EMG-pants in Sports: Concept Validation of Textile-integrated EMG

Measurements

Aljoscha Hermann

and Veit Senner

Professorship of Sport Equipment and Materials, Department of Mechanical Engineering,

Technical University of Munich, Munich, Germany

Keywords: Wearables, Technical Textiles, EMG, Sensors, Textile-Integrated Electrodes.

Abstract: Background: Textile-integrated EMG measurements are attractive for professional or keen amateur athletes

or can be beneficial in rehabilitation or ergonomics. There are several systems commercially available, but

independent validation studies are scarce. This study validates the concept of using textile-integrated

electrodes embedded in garments for EMG recordings of the thigh. Method: a self-produced prototype of

EMG-pants using stainless steel electrodes was developed. 28 participants performed isometric exercises on

an Isomed 2000 device. Measurements of textile-integrated electrodes were compared with measurements

using standard electrodes based on the SENIAM-standard. Results: textile-integrated sensors exhibit linear

behaviour and measurements are comparable to measurements with standard electrodes. Discussion: the

results of this study show that textile-integrated sensors are a valid tool for measuring isometric muscle

activation. For certain research topics or for use in training, the easy application and use of textile-integrated

EMG may be preferable to more scientific tools. Further work is required to address the validity of EMG-

pants for dynamic muscle activation.

1 INTRODUCTION

In the past, measuring muscle activity required

expensive equipment and appropriate experience in

its application. The integration of sensors that

measure muscle activity into garments or wearables

provides individuals with an easy-to-use approach

and thus opens up the technology for new fields of

application in sports, medicine, rehabilitation, and

ergonomics. Electromyography (EMG) is the

standard method for measuring muscle activity. It is

differentiated into two main kinds of EMG: surface

EMG and intramuscular EMG. Only surface EMG

systems are suitable for integration into a sports

garment because intramuscular EMG is invasive. In

surface-EMG, electrodes are applied to the skin

surface to measure muscle activation. For more

details on the principles of muscle activation and the

basics of EMG measurements, the reader is

encouraged to refer to standard literature, e.g. the

recommendations of the Seniam Project (Freriks &

Hermens, 2000) or the EMG-Fible (Konrad, 2011).

https://orcid.org/0000-0003-3168-3273

https://orcid.org/0000-0001-5136-7580

The recently published review paper by Guo et al.

(2020) summarises research into textile electrodes

between 2007 and 2018, listing 41 publications. Of

these, only 14 papers were published in peer-

reviewed journals, showing that textile-integrated

surface EMG and its application remains a niche

subject. Publications relating to sports applications

are even rarer. Excluding papers about the material or

system design, only four publications focus on

garments specifically designed for use in sports (the

authors did not assign the system design paper of

Finni et al. (2007) “sports” category, even though,

the Myontec Mbody pants, shown in Figure 1 (left),

subsequently resulted from it). Of these four papers,

two (Colyer & Mc Guigan, 2018; Tikkanen et al.,

2012) used commercial products (Myontec) in the

study, the other two (Ribas Manero et al., 2016; Shafti

et al., 2017) used the same custom-made pants (same

research group; papers published at the same

conference).

Two examples of commercially available

products for measuring muscle activity in consumer

Hermann, A. and Senner, V.

EMG-pants in Sports: Concept Validation of Textile-integrated EMG Measurements.

DOI: 10.5220/0009982401970204

In Proceedings of the 8th International Conference on Sport Sciences Research and Technology Support (icSPORTS 2020), pages 197-204

ISBN: 978-989-758-481-7

197

Figure 1: Commercially available spots garments for tracking muscle activity of the thighs by Myontec (left) and Athos

(right). Pictures taken from the websites www.myontec.com and shop.liveathos.com on the 21/03/2020.

sports garments by Myontec (Finland) and Athos

(USA) are shown in Figure 1. Both companies

provide pants, as well as shirts, for measuring muscle

activity.

The company Myontec is a spin-off of the

University of Jyväskylä of the group of Prof. Taija

Finni (Juutinen). The company lists 11 publications

on the website backing the design, functionality, and

validity of their product, including the three

publications mentioned above. Athos names one

publication (Lynn et al., 2018) to prove the validity

and reliability of their product. Three of the authors

are employed by the manufacturer.

The publications supporting the applicability of

those products may well be biased due to conflicts of

interest. As these products are health-related,

independent validation is necessary. In general, there

is a lack of studies comparing textile-based EMG

garment solutions with established scientific systems.

The few existing papers largely validate textile

electrodes not yet integrated in a garment. For

example, Li et al. (2011) and Sumner et al. (2013)

confirm that measurements with textile electrodes are

comparable to ones with standard EMG electrodes.

Both studies are limited due to the fact that only six

individuals were tested.

2 VALIDATION STUDY

The purpose of this study is to validate the concept of

using textile electrodes embedded in garments. Self-

designed EMG-measuring pants incorporating

stainless steel electrodes are tested.

2.1 Prototype

In this study self-produced EMG-pants for measuring

quadriceps and hamstrings activation using stainless

steel electrodes are tested. Stainless steel electrodes

are cheaper than the more commonly used silver

electrodes and therefore would be an attractive

alternative. The electrodes consist of a polyester

fabric, with stainless steel wires (Ø=0.05 mm²)

wound twice around the polyester threads. For the

quadriceps, the skin contact area of each of the two

electrodes is 30 mm x 165 mm. For the hamstrings,

the skin contact area of each of the electrodes is

30 mm x 105 mm. The dimensions were chosen to

optimally match the respective muscle regions. A

textile patch was applied to prevent skin contact

(visible in Figure 2 top-left picture) over an additional

area of 30 mm x 30 mm of each electrode. In this area

a button was attached to route the signals to the

outside of the pants for attachment of an EMG sensor

module. The commercially available Myon EMG

Mobile System (Myon, Switzerland) was used as

sensor module for signal amplification and wireless

transmission to the computer. As the measuring-pants

used were not compression pants, standard

compression pants made by 2XU was used as a

second layer to improve electrode-skin contact. The

prototype is shown in Figure 2. Sizes available were

S, M, L and XXL. A pre-test showed good

washability of the electrodes. Three electrodes were

washed in four handwash cycles using chlorine-free

washing solutions. The electrodes showed no change

in electrical resistance. Similarly, the use of

disinfection spray did not alter the electrical

resistance (measured after the pants were dry again).

icSPORTS 2020 - 8th International Conference on Sport Sciences Research and Technology Support

198

2.2 Methodology

2.2.1 Participants

In total 28 participants (9 female/19 male) between 20

and 35 years of age (mean 25.3 years) were enrolled

in the study. Before the start of the experiment all

participants gave their informed consent. The

participants were free of injuries, had not undergone

ACL surgery and had no orthopaedic or

neuromuscular damage to the knee or leg muscles.

Table 1 summarises the activity levels of the

participants.

Participant acquisition was based on the personal

environment and by advertising the study at the

university.

Table 1: Number of days per week with sports activity.

None 1-2

times

3-4

times

> 4

times

Male 2 6 8 3

Female 1 3 2 3

2.2.2 Study Design

Each participant underwent five test scenarios in a

randomised order, with each scenario involving one

of the following electrode settings:

1) Quadriceps: standard electrodes; application

of electrodes as recommended by SENIAM

as shown in Figure 3 in the left picture. Short

pants not interfering with the electrodes.

2) Quadriceps: standard electrodes; row-

application; application of electrodes

corresponding to the arrangement of the

textile electrodes in the measuring pants as

shown in Figure 3 in the middle picture.

Short pants not interfering with the

electrodes.

3) Quadriceps: measuring pants with textile-

integrated electrodes.

4) Hamstrings: standard electrodes;

Application of electrodes on the biceps

femoris and the semitendinosus, as shown in

Figure 3 in the right picture.

5) Hamstrings: measuring pants with textile-

integrated electrodes.

EMG measurements were taken for isometric

muscle activation for five activity levels using a

standard EMG system by Myon (Switzerland) as

shown in Figure 5. The data acquisition rate was set

at 1000 Hz. The knee’s flexion/extension moment

was recorded using an Isomed 2000 dynamometer

(D&R Festl GmbH, Germany). Activity levels were

100%, 80%, 60%, 40% and 20% of maximum

voluntary performance of a flexion/extension knee

moment. Figure 4 shows a typical test protocol. The

participants always started with an application of

100% maximum voluntary performance, followed by

stepwise lowering of the moment and a second

maximum voluntary performance with two more

decreases.

Measurements on the quadriceps were performed

with an angle of 60° between femur and tibia and

measurements on the hamstrings with an angle of 90°

between femur and tibia (180° would be a straight

leg). Those angles were chosen as a compromise

between comfort, optimal angle for the respective

muscle to generate maximum force and the need to

prevent interaction between the seat and the sensor

areas.

A graphical feedback interface helped the

participants to control the activity level. They were

asked to maintain each activity level for 3 sec.

The test procedure steps for each participant were:

1) Selection of the most appropriate size of

pants.

2) Adaption of the settings of the Isokinet to the

participant.

Followed by repetitive steps for the five

randomised test scenarios:

3) Skin preparation according to SENIAM

standards.

4) Application of electrodes/pants.

5) Warm up (flexion/extension movements

against a small force).

6) Test according to the protocol in Figure 4.

2.2.3 Data Processing

All measurements, except the hamstring

measurement of one participant, were completed and

included in the analysis. The EMG data was rectified

and filtered using a second order Butterworth low

pass filter (cut-off frequency 200 Hz) and a root mean

square (window size 20 ms) filter. The sum of the

signals from the three electrode-pairs of the

quadriceps, respectively two electrode-pairs of the

hamstrings (see Figure 3), were calculated to allow

comparison with the signals from the measuring

pants. The signal was cut to segment the activity

levels, using thresholds of 8 Nm for quadriceps and

11.3 Nm for hamstrings measurements, which is

equal to 90% of the smallest respective moment

EMG-pants in Sports: Concept Validation of Textile-integrated EMG Measurements

199

Figure 2: EMG-pants prototype using stainless steel electrodes (top-left). The electrodes are connected via buttons on the

outside of the pants (bottom-left) to a sensor module (top-right) by Myon (Myon EMG Mobile System: amplifier, wireless

transmission). Standard compression pants were used as a second layer (bottom-right) to improve electrode-skin-contact.

Figure 3 Left: SENIAM recommended application for the quadriceps. Middle: row-application; electrodes applied to the

quadriceps corresponding to the electrode layout of the measuring pants. Right: hamstring application (SENIAM).

Figure 4: Typical test record, starting with 100% maximum voluntary performance, followed by stepwise 20% decreases and

a second maximum voluntary performance with two decreases.

measured across all participants. A mean value was

calculated for each activity level measurement. EMG

values and values of flexion/extension moments were

normalised using the higher of the two maximum

voluntary performances of the respective flexion/

extension. For each measurement, a linear regression

(least-square-method) of normalised-EMG vs.

normalised-moment was calculated.

icSPORTS 2020 - 8th International Conference on Sport Sciences Research and Technology Support

200

Average EMG and flexion/extension moment

was calculated for each participant for the SENIAM-

application and the pants. To reduce the influence of

the applied flexion/extension moment, also the

EMG/moment ratio (EMG normalised by the

moment) was calculated. Bland-Altman plots (Bland

& Altman, 1986) were used to compare the two

measurement methods.

2.3 Results

In Figure 5 and Figure 7 the EMG measurements are

plotted against the applied flexion/extension moment

for each activity level of all participants. For

hamstrings and quadriceps, the absolute values of the

pants are smaller than the values of the SENIAM-

electrode-application. The row-application of

standard electrodes on the quadriceps results in the

highest EMG values. As observable in the normalised

data in Figure 6 and Figure 8, all measurement

methods have a similar linear behaviour.

For the hamstrings, the measurement amplitude

for pants is about 60% of the SENIAM amplitude.

In Figure 9, the Bland-Altman plot for the average

EMG values of each participant is shown. The mean

difference between the two systems is 0.237 V, with

a width of the 95% limits of agreement of 0.591 V. In

the Bland-Altman plot in Figure 10 the EMG-

signal/moment ratio is compared. The mean

difference is 0.0043 V/Nm, the width of the 95%

limits of agreement is 0.0102 V/Nm. It is also notable

that the difference between the systems increases with

increasing muscle activity.

3 DISCUSSION

The aim of this study was to validate the concept of

using textile electrodes embedded in garments.

Even though the surface of the textile-integrated

electrodes is bigger, signal amplitudes of the textile-

integrated electrodes are smaller compared to

standard electrodes. Moreover, the Bland-Altman

plots demonstrate an increasing difference between

the signal amplitude of the textile-integrated

electrodes and standard bipolar electrodes for

increasing muscle activation. As both systems

showed a comparable linear behaviour, this

increasing difference does not impact a feasible

textile solution and could easily be addressed by

increased amplification of the signal. Also, by using

gel or wetting the electrodes, signal attenuation could

be decreased (Pani et al., 2019). In this study a very

simple textile prototype was used. Different materials

or more sophisticated textile sensors or improved

integration of the sensors in the garments could also

lead to reduced signal attenuation.

The large electrode surface of our system does not

permit statements about individual muscles, but it

could be even more suitable for the recording of

muscle regions. This is reflected by the fact, that the

row-application of electrodes, which correspond to

the arrangement of textile electrodes in the pants

achieved higher signals than the SENIAM-

application (in both cases the sum of three pairs of

electrodes).

Even though stainless steel electrodes are

washable and corrosion-resistant, they might not

represent the optimum solution for the integration in

pants. Applications using gold, silver, nickel, copper

or carbon-filled silicon can be found (Awan et al.,

2019; Guo et al., 2020; O'Brien et al., 2019). The

choice of material is influenced by multiple and

varying factors such as price, location of integration,

carrier textile, expected mechanical properties (e.g.

flexibility due to the required range of motion), and

biocompatibility. Textile and materials research is

highly active in developing better materials suited for

wearable and textile integration (Gehrke et al., 2019;

Lee et al., 2020). Our electrodes were made of

polyester threads around which metallic threads were

wound. Such threads can break and thus loose

conductivity. Also, they become uncomfortable for

the user because the broken threads penetrate the skin.

There are more suitable production methods, e.g.

coating, for producing electrodes directly on or in the

textile.

When performing EMG measurements, multiple

interfering influences must always be taken into

consideration: crosstalk, individual anatomical and

physiological differences, subcutaneous fat and

fatigue to name but a few. EMG-pants will have to

overcome with all these. In this study, an influence of

the dynamometer electric motor was apparent.

However, because the influence was comparable in

all measurements, we considered it acceptable.

The study cohort generally practice sports, had a

mean age of 25.3 and a mean body-mass-index of

24.2. Therefore, it does not represent the population

in general. Further tests are required especially

involving heavier participants. As surface-EMG is

highly influenced by subcutaneous fat, its

applicability to heavier persons will certainly be

limited. Therefore, wide use of EMG-wearable

textiles in health and rehabilitation will not be without

challenges. For sports applications, especially for

professional or keen amateur athletes, such systems

could be beneficial in training and injury prevention.

EMG-pants in Sports: Concept Validation of Textile-integrated EMG Measurements

201

Figure 5: Hamstrings: regression of non-normalised values

of EMG and flexion moment. Orange: standard electrodes

(SENIAM). Blue: pants.

Figure 6: Hamstrings: regression of normalised values o

EMG and flexion moment. Orange: standard electrodes

(SENIAM). Blue: pants.

Figure 7: Quadriceps: regression of non-normalised values

of EMG and flexion moment. Orange: standard electrodes

(SENIAM). Green: standard electrodes applied in a row.

Blue: pants.:

Figure 8: Quadriceps: regression of normalised values o

EMG and flexion moment. Orange: standard electrodes

(SENIAM). Green: standard electrodes applied in a row.

Blue: pants.

Figure 9: Bland-Altman-Plot of rectified, average EMG

signal of each participant; dots: hamstrings; squares:

quadriceps.

Figure 10: Bland-Altman-Plot of ratio (EMG-values/applie

moment) of each participant; dots: hamstrings; squares:

quadriceps.

icSPORTS 2020 - 8th International Conference on Sport Sciences Research and Technology Support

202

We tested our self-produced EMG-pants in an

isometric setting. Therefore, the interpretation of

results is only valid for the use of such garments in

similar settings e.g. isometric training and

rehabilitation. Nevertheless, there might be some

transferability of results to EMG measurements of

more dynamic movements. Future research must

prove the validity of such measurements because

dynamic measurements are per se subject to more

interfering artefacts (Farina, 2006). In addition to

those artefacts, which also apply to measurements

using standard electrodes, motion artefacts resulting

from electrode-skin displacement might have a strong

impact on the feasibility of garments with textile-

integrated electrodes (Zhang et al., 2011).

4 CONCLUSIONS

The measurements of isometric muscle activation

with our self-produced textile-integrated EMG-

sensors and measurements with standard electrodes

were comparable. Especially when comparing larger

muscle groups, EMG-garments are an easy-to-use

tool for the researcher, athletes or patients undergoing

rehabilitation. Scenarios such as recording the ratio of

quadriceps and hamstrings activation, which do not

need a normalisation procedure, enable athletes to use

the approach without having qualified knowledge

about electrode application procedures and data

processing.

ACKNOWLEDGEMENTS

This research is funded by the Bayerische

Forschungsstiftung, Grant #AZ-1375-19.

REFERENCES

Awan, F., He, Y., Le, L., Tran, L.L., Han, H.D., &

Nguyen, L. P. (2019). ElectroMyography Acquisition

System Using Graphene-based e-Textiles. In 2019

International Symposium on Electrical and Electronics

Engineering: Proceedings : October 10-12, 2019, Ho

Chi Minh City, Vietnam (pp. 59–62). IEEE.

https://doi.org/10.1109/ISEE2.2019.8920937

Bland, J. M., & Altman, D. G. (1986). Statistical methods

for assessing agreement between two methods of

clinical measurement. Lancet (London, England),

1(8476), 307–310.

Colyer, S. L., & Mc Guigan, P. M. (2018). Textile

Electrodes Embedded in Clothing: A Practical

Alternative to Traditional Surface Electromyography

when Assessing Muscle Excitation during Functional

Movements. Journal of Sports Science and Medicine,

17, 101–109.

Farina, D. (2006). Interpretation of the Surface

Electromyogram in Dynamic Contractions. Exercise

and Sport Sciences Reviews, 34(3), 121–127.

Finni, T [T.], Hu, M [M.], Kettunen, P., Vilavuo, T [T.], &

Cheng, S [S.] (2007). Measurement of EMG activity

with textile electrodes embedded into clothing.

Physiological Measurement, 28(11), 1405–1419.

https://doi.org/10.1088/0967-3334/28/11/007

Freriks, B., & Hermens, H. J. (2000). SENIAM: European

recommendations for surface electromyography :

results of the SENIAM project (2nd ed.). Roessingh

Research and Development.

Gehrke, I., Lutz, V., Schmelzeisen, D., Tenner, V., &

Gries, T. (2019). Smart Textiles Production. MDPI.

https://doi.org/10.3390/books978-3-03897-498-7

Guo, L., Sandsjö, L., Ortiz-Catalan, M., & Skrifvars, M.

(2020). Systematic review of textile-based electrodes

for long-term and continuous surface electromyography

recording. Textile Research Journal, 90(2), 227–244.

https://doi.org/10.1177/0040517519858768

Konrad, P. (2011). EMG-FIBEL: Eine praxisorientierte

Einführung in die kinesiologische Elektromyographie.

Version 1.1. http://www.velamed.com/wp-

content/uploads/EMG-FIBEL-V1.1.pdf

Lee, J., Llerena Zambrano, B., Woo, J., Yoon, K., & Lee, T.

(2020). Recent Advances in 1D Stretchable Electrodes

and Devices for Textile and Wearable Electronics:

Materials, Fabrications, and Applications. Advanced

Materials (Deerfield Beach, Fla.), 32(5), e1902532.

https://doi.org/10.1002/adma.201902532

Li, G., Geng, Y., Tao, D., & Zhou, P. (2011). Performance

of electromyography recorded using textile electrodes

in classifying arm movements. Conference

Proceedings: Annual International Conference of the

IEEE Engineering in Medicine and Biology Society.

Annual Conference, 2011, 4243–4246.

https://doi.org/10.1109/IEMBS.2011.6091053

Lynn, S. K., Watkins, C. M., Wong, M. A., Balfany, K., &

Feeney, D. F. (2018). Validity and Reliability of

Surface Electromyography Measurements from a

Wearable Athlete Performance System. Journal of

Sports Science and Medicine, 17, 205–2015.

O'Brien, S., Searle, T., & Alici, G. (2019). Flexible Surface

Electrodes Targeting Biopotential Signals from

Forearm Muscles for Control of Prosthetic Hands: Part

1 – Characterisation of sEMG Electrodes. In 2019

IEEE/ASME International Conference on Advanced

Intelligent Mechatronics, AIM (pp. 1019–1024).

https://doi.org/10.1109/AIM.2019.8868600

Pani, D., Achilli, A., Spanu, A., Bonfiglio, A., Gazzoni, M.,

& Botter, A. (2019). Validation of Polymer-Based

Screen-Printed Textile Electrodes for Surface EMG

Detection. IEEE Transactions on Neural Systems and

Rehabilitation Engineering : A Publication of the IEEE

Engineering in Medicine and Biology Society, 27(7),

EMG-pants in Sports: Concept Validation of Textile-integrated EMG Measurements

203

1370–1377.

https://doi.org/10.1109/TNSRE.2019.2916397

Ribas Manero, R. B [R. B.], Shafti, A [A.], Michael, B.,

Grewal, J., Ribas Fernandez, J. L., Althoefer, K [K.], &

Howard, M. J [M. J.] (2016). Wearable embroidered

muscle activity sensing device for the human upper leg.

Conference Proceedings: Annual International

Conference of the IEEE Engineering in Medicine and

Biology Society. IEEE Engineering in Medicine and

Biology Society. Annual Conference, 2016, 6062–6065.

https://doi.org/10.1109/EMBC.2016.7592111

Shafti, A [Ali], Ribas Manero, R. B [Roger B.],

Borg, A. M., Althoefer, K [Kaspar], & Howard, M. J

[Matthew J.] (2017). Embroidered Electromyography:

A Systematic Design Guide. IEEE Transactions on

Neural Systems and Rehabilitation Engineering : A

Publication of the IEEE Engineering in Medicine and

Biology Society, 25(9), 1472–1480.

https://doi.org/10.1109/TNSRE.2016.2633506

Sumner, B., Mancuso, C., & Paradiso, R. (2013).

Performances evaluation of textile electrodes for EMG

remote measurements. Proceedings of the Annual

International Conference of the IEEE Engineering in

Medicine and Biology Society, EMBS, 6510–6513.

https://doi.org/10.1109/EMBC.2013.6611046

Tikkanen, O., Hu, M [Min], Vilavuo, T [Toivo],

Tolvanen, P., Cheng, S [Sulin], & Finni, T [Taija]

(2012). Ventilatory threshold during incremental

running can be estimated using EMG shorts.

Physiological Measurement, 33(4), 603–614.

https://doi.org/10.1088/0967-3334/33/4/603

Zhang, H., Li, W., Tao, X., Xu, P., & Liu, H. (2011).

Textile-structured human body surface biopotential

signal acquisition electrode. In P. Qiu (Ed.), 4th

International Congress on Image and Signal

Processing (CISP), 2011: 15 - 17 Oct. 2011, Shanghai,

China ; proceedings (pp. 2792–2797). IEEE.

https://doi.org/10.1109/CISP.2011.6100739

icSPORTS 2020 - 8th International Conference on Sport Sciences Research and Technology Support

204