Modifica di Electromyography (sezione)

==General Information==
A signal is, by definition, nothing more than the graphical representation of the temporal trend of a physical quantity. In the case of the surface electromyogram (sEMG), this quantity is the potential difference generated by the muscle during its contraction, which produces an electric current in the tissues and a potential difference that is ultimately recorded on the skin. The graphical representation of this is the electromyogram or electromyographic trace or electromyographic signal or sEMG.

When detecting and recording an sEMG signal, two main aspects that influence the fidelity of the recording must be considered: the signal-to-noise ratio and distortion. The first is defined as the ratio between the energy of the useful signal (i.e., the desired signal) and the energy of the noise. The latter consists not only of actual noise (which we could imagine as the background hiss of old 78 RPM records) but also of any other signal that is simply unwanted, such as cardiac signals, signals from other muscles, or signals due to artifacts. This contamination, although often referred to as noise, should more accurately be termed interference, leaving the term noise for purely thermal noise. Distortion, on the other hand, is an alteration of the useful sEMG waveform that manifests mathematically as an undesired variation in the frequency components of the sEMG signal.

The signal-to-noise ratio and distortion are two problems that, by altering the recorded signal representation, can modify or hide the information the sEMG signal is meant to convey.

It is now well known that the amplitude of the sEMG signal is random in nature and can be represented by a Gaussian distribution. Thus, the sEMG signal is not perfectly predictable a priori, not even by analyzing a segment of the trace immediately preceding the one to be predicted. But this is precisely what one would expect from a signal meant to convey information according to Shannon's theory. It follows, again from information theory considerations, that using the Fourier transform on the sEMG signal is entirely inappropriate.

Perhaps at this point, it is worth commenting on why the Fourier transform does not make sense for random or otherwise stochastic signals. In fact, if we admit that a noisy signal should contain all possible frequencies, then we would expect from it a flat Fourier transform, i.e., one containing all frequencies in the spectrum. However, another noisy signal, temporally different from the first, should then have the same Fourier transform. This would lead to the absurd conclusion that two different signals would have the same Fourier transform. So which signal should be reconstructed from the inverse Fourier transform? It follows that the Fourier transform of random signals is not meaningful for their spectral content, and other methods should be used in this regard.

The amplitude of the sEMG (surface EMG) signal depends on many pathophysiological and technical factors. Excluding the latter, we can consider sEMG signals with a maximum amplitude range (dynamic range in amplitude) from 1.5 mV to 10 mV. The frequencies present in the EMG signal range from 0 to 500 Hz, but the diagnostic and clinically useful band ranges from 50 to 150 Hz. Obviously, only signals in this band with an intensity greater than that of noise in the same band are usable.

===Characteristics of Noise in the sEMG Signal===

====Electronic Noise from the Amplifier====
An unavoidable source of noise is the one intrinsically present in the electronic circuits used for amplifying and conditioning the sEMG signal. This noise has frequencies ranging from direct current (0 Hz) to tens of kHz. To minimize this noise, state-of-the-art amplifier design techniques and high-quality electronic components must be used.

''Electromagnetic Environmental Noise''

Another highly annoying source of noise is environmental noise, originating from electromagnetic radiation (radio, television, cell phones, power distribution lines, electrical and electronic devices, etc.) that continuously inundates the human body when it is in modern urban environments. More accurately, this should be referred to as interference. The most significant type is the so-called 50 Hz noise (60 Hz in the American continent and Japan), caused by electromagnetic emissions from power lines. The 50 Hz noise, also known as "mains hum" or "alternating current hum," is particularly disturbing because it often reaches levels that are 100 to 1,000 times higher than the sEMG signal itself. The fight against mains hum is carried out in various ways, including the design of appropriate amplifiers to minimize the recording of 50 Hz noise, proper electrode placement techniques, and, finally, conducting the recording in specially shielded rooms (Faraday cages).

====Movement Artifacts====
Further disturbances in the faithful recording of an sEMG signal can arise from movement artifacts. This is of particular interest in the recording of surface sEMG because it is evident that movement is inherently generated by the muscle beneath the skin where the electrodes are applied. At least two different types of movement artifacts are described. The first and most obvious is the one that results from a variation in the electrode surface facing the skin. This occurs more easily with large and rigid electrodes rather than small and flexible ones, which can better and more quickly adapt to the changing curvature of the skin over the muscle during contraction. The variation in the electrode surface in contact with the skin produces a sudden change in the electrode's electrical capacitance and, consequently, a variation in the electrode's direct current voltage. The second type is caused by the movement of the cables that connect the electrodes to the amplifier. In this case, the artifact is essentially still due to capacitive variations at the amplifier input, which can be minimized fairly easily with proper design of the amplifier input stages or by shielding the cables. These artifacts typically have a spectrum ranging from 0 to 20 Hz, i.e., outside the useful band for sEMG recording, and can therefore be eliminated by appropriate filtering circuits without significantly altering the useful signal.

====Randomness of the EMG Signal====
The last and less obvious source of noise in the sEMG signal is the quasi-random nature of the sEMG signal itself. This occurs mainly in the 0 to 20 Hz range of the spectrum and is due to the random frequency of motor unit discharges. Motor units, in fact, have an activation frequency precisely in the 0 to 20 Hz range. The unstable nature of these components of the signal should lead them to be considered noise and, therefore, filtered out. This is normally done. Unfortunately, the reader unfamiliar with signal theory may not fully understand this point. Filtering signals in the 0 to 20 Hz band, where the motor unit firing frequencies are present, might seem counterintuitive, leading to the removal of any informative content from the signal.

But a common-life example might help. Imagine listening to loud rock music from the car stereo of a nearby vehicle, both waiting at a traffic light. What you hear is just a rhythmic succession of drumbeats. But if the driver of the nearby car rolls down the window, you can immediately hear the music. Before the window was lowered, you could perceive the rhythm (low frequency) better than the other sounds (which also followed the rhythm of the music but had a higher frequency content), which the rhythm itself prevented you from understanding. In this example, the window had to be lowered (to allow the music through), whereas in the case of sEMG, the rhythm is filtered out (to better see the signal).

===Electrodes===
Everyone knows that the recording of biological electrical signals starts with electrodes, but very few realize the real "necessity" of these. Electrodes seem like something inherent in the recording process, and no one really questions their role.

In reality, the problem is quite simple. The electronic circuits for amplifying and recording sEMG signals are essentially made of electrical wires. These wires are obviously metallic (copper), and common electrical charges of a single type flow through them: electrons. Surely everyone knows that electrons flow in electrical wires. However, few people question whether electrons can also flow in the human body. Certainly, cellular potentials, which are the basis for the potential differences detectable on the skin, cause electric currents, i.e., flows of electric charges. But these charges in the body's tissues cannot be electrons. In fact, it is difficult to find free-moving electrons in the human body, as happens in the metallic lattice of a wire. In our body, we have other carriers of electric charge, which are ions. Ions are "pieces" of molecules with a net electric charge different from zero. They are very different from electrons: they can weigh tens or hundreds of thousands of times more and may have multiple charges compared to an electron and even of opposite sign. Unfortunately, they can only flow in an aqueous environment and certainly not in a wire due to their size. So, the situation is as follows: we have an electric current in the metallic wires of the amplifier, just as we have an electric current in the body's tissues, where the carriers are ions. How can we ensure that the electric charge flows in such a "mixed" circuit? How can we ensure that the carriers exchange electric charge? This is precisely the important role of the electrode. Here, a chemical reaction exchanges electric charges between electrons and ions. The only chemical reaction that does this is the one known as redox (oxidation-reduction). Thus, the purpose of the electrodes is to provide a site for a redox reaction that "closes the circuit" and allows electric charges to flow continuously from the body's tissues to the amplifier and vice versa, thus enabling the biopotentials on the skin to be detected and amplified. It all works as if the electric charge travels on one type of transport (electrons) in one environment and another type of transport (ions) in a different environment. We need a sort of "interchange" where the electric charge can be transferred from one medium to another.

This is why electrodes are so important and not just simple and trivial pieces of wire to connect to the skin.

If no one had yet invented an electrode, one could think of making it as follows. It would seem appropriate to make it in two parts: a metallic part to connect to the wire going to the amplifier and a saline part, attached to the former, capable of participating in the redox reaction. Furthermore, it would be important that the electrode's resistance is as low as possible to avoid excessive voltage drop at the electrode, which would result in a smaller value being measured on the trace. Therefore, a low-resistivity metal (and dermatologically suitable) such as silver should be chosen (not gold, as it is too expensive). For the saline part, a silver salt would obviously be chosen. Which one? Since the electrode is placed on the skin, which is in direct communication with the extracellular fluids of the tissues rich in chloride, silver chloride would be chosen. So the electrode would be made as follows: a small metallic silver plate covered with a layer of silver chloride in the area that comes into contact with the skin. To conclude, a sponge soaked in a silver chloride solution in water could be used to ensure the appropriate mobility of ions. It would be wise to keep the entire setup protected from light since light decomposes silver salts, as you might recall from film photography, which has now disappeared. And so we have "invented" a nice electrode. But how does it work?

The redox reaction that occurs between the electrode and the skin is the following:

<math>AgCl + e^-\Leftrightarrow Ag + Cl^-</math>

and everything seems to work well. In particular, since the reaction is reversible, there is the possibility of current flowing in both directions with the same redox reaction. The electrode is said to be reversible. But what happens if the current flows in only one direction, as in long-duration electromyographic measurements? In this case, the electrode could "wear out," meaning that the chloride layer could dissolve entirely, and the metallic silver would come into direct contact with the skin. Thus, the electrode is said to be consumable. A silver/silver-chloride electrode is both reversible and consumable. The depletion of the electrode is not a positive outcome. To make a measurement with the amplifier, at least two electrodes are needed. Each of them will probably "see" a different concentration of chloride ions in the area where it is placed. This will cause each electrode to generate its own half-cell potential (Nernst equation) different from the other. This potential is also known as the liquid junction potential. Since the two potentials are different, they will not cancel each other out, and thus the measured value will be the muscle potential added to the difference in the half-cell potentials of the electrodes. The muscle electrical potential has values well below a millivolt, while the liquid junction potential has values on the order of volts. This fact makes the measurement somewhat complicated, but it is still possible to manage this phenomenon and obtain good recordings. At least until the electrode is in good condition! Once the chloride is completely depleted, the half-cell potential becomes unpredictable and erratic, depending on other ions present in the area, as well as impurities in the silver. It will be very difficult for the electromyographic amplifier to compensate and overcome this effect. At this point, it is said that the electrode has become polarized and can be discarded without regret.

<gallery mode="slideshow">
File:SnapShot 241014 165659.jpg|'''Figure 1:''' Schematic diagram of the transformation of an ion current (negatively charged) into an electron current (negatively charged) through the exploitation of the redox reaction made possible by the presence of the electrode. 
</gallery>

It would be nice, then, to invent an inexhaustible electrode. One could be made with a plate of metallic platinum. Platinum catalyzes the electrolysis of water (we are obviously in an aqueous environment), and we have the following reaction:

<math>2e^-+2H_2O\rightarrow 2OH^- + H_2</math>

However, this time it is a non-reversible reaction, so if the current direction is reversed, a different reaction occurs:

<math>2H_2O\rightarrow 4H^++O_2+4e^-</math>

Thus, we have an inexhaustible electrode (platinum catalyzes the reaction but does not chemically participate in it, so it does not wear out), but it is irreversible. The production of gas (gaseous hydrogen or gaseous oxygen) during the electrolysis reaction is quite inconvenient because the gas tends to insulate the electrode from the skin, making this type of electrode not particularly useful.

Although there are at least two or three other types of electrodes for electromyography, the Ag/AgCl electrode is the most commonly used and is now sold for just a few dozen cents each.

Historically, an interesting electrode is worth mentioning: the "spray-on" electrode, developed by NASA for monitoring the electrocardiograms of the first astronauts. The spray-on electrode was made by spraying colloidal graphite (carbon powder) onto the skin, effectively painting it. The conductive graphite created an intimate contact with the skin, and a normal metal wire could simply be placed on the "black patch." Today, the spray-on electrode is almost no longer used.