As with most basic bioimpedance technologies, 4 electrodes are generally positioned in a linear array, with either a transverse or linear orientation—the two outer ones supplying the electrical current and the two inner ones measuring the generated voltages [1]. There are two major reasons for using a 4-electrode system rather than a 2-electrode system, which can also be utilized to measure impedance. First, the two-electrode approach is adversely affected by electrode polarization at the electrode-skin interface, adding an additional component to the measured impedance values that has no relevance to tissue condition. The second reason is that the two inner electrodes are measuring the voltages across tissue at a distance from the current electrodes, allowing the electrical current to substantially penetrate the subcutaneous fat into the muscle at the point of measurement. The use of 4 electrodes thus captures more data representing muscle than subcutaneous fat.

Outside of the restriction of using 4 electrodes in a specific order (Current-1, Voltage-1, Voltage-2, and Current-2), EIM can applied to the body using a variety of approaches. For example, in early EIM work, the current emitting electrodes: silver (Ag)—silver chloride (AgCl) adhesive electrodes, were placed at a distance from the voltage measuring electrodes, usually on the feet or hands and the voltage electrodes were placed over a muscle or muscle group of interest (e.g., over the quadriceps or biceps) [4]. An advantage of this approach is that it ensures that most of the electrical current flows through deeper tissues and thus is less affected by fat. In addition, large current-emitting electrodes can be used to ensure a relatively broad application of current throughout the limb. Disadvantages include the clumsiness of the approach (with multiple long cables attaching to the patient), the potential for even small changes in joint angle to have a significant impact on the data (a joint acting as a lens distorting the current flow as it passes through it), and the potential of substantial variations due to hydration status, associated with the current’s flowing through deep veins and arteries.

For these reasons a simple and more convenient approach is taken in which all 4 electrodes are placed within several centimeters of one another [1]. This has obviated many of the problems associated with the initial approach, although it has the disadvantage of a requirement that the patient not move the limb during measurement (since changes in the shape of the muscle will affect the data) and a significant contribution of subcutaneous fat to the measured impedances (although there are approaches for disentangling these effects as described below). This approach also allows measurement of the muscle’s anisotropy, discussed further below, as well as the use of pre-formed electrical arrays, expediting the process of data acquisition. Finally, this approach appears to be much less impacted by hydration status since the current does not readily penetrate deep blood vessels [5].

Using this “near” electrode description still leaves open many aspects of the possible configuration including the best size and shape of the electrodes for a given muscle and the ideal inter-electrode distances [6]. Ongoing work continues to model and improve these electrode configurations attempting to develop arrays that are effective at achieving excellent penetration of muscle while ensuring high repeatability and ease of use.

A basic tenet of all impedance work is that by varying frequency one can learn a great deal more about tissues than by applying current at just a single frequency. As noted above, applying electrical current at multiple frequencies allows one to investigate the actual structure of the tissue being studied, which is not possible with single frequency approaches. Figure 14.2 gives an example of the frequency spectra for reactance, resistance, and phase of a healthy child, and one with Duchenne muscular dystrophy. As can be seen the entire spectrum for all three measures alters in complex ways, although lower reactance and phase values are perhaps the most prominent feature. Moreover, if multiple frequencies are not utilized (e.g., data from only a single frequency is obtained), it is impossible to tell if there is an artifact present in the data. Such distortions are readily identifiable when performing a multifrequency analysis.

A challenge of obtaining multifrequency data, or impedance spectroscopy, is that it is necessary to condense or collapse all the values into a single output value of interest. One approach is by applying simple mathematical operations that attempt to capture this change, such as the determination of linear fits of the data over certain frequency ranges [7] or calculating simple ratios of data from individual frequencies [8]. Another is by using “Cole parameters” which are based on a basic simple impedance circuit model [2]. These parameters include: f _c, the center or characteristic frequency, corresponding to the highest reactance value, and an indirect measure of muscle cell size; α, an index of variation of muscle fiber size (one being perfectly homogeneous); and R₀ and R_inf, corresponding to the resistance at 0 kHz and infinite kHz respectively, providing measures of cell density.

Another aspect of muscle is a property termed anisotropy, or directional dependence of electrical current flow [9]. Since muscle fibers are essentially long cylinders arranged in compact bundles (fascicles), applied electrical current will flow more easily along the fibers than across them [10]. The anisotropic nature of muscle, however, may be disturbed in situations in which there is alteration or destruction of muscle fibers or interposition of other materials into the muscle such as fat and connective tissue. Thus, myopathic diseases, including muscular dystrophies, will manifest a reduction in the normal anisotropic features of muscle [11]. In addition, a given direction of applied current flow may provide a better measure of disease status than another direction.

One challenge of performing impedance measurements is identifying which impedance characteristic is most informative of a specific muscle disease. Given the range of frequencies, the anisotropy issues, and varying electrode designs, it is challenging to know which values are best for assessing disease status. Much early work looked at single frequency measures, such as the 50 kHz phase [4, 12, 13]. While this value appeared repeatable and sensitive to progression in amyotrophic lateral sclerosis, it has proven less useful in other disorders. The reason for choosing this value initially was simply convenience: it was readily obtainable with relatively inexpensive impedance measuring equipment. Since then, other outcomes have been assessed including some of the multifrequency phase parameters we have discussed above, such as slopes, ratios, and Cole parameters. Moreover, in some conditions, rather than focusing on the phase value, the single or multiple frequency reactance or resistance values appear useful [14]. Moreover, of the three parameters, reactance appears to be least correlated with subcutaneous fat thickness [15], but at the same time is the most variable of the three parameters. Thus, for the most part, the question as to the “best” impedance parameter remains unsettled. In choosing any given parameter or set of parameters to use as a biomarker, it will be important that such measures be tested in separate populations and can survive verification by other investigators to ensure that it is truly robust and meaningful. Moreover, for such a parameter to be useful longitudinally, it not only should track disease status and demonstrate outstanding reproducibility, but should also be sensitive to the beneficial impact of an effective drug or other therapy.

Although initial uses of EIM focused on adults with neuromuscular disorders including amyotrophic lateral sclerosis, it became clear early on that since the technique was entirely painless and could be applied at the bedside or in clinic, it could have special value in the pediatric population. A list of the potential advantages and disadvantages to using EIM in children is shown in Table 14.1. To date, the technique has been mainly studied in children with spinal muscular atrophy (SMA) and Duchenne muscular dystrophy (DMD) [7, 16, 17]. However, it was also recognized that in order to assess longitudinal changes in pediatric disease, it would also be necessary to evaluate changes in healthy children with growth. Given the sensitivity of impedance technologies to muscle histology, our a priori anticipation was that we might observe the effect of increasing muscle fiber size with age. Thus, in nearly all our human studies to date, both a healthy cohort of children has been included along with those affected by disease processes.