Electrical Stimulation as a Modality to Improve Performance of the Neuromuscular System

Abstract

Transcutaneous neuromuscular electrical stimulation (NMES) can modify the order of motor unit recruitment and has a profound influence on the metabolic demand associated with producing a given muscle force. Because of these differences, interventions that combine NMES with voluntary contractions can provide beneficial outcomes for some individuals. The adaptations evoked by NMES are not confined to the activated muscle but also involve neural adaptations through reflex inputs to the spinal cord and supraspinal centers.

INTRODUCTION

The possibility to evoke contractile activity by applying an electric current to the neuromuscular system has been known since at least the 18th century. In the second part of the 20th century, electrical stimulation has been used to elicit selected movements by stimulating paralyzed muscles of humans. This clinical application, called functional electrical stimulation (FES), has demonstrated its efficacy in the rehabilitation of stroke and spinal cord-injured patients, especially since the development of programmable multichannel implants.

When used on weakened but otherwise normally innervated muscles, the application of electric current to the muscle is often referred to as neuromuscular electrical stimulation (NMES). Practitioners generally prefer a direct application of the NMES technique whereby electrodes are placed on the muscle, and electric pulses provoke the muscle contraction via the axonal branches, rather than the indirect stimulation of the muscle via a major motor nerve. Moreover, NMES treatments are performed in a transcutaneous manner with the aid of surface electrodes located over motor points (i.e., the innervation zone of the muscle). The use of percutaneous electrodes placed beneath the skin's surface remains infrequent and is reserved for experimental designs or precise pathological circumstances.

The transcutaneous NMES technique has been used for a number of years by physical therapists as a means of limiting the atrophy that occurs with immobilization. Various locomotor pathologies have been effectively treated by NMES, and several investigators have highlighted the value of quadriceps stimulation during rehabilitation after traumatic injury or surgery of the knee. However, it was generally acknowledged that traditional strength-improvement techniques were of greater benefit than NMES in restoring muscular performances and that NMES has to be applied specifically in the early phase of rehabilitation when programs with voluntary contractions are not applicable as a consequence of a neuromuscular inhibition linked to pain or healing phenomena.

This viewpoint was prevalent until the Russian researcher Y. Kots claimed at a 1977 symposium at Concordia University that electrically induced contractions produced 10%-30% greater isometric force than did maximal voluntary contractions and that NMES programs produced strength gains of 30%-40% in highly trained athletes. Moreover, NMES was presented as a technique of choice for improving muscle strength in athletes, able-bodied individuals, and patients (¹⁷). Although poorly documented and described, the Russian technique, which involved a sinusoidal carrier signal of 1500-2500 Hz that was modulated at 50 Hz, was applied by other investigators, and the spectacular results presented by Kots were never confirmed. Nevertheless, the assertions by Kots provided renewed interest in the use of NMES for therapeutic and athletic applications. Many of the early studies, however, are generally characterized by poor experimental design and used a great variety of stimulation conditions and testing procedures. This heterogeneity induced many controversies and unanswered questions, which have resulted in many physical therapists and coaches applying NMES in an empirical manner often based on personal beliefs. The lack of scientific consensus and international recommendations is particularly crucial because stimulators are available to everyone (patients, competitive and leisure sports people, sedentary subjects, etc.), and the commercial pressure from manufacturers has reinforced the almost magical feature of the electrically induced contraction.

Based on our work about motor unit activation pattern (¹⁰) and energetics (^27-29) observed in the electro-stimulated human muscle, we believe that particularities exist in the characteristics of the electrically evoked contraction. In this review, we propose a contemporary and integrative analysis of the following concepts: the activation order of motor unites under NMES, the metabolic demand relative to the force that is electrically evoked, and the central and muscular effects of training by NMES.

TORQUE PRODUCTION

When researchers examine the instantaneous effects of NMES, maximal strength is generally measured during an isometric contraction evoked by transcutaneous stimulation of the quadriceps of able-bodied subjects and athletes. When the protocol involves increasing the current intensity to the greatest level the subject will tolerate, the measured peak torque is expressed as a percentage of the maximal voluntary torque (MVT). The available results range from 25% to 90% MVT, demonstrating that NMES is less effective than voluntary contractions in developing an intense contraction (Fig. 1). Other authors report higher values for stimulated contractions with contractile force exceeding 100% of MVT. However, those results were obtained, either in a single highly trained and motivated subject who can tolerate exceptionally high current intensities or in experimental conditions that are not reproducible in a clinical or athletic setting (indirect stimulation of the femoral nerve or intramuscular stimulation with needles directly implanted in the quadriceps).

Figure 1: A literature review of maximal electrically evoked torque (Max EET), expressed in percentage of maximal voluntary torque (MVT), obtained in the quadriceps femoris muscle by direct and transcutaneous application of neuromuscular electrical stimulations (NMES).

The inability to produce 100% MVT with NMES is mainly due to the difficulty in recruiting all the motor units as the result of the discomfort associated with an intense cutaneous and muscular electric stimulation. Because the force evoked by electrical stimulation largely depends on the subject's tolerance of the current, there is considerable variability across individuals in the stimulated muscle performances and a stronger response to NMES is observed in elite athletes who are more accustomed to the discomfort associated with very high levels of training. To evoke an intense contraction, therefore, it is crucial to motivate the subject to tolerate the highest current intensity and to minimize the skin and muscle nociceptive sensations associated with the application of the current. Currently, there is no consensus of the optimal stimulation conditions, but there is agreement on some current characteristics: most researchers use biphasic rectangular pulses that are symmetrical and last 0.1-0.5 millisecond and are delivered at a pulse rate of 50-100 Hz. In contrast to the early work, it is now known that a medium frequency (~2500 Hz) modulated in low frequency (~50 Hz) is less effective at producing electrically evoked contraction than those achieved with low-frequency stimulation (¹⁹). Despite these agreements, the number, size, and localization of electrodes remain controversial. The lesser torque evoked with NMES compared with a maximal voluntary contraction is also partly attributable to the absence of synergist muscle activity that stabilizes the posture. These muscles are recruited during intentional maximal tests.

Some authors have also compared the torques produced by voluntary and NMES contractions by applying a train of electrical stimuli during a voluntary contraction (superimposition technique). With able-bodied subjects who are physically active, most studies report no difference in the peak torques achieved with the stimulus superimposed on a variety of muscle actions (isometric contractions, shortening contractions, and complex tasks) (²⁴).

GAINS IN MUSCLE PERFORMANCE

In rehabilitation programs, strength training of athletes, and preventive conditioning of sedentary subjects, the main objective of NMES is to improve muscle performances through multiple sessions of stimulation. The long-term adaptations to NMES have been the subject of hundreds of publications that vary in experimental design relative to the fitness status of the subjects, the stimulated muscle, the current characteristics, the number and duration of sessions, and the measured outcomes. Generally, the intervention occurs during a 4- to 5-wk period that involves 20-25 sessions of NMES, with each session lasting 10-30 min. The primary outcome is typically the static or dynamic force produced by the stimulated muscle. Unfortunately, most studies have design flaws, such as no controlled and randomized trials (CRT); in a recent systematic review focusing on quadriceps strengthening by NMES, only 35 CRT of more than 2000 citations could be included in a meta-analysis. Moreover, an independent quality assessment revealed that even the included studies contained methodological biases, such as a small sample size and a large number of dropouts (³).

Despite the quantitative diversity, reliable studies indicate that NMES constitutes an effective way to increase strength in normal and impaired muscle compared with doing no exercise (³). When comparing NMES with traditional strengthening techniques, the initial muscle status has to be considered; with orthopedic postoperative or postinjury subjects, electrically induced exercises may be more effective than voluntary contractions to prevent muscle atrophy during the immobilization period (¹²), whereas NMES is less (⁸) or equally (²³) effective than identical training programs, including volitional contractions, in healthy subjects (Fig. 2). Moreover, it has been demonstrated that the gains observed at the end of the training period depend on the intensity of the electrically evoked contractions (¹⁸). This result highlighted, again, the importance of using appropriate current characteristics to minimize the discomfort and optimize the spatial recruitment of motor units during NMES.

Figure 2: Representative data from the literature showing the respective effects of neuromuscular electrical stimulations (NMES) and voluntary contractions (VC) on muscle maximal performance, during an immobilization period (Gould et al., 1982 (quadriceps)) or during training programs (Duchateau and Hainaut, 1988 (adductor pollicis) and Miller and Thépaut-Mathieu, 1993 (biceps brachii)). In comparison with VC, NMES is more effective at preventing the muscle atrophy caused by immobilization and less or equally effective at improving muscle torque during training.

Based on the efficacy of NMES and voluntary contractions and the differences between the two techniques in the pattern of muscle recruitment and the associated metabolic activity, it is widely accepted that NMES constitutes a complementary tool that should be used in association with voluntary training. However, the optimal protocols remain uncertain. In mixed training sessions, NMES is either simultaneously superimposed (superimposed technique) or separately combined (combined technique) with voluntary exercises. Despite discrepancies in some published results, Paillard et al. (²⁴) concluded that the superimposed technique is an effective therapeutic intervention in rehabilitation. Whatever outcomes are measured (simple measures of strength or complex tasks involving muscle coordination), however, the enhancements observed with the NMES superimposition technique are no greater than those obtained with training programs using only volitional exercises. In contrast, the combined-training technique induces more pronounced effects than traditional training using only voluntary contractions even in the torque and power output of well-trained subjects. The enhanced efficacy of the combined technique likely results from the cumulative effects of both training methods (quantitative aspect) and from the difference in the motor drives that are specifically involved by NMES and voluntary contractions (qualitative aspect).

ACTIVATION ORDER OF MOTOR UNITS WITH NMES

The force produced by a muscle depends on the amount of motor unit activity, varying with the number of motor units that are active (motor unit recruitment) and the rates at which motor neurons discharge action potentials (rate coding). Motor units with low recruitment threshold are composed of slow-twitch muscle fibers, whereas units with high recruitment threshold contain fast-twitch fibers. Motor units are recruited according to the size principle during voluntary contractions. This means that during the graded increase in force, motor units are activated in an order that proceeds from the smallest to the largest. This stereotypical order of recruitment mainly depends not only on the size of the motor neurons but also on their biophysical properties and the distribution of synaptic input. As a consequence of these effects, the change in membrane potential of a motor neuron in response to a given synaptic current is proportional to its input resistance. Because small motor neurons have the greatest input resistance, they experience the largest change in membrane potential for a given synaptic current and, therefore, reach threshold for the discharge of an action potential with lesser synaptic currents. In contrast, motor units are recruited in a r eversed order in preparations when electric shocks are applied directly to nerves. The change in activation order occurs because there is an inverse relationship between axon diameter and axial resistance, which enables the flow of current along the axon to occur at a lower transmembrane current in large-diameter axons. The influence of transcutaneous electrical stimulation on recruitment order in humans is less clear. Although some studies supported a reversed order during NMES (^15,25,29), others found a similar recruitment order as voluntary contractions (^1,4,16). These diverse results could have been related to differences in experimental approaches and muscles that were studied. Furthermore, the activation order of motor units in these studies was examined indirectly.

To analyze the influence of NMES on recruitment order more directly, Feiereisen et al. (¹⁰) compared the recruitment order of pairs of single motor units in the tibialis anterior during graded voluntary contraction and transcutaneous NMES. The results showed a reversal of recruitment order in about one third of the motor unit pairs that were examined (Fig. 3). This recruitment reversal during NMES was not related to the magnitude of the difference in recruitment threshold between motor units but increased slightly when the width of the electrical pulse was reduced (35% vs 28%). The relatively limited influence of NMES on motor unit recruitment order can be explained by the deeper location of larger diameter axons (¹⁴) or by a small difference in size of the axonal branches innervating the different types of motor units in the tibialis anterior (¹⁶). These results indicate, however, that in addition to difference in the nerve axon input resistance, other factors such as the distance between the stimulating electrode and the axons and the morphological organization of the axonal branches within the muscle also play a role in the recruitment order of motor units during transcutaneous NMES. In conclusion, if transcutaneous NMES seems to favor the recruitment of high threshold motor units, the order of motor unit recruitment is not a strict reversal of that observed during voluntary contraction.

Figure 3: Illustration of the recruitment order of two motor units (A, B) from the tibialis anterior with different action potentials (1) and mechanical properties (2) during voluntary (C) and electrically induced (D) contractions. Traces A and B are averaged waveforms (117 and 66 traces, respectively), whereas single trials are shown in (D). In voluntary isometric graded contractions (C), these motor units were recruited according to their sizes (recruitment thresholds: 87 and 203 N for motor unit 1 (MU1) and motor unit 2 (MU2), respectively). In electrically induced contractions (pulse widths of 1 millisecond), a reversal in the order of activation occurred (D), and MU2 was recruited at a lower level of stimulation than MU1 (7 vs 17 mA, respectively). Such recruitment reversal of motor unit pairs occurred in 35% and 28% of the recorded population of units (n = 302) for pulse widths of 0.1 and 1 millisecond, respectively. (Reprinted from Feiereisen, P., J. Duchateau, and K. Hainaut. Motor unit recruitment order during voluntary and electrically induced contractions in the tibialis anterior. Exp. Brain. Res. 114:117-123, 1997. Copyright © 1997 Springer-Verlag. Used with permission.)

MUSCLE ENERGETICS DURING ELECTRICALLY INDUCED CONTRACTIONS

Despite the widespread clinical and athletic use of NMES, few studies have explored the metabolism of electrically contracted muscle in humans. In the 1980s, biopsy samples were taken from human muscles subjected to electrical stimulation, but experimental designs were restrictive (ischemic conditions) and far away from the practical use of NMES. Moreover, the biopsy technique is invasive and does not allow dynamic investigations in the same tissue sample. In contrast, the ³¹P NMR spectroscopy (NMRS) can continuously monitor, in the same representative mass of tissue and with precise acquisition timing, the phosphorylated metabolites involved as energy transducers. There are a limited number of research groups that have used NMRS to study NMES in humans, partly because it is technically difficult to reduce interference between the current generator and the NMR system. Available studies reported dramatic changes in intracellular pH and in inorganic phosphate (Pi) and phosphocreatine (PCr) concentrations during stimulation protocols applied to different muscles (²⁷). An improved understanding of the mechanisms involved during NMES has been gained from newly interleaved NMR techniques. ³¹P and ¹H NMRS have been used to simultaneously monitor the metabolism of high-energy phosphates (via the Pi and PCr peaks) and cellular oxygenation (via the deoxymyoglobin peak) in the electrically stimulated quadriceps (²⁹). Results indicated that the metabolic demand associated with NMES was substantially greater than that observed during voluntary contractions requiring the same force. Furthermore, the greater energy charge was associated with more profound changes in the level of cytoplasmic oxygenation, presumably because of a higher level of local oxygen consumption. The positron emission tomography technique has already shown a disproportionate level in the muscle rate of oxygen metabolism and in the local blood flow during electrically evoked contractions (²⁶).

The greater metabolic demand during NMES could result from additional events in the pattern of muscle contractile activity. First, the stimulation frequency that is regularly used to ensure a maximal tetanic force (50-100 Hz) imposes on stimulated fibers an over activation that is associated with an increased metabolic demand (²²). Second, NMES preferentially recruits axonal branches near the electrode, and this recruitment diminishes proportionally with increasing distance from the electrode (²⁸) (Fig. 4). Third, a decrease in the mechanical response linked to fatigue of the superficial fibers can only be compensated by an increase in stimulation intensity, which depolarizes new fibers at a greater distance from electrode but continues to impose a sustained contractile activity to the superficial ones presenting a certain degree of fatigue (²⁸) (Fig. 4). In the specific case of the quadriceps in which there is a predominance of Type II fibers superficially, NMES evokes a preferential depolarization of glycolytic fibers. This pattern explains that the electrically stimulated quadriceps is more acidotic than the intentionally contracted one (²⁹).

Figure 4: Percentage of activated regions of interest (AROI) (tissue blood flow, >5 mL min⁻¹ 100 g⁻¹) in a thigh that was stimulated at 5% of quadriceps maximal isometric voluntary torque (QMIVT) (mean current intensity = 57 mA) and in a thigh stimulated at 10% of QMIVT (mean current intensity = 84 mA) relative to the distance between the ROI and the electrode (positioned on the vastus lateralis). The range of numbers on the horizontal axis indicates the separation distance (centimeter) between ROI and electrode (all = all ROI). This figure illustrates that muscle recruitment is linked to electrode proximity and that an increase in the stimulation intensity induces the recruitment of new regions situated at a greater distance from electrode. (Reprinted from Vanderthommen, M., J.C. Depresseux, L. Dauchat, C. Degueldre, J.L. Croisier, and J.M. Crielaard. Spatial distribution of blood flow in electrically stimulated human muscle: a positron emission tomography study. Muscle Nerve 23:482-489, 2000. Copyright © 2000 John Wiley and Sons. Used with permission.)

In conclusion, maximal NMES imposes a greater metabolic demand relative to the force that is evoked. This is a consequence of the temporal and spatial recruitment of muscle fibers and constitutes an argument in favor of the combination of NMES with voluntary contractions in the context of rehabilitation or sports training.

CENTRAL EFFECTS OF NEUROMUSCULAR NMES

The observation that substantial strength gain can be obtained in less than 4-5 wk of training, at a time when muscle hypertrophy is limited, has led some authors to suggest that NMES can induce adaptations within the neural system (^7,25). Several lines of evidence support this conclusion. Although the relation between the average rectified electromyogram (EMG) and the neural drive to the muscle is not always linear, this method is classically used to identify the neural mechanisms that contribute to changes in maximal voluntary contraction (MVC) force. For example, it has been found that the EMG during an isometric MVC is significantly increased after 7 wk of NMES in the biceps brachii (⁷). Maffiuletti et al. (²⁰) found a similar increase in the plantar flexor muscles after 16 sessions of NMES, even when the EMG activity was normalized to the M wave.

Alternatively, central adaptations can be inferred by comparing the force exerted during an MVC with the force that can be elicited artificially with electrical stimulation (tetanic contraction) or by assessing the maximality of an MVC with the twitch superimposition (interpolation) method. The latter method involves superimposing one or a few supramaximal electrical stimuli on the motor nerve (indirect stimulation) that supply the target muscle during an MVC. If the stimulus evokes extra force, then the central drive during the contraction is not maximal, and the individual is deemed to be exhibiting a deficit in voluntary activation. A decrease in the voluntary activation deficit after an intervention indicates an increased motor unit recruitment or discharge rate (⁹). Duchateau and Hainaut (⁸) found that at 6 wk of training, the thumb adductor muscles produced only a slightly greater increase in MVC force compared with a maximal tetanic contraction (15.5% vs 9.5%). Maffiuletti et al. (²⁰) reported a significant reduction of the superimposed twitch in the triceps surae that corresponded to an 11.9% increase in voluntary activation after 16 sessions of NMES. Similarly, Gondin et al. (¹¹) found an increased activation of 6% for the knee extensor muscles after 8 wk of NMES training.

Additional evidence of neural adaptation after an NMES training program comes from the observation of its effect on the nonexercised contralateral limb. This adaptation, called "cross-education," has been demonstrated by several studies after a short program of training by NMES as what occurs after training with voluntary contractions (³⁰). In the absence of EMG activity in the contralateral muscles during the training of the other side of the body, and although the underlying mechanisms are not yet elucidated (⁵), this cross-transfer phenomenon during NMES is consistent with the concept of neural adaptations by training with NMES.

UNDERLYING MECHANISMS OF NEURAL ADAPTATIONS

Neuromuscular electrical stimulation of human muscle or peripheral nerve is commonly used to generate muscle contraction primarily by activating motor axons and thereby inducing an adaptation in the muscle. However, the electrical stimulation can also generate action potentials in sensory axons, and the input can activate spinal motor neurons in an order that corresponds to the size principle (²⁵). Because a program of NMES does not change spinal reflexes (Hoffmann reflex or H-reflex), although voluntary activation was increased, Maffiuletti et al. (²¹) suggested that the adaptations were mainly located at the supraspinal level. Nevertheless, as suggested by the authors, a possible role for the cortex in the development of central torque after training with NMES is not excluded because FES can cause long-lasting increases in cortical excitability through long-loop pathways. Furthermore, a recent functional neuroimaging study has shown substantial cerebral cortex activation during electrically evoked contractions of human wrist extensor muscles (¹³). As for training with voluntary contractions, additional studies are needed to locate more precisely the mechanisms (spinal cord vs cortex) responsible for the increase in neural activation.

The relative involvement of sensory feedback in NMES can be manipulated by changing the pulse width (²). When electrical stimulation is delivered using wider pulse widths (~1 millisecond) than classically used for NMES stimulation (0.1-0.5 millisecond), neural adaptations can be maximized because the electrically evoked sensory volley can contribute to the recruitment of spinal motor neurons. Baldwin et al. (²) showed that extra torque can be developed in the plantar flexor and extensor muscles and the wrist flexor muscles, during submaximal tetanic contraction at high frequency (100 Hz) with long pulses. Because this additional torque can also be evoked at intensities below motor threshold, it represents the central contribution from the recruitment of the spinal motor neurons by the evoked afferent volley (⁶). Although a possible role for the cortex in the development of central torque is not excluded (see above), experimental data suggest that this recruitment arises mainly from the development of persistent inward currents in spinal motor neurons or interneurons (i.e., plateau potential) (⁶) or potentiation of neurotransmitter release from Ia afferents. Many factors are yet to be investigated with respect to the optimal stimulation parameters in NMES applications, but the use of submaximal tetanic contractions induced by wide pulse widths and high frequencies of stimulation is promising for rehabilitation.

CONCLUSIONS AND PERSPECTIVES

In summary, long-term NMES can cause beneficial adaptations in the neuromuscular system. These adaptations are mediated not only by muscular but also by neural mechanisms, and their relative importance seems to be specific to the modalities of the electrical stimulation and characteristics of the stimuli. Tetanic contractions elicited by pulses of high intensity and short duration induce a high metabolic stress in the muscle, contribute to the reversal of motor unit recruitment, and improve the maximal capability of the neuromuscular system primarily not only through increased force-generating capacity of the muscle but also through intensified voluntary activation. In contrast, tetanic contractions induced by pulses of low intensity and long duration favor the normal recruitment of motor units (size principle) (²⁵) and neural adaptations through reflex inputs to the spinal cord and supraspinal centers. Maximizing the muscular and neural adaptations specifically to the requirements for treatments or training regimes may be beneficial for patients and athletes. Therefore, therapists and coaches need more precise and scientifically validated indications concerning practical modalities of stimulation (number, size, and localization of electrodes), as well as pulse intensity and width for the most often stimulated muscles. Furthermore, better-designed studies are needed to compare the influence of the combined and superimposed techniques on muscle performance.