4

CHAPTER 2 PHYSIOLOGY OF EMG

Electromyography (EMG), also referred to as myoelectric activity, measures the electrical impulses of muscles at rest and during contraction. As with other electrophysiological signals, an EMG signal is small and needs to be amplified with an amplifier that is specifically designed to measure physiological signals. This signal can be recorded or measured with an electrode, and is then displayed on an oscilloscope, which would then provide information about the ability of the muscle to respond to nerve stimuli based upon the presence, size and shape of the wave the resulting action potential. While the electrode could be inserted invasively into the muscle (needle electrodes), a skin surface electrode is often the preferred instrument, because it is placed directly on the skin surface above the muscle without employing the method of pinch insertion into the test subject. When EMG is measured from electrodes, the electrical signal is composed of all the action potentials occurring in the muscles underlying the electrode. This signal could either be of positive or negative voltage since it is generated before muscle force is produced and occurs at random intervals. The EMG signal is first picked up by electrode and amplified. Frequently more than one amplification stages are needed, since before the signal could be displayed or recorded, it must be processed to eliminate low or high frequency noise, or any other factors that may affect the outcome of the data. The point of interest of the signal is the amplitude, which can range between 0 to 10 millivolts (peak-to-peak) or 0 to 1.5 millivolts (rms). The frequency of an EMG signal is between 0 to 500 Hz. However, the usable energy of EMG signal is dominant between 50-150 Hz [7]. The extended study of EMG signal characteristics can be found at [6] and [8]. In order to obtain a signal that yields the maximum information, the method employed and the implementation device has to be considered. There are many dependent factors that could affect a surface EMG since the signal is susceptible to noise interference such as hum, signal acquisition such as clipping and baseline drift, skin artifacts, processing

5 errors, and interpretation problems. For example, the contact of electrode to the skin could distort a recording signal. The inadequate amplification of the signal could cause a recorder detection problem. A wrong filter could efface some of desirable information of a signal. Moreover, there are other factors such as the distance between electrodes as well as the recording times used in the experiment. The device utilized in the measuring of the signal must also be considered since low-level input into a recording device could also affect data and yield inaccurate results.

2.1 EXCITABLE TISSUE AND ACTION POTENTIAL

There are two main types of tissue in the nervous system: excitable tissue and nonexcitable tissue. The excitable tissue, which is composed of neurons, responds to and transmits nerve stimuli. The non-excitable tissue, composed of glial cells, does not response to voltage or any other conventional stimulus, since glial cells are non-conducting, and function only as support cells in the nervous system. Excitable tissue can be divided into four components: sensory receptors, neuron cell bodies, axons, and muscle fibers [45]. In a situation involving a harmful stimulus such as contact with a sharp pebble or a hot surface, the resulting pain and pressure are transmitted by sensory receptors. The pain is by a receptor potential, which is the transmembrane potential difference of a sensory cell. Produced by sensory transduction, a receptor potential results from inward current flow, which will bring the membrane potential of the sensory receptor toward the threshold to trigger the neuron into generating a rapid burst of voltage pulses called the action potential (APs). As shown in Figure 2.1, triggered by a constant high-pressured stimulus, the sensory receptor generates an initially high receptor potential that rapidly decreases to a much lower, steady level. This decrease in the receptor potential is called adaptation. The action potential produced by the neuron has a magnitude of 0.1 volts [45], which is a value that is shared by every animal, from a squid to a human. To step off that pebble, the neuron sends a message along a nerve axon to the base of the spinal cord. The axon, or nerve fiber, is the slender projection of a neuron that conducts electrical impulses away from the soma, or the nerve cell body. There are two types of axons: the afferent axon and efferent axon. The afferent axon, or sensory axon, leads to the central

Figure 2.1: Typical time variations associated with a sudden, steady stimulus. nervous system, and carries messages from sensory receptors at the peripheral endings to the spinal cord or brain. The efferent axon, or motor axon, originates at the spinal cord and carries information through the body parts, synapse with muscle fibers to stimulate muscular contraction as well as the muscle spindles to alter proprioceptive sensitivity, which is a key factor in muscle memory and hand-eye coordination. Because these two types of axons are designed to relay high-speed messages, their diameter is between 0.001 and 0.022 millimeters, which is longer than ordinary axons, which have a diameter between 0.0003 and 0.0013 millimeters [45]. When compared with the ordinary axons, the efferent and afferent axons also have a thicker layer of myelin, an electrically insulating fatty layer that increases the speed of impulses by means of saltatory conduction. Therefore, by inhibiting charge leakage, myelinated axons propagate action potentials that recur at successive nodes rather than waves, and thus hop along the axon, thereby increasing the speed of the impulse. With a large diameter and thick of myelin sheaths, all signals can thus travel through the afferent and efferent axons at speeds as high as 120 meters per second, or 270 miles per hour [45]. On the other hand, the ordinary axons, which are solely responsible for simple activities such as

7 reporting pain and temperature changes, have small diameters and unmyelinated fibers, which are adequate to carry slow-speed signals. As shown in Figure 2.2, the afferent axon carries the action potential burst from the neuron to the interneuron, a neuron that communicates only to other neurons, or to the motor neuron. This causes a chemical transmitter to be released across a narrow fluid gap called synapses. The latter are specialized junctions that allow neurons to signal to their target cell, which could be another neuron or a non-neuronal cell such as a muscle or gland. The action potential crosses this junction to either another interneuron or a motoneuron, triggering another action potential burst as the process repeats until the message reaches the efferent axon, which then carries the action signal back down to the leg muscle. Once the signal reaches the muscle tissue, the message instructs the muscle to contract, resulting in lifting the foot off the pebble.

Figure 2.2: Excitable tissue called into play when a person steps on a sharp pebble.

2.2 GENERATION OF ACTION POTENTIAL

As stated earlier, after a sensory receptor generates information, this electric signal is transmitted to its intended target by traveling through an axon. However, an axon is a relatively poor conductor because it rapidly attenuates the electrical signal. The potential can decrease to 37 % of its original value after traveling a distance of only 0.15 millimeters along an axon, resulting in an unusable potential value [46]. This distance in which the potential becomes unusable is called the length constant. The length constant is dependent upon the size of the axon, as it is proportional to the square root of an axon diameter. To overcome this tendency of signal attenuation, the nervous system uses a method to increase the strength of the electric signal. When the potential decreases to a threshold level, such as eight millivolts, the neuron will fire another 100 millivolts action potential [46]. However, the action potential will keep decreasing after travel through the axon, which in effect will stimulate the neuron to fire one burst of action potential after action potential, a process that is referred to as frequency modulation. For example, in order to make a potential increase to 10 millivolts, the neuron might fire ten times per second, although the neuron is also able to extinguish voltage in order to end action potential. To get more insight in this process, one must understand the structure of the axon. There are ions arranged in constant random thermal motion inside an axon, with protein molecules being one of the main components of the axon membrane. Under normal conditions, sodium (Na+) and calcium (Ca2+) are more concentrated in the extracellular fluid, while potassium (K+) is more concentrated within the cell. In effect, K+ is the key determinant of the resting membrane potential, since the resting cell membrane is more permeable to K+ than to the Ca2+ and Na+ molecules [14]. However, while it plays a small part in the resting membrane potential, Na+ is a key player in the generation of electric signals. When a cell goes from a resting to an excited state, or firing level, the cell increases its Na+ permeability. This causes Na+ molecules to enter the cell through voltage-gated channels, thus moving down its chemical gradient. This addition of the positive charge of Na+ to the intracellular fluid causes the cell to become depolarized and initiates an action potential. (M.S. Gordon, 1972). The extinguishing level that marks the falling phase of the action potential is the result of an increase in K+ permeability in the cell. However, the closing and opening of the voltage-gated channels is regulated by the jostling of the atoms

9 within the cell, which results in randomness in the train of the generated action potentials. Consequently, any undesired departure from a perfectly ordered system may give rise to socalled noise. Higher receptor potential will initiate less noisy action potentials. However, a noisy system is not always bad, as it enables living things to be able to adjust themselves to changing environment [46]. The action potential is not only in the shape of narrow spike. Alan L. Hodgkin and Andrew F. Huxley (1952) also suggest another model of action potential as illustrated at the top of Figure 3-6 [46]. They applied a +20 millivolts trigger at zero time to the giant axon of the squid, and found that during an action potential, an ion moves to the axon membrane by using its protein molecules to create bridge to the membrane. The protein molecules are unique for each ion species, with the Na+ ions trying to diffuse into the axon while the K+ ions are trying to diffuse out of the axon membrane.

Figure 2.3: Action potential model described by Alan L. Hodgkin and Andrew F. Huxley.

10 From Figure 2.3, the sodium in curve is negative because the current that flows into the axonplasm is defined as negative. On the other hand, the current that flows out of the axonplasm is defined as positive. The sodium carrier proteins convey Na+ ions into the axon in accordance with the sodium in curve. When crossing the membrane, there is only low voltage left to drive the sodium ions into the bridges of the transport proteins. Consequently, the dip of the sodium curve exists at the peak of action potential curve. On the opposite side of the sodium curve, the potassium carrier proteins convey K+ ions out of the axon in accordance with the potassium out curve. To the left of the sodium dip, the Na+ current in is much greater than the K+ current. As a result, the voltage rapidly rises to 100 millivolts above the resting potential. To the right of the dip, the potassium ions are small excess to the sodium ions, which marks the slow drop in voltage.

2.3 PROPAGATION OF ACTION POTENTIAL

In this section, the focus will be on how unmyelinated and myelinated axons generate their action potentials. Regenerating nerve fibers could either be unmyelinate or myelinated [47]. As stated earlier, unmyelinated fibers have thin membranes to carry slow-speed signals. On the other hand, myelinated axons have thick membrane allowing them to carry highspeed signals. In unmyelinated fibers, the AP propagates in the form of an ocean wave. When the voltage across the membrane rises above the threshold level of eight millivolts, a regenerating axon starts to generate an action potential. The thin membrane of the unmyelinated axon allows ions to easily move across the membrane. Figure 2.4 shows the AP waveform initiated by an unmyelinated axon. Please note that the net ion current of Figure 2.4 is based on the model of the impulse of an RC cable, which is a simpler and less accurate model than the Hodgkin and Huxley model. Therefore, the net ion current curve in this section is different from that in section 2.2. Figure 2.4 (a) depicts the net ion current that generates the action potential. The voltage rises to its peak at 100 millivolts before decreasing in value. Figure 2.4 (b) shows the curve of voltage and current versus distance. It is interesting to see the distance and time curve are mirror image of each other. While generally the distance curve can differ from the

11 time curve, the myelinated axon can generate the distance curve and time curve of the action potential that has the same shape and amplitude [47].

Figure 2.4: Net ion current curve and action potential curve of an unmyelinated axon. In comparison to unmyelinated axons, myelinated axons have a thicker wall. This makes it impossible for sodium and potassium ions to move across the axon membrane. As a result, regeneration of the action potential cannot occur. However, the thick membrane also enables the axon to carry high voltage messages without breaking down. (P. Morell and W.T. Norton, 1980) The thick, non-regenerating myelin membrane is offset by periodic nodes, which are also known as nodes of Ranvier [47]. The nodes are 100 outside diameter apart, as illustrated at the top of Figure 2.5. From Figure 2.5, the upper-wave form is an action potential, initiated by the first node on the left. Then, theoretically, the action potential would fall to the dashed curve as showed in the bottom of the picture. Instead, when the action potential reaches eight millivolts, as indicated by the . (dots), the second node fires to generate a new action potential. as shown in the lower wave form. This generation at the node is the same as that of the unmyelinate axon.

12

Figure 2.5: Nodes in the myelinated axon and action potential that generated.

2.4 EMG ELECTRODES

The EMG electrode could be explained by a receiving antenna concept. A receiving antenna is an electrical device that detects oscillating magnetic fields, which are generated from various sources. Then the signal is transmitted through the air from source to the receiving antenna, a concept that is used to engineer the design of electrode. In terms of recording the EMG signal, the muscle fiber is a biological signal generator, spreading out over voltage fields to the volume-conductivity surrounded by fluid [18]. This fluid serves to convey an EMG signal to an electrode, like air carry signals to an antenna. The EMG recording starts from the beginning of the bioelectrical events as shown in Figure 2.6. The changing conductivities in the membranes will make action currents flow across the membranes as well as into the extracellular fluids around active cells. The extracellular currents will then generate potential gradients as they flow through the resistive extracellular fluids. The changing potential gradients, subsequently, will produce electrical currents in the electrode leads by capacitive conductance across the metal/electrolyte

13 interface of the electrode contacts. These weak currents will then flow through the highimpedance circuits of the amplifier input stages, which will then convert these currents into large output voltages.

Figure 2.6: The series of bioelectrical events. The EMG electrodes can be classified by using its geometry. There are six classes of EMG electrodes: monopolar electrode, bipolar electrode, tripolar electrode, multipolar electrode, barrier or patch electrode, and belly tendon electrode. A monopolar electrode takes potential from electrode and ground as the inputs to the differential amplifier. When measuring, only a bare electrode is placed, without utilizing other electrical connection. Because the ground yields a negative input to differential amplifier, the potential from electrode is always based on ground.

14 A bipolar electrode is used to measure the voltage different between two specific points. It generally must be used with a differential amplifier. A bipolar electrode has two contacts that are not connected to each other. Therefore, one node will be used for positive input, and the other will be used for a negative input for the differential amplifier. Because the differential amplifier treats both inputs equally, it will yield an accurate output. However the distance between the electrodes could affect the measurement result. Placing the electrodes too far from one another could yield a weak signal. On the other hand, placing them too close may also result in unusable data, since the amplifier preprocesses each inputs signal separately before subtracting those signals for output. In addition, there must be another contact used as a reference point for these two inputs. A tripolar electrode has three electrodes that are placed at equal intervals along a straight line. The central electrode is usually connected to the positive input of a differential amplifier, while the electrodes on the sides are usually connected to the negative input of a differential amplifier. This configuration also requires another electrode to serve as a reference. The tripolar electrode is often used to record nerve potentials, as its configuration holds the advantage of being able to reject some forms of biological noise. A multipolar electrode consists of rows of bipolar electrodes where an equal lead is connected each side of bipolar electrodes to serve as a positive and negative input for a differential amplifier. Besides, another electrode must be applied as a reference point. The multipolar electrode is often used to record the activity of certain motor units based on idiosyncrasies in their fiber locations. The barrier or patch electrodes are typical bipolar electrodes that are closely connected to a dielectrical barrier. The dielectrical is a non conductive substance that is placed between the electrodes. This configuration redirects currents in extracellular flowing around the tissue nearby. The patch also keeps the currents that are generated from tissues on each side to prevent them from spreading into each other. Consequently, the potential gradient of a desired action is larger, and the potential gradient of an undesired action is smaller. A belly tendon electrode is one of the fields of interest in the clinical EMG. Its geometry is an interesting hybrid of the monopolar and bipolar approaches. In this technique, the first electrode is placed in or over the middle point of the muscle of the belly, which

15 serves as the positive input to the amplifier. The second electrode is placed over the tendon of the same muscle, which is usually about the end of contractile elements, and serves as the negative input to the amplifier. This arrangement gives a clean leading negative waveform, since there is no virtual active contribution from tendon electrode. A belly tendon electrode is employed specifically for tendon applications although it is not used for measuring a selective muscle EMG recording during physiological activity. All of electrode geometries discussed above could be considered as a dipole antenna in term of electrical behavior. Monopolar electrodes are used to measure the EMG signal of very small muscle. This is a good approach for sampling a signal that occurs near the surface of an active single fiber. On the other hand, tripolar electrodes and multipolar electrodes are used for sampling some large muscles.

2.5 AMPLIFIER OF EMG SIGNALS

The amplifier is an electronic device that serves to boost low power signal to higher power signal that is usable to perform work. There are two reasons to amplify the signal [19]. First, amplification increases the level of signal enough to protect an electrical interference during transmission. Second, the signal is amplified so that it could be stored in a storage device, or displayed by a measurement device like oscilloscope. In case of an EMG signal, an amplifier is necessary. There are no such devices that can measure EMG signal without amplification. A differential amplifier is used to amplify an EMG signal, as it has the ability to eliminate the noise from the signal [7]. As shown in Figure 2.7 the differential amplifier takes two inputs, subtracts them and amplifies the different. In this case, if there is noise interference through the input wires, the noise could be circuitry canceled out so long the transmission of the two inputs is completely symmetrical manner. It is difficult to make an amplifier with perfect subtraction. The Common Mode Rejection Ratio (CMRR) could measure the accuracy of subtraction in each amplifier. It is suggested to have a CMRR value at 90dB in order to sufficiently discard a contaminated noise [7]. Yet with modern technology, the differential amplifier could make a CMRR value of 120dB. However, even though a differential amplifier has the ability to reduce unwanted noise signals that occur from both sides of the input wires, contaminated noises could still

16 exist. This noise could have been injected into the signal by a stray capacitance that has been amplified, and thus, degrading the signal [19].

Figure 2.7 A schematic of the differential amplifier configuration. The EMG signal is represented by m and the noise signal by n. Every electronic component or even an amplifier itself behaves as an effective filter, since there are no such electronic devices that can transfer all frequency range. The electrode itself tends to have lower impedance for a higher frequency and have higher impedance for a lower frequency [19]. The connection of electrode, cable and amplifier creates an implicit filter effect. The electrode contacts are connected in series to an amplifier; they function similar to that of a capacitor, while an impedance of amplifier is similar to that of a resistor. This connection visualizes a High-Pass filter circuit. The low frequency voltage tends to be attenuated and drop the highest voltage across the electrode contacts rather than the amplifier. On the other hand, the cables that connect electrodes to an amplifier have a stray capacitor behavior. It is considered that this capacitor is connected to ground, which simulates a Low-Pass filter circuit. The stray capacitor will provide low impedance, at which the high frequency picked up by electrodes tend to drop their voltage here. Therefore an amplifier will see an attenuation of high frequency. The implicit filter could cause signal problems if it is not considered carefully in the design of an amplifier. The explicit filter with real components (resistor and capacitor) functions by using the same concept of an implicit filter, and could help in increasing the signal-to-noise ratio. Since a signal is desired to be within in some frequency range, it is good

17 idea to have an explicit filter for that particular band. Therefore, noise with the frequency outside a desired frequency band will be distorted. (see Figure 2.8)

Figure 2.8 Implicit RC filter. The implicit filter is important in designing a differential amplifier. To reduce an implicit capacitance effect, the electrode contact should be placed close to an amplifier. In other word, an amplifier should be located as close to the signal source as possible. A raw EMG signal is an AC signal. Its bandwidth could occur anywhere from a few tens of cycles per second to 3000 cycles per second. Therefore, sometime a large amount of DC voltage appeared at the output of preamplifier [19]. A DC offset can be removed by adding a series capacitor to the output, which also allows the AC signal to pass through. The DC offset can also happen between the recording electrodes themselves, especially if they are made from different materials. A battery-powered source sometime could overpower a sensitive preamplifier, which could also disgrace the contact of electrodes or damage surrounding tissue. Placing a capacitor between inputs of AC coupling could prevent this

18 problem. (In chapter 3, Figure 3.6, a small capacitor is placed at the inputs to the preamplifier to prevent this problem.) A DC-offset-adjust potentiometer is often added to sensitive equipment, since it is easy to adjust the DC-offset. However, the DC offset also dependents on temperature [19]. Therefore, it is better to power up the equipment for sometime before adjusting the DC offset.

2.6 PROBLEM WITH NOISE AND ARTIFACTS

In the process of recording an EMG signal, the source of the generated signal is not only from bioelectrical generator or active cell, but also from any electrical fields that occur around an electrode and lead cables. These electrical fields produce some signals that could also be added to an EMG signal, causing a form of interference that is called noise. Interference noise can be produced from anything that has an electrical field such as power lines, computer monitors, transformers, or EMG amplifier itself. Once noise has occurred, it could cause problem in recording an EMG signal. Therefore while planning out the design of the amplifier device and recording the EMG signal, noise factors should be taken into consideration, since noise could come from a variety of sources such as electronic components, recording devices, ambient noise, motion artifacts, or inherent instability of the signal [7]. Any electronic devices can produce noise. The noise frequency is range between 0 Hz to a 1000 Hz [7]. This kind of noise cannot be eliminated. Using an intelligent circuit design and a good quality of electronic components to construct the device can only reduce the noise. Ambient noise could be generated from any electronic device that created an electromagnetic field such as televisions, computer monitors, motors, electrical power lines, fluorescent lamps or light bulbs. In fact there are radio waves and magnetic fields floating all over our body. It is virtually impossible to drain these radiators to ground (earth surface). The ambient noise also cannot be avoided. The ambient noise frequency occurs primarily within the range 50 Hz or 60 Hz [7] [44], while the amplitude of an ambient noise is about one to three times grater than that of an EMG signal.

19 Motor artifacts come from two sources [7]; first, from the contact of an electrode to skin; second, from the connection of cable from electrode to the amplifier. When performing a grasping experiment, a movement of the wire alone could cause a noise problem. This electrical noise from both sources has the frequency range between zero to twenty hertz [7]. However, a proper design of circuitry with a good connector and stable electrode contact could reduce this motion artifact problem. Inherent instability of the signal is caused by a nature of EMG signal. The amplitude of the EMG signal frequency range between zero hertz to twenty hertz is particularly unstable due to the quasi-random nature of the firing rate of motor units. It is suggested to consider an EMG signal frequency in this range as an unwanted noise signal [7]. There is no boundary in what amplitude of an EMG signal is good for yielding accurate recordings. While a certain amount of noise could be tolerated, the question is exactly how much noise could be allowed. In other words, there is a question about the tolerable levels of the signal-to-noise ratio. The signal-to-noise ratio is determined by taking the ratio of the amplitude of desire signal over the amplitude of added noise signal. The main concern now is in the lever of the signal-to-noise ratio that could degrade an analysis result. If the noise is produced from thermal motion, then twice of desired signal amplitude over the noise amplitude is good enough. However, if the noise occurred periodically (pause like) and forms a pattern similar to the desired signal, it is suggested that ten times grater of the desire signal amplitude than the noise amplitude would be clarified a confusing event [20]. In order to determine signal-to-noise ratio, the study of noise behavior alone may be required to see how noise could affect the real signal.