**5. EMG signal**

Electromyogram (EMG) is the electrical activity produced by a contracting muscle. EMG signal is extensively used in the field of rehabilitation, biomechanics, orthopedics, ergonomic product design, and prostheses. Due to the fact that EMG allows directly looking into the muscle and measuring the muscular performance, it also helps in decision-making both before and after surgery. The basic functional element that is responsible for producing the EMG signal is called the motor unit (MU). The MU consists of an a - *motorneuron* that has cell body in the spinal cord and extends its axon from the spinal cord to the skeletal muscle fiber (as shown in the **Figure 2**), where it innervates and forms a junction, usually called motor end-plates.

The signal from the a - motorneuron causes depolarization in the muscle fiber that travels in either direction from the junction and creates a potential difference. This difference is measured by the electrodes. The muscle fibers connected to a single neuron react together and hence termed as motor unit. The signal generates by the MU is called motor unit action potential (MUP). The resulting EMG signal is the sum of all activated MUs during the contraction of that muscle. When the signal is acquired using surface electrodes, the signal has to travel through the remaining tissues before reaching the electrodes. This traveling of the signal results in the decaying of the signal amplitude.

**Figure 2.**

*A typical* a - *motorneuron extends from the spinal cords and ends at the motor end-plates in the skeleton muscle fiber [4].*

#### **6. Acquisition of the EMG signal**

The signal generates due to the contraction of the muscle can be detected by placing the electrode on the surface of the muscle. The electrode placement is the main challenge in the acquisition of the surface EMG (sEMG), as the strength of EMG signal varies significantly if the electrodes are slightly displaced from the previous position [5]. In this research, the electrodes have been placed on the triceps and biceps muscles. The placement on the triceps muscle is quite a challenge due to its small size. We searched the spot on the subject for the best EMG signal acquisition. The signal detects by the surface electrodes can be seen in the **Figure** 3**a** and termed as a raw EMG signal.

The raw EMG signal is then passed through an instrumentation amplifier with high common mode rejection ratio (CMRR). The instrumentation amplifier is configured in differential amplification mode to eliminate the noises using the CMRR feature. The gain of the instrumentation amplifier is set as high as 1000 to amplify the minute EMG signal, as the typical EMG signal amplitude measures around 100 mV. After successfully eliminating the common noises and amplification, the amplified EMG signal is passed through a bandpass filter of a low cutoff frequency of 450 Hz and a high cutoff filter of 10 Hz for further filtration. This filtered EMG signal is then passed through a notch filter to eliminate the 50 Hz line frequency. The final filtered EMG signal is then rectified and passes through a smoothing filter to obtain the processed EMG signal. This signal is then used by the controller to execute the control algorithm. **Figure 4** shows the flow of the EMG signal acquisition and process while **Figure 5** shows the EMG signal obtained at each step of the EMG signal acquisition, that is, after the acquisition, then filtration, then rectification, and then smoothing process (moving average filter and RMS).

**15**

**Figure 3.**

**Figure 4.**

*(b) EMG signal after filtration process.*

**6.1 Electrode placement for EMG signal**

*sEMG signal acquisition flow chart.*

Since 1980, researchers are working on the development of protocols and standards for the surface electromyogram (sEMG) electrode placement procedures. Initial attempt was done by the International Society of Electrophysiological Kinesiology in 1980 [6]. After few years, more detailed and in-depth report was published as surface EMG for a non-invasive assessment of muscles (SENIAM).

*The EMG signal at every step of the EMG signal acquisition. (a) Raw EMG of three biceps contraction;* 

*Muscle Mechanics and Electromyography DOI: http://dx.doi.org/10.5772/intechopen.93282* *Recent Advances in Biomechanics*

**6. Acquisition of the EMG signal**

The signal generates due to the contraction of the muscle can be detected by placing the electrode on the surface of the muscle. The electrode placement is the main challenge in the acquisition of the surface EMG (sEMG), as the strength of EMG signal varies significantly if the electrodes are slightly displaced from the previous position [5]. In this research, the electrodes have been placed on the triceps and biceps muscles. The placement on the triceps muscle is quite a challenge due to its small size. We searched the spot on the subject for the best EMG signal acquisition. The signal detects by the surface electrodes can be seen in the **Figure** 3**a** and termed as a raw EMG signal. The raw EMG signal is then passed through an instrumentation amplifier with high common mode rejection ratio (CMRR). The instrumentation amplifier is configured in differential amplification mode to eliminate the noises using the CMRR feature. The gain of the instrumentation amplifier is set as high as 1000 to amplify the minute EMG signal, as the typical EMG signal amplitude measures


V. After successfully eliminating the common noises and amplifica-

tion, the amplified EMG signal is passed through a bandpass filter of a low cutoff frequency of 450 Hz and a high cutoff filter of 10 Hz for further filtration. This filtered EMG signal is then passed through a notch filter to eliminate the 50 Hz line frequency. The final filtered EMG signal is then rectified and passes through a smoothing filter to obtain the processed EMG signal. This signal is then used by the controller to execute the control algorithm. **Figure 4** shows the flow of the EMG signal acquisition and process while **Figure 5** shows the EMG signal obtained at each step of the EMG signal acquisition, that is, after the acquisition, then filtration, then rectification, and then smoothing process (moving average filter

**14**

and RMS).

around 100

**Figure 2.** *A typical* 

a

*muscle fiber [4].*

m

**Figure 3.**

*The EMG signal at every step of the EMG signal acquisition. (a) Raw EMG of three biceps contraction; (b) EMG signal after filtration process.*

**Figure 4.**

*sEMG signal acquisition flow chart.*

### **6.1 Electrode placement for EMG signal**

Since 1980, researchers are working on the development of protocols and standards for the surface electromyogram (sEMG) electrode placement procedures. Initial attempt was done by the International Society of Electrophysiological Kinesiology in 1980 [6]. After few years, more detailed and in-depth report was published as surface EMG for a non-invasive assessment of muscles (SENIAM).

This report was published with the support of European Union. SENIAM was reviewed and further deliberated by a number of laboratories around the world and more refined booklet was published as SENIAM 8: European Recommendations for Surface Electromyography, 1999 [7]. Few of SENIAM recommendations are as follows:


**Figure 5.** *The EMG signal (a) after band pass filter (b) after rectification in blue and EMG signal after smoothing filter in black.*

**17**

**Figure 6.**

*A typical operational amplifier in differential amplifier configuration.*

Section 6.4.

*Muscle Mechanics and Electromyography DOI: http://dx.doi.org/10.5772/intechopen.93282*

resistors as shown in **Figure 6** are same.

Amplifier gain plays an important role in the acquisition of quality and noisefree sEMG signal after the selection of electrode and its placement. The sEMG signal of a normal adult ranges from few hundreds microvolts to 2 mV. Contrary, in athletes, the sEMG signals 5 mV is recorded during maximum voluntary muscle (MVC) [8]. Selection of the amplifier gain considerably depends on the application and requirement of the system. However, for 2 mV input sEMG signal, a gain of 1000 can be used. This gain will results in an output of 2 V. This amount of amplification will also lead in the magnification of noises which can be ignored otherwise. To overcome the amplification of noises acquired with the sEMG signal, one can use the electrodes in differential configuration. In a typical differential amplifier, two electrodes are used with one common or ground electrode. The two electrodes are connected with the positive and negative terminal of an amplifier as shown in **Figure 6**. A typical differential configuration of operational amplifier (op-amp) is shown in **Figure 6**. The gain of the op-amp in differential configuration can be set using the Eq. (1). Whereas the output voltage *Vout* can be calculated with reference to the input voltages *V*1 and *V*2 using Eq. (2) if both the input and feedback

2 1

( 2 1 ) *<sup>f</sup>*

*V VV* = = - (1)

*<sup>R</sup>* = - (2)

*in out*

*out V V <sup>A</sup>*

*R V VV*

*i*

The main advantage of using differential configuration op-amps is to eliminate the noises common to both the input pins. The common noises will be canceled out due to the positive and negative input terminals resulting in a phenomenon known as common mode rejection ratio (CMRR) which is discussed in detail in the

*v*

*out*

**6.2 Selection of amplifier gain**

## **6.2 Selection of amplifier gain**

*Recent Advances in Biomechanics*

recommendations are as follows:

this catalog.

and fixation and testing of the connections.

This report was published with the support of European Union. SENIAM was reviewed and further deliberated by a number of laboratories around the world and more refined booklet was published as SENIAM 8: European Recommendations for Surface Electromyography, 1999 [7]. Few of SENIAM

• sEMG sensor—In sEMG sensor, the SENIAM recommend the electrode shape, size, and constructions for different muscle size and volume. This category also

• Sensor placement—The sensor placement is the most critical part of the sEMG signal acquisition during the preparation of the subjects' skin and selection of sensor location. This category also includes the standards for the placement

• Sensor locations—The last and final standardized category in the SENIAM is the optimal sensor location for the best sEMG signal with minimal crosstalk and artifacts. Almost all the upper and lower body details can be found in

*The EMG signal (a) after band pass filter (b) after rectification in blue and EMG signal after smoothing* 

includes details of electrode material and electrode inner distance, etc.

**16**

**Figure 5.**

*filter in black.*

Amplifier gain plays an important role in the acquisition of quality and noisefree sEMG signal after the selection of electrode and its placement. The sEMG signal of a normal adult ranges from few hundreds microvolts to 2 mV. Contrary, in athletes, the sEMG signals 5 mV is recorded during maximum voluntary muscle (MVC) [8]. Selection of the amplifier gain considerably depends on the application and requirement of the system. However, for 2 mV input sEMG signal, a gain of 1000 can be used. This gain will results in an output of 2 V. This amount of amplification will also lead in the magnification of noises which can be ignored otherwise. To overcome the amplification of noises acquired with the sEMG signal, one can use the electrodes in differential configuration. In a typical differential amplifier, two electrodes are used with one common or ground electrode. The two electrodes are connected with the positive and negative terminal of an amplifier as shown in **Figure 6**. A typical differential configuration of operational amplifier (op-amp) is shown in **Figure 6**. The gain of the op-amp in differential configuration can be set using the Eq. (1). Whereas the output voltage *Vout* can be calculated with reference to the input voltages *V*1 and *V*2 using Eq. (2) if both the input and feedback resistors as shown in **Figure 6** are same.

$$A\_v = \frac{V\_{in}}{V\_{out}} = \frac{V\_{out}}{V\_z - V\_z} \tag{1}$$

$$\mathbf{V}\_{\rm out} = \frac{\mathbf{R}\_f}{\mathbf{R}\_i} (\mathbf{V}\_z - \mathbf{V}\_\mathbf{z}) \tag{2}$$

The main advantage of using differential configuration op-amps is to eliminate the noises common to both the input pins. The common noises will be canceled out due to the positive and negative input terminals resulting in a phenomenon known as common mode rejection ratio (CMRR) which is discussed in detail in the Section 6.4.

**Figure 6.** *A typical operational amplifier in differential amplifier configuration.*

#### **6.3 Input impedance**

Human skin behaves different to different things (such as gel, oil, cream, drugs), especially current and voltages of different frequencies and amplitudes [9]. The resistance experienced by the electrodes during the acquisition of the sEMG signal can be reduced by using the gel in between the electrodes and skin [10].

#### **6.4 Common mode rejection of noises**

The amplification circuit has an ability to reject the common signal to both of its input if the signal is in phase and has same amplitude. This property of the amplifier is referred to as common mode rejection ratio (CMRR). In the acquisition of the biopotentials, like EEG, ECG, and sEMG, this property of the amplifier comes in handy if it is used in the differential configuration. Almost, all the op-amps inherit this property with the variation of the CMRR value. Ideally, an op-amp should have infinite amount of CMRR, but in real, it is limited to the range of 70–120 dB [11]. The typical formula to calculate CMRR is shown in Eq. (3). Where *ADM* is the *differential mode gain* and *ACM* is the *common mode gain*. Eq. (4) shows the CMRR formula to convert it into *deci Bell*.

$$\text{CMRR} = \frac{A\_{\text{DM}}}{A\_{\text{CM}}} \tag{3}$$

$$\text{CMR}(dB) = 20\log\_{10}\left(\text{CMRR}\right) \tag{4}$$

#### **6.5 Crosstalk of the EMG signal**

It is quite difficult to identify the muscle when the sEMG signal acquired from the skin surface which contains a bunch of muscles underneath especially the forearm muscles. A cross-sectional view of the forearm is shown in **Figure 7**. The sEMG signal acquired from the forearm will be resulting from the contraction of multiple muscles due to the reason that the muscles are bundled and overlap each other in the forearm area. It is quite impossible to avoid crosstalk in this area as the active small motor unit range is around 0.5 cm and large motor unit range is around 1.5 cm [12]. Therefore, the electrode placed on the surface will acquire the sEMG signal produced by the contraction of multiple muscles' motor units which are under it. This crosstalk can be minimized by the following techniques:


Manual identification of the muscle resistance is not feasible in all cases. Therefore, the most successful technique is to use the cross-correlation method [13]. A general formula for the detection of cross-correlation is shown in Eq. (5); however, detail discussion is beyond the scope of the book.

$$R\_{\rm xy} \left( \tau \right) = \frac{\frac{1}{T} \int\_{\circ}^{T} \varkappa(t) \, \varkappa(t - \tau) dt}{\sqrt{R\_{\rm xx} \left( 0 \right) R\_{\rm y} \left( 0 \right)}} \tag{5}$$

**19**

**Figure 8.**

*Muscle Mechanics and Electromyography DOI: http://dx.doi.org/10.5772/intechopen.93282*

**7. EMG signal processing**

*Cross-sectional view of human forearm, image courtesy [14].*

**Figure 7.**

• Filters (bandpass)

• Rectification

• Smoothing filters

of 10–250 Hz may be used.

*The signal processing blocks in a typical sEMG acquisition system.*

**7.1 Filters—bandpass**

• Notch Filter (50 Hz)

After the acquisition and amplification of the sEMG signal, the signal processing has been started as shown in **Figure 4**. The following blocks are essential for the signal processing of the sEMG signal, and the sequential process is shown in **Figure 8**.

Not only the amplitude of the sEMG signal depends on the site of acquisition but also the frequencies vary with it. Other factors depend on the subject physical health as the sEMG signal of athletes has higher amplitude and frequencies as compared with the normal subject [8]. The recommended cutoff frequencies setting for the bandpass filter are from 10 to 500 Hz [7]. However, in some cases, a tight band

### *Muscle Mechanics and Electromyography DOI: http://dx.doi.org/10.5772/intechopen.93282*

#### **Figure 7.**

*Recent Advances in Biomechanics*

**6.4 Common mode rejection of noises**

formula to convert it into *deci Bell*.

**6.5 Crosstalk of the EMG signal**

crosstalk can be minimized by the following techniques:

ever, detail discussion is beyond the scope of the book.

• Using cross-correlation technique of signal processing.

( )

t

*xy*

Manual identification of the muscle resistance is not feasible in all cases. Therefore, the most successful technique is to use the cross-correlation method [13]. A general formula for the detection of cross-correlation is shown in Eq. (5); how-

> 0 1 *<sup>T</sup>*

*x t y t dt <sup>T</sup> <sup>R</sup> R R*

( ) ( )

t

<sup>=</sup> <sup>ò</sup> (5)


(0 0 ) ( )

*xx yy*

• Manually checking the muscle resistance

Human skin behaves different to different things (such as gel, oil, cream, drugs),

The amplification circuit has an ability to reject the common signal to both of its input if the signal is in phase and has same amplitude. This property of the amplifier is referred to as common mode rejection ratio (CMRR). In the acquisition of the biopotentials, like EEG, ECG, and sEMG, this property of the amplifier comes in handy if it is used in the differential configuration. Almost, all the op-amps inherit this property with the variation of the CMRR value. Ideally, an op-amp should have infinite amount of CMRR, but in real, it is limited to the range of 70–120 dB [11]. The typical formula to calculate CMRR is shown in Eq. (3). Where *ADM* is the *differential mode gain* and *ACM* is the *common mode gain*. Eq. (4) shows the CMRR

> *DM CM*

*<sup>A</sup>* <sup>=</sup> (3)

*CMR dB CMRR* ( ) = 20log<sup>10</sup> ( ) (4)

*<sup>A</sup> CMRR*

It is quite difficult to identify the muscle when the sEMG signal acquired from the skin surface which contains a bunch of muscles underneath especially the forearm muscles. A cross-sectional view of the forearm is shown in **Figure 7**. The sEMG signal acquired from the forearm will be resulting from the contraction of multiple muscles due to the reason that the muscles are bundled and overlap each other in the forearm area. It is quite impossible to avoid crosstalk in this area as the active small motor unit range is around 0.5 cm and large motor unit range is around 1.5 cm [12]. Therefore, the electrode placed on the surface will acquire the sEMG signal produced by the contraction of multiple muscles' motor units which are under it. This

especially current and voltages of different frequencies and amplitudes [9]. The resistance experienced by the electrodes during the acquisition of the sEMG signal

can be reduced by using the gel in between the electrodes and skin [10].

**6.3 Input impedance**

**18**

*Cross-sectional view of human forearm, image courtesy [14].*
