* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

# Download 1 Basic Electronics in Clinical Neurophysiology

Oscilloscope history wikipedia , lookup

Analog-to-digital converter wikipedia , lookup

Radio transmitter design wikipedia , lookup

Josephson voltage standard wikipedia , lookup

Nanofluidic circuitry wikipedia , lookup

Nanogenerator wikipedia , lookup

Crystal radio wikipedia , lookup

Spark-gap transmitter wikipedia , lookup

Regenerative circuit wikipedia , lookup

Galvanometer wikipedia , lookup

Schmitt trigger wikipedia , lookup

Power electronics wikipedia , lookup

Electrical ballast wikipedia , lookup

Power MOSFET wikipedia , lookup

Operational amplifier wikipedia , lookup

Index of electronics articles wikipedia , lookup

Current source wikipedia , lookup

Surge protector wikipedia , lookup

Valve RF amplifier wikipedia , lookup

Switched-mode power supply wikipedia , lookup

RLC circuit wikipedia , lookup

Resistive opto-isolator wikipedia , lookup

Current mirror wikipedia , lookup

Network analysis (electrical circuits) wikipedia , lookup

1 Basic Electronics in Clinical Neurophysiology Christopher M. Sinclair, Mason C. Gasper, and Andrew S. Blum Summary A basic understanding of simple electronics is vital for the student of clinical neurophysiology to better understand how we begin to analyze neurobiological systems. The elements of basic circuits have relevant and tangible application to the way in which we model the behavior of neural systems in the laboratory. This chapter helps to define and assemble these varied circuit elements for the student. This base of understanding is then used to illustrate how simple electronic circuits can filter and amplify biological data. The composition and behavior of commonly used electrodes are discussed, as are the varied montages we use to record and/or display the measured data, as in an EEG. Attention is devoted to digital signal analysis because modern clinical neurophysiology increasingly relies on digital sampling for ease of data analysis and storage. Lastly, electrical safety issues are considered, particularly as they apply to the clinical neurophysiology arena. Key Words: Amplifier; circuit element; digital conversion; electrical safety; electrode; electronic filter; montage. 1. GENERAL PRINCIPLES An understanding of the nature of electricity and the behavior of charged particles begins with one fundamental principle—like charges repel and opposite charges attract. If a collection of charges, whether positive or negative, are unevenly distributed, there is an inherent drive for those charges to redistribute to achieve electrical neutrality. This drive may be considered the electrical potential. The MKS (meter-kilogram-second) unit of energy (E) is the joule (J). One joule is defined as the energy required to accelerate a 1-kg mass by 1 m/s2 over a distance of 1 m. The unit of charge (Q) is the coulomb (C). One coulomb is defined as 6.24 × 1018 individual units of charge, where a single electron carries one unit of negative charge. Separated charges (that have not achieved electrical neutrality) are a form of stored or potential energy, and this energy will be expended as the charge separation is neutralized. In the MKS system, 1 J of energy is needed to separate 1 C of charge against an electrical potential of 1 V. Stated more concisely, 1 J = 1 V × 1 C. 2. CURRENT The flow of electrons in response to an existing or applied electrical potential, or voltage, is known as current. Current (I) is simply some quantity of charge (Q) moving in some quantity of time (t). Mathematically, this is expressed as: I = Q/t From: The Clinical Neurophysiology Primer Edited by: A. S. Blum and S. B. Rutkove © Humana Press Inc., Totowa, NJ 3 4 Sinclair, Gasper, and Blum where I is the current in amperes (A), Q is the quantity of charge in coulombs, and t is the time in seconds required for the transfer of charge. The current must travel through a medium that consists of other particles, and this medium may interfere with the efficient flow of charge; it presents resistance (R) to that flow. Thus, the current is not only affected by the applied potential but also by the amount of resistance in the conducting medium. Various media conduct electricity with variable efficiency. Metals conduct very well because of their abundant free electrons and, thus, are termed conductors. Conversely, materials that lack free electrons to facilitate the flow of charge resist this flow, and are known as insulators. Although the flow of electricity is achieved through the movement of electrons, current is conventionally described to flow from the positive pole to the negative pole. Thus, the direction of current refers to the movement of positive rather than negative charge. Current may consist of other forms of charge apart from electrons. Current may also be conveyed by ions (regardless of charge polarity) in a tissue or solution, as is the case in the conduction of muscle or nerve potentials. 3. CIRCUIT ELEMENTS 3.1. Resistors Under everyday conditions, current meets with some resistance to flow, much as friction opposes the movement of an object over a surface. Some energy or force is expended in overcoming this resistance. The voltage (or potential) difference across a given resistance is known as the voltage “drop,” and the relationship between these parameters and the resultant current is given by Ohm’s law: V = IR The unit of resistance is the ohm (Ω), which is defined as the resistance (R) that will dissipate 1 J of energy when a current of 1 A flows through it for a period of 1 s. Practically speaking, resistors are made from materials that do not easily allow the free movement of electrons, such as carbon. Very high resistance materials that are the most restrictive toward the movement of electrons, such as air, rubber, or glass, make the best insulators. The greater the distance that current must traverse through a resistive material, the more resistance to flow there will be. It is, thus, useful to alter the length of a resistive material to vary the current flow. As given by Ohm’s law, resistance and current vary inversely with one another (R = V/I). Therefore, a reduction in the length of a resistive medium by half will lead to a doubling of the current. The potentiometer (voltmeter) uses this principle by providing a way to vary the length of a resistor (and thereby vary the current flow) to advantage. Resistance in the acquisition of a biological test, such as an EEG, does not only derive from the material of the wiring in use. Resistance derives from any material through which current must pass. For example, resistive elements in the EEG include not just the electrode wiring but also the scalp–electrode interface and the internal circuitry of the machine. Resistance is provided by anything that lies between the positively charged terminal of a circuit (the cathode) and the negatively charged terminal (the anode). If the resistance is infinitely large, then the current becomes infinitely small (or ceases). This produces a circuit that is “open.” Circuit breakers act in this way to ensure the safety of an electrical system. If resistance is reduced to a miniscule value, this permits a relatively large current, and is deemed a “short circuit.” Any resistance between the anode and cathode that allows current to flow but is neither infinitely large nor extremely small is a “closed circuit.” Basic Electronics in Clinical Neurophysiology 5 Fig. 1. This figure contrasts the organization of a series circuit (A) and a parallel circuit (B). In a series circuit, equal current must flow through each resistor in turn. Therefore, the resistors function as a voltage divider. The resistance, Rcomb, is given by R1 + R2 + R3. The parallel circuit functions as a current divider, with equal voltage across each resistor. The combined resistance is given by 1/Rcomb = 1/R1 + 1/R2 + 1/R3. Each element in a circuit contributes its own resistance. If multiple resistive elements exist in a succession along a circuit, they are said to be in series (Fig. 1A). If they are configured to allow current to travel in multiple alternate paths, they are said to be in parallel (Fig. 1B). Because the series configuration fractionates the total voltage across each of the resistive elements, it is also known as a voltage divider. Addition of these resistive elements creates a resistor of greater length that is equivalent to the sum of all the component resistances. Therefore, the equivalent resistance (Req) for a series circuit may be obtained by summing the individual resistances in the circuit as such: Req = R1 + R2 + R3 By contrast, a parallel circuit will allow current to fractionate and travel any of a number of paths, and, therefore, is known as a current divider. The several routes that the current may travel effectively reduces the total resistance to flow to less than that of any of the component resistances in the circuit. This is represented by the following relationship: 1/Req = 1/R1 + 1/R2 + 1/R3 In considering a complete circuit, there are two other applicable laws. Kirchoff’s current law states that the sum of current flowing into and out of any circuit node must be zero. Kirchoff’s voltage law states that the sum of all voltage steps (voltage sources and drops) around a complete circuit must be zero. 3.2. Capacitors A capacitor is a device that permits the storage of charge. It consists of two parallel conducting plates closely apposed to one another but separated by a small distance and an interposed insulating material, the dielectric. The gap between the plates provides a large resistance to the flow of current from plate to plate. As such, when a potential is applied across a circuit containing a capacitor, positive charge will accumulate on the positive plate, attracting negative charge to the opposite plate. Current flows between the plates via the circuit without charge actually crossing the dielectric gap between the plates. The accumulation of separated charge creates a potential difference across the plates that eventually balances the potential applied across the circuit, and current flow then ceases. Several factors affect the magnitude of charge, or capacitance, that may be stored by a capacitor. This is proportional to the size of 6 Sinclair, Gasper, and Blum the plates of the capacitor, inversely proportional to the distance between those plates, and is affected by the dielectric material between the plates. The MKS unit for capacitance is the farad (F). A farad will store 1 C of charge on the plates of a capacitor with an applied potential difference of 1 V. This is mathematically expressed as: C = Q/V where C is the capacitance in farads, Q is the charge in coulombs, and V is the voltage in volts across the plates. In practice, most circuits use capacitance on the order of microfarads or picofarads. If you differentiate both sides of the above capacitance equation with respect to time and rearrange the result, you obtain the following relation: I = C × dV/dt or current is equal to capacitance multiplied by the change in voltage with respect to time. Thus, if the voltage is unchanging (dV/dt = 0), then current flow becomes zero. This is the case with a direct current (DC) circuit, wherein current flows directly between the anode and cathode with an invariant voltage. Once the potential difference between the plates of the capacitor has equaled that applied constant voltage, current flow ceases. Conversely, a continually varying potential will be able to maintain current flow across a circuit that includes such a capacitive element. This is the effect produced by alternating current (AC) that, as the name implies, is constantly oscillating between two alternating poles. (AC will be described in more detail later.) Thus, a capacitor will pass AC flow but will block DC flow. This impeding effect of the capacitor is known as capacitive reactance and is defined as follows: XC = 1/(2πƒC) where XC is in ohms, ƒ is the frequency of the current in hertz, and C is in farads. One can see that as the frequency of the current approaches zero (as in DC), the capacitive reactance (resistance to flow) becomes infinitely large. Capacitance is crucial to any system that can maintain separated charge and, thereby, store potential energy for use in doing work. The lipid bilayer membrane of nerve tissue is a superb capacitor, which both permits and restricts the flow of ionic currents. It is these intermittent fluctuations in biological currents that ultimately produce the potentials of interest in clinical neurophysiology, such as in EEG. However, other sources of biological capacitance can also interfere with these signals, such as the capacitive resistance in the cerebrospinal fluid, skull, and scalp. As the equation for XC predicts, these will affect differing neuronal frequencies to different degrees. For example, 3 Hz activities through 2 µF of capacitance will have an XC = 1/[2 · 3.14 · 3 Hz · (2 × 10_6) F] = 26.5 kΩ, which is much larger than the 4.4 kΩ reactance seen by 18-Hz beta frequencies. This illustrates how much more capacitive reactance there is to low frequencies vs higher frequencies with scalp recordings. Multiple capacitors in a circuit interact in a manner that is opposite to the behavior of resistors. When arranged in parallel, there is an additive effect as such: Ceq = C1 + C2 + C3 and when arranged in series, the equivalent capacitance is less than any of the individual values, as such: 1/Ceq = 1/C1 + 1/C2 + 1/C3 Basic Electronics in Clinical Neurophysiology 7 Fig. 2. This figure illustrates a transformer, which is based on the principle of induction. An alternating current (AC) with voltage, V1, is applied to an inducer, represented by the coil with N1 turns. Another coil with N2 turns shares the same rod. AC flowing through the coil induces a magnetic field that then induces a reciprocal electrical field (voltage) in the second coil. The ratio of coil loops determines the change in voltage in the second circuit; fewer turns leads to a proportionately reduced voltage in the second circuit. 3.3. Inductors An inductor consists of a continuous coil of wire called a solenoid. Current flowing in this coil generates a magnetic field whose axis passes through the coil (with directionality dictated by the right hand rule). Because of the equivalence of electricity and magnetism (i.e., Maxwell’s equations), this magnetic field can induce an electromotive force (emf, ε) in a nearby conductor, if the magnetic field is variable over time. The magnetic field can vary if the current flow in the coil varies. The relationship of this emf to the current is: ε = – L × dI/dt, where L is a constant called the inductance of the device. The negative sign in the equation indicates that the changing current (dI) induces an emf that opposes that change. The unit of inductance (L) is the henry (H). The inductance of a solenoid is proportional to the number of turns in the coil. A changing current (i.e., AC) passing through a coil will generate a changing magnetic field that passes through its core. If a second coil of wire is wrapped around a nearby section of this core, the changing magnetic field will generate a reciprocal emf and current in the second coil. One can tap this feature to step voltage from one value to another, as in a transformer (Fig. 2). Because inductance (L) depends on the number of turns (N) in the coils, if the number of turns in the first coil (N1) is greater than in the second coil (N2), then the inductance will decrease in the second coil. From the above equation, if L decreases, then dI/dt will increase proportionately. The induced emf (or voltage) in circuit two will decrease in proportion to the drop in inductance. Therefore, voltage varies directly with L and current varies inversely with L, whereas the total energy (power) in the system is conserved. As current steps up, voltage steps down. These vary according to the ratio L1/L2, which is directly related to N1/N2. Inductance is similar to resistance in that it poses an impediment to the motion of charge generated by another source. For example, an AC source, with its associated emf, providing a current through a circuit with an inductor, will be opposed by the emf generated by that inductor. The inductor’s emf is, in effect, subtracted from that of the circuit to determine the net potential. This property is known as the inductive reactance (XL), which is: XL = 2πƒL where XL is in ohms and the frequency (ƒ) is in Hz. 8 Sinclair, Gasper, and Blum 4. POWER Energy is simply charge moving across some potential energy gradient (E = QV). Power is the rate of transfer of this energy, or mathematically: P = E/t A useful permutation of power for use in electrical circuits is as follows: P = Q·V/t = (Q/t)·V = IV where I is the current and V is the voltage. The SI unit of power is the watt (W), which is equivalent to 1 J/s (energy per time). Recall the previous transformer discussion. As the voltage climbs, the current drops proportionally, and the product of these (the power) will remain constant. Of course, this is an ideal, and a transformer in the real world will lose something in the transfer (albeit not much). Their typical efficiency is on the order of 90 to 99%. The intensity of power (β) is often represented as a ratio with a second power level on a normalized, logarithmic scale with units in decibels. The decibel ratio of two power levels is: β = 10log(P2/P1) 5. ALTERNATING CURRENT AC has very useful properties, particularly in circuits involving capacitors and inductors. In the previous treatment of inductors, we saw that a current generated in a coil around a magnetic material could induce a magnetic flux that, in turn, would result in an emf (voltage) across the circuit. An AC generator operates on a similar principle with a minor difference. That is, a magnetic field across a rotating wire will cause a changing field in that wire that will, in turn, induce an alternating emf and current in the wire. The wire is part of a circuit that must be rotated by some external energy source. Practical examples of this are wind, falling water at a hydroelectric plant, or burning coal. The resulting rotation will produce a sinusoidal flow of AC with a characteristic amplitude and frequency. As is apparent from the sinusoidal nature of AC, the average current over any given complete cycle is zero. The quantity of current delivered, however, relates to the amplitude of the sine wave. Because AC is only at its maximum amplitude for an instant, it will not produce the same heating effect as an equivalent DC, nor will it produce an equivalent current. It will instead produce an effect that is similar to a DC of lesser quantity. The effective current in an AC circuit across a resistor is given by the root mean square (rms) value. This can be shown to be Irms = Im/√2 = 0.707Im, where Im is the maximum amplitude and Irms is the effective equivalent to DC. The direction of this current makes no difference in the power of the system because P = IV, which is equivalent to I 2R. 6. IMPEDANCE Impedance (Z) is the term used for the combined effects of resistance along with capacitive and inductive reactance in an RC circuit (a circuit that includes a resistor and capacitor in series) passing AC current: Z = [R2 + (XC _ XL)2]1/2 Basic Electronics in Clinical Neurophysiology 9 Fig. 3. This figure depicts a simple resistance/capacitance (RC) circuit. In (A), VR is the voltage drop across the resistor and VC is the voltage drop across the capacitor. In (B), the behavior of such a circuit in response to an applied square wave pulse (Vapplied) is illustrated. The voltage across the capacitor, VC, gradually builds as charge accrues on the capacitor. This process follows a logarithmic function and has a time course shown in VC. Eventually, VC opposes the flow of current in the circuit. When the applied voltage is zero at the end of the pulse, the capacitor discharges in a reciprocal fashion. By contrast, VR is maximal at the onset of the applied pulse, but as the capacitor charges and opposes the source voltage, current flow decreases and then stops; VR then reaches zero because current, I, becomes 0 (V = IR). Note that VR and I behave similarly because they are directly proportional to one another. In this way, VC behaves as a high-frequency (low-pass) filter, whereas VR behaves as a low-frequency (high-pass) filter. The inductive reactance is subtracted from the capacitive reactance because they have opposite phase. In an AC circuit, Ohm’s law takes the form V = IZ, where Z is the term for resistance in this type of circuit. This is analogous to Ohm’s law as applied to DC circuits (V = IR). 7. TIME CONSTANTS With this understanding of basic circuit elements, we can now examine how simple circuits behave and permit basic electronic filtering of waveform data. An RC circuit is shown in Fig. 3A. When voltage is applied to the circuit, current flows across the resistor and begins to accumulate on the capacitor. As the capacitor becomes fully charged, it accrues a voltage that opposes further flow of current through the circuit. If the power source is turned off, the capacitor discharges in the opposite direction of current flow as it charged. The charging (and discharging) behavior of a capacitor over time is exponential. Its kinetics are described using a time constant, τ, which is that time required for the capacitor to reach approx 63% of its charge. This is 1 – 1/e, where e is the base of the natural logarithm (~2.718). This time is independent of the applied voltage, but rather depends on the resistor and capacitor combination. In an RC circuit, the time constant can be calculated as: τ=R×C A larger resistor permits less current to flow to fill the capacitor, thereby prolonging the time constant of an RC circuit. Similarly, a larger capacitor takes longer to charge, thus, prolonging τ. 8. FILTERS Let us return to the RC circuit and apply a voltage square wave pulse (Fig. 3B). At the outset, the voltage change is seen across the resistor, but there is a lag in the appearance of 10 Sinclair, Gasper, and Blum voltage across the capacitor. At steady state (with no current flow after the capacitor is fully charged), there is no measurable voltage drop across the resistor and there is maximal voltage across the capacitor. The sum of the voltages across these elements always equals the input voltage. Thus, the voltage output across the resistor is very sensitive to sudden changes in input voltage (high frequencies) but insensitive to relatively unchanging voltage (low frequencies); the opposite is true of the capacitor. Therefore, these circuit elements form the basis of high- and low-frequency electronic filters of variable input waveforms (as we record in EMG and EEG). That is, the resistive element serves as a low-frequency filter, and the capacitive element serves as a high-frequency filter. The low-frequency (or high-pass) filter is helpful in EEG, for instance, in blunting slow DC potentials that are of lesser interest. A shorter time constant makes for a more stringent lowfrequency filter (higher cutoff frequency). The relationship between τ and the low-frequency filter is given by: Fcutoff = 1/(2πτ) ≈ 0.16/τ, where Fcutoff is that frequency above which greater than 70% of the input amplitudes will pass. The high-frequency (low-pass) filter is based on the capacitive element and is useful in EEG, to attenuate undesired frequencies that may stem from muscle activity near the scalp leads. The cutoff frequency is defined similarly as in the low-frequency filter. The combination of the low- and high-frequency filters defines the operative bandwidth in use. This is the range of input frequencies that will be allowed through for further analysis. With such RC circuit-based filters (e.g., as in older analog EEG machines), input data is not filtered in an all-or-none fashion, but rather there is a roll-off in the restriction of input frequencies above and below the low- and high-frequency filter settings (Fig. 4). By combining such circuits, one can obtain filters that are more specific, such as the 60-Hz “notch” filter, which more dramatically blunts 60-Hz inputs, a common source of artifact in typical recording environments because of ambient electrical noise. This discussion notwithstanding, modern digital EEGs filter input data using different methodologies than described, with much steeper frequency response characteristics than available with simple RC circuits. It is important to note that overly stringent filtering can distort the output data, for instance, making waveforms seem less sharp than in reality. This can become clinically relevant to the interpretation of the data, for instance, in the recognition of subtle notched morphologies on EEG. 9. AMPLIFIERS Amplifiers are electronic devices that serve to multiply an input signal by a constant. This amplification factor is called gain and is related to the ratio, Vout/Vin. It is common to express gain in decibels as 20 × log10(Vout/Vin). The dynamic range of an amplifier refers to the voltage range over which the amplifier behaves linearly. The sensitivity control on an EEG machine helps to modify the dynamic range of the amplifier. Sensitivity is expressed as microvolts per millimeter and refers to the size of the deflection on the paper or screen that represents this voltage. Typical sensitivity settings for EEG are 7 µV/mm. Increasing the amplifier gain requires lowering the sensitivity; they are inversely related. EEG amplifiers have multiple circuit elements that include voltage regulators, filters, and calibration circuits, among other elements. The heart of the EEG machine is the differential amplifier. The difference between the input voltages from two electrodes relative to a reference electrode (ideally close to the Basic Electronics in Clinical Neurophysiology 11 Fig. 4. Frequency response of filters. This figure shows the percentage of input voltage that is allowed as a function of frequency in relation to applied filters based on analog resistance/capacitance (RC) circuits with various time constants, τ. Such filters exhibit a “roll off” in their attenuation of input frequencies. As τ shortens, the curves for both the high- and low-pass filters are shifted toward higher frequencies. The cutoff frequency is inversely related to τ. For a given filter circuit, it is the frequency above which approx 70% (0.16/τ) of the input amplitudes will pass through for analysis. The notch filter is designed to specifically filter out 60-Hz inputs, because these are frequently artifactual in origin. recording leads) is amplified and serves as output. This method serves to subtract common artifactual noise that may be contaminating both input electrodes. One example is ambient 60-Hz noise from the local recording environment. This subtraction of common noise is called common mode rejection. The capacity of an amplifier to perform common mode rejection is described by the common mode rejection ratio, which is equal to the common signal voltage divided by the nonamplified output voltage. The common mode rejection ratio for many amplifiers in modern EEG devices is 10,000. In a differential amplifier, by convention, if input 1 is negative with respect to input 2, then the pen deflection is upward. If input 1 is positive with respect to input 2, then the deflection is downward. 10. ELECTRODES The above principles can now be applied to the acquisition of neurophysiological data. This begins with the electrode and the interface between the subject and the electrode. Electrodes connect the patient to the circuits of the neurophysiological recorder. They serve to detect and conduct electrical potentials from the patient to the machine. They are metallic, and an electrolyte paste is used to help conduct current and reduce movement artifacts. Electrodes may be nonreversible (polarized) or reversible (nonpolarized). Polarized electrodes are prone to develop significant capacitance, and this may interfere with the faithful transmission of underlying biological signals (the electrode behaves like a low-frequency filter). Reversible electrodes, such as those of silver chloride, are preferred for common neurophysiological applications. Polarization is avoided because the chloride ion is common to both the electrode and the electrolyte. Other metals can be used, such as gold or platinum, but may be costly. Because the electrode–electrolyte interface resembles a simple RC circuit, it is important for the electrode impedance to be as low as possible, typically less than 5 kΩ, for scalp electrodes. 12 Sinclair, Gasper, and Blum Large electrode impedances foster large artifact potentials caused by even small local currents induced by ambient electric fields (Ohm’s law: V = IR, V = IZ). This leads to troublesome electronic noise. Preparing the site (e.g., with alcohol) to remove oils helps to lower the impedance to desired levels. Because inputs from neighboring contacts are compared using differential amplifiers in EEG, their various input impedances must be comparable. If there is an impedance mismatch, the signal fidelity will be degraded, because common sources of electronic artifact will not be efficiently subtracted away. Higher impedances favor lower than true amplitudes and loss of lower frequencies. Excessively low impedances (e.g., a “salt bridge” caused by a smeared electrode gel between contacts) can also cause erroneously low amplitudes. Calibration maneuvers using square wave inputs and biocalibration signals help to assay the fidelity of the recording setup. Many other types of electrodes exist. Subdermal needle electrodes have generally a smaller area of contact, hence, higher input impedances, and, therefore, more susceptibility to noise. Sphenoidal leads are long, thin leads made of stainless steel or platinum placed to position the tip lateral to the foramen ovale. They may be more sensitive to potentials from the anterior temporal lobe than surface electrodes; this view has been, however, disputed. Stereotactically implanted depth electrodes made of stainless steel or platinum may also be used to detect activity from deeper contacts, such as the amygdala, hippocampus, or cingulum, among others. These leads have low impedance and can remain indwelling for weeks, if needed. They circumvent other problems of surface recording, such as muscle artifact and filtering effects of the dura and scalp. Subdural strips and grids made up of stainless steel or platinum discs embedded in a sheet of plastic can be surgically placed to cover the cortex. These permit not only the recording of brain activity, but also the mapping of eloquent cortex (e.g., language cortex) by stimulation of the underlying cortex with various testing paradigms. 11. EEG MONTAGES In the recording of the EEG, electrodes are typically placed on the scalp using the 10-20 system (Fig. 5). In this standardized method, contacts are named by their location (frontopolar, frontal, central, parietal, temporal, occipital, and auricular). They are also numbered with odd numbers over the left hemisphere, even numbers over the right, and z referring to the midline. The particular sequence in which the EEG data is displayed is called the montage. Montages may be bipolar or referential. Bipolar montages involve a comparison of voltages recorded from (usually adjacent) active electrodes in a chain-like fashion, (e.g., front-to-back or side-to-side). By contrast, referential recordings involve a comparison of each electrode to an (ideally) inactive electrode. Examples of reference sites include the ear, the mastoid, the vertex, or an average of many active leads (“average reference”). Bipolar montages visually map the peak of voltage negativity over the scalp owing to the property of phase reversal that emerges from the serial comparisons between adjacent electrodes along a chain. As one ascends and descends the underlying voltage “hill” with each bipolar comparison along a chain, the sign of the comparison flips from positive to negative (Fig. 6). Referential recordings do not exhibit phase reversals, but may show a truer picture of the relative amplitudes of voltage at each electrode. Negativity at the “active” lead is defined as an upward deflection in the display. In addition to displaying the channels based on comparisons among the electrodes, other channels are often used to permit comonitoring of cardiac rhythm, eye movement, respiration, and EMG activity as needed for the study at hand. Basic Electronics in Clinical Neurophysiology 13 Fig. 5. The International 10-20 system of electrode placement. This figure depicts an overhead view of the most commonly used array of standardized electrodes in EEG. Odd numbers refer to the left, even numbers to the right. Electrodes are named according to location: F, frontal; C, central; P, parietal; T, temporal; Fp, frontopolar; O, occipital; and A, auricular. Sphenoidal leads (Sp) and supplemental temporal (T1, T2) leads are not shown. Fig. 6. Bipolar EEG recordings and phase reversal. A hypothetical region of scalp negativity is illustrated topologically, as a voltage “hill” (A). The higher the hill, the greater the negative field potential at that point along the anterior–posterior axis, with electrode A most anterior, and electrode E most posterior. The hill peaks at C. In (B), the waveforms derived from bipolar comparisons between adjacent electrodes are illustrated, as in a bipolar recording. By convention, if contact 1 is more negative than contact 2, the deflection is upward. Because these comparisons are performed serially along the anterior–posterior axis, the difference between adjacent channels flips polarity as one traverses the peak of the hill. The flip in polarity leads to the highly visible phase reversal on a bipolar tracing, denoted by the star in (b). The shared contact of the phase reversal maps the voltage peak, electrode C in this case. 14 Sinclair, Gasper, and Blum Fig. 7. Analog vs digital. The upper curve depicts an analog signal with variable amplitude over time. Digitization relies on using discrete intervals along both the x- and y-axes and making sampling measurements of the analog signal at discrete time intervals and assigning discrete amplitude values. The ability of a digitized output signal to faithfully resemble its analog input depends on the fineness of these discrete steps (resolution) along these axes. Sampling at greater than twice the Nyquist frequency ensures an adequate degree of digital sampling to minimize aliasing. 12. DIGITAL SIGNAL ANALYSIS As elsewhere in our lives, the digital revolution has become firmly entrenched within the arena of clinical neurophysiology. Increasingly, our neurophysiological recording systems and our displays of recorded waveforms are based on computers. This affords numerous advantages, such as the reformatting of EEG data into different montages, refiltering data, and/or applying spike and seizure detection systems to accrued data sets. Storage and sharing of data is simpler; no longer are rooms dedicated to the storage of reams of paper tracings. Computers require that data be digitized. However, the input waveforms that we strive to faithfully record occur in an analog universe. Analog refers to continuous data with respect to time and with respect to value. In EEG, the scalp voltages we record are generated continuously over time and with a continuous range of amplitudes. Digital, on the other hand, refers to the intermittent sampling of data. Measurements are sampled at discrete moments and amplitudes are assigned to nearest discrete values. The degree of fidelity with which the digitally sampled data resembles the original analog data depends on how finely it is sampled, that is, how many discrete steps there are along the time (x) axis and the amplitude (y) axis (Fig. 7). Analog signals are first amplified, then filtered, then subjected to digital conversion. The analog-to-digital converter (ADC) is at the core of the generation of digital signals. The ADC samples the continuous data and assigns a binary code to represent the voltage at that moment in time. Many modern recorders have a sampling rate of 256-Hz per channel. The total throughput refers to the number of channels times the sampling rate. For a 21-channel EEG, this leads to a throughput of 5276 samples/s. One can have a sampling system race through all of these in sequence, or parallel samplers can analyze each channel singly. Basic Electronics in Clinical Neurophysiology 15 It is important to have an adequate sampling rate to avoid undue distortion of the input waveforms. The Nyquist frequency is the highest input frequency component of interest to be recorded. To avoid distortion, the sampling rate should be greater than twice the Nyquist frequency. If one samples at less than twice the Nyquist frequency, aliasing will occur. Aliasing refers to the artifactual creation of low frequencies in ADC, by undersampling high-frequency inputs. The alias frequency is mathematically defined as the sampling frequency minus the Nyquist frequency. In addition, another means to deter such ADC artifact is to apply a high-frequency filter to the input data to cut off undesirable high-frequency elements. The dynamic range of a recording system refers to the range of measurable values, from highest to lowest; in EEG, this refers to the recordable voltage range. The dynamic range is the log10(Vh/Vl). The resolution of a digital recording system depends on the range of recordable values divided by the number of digitally assignable increments. This, in turn, depends on the number of available digital bits in the recording system. With n-bit data, one has 2n possible values; for a 12-bit EEG recording system, this is 212 = 4096 available voltage increments. This can be regarded as plus or minus 2048 units, 0 to 2047. The available dynamic range available is log10(2047/1) = 2.8. Obviously, more bits lead to finer resolution. In an analog system, values at the extremes may be clipped because of inadequacies in the recording or display capacity. Generally, digital systems afford better dynamic range, and much less risk of loss of information at the extremes of the recordable range. Although digital recording systems typically have very good resolution and dynamic range, our ability to display this data depends on the digital resolution of our monitors. When the monitor’s resolution becomes limiting, we can improve our viewing resolution by restricting our analysis to a subset of the data. As can be seen, digitization of neurophysiological data affords numerous advantages to previous analog methods. For instance, digitization has improved our EEG analysis by permitting switching the viewing montage, permitting the adjustment of filters and sensitivity settings post hoc, improving the dynamic range and preservation of data at the extremes of the recording range, lowering the cost for data storage and data sharing, and facilitating the recall of previous data for comparison. In the arena of long-term video EEG monitoring, digitization has been a tremendous boon. It is now simpler to integrate the video and EEG data stream, easier to locate events of interest, and simpler to store the data over time. Computerized methods now exist to apply search techniques to automate the detection of spikes and seizures within a digital data set. Algorithms exist and are being perfected to allow for frequency analysis of data (Fourier transformation—spectral analysis), current source mapping, and real-time seizure detection and even seizure prediction technology that may one day revolutionize the treatment of epilepsy. 13. ELECTRICAL SAFETY Electrodiagnostic procedures are extremely safe, however, injuries can occur and may include pain, dermal burns, seizure, and ventricular fibrillation. A current of 1 mA at 60 Hz applied to dry skin is the threshold for pain sensation. Lethal current injury causing ventricular fibrillation exceeds 100 to 300 mA. Be aware of low-resistance pathways between external current sources and the heart (e.g., indwelling cardiac catheters), because currents less than 100 mA can cause serious cardiac arrhythmias under such circumstances. Following a few simple rules will minimize the risk of injury while recording an EEG or EMG. 16 Sinclair, Gasper, and Blum High-risk populations among those undergoing EEG evaluation include neonates and patients with intravascular catheters or instruments. Lower risk populations include nonneonates without implanted medical instrumentation and bystanders who are near instrumentation, possibly in contact with an electrical device. The overriding safety issue in recording an EEG is the risk of exposing the patient to current. The amount of current is the most important predictor of electrical injury. A secondary concern, especially with older EEG equipment in which small sparks are generated when the power switch is activated, is prevention of flammable gas explosion. To reduce the risk of patient exposure to current, proper grounding must be ensured. The electrical circuit in EEG consists of a current provided by the “hot” contact in a wall outlet, which travels through the EEG machine and returns to the outlet’s neutral contact. Excess current can accrue from various sources that escape this circuit (i.e., a short circuit within the instrument that directs current to the covering of the EEG machine). This potentially dangerous leaked current can fortunately be shunted to the ground contact, which is the path of lowest resistance. However, should ground failure occur, the leak current may flow preferentially through a low-resistance pathway, which may include the patient. In intensive care units or surgical operating rooms, multiple machines may often be attached to a patient and may lead to the use of multiple grounds. A ground loop may be formed if each of these several grounds has a slightly different potential, allowing excess current to flow from the higher resistant ground to the lower resistant ground. This may be hazardous if this ground loop includes a nearby person or the patient. A simple rule to avoid this hazard is to connect the patient to only one common ground that serves all of the devices in use. Basic requirements for proper grounding include: 1. EEG equipment should always be connected to a three-pronged wall receptacle, never a twopronged plug. Three-pronged plugs include a black “hot” connector, a white neutral connector, and a green ground connector. The ground is absent in two-pronged plugs. The black connector has AC current at 110 V. The neutral connector is the reference for the “hot” line but is not necessarily at 0 V. The ground is the building ground connection. 2. A working fuse within the machine to break the circuit, should high current travel through the ground, alleviating a potential safety risk to the patient. 3. Adequate power outlets for patient-related equipment, typically marked by a green dot, indicating a higher standard of safety. 4. Regular evaluation and testing of equipment. Equipment and plug grounds must be periodically tested with an ohmmeter, which should show a resistance of less than 0.1 ohms when assessing the drop in current from cover to ground connector. 5. Ensure that the ground plug is firmly in place in the socket. 6. Connect all devices into the same outlet using a grounding bar. This will ensure a common low resistance ground pathway for all stray currents. In the case of multiple devices in use, placing a ground on the patient may actually increase the danger of electrical shock. Because another purpose of the ground in EEG and EMG is to reduce ambient 60-Hz interference, avoidance of multiple grounds will actually improve the quality of the recording. Excessive current that requires shunting arises typically because of a short circuit within the equipment or a leakage current. Leakage currents from stray conductive and stray inductive sources mainly arise from the power cable. The cable contains three wires covered by insulation, which creates a type of capacitor in which currents can flow between wires (from “hot” to the others). Longer cords create greater capacitance, with a larger potential for current leak (up to 70 µA) into the instrument. Basic Electronics in Clinical Neurophysiology 17 Unwanted inductive current is a less significant source of current leak than cable capacitance. Inductive currents are generated by the magnetic field surrounding a wire that creates current in other wires. This may arise when wires are coiled. Steps to avoid generation of excessive currents are: 1. Power cords are rated in capacitance per foot and selected by instrument manufacturers to ensure minimal leakage. 2. Extension cords should never be used. A leakage current produced by adding a 6-foot power extension cord will range from a minimum of 7 mA to 60 mA. 3. Keep wires from excessive coiling. SUGGESTED READING American EEG Society. Guidelines in EEG. J Clin Neurophysiol 1986;3:1–147. American EEG Society. Guidelines for standard electrode position nomenclature. J Clin Neurophysiol 1991;8:200–202. Hughes JR. EEG in Clinical Practice. Butterworth-Heinemann, Boston, MA, 1994. Laks MM, Berson A. Electrical safety measures for physicians using electromedical equipment. Sem Neurol 1995;15:311–316. Litt B, Fisher RS. EEG engineering principles. In: Current Practice of Clinical Electroencephalography, 2nd ed. (Daly DD, Pedley TA, eds.) Raven, New York, NY, 1990. Misulis KE. Basic electronics for clinical neurophysiology. J Clin Neurophysiol 1989;6:41–74. Misulis KE, Head TC. Essentials of Clinical Neurophysiology, 3rd edition. Butterworth-Heinemann, Boston, MA, 2003. Niedermeyer E, Lopes da Silva FH, eds. Electroencephalography: Basic Principles, Clinical Applications and Related Fields, 3rd ed. Williams and Wilkins, Baltimore, MD, 1993. Seaba P. Electrical safety. Am J EEG Tech 1980;20:1–13. REVIEW QUESTIONS 1. In a simple circuit with a single resistor, if the applied voltage is 10 V, and the resistance is 5 Ω, what is the current? 2. In a series circuit with two resistive elements, R1 and R2, what is the derived combined resistance in the circuit? 3. In a parallel circuit with two resistors, R1 and R2, what is the resultant combined resistance in the circuit? 4. How does the resistance to current flow attributable to a capacitor (capacitive reactance) relate to the frequency of the AC? 5. In a transformer with a coil (N1) of 20 turns and an applied voltage of 100 V, what will be the resultant voltage in its associated secondary circuit if its coil (N2) has only 5 turns? 6. What is impedance? 7. In an RC circuit, what is its derived time constant? What is a time constant? 8. Which element in an RC circuit behaves as a low-pass (high-frequency) filter? 9. What is the Nyquist frequency? If the Nyquist frequency for EEG is considered to be 100 Hz, what should the sampling frequency be to avoid aliasing? 10. What is a ground loop? REVIEW ANSWERS 1. 2. 3. 4. Using Ohm’s law, V = IR, the derived current is 2 A. The combined resistance in a series circuit is Rcomb = R1 + R2. This is a current divider. The resultant combined resistance is 1/Rcomb = 1/R1 + 1/R2. Capacitive reactance is inversely related to the applied frequency of AC. 18 Sinclair, Gasper, and Blum 5. In a transformer, the ratio of the coil number defines the step-down or step-up relationships. In this instance, N1/N2 = 20/5 = 4. Therefore, the induced voltage will be 1/4 of the original voltage, or 100 V/4 = 25 V. 6. Impedance is the combined resistances to current flow in a circuit passing AC. It may include not only resistance from a resistor, but also capacitive and inductive reactances. 7. The time constant, τ, of a circuit with resistance R and capacitance C is defined by: τ = R × C. The time constant, τ, refers to the time needed for a capacitor in an RC circuit to reach 63% (1 – 1/e) of its full charge. 8. The voltage reading across the capacitor in an RC circuit behaves as the high-frequency filter. Adjustment of this element may serve to help filter out higher-frequency artifact, such as EMG activity in the recording of an EEG. 9. The Nyquist frequency is the highest input frequency that is deemed desirable to detect. If the Nyquist frequency is 100 Hz, then the sampling frequency should be at least 200 Hz to avoid aliasing. 10. A ground loop occurs when multiple grounds are applied to a subject because of various electrical devices, such as in an intensive care unit setting. If these various grounds are not identical in voltage, an electrical pathway can exist that includes the patient, through which stray currents may travel, thus, endangering the patient. A common ground can avoid this hazard.