Electrogram Recording and Signal Processing

Extracellular Current Recording

The basis of EGM recording is passage of electrical charge between electrodes and electrolytes. Whenever we record signals from a catheter, pace the heart, deliver radiofrequency (RF) energy for ablation, or perform cardioversion/defibrillation, electricity passes through the electrode-tissue interface. Electrodes used in intracardiac catheters are made of inert metals that quickly equilibrate their potential with the surrounding charge. ^, A polished platinum-iridium (Pt-Ir) alloy is commonly used for EP catheter electrodes because platinum is an inert and biologically safe metal with excellent electrical properties; the addition of iridium provides mechanical strength to the electrode without introducing electrical interference. Inert metal electrodes do not transfer electrons to surrounding atoms to create ions. The opposite is also true; surrounding ions do not transfer electrons to the electrode. As a result, there is no direct transfer of current between the electrode and the electrolyte.

When an electrode is immersed in an ionic solution, equilibration of potentials between the electrode and the electrolyte occurs via capacitance currents. That is, if there is a sudden rise in the aqueous solution charge, a layer of soluble, positively charged cations builds up on the aqueous side of the electrode-electrolyte interface. In response, electrons within the electrode are attracted to the surface. This forms a negatively charged outer electrode layer and a relatively positively charged inner layer. This transfer of capacitance current between the electrode and the electrolyte results in electrode polarization. The positive potential inside the electrode (which ideally reflects the potential of the aqueous solution) is what is then transmitted to the amplifier. The goal of any recoded EGM is to transmit and reflect this extracellular potential with sufficient fidelity and also reduce extrinsic artifacts including external noise or physiologically driven signals that are not relevant (e.g., baseline wander because of respiration and distracting information, such as local T waves).

The kinetics of current interchange between the electrode and the electrolyte (i.e., how quickly and accurately electrical gradients in the electrode reflect the surrounding solution) depend on the ease with which electrons flow within the metal electrode. This is determined by the impedance of the electrode, a quantity that reflects electrode resistance and capacitance. Higher electrode impedance results in slower electrical kinetics. In inert metals (such as Pt-Ir alloy), a larger surface area results in lower impedance and, consequently, faster kinetics (thus the kinetics of an 8-mm-tip catheter are better than a 4-mm-tip catheter). Furthermore, the impedance of metal alloy electrodes depends on the frequency of the recorded signal; low signal frequencies have higher impedances (because of the capacitive nature of the electrodes), resulting in a lower signal fidelity and vice versa. Electrode capacitance is itself influenced by extrinsic factors, such as pacing currents. The effect of these currents is larger in smaller electrodes that have higher impedance and slower kinetics. This is discussed in more detail later in this chapter.

The challenges of high impedance and slow kinetics posed by the polarizing Pt-Ir metal alloy electrodes can be reduced by using electrodes made of silver-silver chloride (Ag/AgCl) composite. Unlike platinum, silver is a reactive metal that can exchange electrical charge with aqueous ionic solutions. As silver releases electrons into the circuit, the silver cations (Ag ⁺ ) combine with Cl ^– from the bath solution to form silver chloride. Silver chloride is insoluble and deposits on the surface of the electrode. This direct flow of charge that traverses the electrode-solution interface is called the Faradaic current. Because of this property, electrode polarization and capacitance currents are minimal. Thus the major advantage of these electrodes is their lower impedance that is not impacted by signal frequency. This results in high-fidelity measurements for lower-frequency signals (e.g., body surface electrocardiogram [ECG]). Nevertheless, despite these advantages, Ag-AgCl electrodes have limited use for intracardiac catheters because of their potential toxicity. The exception to this rule is the Franz catheter, which is designed to measure low-frequency signals, such as the monophasic action potentials (MAPs; discussed later).

Amplifier

The role of the instrumentation amplifier is to magnify the small differential signal recorded in the heart and minimize the amplification of undesired signals (noise). These devices are high-precision amplifiers with ultralow offset and drift, low gain error, and an excellent ability to reject noise. Although the circuitry of a modern amplifier is complex and goes well beyond the scope of this chapter, basic knowledge of amplifier circuitry can be helpful in understanding the effects that electrode size, signal amplitude, and filters have on EGM characteristics.

Medical precision amplifiers have high input impedance and low output impedance. The signal arriving into the amplifier input is in the form of a miniscule current that travels from one electrode to another. The current passes over an extremely high impedance (amplifier input impedance). The higher the input impedance, the lower the current and the smaller the perturbations introduced into the system. Therefore high input impedance maximizes the source signal that passes into the amplifier and minimizes the signal that drops out because of source resistance. This is particularly important when working with small electrodes that have larger impedances. As the ratio between electrode impedance to amplifier input impedance rises, signal fidelity decreases. A low output impedance minimizes the voltage loss between the amplifier and the output source (recording system).

Theoretically, modern instrumentation amplifiers have gigaohm input impedances; however, the practical input impedance of a mapping system is limited by the presence of additional circuitry that is connected in parallel to the amplifier (e.g., ablation protection circuits and pacing routing electrodes). These components reduce the effective input impedance to the order of megaohms, which is sufficient for most intracardiac catheter electrodes and results in signals that are several orders of magnitude higher than electrode noise. In special circumstances, however, such as in very small metal alloy electrodes (approximately ≤0.5 mm ² ), this input impedance may not be sufficiently high and thus may confer higher noise levels.

Amplifier gain is a measure of how much an amplifier magnifies the input signal and is defined as the ratio of output to input signal. The voltage gain of EP amplifiers is often very large because their starting point is a microvolt input signal. For example, a 100 μV His bundle signal is amplified by a factor of 5000 to an amplitude of 500 mV. Fig. 127.1B illustrates the concept of amplifier gain. The tip and ring electrodes of an intracardiac catheter detect intracardiac potentials V _Tip and V _Ring , respectively (the difference being V _Bipolar ). The potentials detected at the amplifier input terminals (V ₊ and V _– ) are affected by electrode impedances (R _Tip and R _Ring , respectively). The amplifier input signal (V _in = V ₊ – V _– ) is then magnified by a factor G (i.e., amplifier gain) to the final output voltage of V _O . Ideal amplifier gain should be sufficient to increase microvolt signals and should not amplify noise. It should also remain constant across varying values and frequencies of input signals. In this regard, every differential amplifier is designed to suppress ambient radiated signals (such as 50/60 Hz signals) from the electrical mains.

The ratio between the amplified and the suppressed signals is called the common-mode rejection ratio (CMRR). High-performing differential amplifiers have very high CMRR ratings of as much as 100,000 to 1 and can maintain these high CMRR values over a wide bandwidth. The noise that will be rejected must be equally present (balanced) on both input channels. Noise that is sensed only on one electrode will bypass the CMRR and be amplified rather than suppressed. Common examples of noise on the ablation electrode include those from delivery of RF energy or connecting a pacing stimulator to the distal electrode. The noise on the distal bipolar EGM is because of the imbalance that is produced by attaching a current source to the distal electrode (even if no pacing current is being delivered). Turning off the pacing circuit restores the EGM appearance by reestablishing the balance between the two inputs. Similarly, if there are major differences between the physical properties of the electrodes (especially differences in electrode size, as is typical for ablation catheters), ambient noise will be picked up differently on each input line. Such imbalanced noise will be amplified, rather than suppressed by the CMRR. It is worth noting that the amplifiers included in mapping systems often provide an analog output for direct connectivity with other modalities, such as recording systems. In this case the analog output is already gained up and may include some undesired additive noise, as previously discussed. Any additional amplification in the recording system will increase the signal of interest and the noise equally (i.e., without the ability to reject the noise).

Analog-to-Digital Converter and Signal Processing

An analog-to-digital converter (ADC) converts a continuous voltage amplitude over continuous time signal to a digital signal with a discrete amplitude at each discrete point in time. This conversion involves the quantification of the input at predefined time intervals, which inherently introduces a small amount of error or noise. Two important parameters that determine the performance of an ADC are resolution and sampling frequency. The resolution of the converter indicates the number of discrete values it can assign over the range of analog values. Thus resolution determines the magnitude of the quantification error and therefore determines the maximum possible average signal-to-noise ratio. For example, an ADC with a resolution of 8 bits can encode an analog input to one in 256 different levels (2 ⁸ = 256), and an ADC with a resolution of 12 bits assigns up to 2 ¹² = 4096 values to each voltage measurement. This is adequate for most EP purposes. An ADC device can only convert signals that are within a certain dynamic amplitude range; in other words, signals exceeding a certain maximum will not be appropriately digitized. The required dynamic range is designed according to the maximum amplitude of the expected physiologic biopotential signals. To balance between the two opposing requirements (high signal gain by the amplifier on one hand and limitations imposed by the ADC dynamic range on the other hand), signals are often prefiltered before their conversion to eliminate possible saturation within the ADC.

The sampling frequency of the ADC refers to the number of voltage measurements recorded per second. The Nyquist-Shannon sampling theorem implies that a faithful reproduction of the original signal is only possible when the sampling rate is at least double the highest frequency of the desired sampled signal. ^, Modern ADCs usually sample individual intracardiac tracings with kilohertz frequencies (>1000 samples/sec). This is adequate for physiologic myocardial activity, with the exception of the conduction system, which requires a sampling rate at least 15 kHz rates. Inadequately low sampling frequencies (undersampling) of high frequency information results in aliasing (i.e., loss of certain signal frequencies in exchange for the creation of “new” artificial signals). One must also be aware that ADCs are typically multiplexed, meaning that multiple recorded channels share one ADC, such that the sampling frequency is shared across multiple channels. As a result, if a single ADC is asked to record too many channels, individual channel sampling frequencies may be reduced to an inappropriately low value. In this regard, LabSystem Pro (Boston Scientific) has 160 channels, a 16-bit resolution, and a sampling rate of up to 4 kHz per channel. In comparison, CardioLab (GE Healthcare) has 128 channels, a similar sampling rate of up to 4 kHz, and a 16-bit resolution. Some mapping systems, such as CARTO 3 (Biosense Webster), have a 16-bit ADC with a higher sampling rate of 8 kHz per channel.

Once the analog signal has been converted to a digital signal, it can be fed in parallel or in series into other EP systems.

Electromagnetic Interference and Noise in the Electrophysiology Lab

From an electromagnetic standpoint, EP laboratories are a noisy environment, and this noise can easily contaminate EGM signals. ^, The electromagnetic interference may arise from medical devices directly connected to the patient’s body (e.g., blood pressure and pulse oximeter monitors, ECG machines, mapping systems, ultrasound machines, external defibrillators, ventilators, and intracardiac catheters). Other sources of interference include devices that are in close proximity to the patient but not in direct contact with the patient’s body (e.g., fluoroscopy systems, warming systems, hospital computers, and backup power supplies). Finally, lab infrastructure, including air conditioning system, lights, Wi-Fi, and vicinity to other hospital infrastructure (e.g. computed tomography [CT] lab, generator room), can also introduce noise. Knowing the possible root causes of noise can help implement strategies to decrease or eliminate it.

Furthermore, the patient’s body acts as an antenna, being both capacitively and inductively coupled to the electrical mains. As such, the patient can pick up a substantial amount of RF noise from wireless headsets, mobile phones, and Wi-Fi systems because these devices have some current leak (i.e., a current that “couples” or transfers to the patient’s body) at a fundamental mains frequency of 50 or 60 Hz with significant harmonics (integer multiples of the baseline frequency), extending to several thousand Hertz.

Overall, this current leakage introduces noise that interferes with our ability to process extracardiac and intracardiac signals. Although some of this noise can be filtered out by the CMRR, the fact that the interference is heterogeneous and not uniformly picked up by each of the two amplifier input ports (e.g., by the intracardiac electrode and by the Wilson central terminal [WCT]) makes noise a sizable challenge for high-fidelity recording of physiologic data. The optimal approach to minimizing interference is to minimize or eliminate sources of electric interference. Using signal processing (e.g., filtering) for this task is suboptimal because, in addition to reducing interference, it may also reduce important physiologic information or produce false potentials.

Shielding, Grounding, and Isolation

The lead shielding of EP laboratories is designed to minimize escape of x-rays to the outside. Fortunately, such shielding works in the opposite direction as well: it helps to minimize the leakage of RF interference from the outside into the EP lab. The strongest potential sources of interference within the EP lab are power sources, which must be kept away from the rest of the circuitry. Unplugging unused electronic equipment can help reduce interference; however, even unplugged electronic devices can act as antennas and cause interference.

Electrical isolation of equipment can help reduce circulating mains currents (50/60 Hz). Electrical isolation is often designed into devices by the manufacturer. In theory, however, minimizing the effect of interference between pieces of equipment can also be achieved by deliberately plugging all devices into wall sockets that are spaced distantly apart and that share a common ground. In practice, however, this technique may paradoxically worsen the situation by exposing the power cables to heterogeneous electromagnetic environments. Ultimately, a trial-and-error approach may be needed to determine the optimal configuration of electrical instruments.

Even in the absence of electrical mains contamination, crosstalk can occur between any two cables that carry physiologically meaningful signals. For example, the appearance of the same signal on both input lines of an amplifier can result in its inappropriate suppression by the CMRR. Such interaction can be minimized by avoiding parallel orientation or coiling of cables; instead, cables should cross in a perpendicular fashion. Unfortunately, this is not always practical and signal wires that are connected to diagnostic catheters are usually housed within the same cable line and tightly wrapped around one another. For this reason, braided cable shielding is used to ground any interference. It is important to note that, when cables are repeatedly resterilized, the shielding can be damaged. This can result in faulty shielding of the interconnect cable between the catheter and junction box, which is a common cause of interference. Replacing the cable can often solve the problem. Additional strategies to reduce signal noise include minimizing cable length and employing fiber optics whenever possible.