 25.0 8.2 28.5
SD 3.4 5.9 10.4 1.1 10.7 1.3 2.7
Skin fold was measured above the biceps brachii muscle at the TMG measuring point. From transversal 
response initial muscle belly stiffness (Di) was assessed at relaxed muscle, relative peak-to-peak distance 
(Dpp) and contraction time (Tc10-90) were calculated by normalizing Dpp to Dm and as shown in Figure 2, 
respectively from maximal transversal response.
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
69
Figure 3: Transversal displacement (upper) and torque (lower) twitch muscle biceps 
brachii responses on graded electrical stimulation.
Figure 4: With model and experimentally determined relationship between normalized 
peak transversal displacement and corresponding peak torque twitch amplitude.
Parabolic trend curve models the relationship. The data were pre- normalized between threshold (0) and 
maximal (1) values for each subject and than averages with standard deviations presented.
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
70
The relationship of the peak amplitude (Dm) of the transversal and torque responses 
is presented in Figure 4. Since the magnitudes of the transversal response and the torque 
response differ signifi cantly from subject to subject, they were normalised between the 
threshold and the maximal values and classifi ed into ten levels. The black circles rep-
resent experimental data with standard deviation at ten levels of different transversal 
amplitudes. At just-above-threshold stimulation we could notice a small increase in 
the torque compared to the transversal following with a comparable increase of both 
responses. Parabolic trend was found at 13 and a linear at two subjects, respectively. 
Parabolic curve appropriately models the relationship.
Different mechanisms infl uence transversal and torque twitch responses. Td was 
signifi cantly longer at just-above-threshold transversal response than in torque re-
sponse. At higher electrical stimulation amplitudes, Td in transversal response stabi-
lises at lower value, those comparable to torque Td. Torque Td was found independent 
from electrical stimulation amplitude (Figure 5).
Figure 5: Delay time (Td) for transversal displacement (TD) and torque at different 
normalised response magnitudes (Dm).
Exponential trend for Td in TD response (yTD, circles and full line) and linear trend of torque response 
(yTorque, diamonds and dotted line) describing relation of the parameter Td values regarding to increasing 
magnitude of the responses. Statistical signifi cant differences are presented with ‘*’ symbol.
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
71
Figure 6: Half relaxation time (Tr) for transversal displacement (TD) and torque at 
different normalised response magnitudes (Dm).
Linear trends for Tr in TD response (yTD, circles and full line) and torque response (yTorque, diamonds 
and dotted line) describing relation of the parameter Tr values regarding to increasing magnitude of the 
responses. Statistical signifi cant differences are presented with ‘*’ symbol.
Figure 7: Contraction times (Tc1_100 and Tc10_90) for transversal displacement (TD) 
and torque at different normalised response magnitudes (Dm).
a.) Double exponential and logarithmic trends for both contraction times Tc10_90 (diamonds, full line) and 
Tc1_100 (circles, dotted line) for transversal displacement and b.) for torque, regarding to increasing mag-
nitude of the responses.
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
72
Tr was longer in the transversal response than in the torque response for all subjects. 
Increasing electrical stimulation amplitude signifi cantly prolongs Tr of both transversal 
and torque responses (Figure 6). The rate of increase of the Tr with increasing nor-
malized displacement is about twice as large for the transversal than for the torque 
response.
Two different defi nitions of the twitch response contraction time were investigated 
Tc1-100 and Tc10-90. Both Tc1-100 and Tc10-90 in transversal response were the shortest at the 
just-above-threshold electrical stimulation and were prolonged with increasing stimu-
lation amplitude reaching maximum at approximately 40% of the maximal response 
(Figure 7a). With further increase of the transversal response magnitude at above 80%, 
both of them decrease and later stabilise (Figure 7a). These times could not be corre-
lated to the torque contraction times, which are continually increasing with increasing 
stimulation amplitude (Figure 7b).
Absolute Tc10-90 in transversal response (being between 20 and 40 ms, average at 
maximal response 28.7 ms) was signifi cantly smaller than in torque Tc10-90 (30 to 75 ms, 
average at maximal response 46.8 ms). Similar differences were found in Tc1-100 being 
from 45 to 70 ms (average at maximal response 64.0 ms) in transversal and from 70 to 
120 ms (average in maximal response 101.6 ms) in torque.
From Figure 8 we did not observe a signifi cant correlation (Figure 8a; r = -0.12) 
between Tc1-100 calculated from torque and transversal maximal responses in every par-
ticipant. On the other hand we found signifi cant correlation (Figure 8b; r = 0.82; p < 
0.001) between torque Tc1_100 and the summation of transversal Tc1_100 and Tpp (occur-
rence of the second peak in transversal response) in fourteen participants with second 
peak occurrence.
Figure 8: Correlation between torque and transversal displacement (TD) contraction 
time (Tc1_100) calculated from maximal biceps brachii response.
a.) Non signifi cant correlation coeffi cient (N = 15; r = -0.12) between Tc1_100 of torque and TD maximal 
responses and b.) signifi cant correlation (N = 14; r = 0.83) between Tc1_100 of torque and time of the TD 
second peak occurrence (Tc1_100 + Tpp) maximal responses.
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
73
DISCUSSION
The aim of this work was to check the potential differences in contractile proper-
ties being measured from the transversal and longitudinal direction of in-vivo twitch 
muscle contraction. We have found and explained signifi cant differences between con-
tractile parameters estimated from transversal and torque twitch response.
The sliding of actin and myosin fi laments decreases the length of contractile ele-
ments refl ecting in increased tension at the end of the serial elastic elements. In iso-
metric twitch contraction both, tendon tension and changes of muscle belly diameter 
take place to compensate constant muscle volume. Zhang, Frank, Rangayyan & Bell 
(1996) explain that diameter changes are a mechanical reaction to the production of the 
muscle force but are not directly related to it. Muscle diameter changes have been thus 
far investigated with various transducers and could be explained by a summation of: 
slow bulk lateral muscle belly movement regarding the different regional distribution 
of the contractile elements; the excitation into fast ringing of the muscle fi bres at its 
own resonant frequency, and pressure waves generated by the dimensional changes of 
the active fi bres (Orizio, 1993).
The measurement technique for transversal response detection used in our inves-
tigation is very similar to the technique using a laser beam sensor with one important 
distinction. The high precision displacement sensor used in this study is a contact sen-
sor and uses a spring to compress a relaxed muscle belly towards the bone where it 
compresses subcutaneous fat tissue as well as the muscle belly. By using this principle 
the amplitude of the measured transversal response is increased and the variability of 
the method is reduced (Križaj et al., 2008; Tous-Fajardo et al., 2010).
As demonstrated in Figure 1, at least two simultaneous phases of the transversal 
response can be identifi ed during isometric contraction. In the fi rst phase the muscle 
fi bres’ longitudinal force pulls the fi bres into the horizontal position where the fi bres 
push onto the sensor tip. In the second phase the sensor tip is further pushed due to the 
thickening of the muscle fi bres.
Two obvious peaks of the transversal response (Figures 2 and 3) were not detected 
in other mechanomyographic investigations using laser beams (on isolated rat muscle, 
Orizio et al., 1999 and 2000). Knowing that contact sensors’ mechanical properties 
could infl uence the results we fi rst tried to seek out an explanation within them. The 
fi rst peak was detected at contractions typically larger than 40 % of the supramaxi-
mal response (Figure 3). We did not observe any separation of the sensors’ tip from 
the skin and no signifi cant correlation between the skin fold and the relative Dpp. We 
further performed experiments in which the contact displacement sensor was used si-
multaneously with a laser beam sensor or a high-speed camera (1 kHz) (unpublished 
observations). Both methods clearly showed the occurrence of two peaks in transversal 
response of the human biceps brachii muscle. These observations are in contrast to ex-
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
74
periments on animals made by Orizio et al. (1999 and 2000). On the other hand, Wetzel 
& Gros (1998) observed large differences in time to peak in isolated twitch response 
of fast twitch motor unit (29 ms) and slow twitch motor unit (151 ms) of rat muscles. 
Through modelling of these two fi bre type characteristics, we anticipate that transversal 
response summates bi-modular muscle contraction, where in the fi rst phase fast twitch 
fi bres and in the second phase slow twitch fi bres contribute to overall muscle transver-
sal response (Šimunič, 2003).
A parabolic model was used to correlate the amplitude of the transversal response 
and the torque response. A similar model was found also by other investigators (Orizio 
et al., 1999) using a laser sensor. The correlation could be split into two components; 
the fi rst with a large increment in transversal and low in force Dm and the second with 
small increment in transversal and large in force Dm. The latter occurs at approaching 
the maximal response (Figure 4). These phenomena could be explained with a “toe” 
region of stress-strain relationship in the soft tissues. Orizio et al. (1999) (demonstrated 
on isolated cat gastrochnemius muscle) that force generation takes place after muscle 
belly deformation at low activation level. However, Von Donkelaar, Willems, Muijtjens 
& Drost (1999) showed that during muscle-tendon complex isometric contractions in 
a rat muscle, muscle belly thickening is at some point linearly related to its shortening.
In agreement with Orizio et al. (1999) was also our fi nding on Td (Figure 5). Force 
twitch response at just above threshold stimulation amplitude has signifi cantly shorter 
Td. With increasing stimulation amplitude Td from both methods equalises and stabi-
lises.
Contraction times of the transversal and the torque response show different trends 
at increasing stimulation amplitude. Tc1-100 and Tc10-90 have the shortest values at just 
above threshold stimulation amplitude, largest at 40% of the stimulation amplitude 
and reduces at the maximal stimulation amplitude (Figure 7a). The same behaviour 
was found consistently in our previous study (Dahmane et al., 2005) in nine examined 
human muscles and was explained as a consequence of inhomogeneous fi bre type dis-
tribution. A higher percentage of type 2 muscle fi bres on the muscle surface were also 
reported by other authors (Edström & Kugelberg, 1968; Polgar, Johnson, Weightman & 
Appleton, 1973; Elder, Bradbury & Roberts, 1982).
Recruitment order at nerve and transcutaneous electrical stimulation is still very 
contradictive. Some experiments support reversal motor unit activation order (Heyters, 
Carpentier, Duchateau & Hainaut, 1994; Trimble & Enoka, 2001), while others are 
rather confl icting (Binder-Macleod, Halden & Jungles, 1995; Feiereisen, Duchateau & 
Hainaut, 1997; Knafl itz, Merletti & DeLuca, 1990). Similar to Dahmane et al. (2005) 
and Feiereisen, Duchtateau & Hainaut (1997) we assume that the nerve axons are di-
rectly activated and those that experience a higher electric fi eld should be recruited 
more easily than those deeper in the muscle. However, contraction time of torque twitch 
response was found to have an incremental trend indicating that increased activation 
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
75
of motor units results in longer contraction times (Figure 7b). Assuming reversed re-
cruitment order this could be explained through additional incremental recruitment of 
slower motor units. However, if we assume much more random recruitment order, then 
we must seek the explanation in non-isometric conditions within the whole muscle-
tendon system during twitch contraction. Increased muscle activation results in short-
ening of the muscle belly. Non-isometric conditions during muscle contraction triggers 
muscle mass movement with torsion between active and the surrounding tissue and a 
consequential dumping effect of longitudinal mechanical muscle response. Such effects 
modify (right-shift) the force/torque mechanical response of intrinsic muscle belly.
Contraction times of transversal responses were found to be signifi cantly shorter 
(more than two-fold) than those of torque twitch responses (Figure 7). This result is be-
lieved to indicate important differences in the dynamics between the investigated meth-
ods. Transversal twitch response carries more information on the velocity of the muscle 
contraction as it reaches peak values twice as soon as the torque twitch response. Ex-
periments on isolated muscles using laser beams and a force transducer (Orizio et al., 
1999 and 2000) or video analysis (Van Donkelaar et al., 1999) do not demonstrate such 
delay, which is believed to be due to a more controlled environment without interfer-
ence from the surrounding tissue.
Our investigation indicates that the commonly proposed idea that a longer Tr in-
dicates lower recruitment levels (Fang & Mortimer, 1991) is not necessarily correct. 
Savelberg (2000) also demonstrated that the composition of fi bre type in a muscle af-
fects Tr. Our fi ndings are in agreement with Savelberg (2000). We both found that Tr 
increases with recruitment level when muscle fi bres are orderly recruited (Figure 6). 
Tr increases with the recruitment rate in both approaches but with different slopes. In 
transversal response the Tr increase was steeper than in torque responses. Differences 
in absolute values were statistically signifi cant for all recruitment levels.
In maximal transversal response we typically distinguished two peaks (Figure 2). 
The time occurrence of the second peak was found to correlate signifi cantly to the 
(single) peak in the torque twitch response (Figure 8b). However, transversal responses 
comprise an additional (fi rst) peak that does not correlate to the torque contraction time 
(Figure 8a). This probably demonstrates the loss of information in the torque twitch 
response and raises new questions for explanations of the double-peaked transversal 
twitch response. We previously modelled double peak occurrence in transversal re-
sponse (Šimunič, 2003) and proposed a probable explanation of both peaks by integra-
tive response of the fast- and slow- twitch fi bres. 
CONCLUSIONS
The aim of this investigation was to fi nd explanation(s) for different mechanisms 
affecting the results of the transversal (tensiomyographic method) and the torque re-
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
76
sponse on twitch contraction. Torque measurement is a well established method, widely 
used in the study of muscle physiology while a TMG method is a relatively novel meth-
od based on mechanomyographic methods. The results indicate different mechanisms 
that affect longitudinal and transversal twitch skeletal muscle deformations. Contractile 
properties extracted from the transversal response enable alternative insights in the ki-
netics of skeletal muscle contraction mechanics. The TMG method offers an additional 
insight into muscle physiology and can, with further development and analyses, be-
come an alternative method for the evaluation of skeletal muscle contractile properties. 
Applications and areas where the method can be applied are wide; from studying mus-
cle atrophy kinetics, hypertrophy after trauma or space travel, developmental muscle 
physiology, exercise physiology, etc.
Acknowledgements
This study was partly funded by the Ministry of Higher Education, Science and 
Technology of the Government of the Republic of Slovenia. We are grateful to Profes-
sor V. Valenčič (University of Ljubljana, Slovenia) and Srdjan Djordjević (University 
of Primorska, Slovenia) for supporting this study. We are indebted to TMG-BMC Ltd. 
Company for Biomedical Engineering for supporting us with a TMG device. We are 
thankful to our study participants for their willingness to participate in a study.
REFERENCES
Allen, D. G., Lee, J. A., & Westerblad, H. (1989). Intracellular calcium and ten-
sion during fatigue in isolated single muscle fi bers from Xenopus laevis. Journal of 
Physiology, 415, 433–458.
Barry, D. T., Geiringer, S. R., & Ball, R. D. (1985). Acoustic myography: a noninva-
sive monitor of motor unit fatigue. Muscle & Nerve, 8(3), 189–194.
Bernardi, M., Solomonow, M., Nguyen, G., & Baratta, R. (1996). Motor unit re-
cruitment strategy changes with skill acquisition. European Journal of Applied 
Physiology, 74, 52–59.
Binder-Macleod, S. A., Halden, E. E., & Jungles, K. A. (1995). Effects of stimulation 
intensity on the physiological responses of human motor units. Medicine & Science 
in Sports & Exercise, 27(4), 556–565.
Botterman, B. R., Iwamoto, G. A., & Gonyea, W. J. (1986). Gradation of isometric 
tension by different activation rates in motor units of cat fl exor carpi radialis mus-
cle. Journal of Neurophysiology, 56, 494–506.
Burger, H., Valenčič, V., Marinček, Č., & Kogovšek, N. (1996). Properties of muscu-
lus gluteus maximus in above-knee amputees. Clinical Biomechanics, 11(1), 35–38.
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
77
Burke, R. E., Levine, D. N., Zajac, F. E., Tsairis, P., & Engel, W. K. (1971). Mam-
malian motor units: physiological-histochemical correlation in three types in cat 
gastrocnemius. Science, 174, 709–712.
Bülow, P. M., Norregaard, J., Danneskoild-Samsøe, B., & Mehlsen, J. (1993). 
Twitch interpolation technique in testing of maximal muscle strength: infl uence of 
potentiation, force level, stimulus intensity, and preload. European Journal of Ap-
plied Physiology and Occupational Physiology, 67, 462–467.
Charles, T., Leonard, P. T., Jennifer, U., Stephens, P. T., Sheri, L., & Stroppel, P. 
T. (2001). Assessing the spastic condition of individuals with upper motoneuron 
involvent: validity of the mytonometer. Archives of Physical Medicine and Reha-
bilitation, 82, 1416–1420.
Close, R. I. (1972). Dynamic properties of mammalian skeletal muscle. Physiological 
Reviews, 52, 129–197.
Dahmane, R., Valenčič, V., Knez, N., & Eržen, I. (2000). Evaluation of the ability 
to make non-invasive estimation of muscle contractile properties on the basis of 
the muscle belly response. Medical & Biological Engineering & Computing, 39, 
51–55.
Dahmane, R., Djordjevič, S., Šimunič, B., & Valenčič, V. (2005). Spatial fi ber type 
distribution in normal human muscle histochemical and tensiomyographical evalu-
ation. Journal of Biomechanics, 38(12), 2451–2459.
Dahmane, R., Djordjevič, S., & Smerdu, V. (2007). Adaptive potential of human 
biceps femoris muscle demonstrated by histochemical, immunohistochemical and 
mechanomygraphical methods. Medical & Biological Engineering & Computing, 
45(3), 323–324.
Edström, L., & Kugelberg, B. (1968). Histochemical composition, distribution of fi b-
ers and fatigability of single motor units. Journal of Neurology, Neurosurgery & 
Psychiatry 31: 424–433.
Elder, G. C. B., Bradbury, K., & Roberts, R. (1982). Variability of fi ber type distribu-
tions within human muscles. Journal of Applied Physiology, 53(6), 1473–80.
Fang, Z. P., & Mortimer, J. T. (1991). Selective activation of small motor axons by 
quasi-trapezoidal current pulses. IEEE Transactions in Biomedical Engineering, 
38(2), 175–179.
Feiereisen, P., Duchateau, J., & Hainaut, K. (1997). Motor unit recruitment order 
during voluntary and electrically induced contractions in the tibialis anterior. Ex-
perimental Brain Research, 114, 117–123.
Grabljevec, K., Šimunič, B., Kerševan, K., Križaj, D., Košorok, V., & Gregorič, M. 
(2004). Detection of contractile properties of chronically spastic muscles in subjects 
after traumatic brain injury with Tensiomyographic (TMG) method. International 
Journal of Rehabilitation Research, 27(1), 132–133.
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
78
Hainaut, K., & Duchateau, J. (1992). Neuromuscular electrical stimulation and vol-
untary exercise. Sports Medicine, 14, 100–113.
Hamada, T., Sale, D., MacDougall, J. D., & Tarnopolsky, M. A. (2000). Postactiva-
tion potentiation, fi ber type and twitch contraction time in human knee extensor 
muscles. Journal of Applied Physiology, 88(6), 2131–2137.
Heyters, M., Carpentier, A., Duchateau, J., & Hainaut, K. (1994). Twitch analysis 
as an approach to motor unit activation during electrical stimulation. Canadian Jour-
nal of Applied Physiology, 19, 451–461.
Hill, A.V. (1948). The pressure developed in muscle during contraction. Journal of 
Physiology 107, 518–526.
Hill, A.V. (1953). The plateau of full activity during a muscle twitch. Proceedings of 
the Royal Society of London, Series B. Biological Sciences, 141(905), 498–503.
Hoyle, G. (1983). Muscles and their neural control. New York: Wiley.
Kawakami, Y., & Lieber, R. L. (2000). Interaction between series compliance and 
sarcomere kinetics determines internal sarcomere shortening during fi xed-end con-
traction. Journal of Biomechanics, 33: 1249–1255.
Kerševan, K., Valenčič, V., Djordjevič, S., & Šimunič, B. (2002). The muscle adapta-
tion as a result of pathological changes or specifi c training procedures. Cellular and 
Molecular Biology Letters, 7(2), 367–369.
Knafl itz, M., Merletti, R., & DeLuca, C. J. (1990). Inference of motor unit recruit-
ment order in voluntary and electrically elicited contractions. Journal of Applied 
Physiology, 68(4), 1657–1667.
Križaj, D., Šimunič, B., & Žagar, T. (2008). Short-term repeatability of parameters 
extracted from radial displacement of muscle belly. Journal of Electromyography 
and Kinesiology, 18(4), 645–651.
Lindström, L., Magnusson, R., & Petersén, I. (1970). Muscular fatigue and action 
potential conduction velocity changes studied with frequency analysis of EMG sig-
nals. Electromyography, 10, 341–356.
Maffi uletti, N. A., & Martin, A. (2001). Progressive versus rapid rate of contraction 
during 7 wk of isometric resistance training. Medicine & Science in Sports & Ex-
ercise, 33(7), 1220–1227.
Maton, B., Petitjean, M., & Cnockaert, J. C. (1990). Phonomyogram and electromy-
ogram relationships with isometric force reinvestigated in man. European Journal 
of Applied Physiology, 60, 194–201.
McComas, A. J., & Thomas, H. C. (1968). Fast and slow twitch muscles in man. 
Journal of the Neurological Sciences, 7, 301–307.
Merletti, R., LoConte, L. R., & Orizio, C. (1991). Indices of muscle fatigue. Journal 
of Electromyography and Kinesiology, 1, 20–33.
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
79
Orizio, C., & Veicsteinas, A. (1992). Soundmyogram analysis during sustained maxi-
mal voluntary contraction in sprinters and long distance runners. International Jour-
nal of Sports Medicine, 13, 594–599.
Orizio, C. (1993). Muscle sound: bases for the introduction of a mechanomyographic 
signal in muscle studies. Critical Reviews in Biomedical Engineering, 21, 201–243.
Orizio, C., Baratta, R., Zhou, B. H., Solomonow, M., & Veicsteinas, A. (1999). 
Force and surface mechanomyogram relationship in cat gastrocnemius. Journal of 
Electromyography and Kinesiology, 9, 131–140.
Orizio, C., Baratta, R. V., Zhou, B. H., Solomonow, M., & Veicsteinas, A. (2000). 
Force and surface mechanomyogram frequency responses in cat gastrocnemius. 
Journal of Biomechanics, 33, 427–433.
Orizio, C. (2002). Comments on the letter “Accelerometer and mechanomyogram”. 
Journal of Biomechanics, 3, 385.
Parker, P., Körner, L., & Kadefors, R. (1984). Estimation of muscle force from intra-
muscular total pressure. Medical and Biological Engineering and Computing, 22, 
453–457.
Pišot, R., Narici, M. V., Šimunič, B., De Boer, M., Seynnes, O., Jurdana, et al. 
(2008). Whole muscle contractile parameters and thickness loss during 35-day bed-
rest. European Journal of Applied Physiology, 104(2), 409–414.
Polgar, J., Johnson, M. A., Weigthman, D., & Appleton, D. (1973). Data on the 
distribution of fi ber types in thirty-six human muscles. Journal of the Neurological 
Sciences, 19, 307–318.
Savelberg, H. H. C. M. (2000). Rise and relaxation times of twitches and tetani in 
submaximally recruited, mixed muscle: a computer model. In: Skeletal Muscle Me-
chanics: From Mechanism to Function, edited by Herzog W. Chichester England: 
Wiley & Sons Ltd.
Sejersted, O. M., Hargens, A. R., Kardel, K. R., Blom, P., Jensen, O., & Her-
mansen, L. (1984). Intramuscular fl uid pressure during isometric contraction of 
human skeletal muscle. European Journal of Applied Physiology, 56, 287–295.
Šimunič, B. (2003). Modeliranje vzdolžnih skrčkov in prečnih deformacij skeletnih 
mišic. [Model of longitudinal contractions and transverse deformations in skeletal 
muscles]. Doctoral thesis. Ljubljana: University of Ljubljana, Faculty of Electrical 
Engineering.
Tous-Fajardo, J., Moras, G., Rodríguez-Jiménez, S., Usach, R., Doutres, D. M., 
& Maffi uletti, N. A. (2010). Inter-rater reliability of muscle contractile property 
measurements using non-invasive tensiomyography. Journal of Electromyography 
and Kinesiology, DOI:10.1016/j.jelekin.2010.02.008.
Trimble, M. H., & Enoka, R. M. (2001). Mechanism underlying the training effects 
associated with neuromuscular electrical stimulation. Physical Therapy, 71, 273–
282.
ANNALES KINESIOLOGIAE • 1 • 2010 • 1
Boštjan ŠIMUNIČ, Dejan KRIŽAJ, Marco NARICI, Rado PIŠOT: TWITCH PARAMETERS IN TRANSVERSAL ..., 61–80
80
Valenčič, V., & Knez, N. (1997). Measuring of skeletal muscles dynamic properties. 
Artifi cial Organs, 21, 240–242.
Van Donkelaar, C. C., Willems, P. J. B., Muijtjens, A. M. M., & Drost, M. R. (1999). 
Skeletal muscle transverse strain during isometric contraction at different lengths. 
Journal of Biomechanics, 32, 755–762.
Wetzel, P., & Gros, G. (1998). Decay of Ca2+ and force transients in fast- end slow- 
twitch skeletal muscles from the rat, mouse and Etruscan shrew. Journal of Experi-
mental Biology, 201, 375–384.
Wong, Y. M. (2001). Letter to the editor: Accelerometer and mechanomyogram. Jour-
nal of Biomechanics, 34, 557.
Zhang, Y., Frank, C. B., Rangayyan, R. M., & Bell, G. D. (1992). A comparative 
study of simultaneous vibromyography and electromyography with active human 
quadriceps. IEEE Transactions in Biomedical Engineering, 39, 1045–1052.
Zhang, Y., Frank, C. B., Rangayyan, R. M., & Bell, G. D. (1996). Relationships of 
the vibromyogram to the surface electromyogram of the human rectus femoris mus-
cle during voluntary isometric contraction. Journal of Rehabilitation Research and 
Development, 33(4), 395–403.