**2.3. Proceeding**

Computational Intelligence in Electromyography Analysis – 72 A Perspective on Current Applications and Future Challenges

modified.

of antagonist muscles.

**2.1. Subjects** 

**2. Material and methods** 

participating in this research.

per unit area to produce large deformation. Hysteresis: represents the energy dissipated between the deformation and recovery of the original shape. Resilience: the amount of energy returned by the deformed material during the discharge phase. Viscosity of a liquid or semiliquid substance is the resistance of a body to deformation in response to a load.

The duration of each phase of the jogging stride (support, flight, and swing) depends on the control of the muscles of each leg executed by the Nervous System: suprasegmental centres, spinal networks, and afferent information from the different senses and from the osteoarticular system. If the information from the feet changes due to modifications on the interface between the foot and the floor, it is expected that the muscular activity will be also

The recording of electrical activity obtained during muscle contraction or electromyogram (EMG) reflects the muscle involvement in the movement of the joints and therefore in the kinetic response of the ground or reaction force. The EMG amplitudes are related both to the nervous system and to muscle tension, although the response of the mechanical system is not directly related to the nervous system signal (Bouisset 1973), as the mechanical response depends on more variables than the muscle activity (such as the length of the muscle, the rate of change of length, time of contraction, and the lever arm magnitude). The relationship between muscle activity and force is not straightforward; however, EMG amplitude, duration, and coordination among different muscles can provide information about the neural and mechanical systems. Therefore, the electromyography is a powerful tool in the study of the neuromusuclar control of movement. The EMG signal is not easily recorded, as it is very susceptible to interferences and cable movement, and it is quite small, varying its amplitude between microvolts and millivolts (Kleissen et al. 1998) being its maximum

amplitude (peak-to-peak) only 5 mV when using surface electrodes (Winter 1979).

electrical activity will provide information about its contribution.

During locomotion, the muscles of the legs are used to meet the following mechanical demands: progression, dynamic stability, and improving the impact and energy conservation (Inman et al. 1981); if any of these tasks is altered, the record of the muscles

The aim of our present study was to evaluate the influence of footwear on the electrical activity of muscles of both legs when running barefoot and running with two different types of sports footwear, in order to assess the effects on: the extent of muscle electrical signal, the profile of muscular electrical activity, the order of muscle involvement, and the coactivation

Ten (six male and four female) healthy subjects, 19 and 20 years old, with an average height of 1.73±0.10 m, with no history of neurological or musculoskeletal dysfunction, voluntarily participated in the study. All of them gave their written informed consent before Prior to obtaining EMG recordings, subjects got used to carrying the electrodes and contact sensors (foot switches –FS) by walking freely in the laboratory until obtaining a normal gait.

For each individual there have been five successive records barefoot at spontaneous speed (no specific speed requested), five with a standard sports shoes and five with his/her own shoes. Subjects ran at ground level at their preferred speed in both the first registration and in the remaining four. When subjects were shod, they were given also some time to get habituated to the shoes, and moved at a freely chosen speed.

Each record registered the EMG corresponding to the cycles required to cross a distance of 10 m. From the cycles registered, clearly identified by the FS signal, only the central 2 were further analyzed; thus avoiding the effect of acceleration and deceleration on muscle activity. In order to calculate the average speed (in ms-1), the time taken to cross the 10 m has been timed. Between each of the five records made for each condition, a pause of one minute was given.

The eight FSs (B & L, U.S.A.) facilitated the identification of the phases of the stride for each of the lower extremities. These sensors are flexible disks of two sizes: 18mm and 30mm. They were placed under the heel (30mm FS) and on the heads of the first and fifth metatarsal and toe tip (18mm FS). When subjects were shod, the FSs were placed on the bottom of the shoes at the sites corresponding to the outer edge of the heel, first and fifth metatarsal, and toe tip. A FS is activated when a pressure greater than 150 g is applied on it.

#### **2.4. EMG**

Visual monitoring of the signal from the FS eased the removal of stride records with deficiencies. The simultaneous recording of signals originating in the FS and the EMG has also allowed identifying the cycle phases (support, swing, or flight) in which the muscles were active. Prior to the start of the records, we checked the signals obtained through the electrodes and the FSs.

The electrode characteristics are presented in **Table 1**. The surface electrodes used were active, equipped with pre-amplifiers providing a gain of x320. EMG signal thus obtained is better than that achieved with passive electrodes respect to the level of noise. They are composed of three stainless steel electrodes, acting two as active electrodes and one as a common ground.


The leads used were bipolar, recording the difference signal between the two active electrodes ends. A general reference electrode was located on the right forearm of each subject.

Comparison by EMG of Running Barefoot and Running Shod 75

**Cross section area [cm2]** 

**FL/ML** 

Crural Nerve L2-L4

L4-S1

**Muscle Mass** 

records.

**Tibialis Anterior** 

**Gastrocnemius** 

and bandage.

**(g)** 

**Muscular Length (ML) [mm]**

**Fibre Length (FL) [mm]**

**Rectus Femoris** 84,3±14 316±5.7 66.0±1.5 5.0±0.0 12.7±1.9 0.209±0.002 **Vastus Medialis** 175±41 335±15 70.3±3.3 5.0±0.0 21.1±4.3 0.210±0.005 **Tibialis Ant.** 65,7±10 298±12 77.3±7.8 5.0±0.1 9.9±1.5 0.258±0.015 **Biceps Femoris** 128±28 342±14 85.3±5 0.0±0.0 12.8±2.8 0.251±0.022 **Semitendinosus** 76.9±7.7 317±4 158±2 5.0±0.0 5.4±1.0 0.498±0.0 **Gastrocnemius** 150±14 248±9.9 35.3±2 16.7±4.4 32.4±3.1 0.143±0.010

The location of the electrodes was done by orienting the surfaces of the electrodes with respect to the direction of muscle fibres (Testut 1971, Wickiewiz et al. 1983, Lieber 1992), in order to obtain a signal of greater amplitude and frequency (De Luca 1997). The electrodes were placed following the recommendations of SENIAM (1999). Electrodes location was verified by performing specific muscular contractions before carrying out the

**Muscles Joint Function (at each Plane) Innervation** 

**Rectus Femoris** hip flexor abductor Crural Nerve L2-L4

**Biceps Femoris** hip extensor adductor external rotator Sciatic Nerve L4-S2 knee flexor external rotator **Semitendinosus** hip extensor adductor internal rotator Sciatic Nerve L4-S2 knee flexor internal rotator

subtalar inverter adductor

**Table 3.** Muscles targeted in this study: function and innervation (Kendall 2000).

knee extensor

**Vastus Medialis** knee extensor

tibiofibular-

**Sagittal Coronal Transversal**

talar flexor Tibial Nerve

knee flexor Tibial Nerve

ankle extensor L4-S3

The optimum recording of action potentials require excellent preparation of the skin before placing on it the electrodes; i.e., waxing the area, removing debris with alcohol, and letting it to air dry. Each electrode was attached to the skin via hypoallergenic tape

**Table 2.** Architectural characteristics of the studied muscles (from Wickiewiz et al. 1983).

**Pennation Angle [º]** 

**Table 1.** Technical specifications of the active, surface electrodes used in this study.

The electrodes chosen for the study were of surface type, because they have the following advantages over intramuscular ones: do not cause pain or bleeding, are easier to apply, and as shown by Kadaba et al. (1985), they provide a more reproducible signal than that obtained with intramuscular electrodes.

The subject carried on his back a box with 14 channels (12 for EMG signal input from the 12 target muscles) and 2 for the signal from the FSs. Each channel had an additional gain range of 1 to 8. The 14 signals were transmitted through a optic fibre cable from the junction box to the electromyograph, a Motion Lab MA200 system equipped with a Pentium PC 64 MB of RAM, a CODAS acquisition card PGH DI 400, with 16 channel and 12 bit resolution. Through the Motion Lab software 900, reports of the actual electrical activity have been obtained and the linear envelop (LE) has been calculated. Afterwards, each LE has been expressed with respect to the normalized gait cycle. The display of the EMG signal in real time while being recorded has allowed us to assess the quality of the recording during movement, ensuring a good contact between the electrode and the skin and the absence of artefacts. The acquisition system (200 Motion Lab) has the following characteristics: level of signal output ± 5 V, bandwidth of 20 Hz to 1000 Hz, CMMR of 40 dB, and input impedance of 100 Ω.

The 12 muscles whose activity has been recorded have been forcibly superficial, as we were using surface electrodes. Four muscles were proximal: rectus femoris (RF), vastus medialis (VM), biceps femoris (BF) long portion, semitendinosus (ST); and two distal: tibialis anterior (TA) and lateral gastrocnemius (GN, or calf). These muscles were selected for their synergistic action and agonist-antagonist relationships. Biarticular muscles: rectus femoris, long head of biceps femoris, semitendinosus, and gastrocnemius; monoarticular: vastus medialis; through one of the joints examined: tibialis anterior. Architectural features of the muscles under study are shown in **Table 2**, and their function and innervation are shown in **Table 3**.


**Table 2.** Architectural characteristics of the studied muscles (from Wickiewiz et al. 1983).

Computational Intelligence in Electromyography Analysis – 74 A Perspective on Current Applications and Future Challenges

Body size 50mm x 18mm x 7mm

Connector LEMO, 5 pin male style Cable Length 60 inches (1.5 metres) Input Impedance Greater than 100 MΩ CMRR >100dB at 60Hz

Bandwidth 10 Hz to 30KHz (-3dB)

Power Requirements ±4V to ±14V at 200μA

Gain at 1kHz 320

obtained with intramuscular electrodes.

of 100 Ω.

**Table 3**.

Weight 30 grams (including connector)

**Table 1.** Technical specifications of the active, surface electrodes used in this study.

The leads used were bipolar, recording the difference signal between the two active electrodes

The electrodes chosen for the study were of surface type, because they have the following advantages over intramuscular ones: do not cause pain or bleeding, are easier to apply, and as shown by Kadaba et al. (1985), they provide a more reproducible signal than that

The subject carried on his back a box with 14 channels (12 for EMG signal input from the 12 target muscles) and 2 for the signal from the FSs. Each channel had an additional gain range of 1 to 8. The 14 signals were transmitted through a optic fibre cable from the junction box to the electromyograph, a Motion Lab MA200 system equipped with a Pentium PC 64 MB of RAM, a CODAS acquisition card PGH DI 400, with 16 channel and 12 bit resolution. Through the Motion Lab software 900, reports of the actual electrical activity have been obtained and the linear envelop (LE) has been calculated. Afterwards, each LE has been expressed with respect to the normalized gait cycle. The display of the EMG signal in real time while being recorded has allowed us to assess the quality of the recording during movement, ensuring a good contact between the electrode and the skin and the absence of artefacts. The acquisition system (200 Motion Lab) has the following characteristics: level of signal output ± 5 V, bandwidth of 20 Hz to 1000 Hz, CMMR of 40 dB, and input impedance

The 12 muscles whose activity has been recorded have been forcibly superficial, as we were using surface electrodes. Four muscles were proximal: rectus femoris (RF), vastus medialis (VM), biceps femoris (BF) long portion, semitendinosus (ST); and two distal: tibialis anterior (TA) and lateral gastrocnemius (GN, or calf). These muscles were selected for their synergistic action and agonist-antagonist relationships. Biarticular muscles: rectus femoris, long head of biceps femoris, semitendinosus, and gastrocnemius; monoarticular: vastus medialis; through one of the joints examined: tibialis anterior. Architectural features of the muscles under study are shown in **Table 2**, and their function and innervation are shown in

ends. A general reference electrode was located on the right forearm of each subject.

Inter-electrode distance Three 1/2" (1.27 cm) disks on 11/16" (0.16 cm) centres

The location of the electrodes was done by orienting the surfaces of the electrodes with respect to the direction of muscle fibres (Testut 1971, Wickiewiz et al. 1983, Lieber 1992), in order to obtain a signal of greater amplitude and frequency (De Luca 1997). The electrodes were placed following the recommendations of SENIAM (1999). Electrodes location was verified by performing specific muscular contractions before carrying out the records.


**Table 3.** Muscles targeted in this study: function and innervation (Kendall 2000).

The optimum recording of action potentials require excellent preparation of the skin before placing on it the electrodes; i.e., waxing the area, removing debris with alcohol, and letting it to air dry. Each electrode was attached to the skin via hypoallergenic tape and bandage.
