During the last decades, increasing attention has been drawn to electromyostimulation (EMS) for rehabilitation in patients (Enoka, 1988; Morrissey, 1988) and strength training in healthy subjects (Dudley and Stevenson, 2003; Maffiuletti et al., 2002). It has been shown that the application of EMS results in maximal voluntary strength gains (Maffiuletti et al., 2002). Various studies have proposed that these strength gains can be attributed to neural adaptations (Colson et al., 2000; Maffiuletti et al., 2000a). When a muscle is artificially activated, by means of EMS, the stimulus evokes action potentials in different structures, for example in intramuscular nerve branches and cutaneous receptors. Thus force is generated directly by activation of motor axons and indirectly by “reflex” recruitment of spinal motoneurons (Collins et al., 2001). Another possibility to induce muscle activation artificially is peripheral magnetic stimulation (Lotz et al., 1989). The cellular mechanism of stimulation is the same for both techniques. In both cases, charge flows into an excitable cell membrane causing a change in transmembrane potential. If an adequate stimulus is applied, it can cause the depolariztion of the membrane and the initiation of an action potential (Barker, 1999).

Peripheral magnetic stimulation stimulates intramuscular motor axons which evoke muscle contraction (Machetanz et al., 1994) and, in contrast to EMS, peripheral magnetic stimulation is characterized by a great dePT_H of penetration, the nonexistence of current flowing through the skin and the absence of local high current densities (Barker, 1999). This avoids the activation of cutaneous receptors like nociceptors and mechanoreceptors (Struppler et al., 2003). Repetitive peripheral magnetic stimulation (RPMS) applied to the muscle induces proprioceptive inflow to the central nervous system (CNS) in two different ways: Firstly, the indirect activation of mechanoreceptors due to stimulation induced rhythmic contraction and relaxation as well as vibration of the muscle. This includes the depolarization of fibre groups Ia, Ib and II. Secondly, the direct activation of sensorimotor nerve fibers with orthodromic and antidromic conduction (Struppler et al., 2004). It has been shown that low-intensity magnetic stimulation of gastrocnemius and soleus is able to activate terminal nerve afferents directly and is therefore capable to induce cerebral potentials (Zhu et al., 1996). Furthermore, it has been demonstrated that RPMS is a potential treatment option for spasticity (Krause and Straube, 2005; Marz-Loose and Siemes, 2009) and that it is able to activate modulatory as well as plastic processes in the CNS (Struppler et al. , 2003). Therefore, the purpose of the present study was to examine the effects of RPMS of the m. soleus (SOL) on the amplitude and latency of its maximal H-reflex (H_max, H_max onset latency) and maximal M response (M_max, M_max onset latency) evoked by electrical stimulation of the posterior tibial nerve. Furthermore we investigated the potential RPMS-induced change of plantar flexor peak twitch torque associated with H_max (PT_H).

Twenty-four healthy male subjects with no history of neurological disorders or injuries participated in this double blind study. The subjects were randomly assigned to a treatment group (TG: n = 13, age 24.1 ± 2.0 years, weight 77.8 ± 7.3 kg, height 1.81 ± 0.07 m, training status 9.5 ± 4.2 h/week) and a control group (CG: n = 11, age 23.0 ± 1.6 years, weight 79.6 ± 11.7 kg, height 1.82 ± 0.08 m, training status 8.9 ± 4.5 h/week). All test persons signed informed consent prior to investigation. The study was conducted according to the declaration of Helsinki and approved by the university ethics committee.

All subjects were examined before and after RPMS. The dependent variables were peak-to-peak electromyographic amplitude of M_max (mV) and H_max (mV) for SOL, the H_max/M_max-ratio, H_max onset latency (ms), M_max onset latency (ms) as well as peak twitch torque associated with H_max (PT_H in N·m).

The subjects were placed on the CYBEX NORM Testing & Rehabilitation System (Computer Sports Medicine®, Inc., Stoughton, MA) in a standardized prone position. The subjects were asked to relax their muscles and not to voluntarily contract during testing and stimulation. Their right foot was fixed to the device in order to assess peak twitch torque associated with the maximal H-reflex, produced by plantar flexion as a result of electrical stimulation of the posterior tibial nerve. The knee and ankle joint angles of 180° and 78° were consistently maintained during the experiment. The subjects remained in the same standardized prone position and stayed inactive during the whole experiment. In order to avoid the influence of acoustic signals on amplitude of the H-reflex (Palmieri et al., 2004) the test persons wore ear protection (Bilsom Thunder T3, signal-to-noise ratio = 36 dB).

RPMS was delivered to SOL of the right leg using a MagProR30 stimulator (Medtronic, Denmark) and a MC-125 circular coil. In order to avoid overheating the circular coil was cooled in a water bath previous to muscle stimulation. RPMS took place once. The stimulation protocol included the application of symmetric biphasic pulses with an impulse duration of 290 µs, a frequency of 15 Hz, 100 stimuli per train (train duration 6.6 s), 20 trains, an inter train interval of 2 s and 40 percent stimulator output (60 A/µs magnetic field gradient). A total of 2000 stimuli were applied to SOL which lasted 172 s. The stimulation was applied in a standardized manner to the posterior, lateral and medial part of SOL of TG to reach a deep penetration of the muscle. In this process the different parts of SOL were stimulated successively. This means that the posterior part of SOL was stimulated first, afterwards the lateral part, followed by the medial part. The 2 s inter train intervals were used to change the position of the magnetic coil. The stimulation site was changed till the 20 trains were completed. At the end the posterior, lateral and medial part of SOL received 7 trains, 7 trains and 6 trains, respectively. The stimulation evoked unfused tetanic contractions. CG received a placebo RPMS with 0 percent stimulator output.

The percutaneous electrical stimulus was a square-wave pulse of 500 µs duration delivered by a Digitimer® stimulator (DS7A, Hertfordshire, UK). Due to the subject position, which made unipolar stimulation not feasible, H-reflex was evoked using a bipolar setup (Pierrot-Deseilligny and Mazevet, 2000). SOL H-reflex was elicited with the cathode placed over the posterior tibial nerve proximal to the anode (both Ambu® Blue Sensor N surface electrodes, 2 cm diameter) in the popliteal fossa to avoid anodal block (Fischer, 1992). In order to ensure constant inter stimulus intervals of 10 s (i.e. 0.1 Hz) a Digitimer® train/delay generator (DG2A, Hertfordshire, UK) was used.

The testing procedure included quasi-random stimulation with different current intensities (5-60 mA), resulting in a recruitment curve, until identification of peak-to-peak H_max and M_max. Afterwards H_max was elicited ten times and then averaged in order to achieve reliable results. M_max was elicited using a supramaximal stimulus (~110%). Testing was performed immediately before and 2 min after RPMS. The time between the treatment and the post test was used to refit the cables required for H-reflex stimulation and recording. Onset latencies of H_max and M_max were measured as the time interval between the stimulus artifact and the first deflection of the H or the M wave, respectively.

Torques induced by the ten stimulations of the posterior tibial nerve to elicit H_max were recorded with a sampling frequency of 100 Hz. The generated peak torques were averaged offline. The bipolar EMG recording used Ambu® Blue Sensor N surface electrodes (2 cm diameter) which were firmly attached to the abraded and cleaned skin over SOL of the right leg. The electrodes were applied with a center-to-center distance of 2 cm, lateral to the calcaneal tendon, ~ 17 cm above the calcaneus and below the muscle fibres of the gastrocnemius. The recording electrodes were in line with the presumed direction of the underlying muscle fibres. The resistance between electrodes was kept below 5 kθ© to guarantee high quality EMG recordings. Signals were amplified (amplification factor 1500) with a band-with frequency ranging from 10 Hz to 1.3 kHz and simultaneously digitized online with a sampling frequency of 3 kHz.

Differences between the values before and after RPMS were tested for significance by analysis of variance (ANOVA) for repeated measurements with groups (2) as factor. Differences in means between groups were tested with two-sided student’s unpaired t-tests. In each case the level of significance was established at p ≤ 0.05. SPSS 15.0 (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. Furthermore, effect size (θµ, Cohen’s d) was calculated. Data are presented as group mean values ± standard error of the mean.

At pre-test, no significant difference was observed between the treatment and control group in terms of their variables. After RPMS H_max/M_max was significantly enhanced in the treatment group (TG: 0.585 ± 0.176 vs. 0.630 ± 0.142, p = 0.015, θµ = 0.61) whereas H_max/M_max of the control group remained unmodified (CG: 0.516 ± 0.195 vs. 0.519 ± 0.214, p = 0.843) (Figure 1.). The increment of H_max/M_max of TG after RPMS was not generated by enhancement of the amplitude of H_max (3.68 ± 1.91 mV vs. 3.64 ± 1.68 mV, p = 0.758, θµ = 0.09) but on account of a significantly decreased maximal M wave (6.11 ± 2.10 mV vs. 5.64 ± 1.95 mV, p = 0.027, θµ = 0.49) (Figure 2. and 3">3.). The stable value for H_max and the declined value for M_max of TG are displayed exemplarily in Figure 4., where H-reflex recruitment curves of one subject before and after RPMS are displayed. For each subject, the amplitude of the small M wave preceding the H_max response (MH_max) was comparable for both pre- and post-test (Table 1.).

CG revealed a slight decrease in the crude amplitude of H_max (3.59 ± 1.77 mV vs. 3.48 ± 1.91 mV, p = 0.505), though not significant. Maximal M wave of CG showed no significant difference before and after placebo RPMS (6.78 ± 1.53 mV vs. 6.49 ± 1.51 mV, p = 0.061) but was slightly diminished as well (Figure 2. and 3">3.). Due to the simultaneous insignificant decline of the peak-to-peak amplitude of H_max and M_max the ratio of CG remained stable (0.516 ± 0.195 vs. 0.519 ± 0.214, p = 0.843) (Figure 1.). The onset latency of H_max was significantly increased after RPMS for both groups, while M_max onset latencies did not change significantly (Table 1.). Peak twitch torque values associated with H_max were not significantly different before and after RPMS for the treatment group and the control group (Table 1.).

The 15 Hz RPMS protocol used in this study had no significant effect on the peak-to-peak amplitude of H_max and the corresponding peak twitch torque. Alterations of H-reflex amplitude are normally ascribed to modifications in alpha motoneuron excitability and/or changes in presynaptic inhibition (PSI) of primary muscle spindle afferents (Zehr, 2002). However, we should take into account that H-reflexes evoked during actual voluntary contractions (Aagaard et al., 2002) and/or during performance (Voigt et al., 1998) probably represent a more functional assessment of treatment-induced adaptations at the spinal cord level. Nevertheless, these aspects were not investigated with the present methodology as H-reflex measurement was performed while muscles were relaxed.

The results indicate that 15 Hz RPMS did not have the ability to change the H-reflex amplitude of SOL significantly. However, M_max of TG was significantly reduced after the intervention. The slight decline of H_max and the significant reduction of M_max led to a significant increment of the H_max/M_max-ratio. Similar results were produced by EMS treatment which was applied once as well and led to a significant decline of M_max, while the H_max/M_max-ratiovaried inconsistently (Tanino et al., 2003). The phenomenon of simultaneous decreasing maximal H-reflex and M wave amplitudes to the same extent typically occurs during long lasting H-reflex experiments (Crone et al., 1999) and could be seen as an explanation for the results of CG but not for TG.

The decline of M_max can originate from various different sources and their associated physiology. The explanations include: (I) altered position of stimulation and recording electrodes, (II) temperature, (III) duration of the experiment and (IV) alterations in the peripheral structures below the point of stimulation (Enoka, 2008).

First, changes in the amplitude of M_max can be generated when the stimulating electrode shifted from its original position or the recording electrodes have moved relative to the muscle (Palmieri et al., 2004). However, these sources of error can be excluded because stimulation and recording electrodes remained in a constant position and were fixed to the skin. The electrode positions were not altered during entire testing and stimulation procedure. Furthermore, the subjects remained in the same standardized prone position and were inactive during the experiment. Therefore, the amplitude of the maximal M wave was probably not influenced by electrodes which have moved relative to the muscle due to changed muscle architecture or altered muscle length. The second explanation for a decreased maximal M wave is the alteration of tissue temperature. It has been shown that local cooling of the abductor digiti minimi induced a significant increase in maximal M wave (Denys, 1990). In contrast, Dewhurst et al., 2005 have revealed a decreased M_max of SOL while warming. The underlying mechanism is supposed to be a change in temperature which alters the opening and closing time of the sodium voltage-gated channels along the nerve axon. Cooling slows down the depolarization time and allows more sodium to enter the cell. This results in an increase of the depolarization amplitude, while warming has the opposite effect (Rutkove, 2001). The present decline of maximal M wave is probably not an effect of temperature change because the research environment was stable during the course of the experiment. Furthermore, M_max onset latencies of both groups were not altered. If RPMS had a significant effect on tissue temperature one would expect a visible discrepancy between the onset latencies of both groups as well as between pre- and post-measurement of TG which was not observable. Repetitive muscle actions typically increase intramuscular temperature (Enoka, 2008) but in accordance with Dewhurst et al., 2005 a rise in temperature had led to decreased onset latencies which were not evident. Therefore the temperature effect is not suitable to explain the difference in the amplitude of M_max due to the treatment. Third, as mentioned before a decline H_max and M_max was observed during the course of an experiment (Crone et al., 1999). Therefore, it is important to look at the H_max/M_max-ratio in order to interpret results of H-reflex studies. In fact, this could be an explanation for the slight decline of H_max and M_max of CG but not a clarification for the significant reduction of M_max and the continuance of H_max of TG after RPMS. Fourth, a decline of maximal M wave is observable in relation with alterations in the peripheral structures below the point of stimulation which have the ability to modify neuromuscular propagation (Enoka, 2008). The alteration of neuromuscular propagation is probably an explanation for the results of this study. The relevant locations of concern are (I) impulse propagation along the motor axon into axon branches, (II) neuromuscular transmission and (III) propagation of action potential along the muscle fibre. Krnjevic and Miledi (1958; 1959) suggested that the failure of propagation regarding the motor axon might occur at either the point of branching of the fibre, or in the region where the fibre expands to form the nerve ending. The basic cause is probably the change of ionic gradients associated with loss of excitability (Sieck and Prakash, 1995). Another explanation is the modulation of neuromuscular transmission, due to lowered number of vesicles fused with the presynaptic membrane and quantal size of acetylcholin (ACh) per vesicle (Reid et al., 1999). In this context, the local calcium dynamic is of interest (Gardiner, 2001) because calcium plays an important role in transmitter release, probably by allowing the fusion of vesicles with the presynaptic membrane (Frotscher, 1996). Furthermore, at this level the desensitization of ACh receptors should be taken into account (Giniatullin et al., 1989; Thesleff, 1959). The third level, which is associated with a decline of maximal M wave, refers to changes in the muscle membrane (Shields et al., 1998). That means the alteration of conduction of the action potential along the sarcolemma due to changed ionic gradients (McKenna et al., 2008). The H-reflex response is mainly generated due to the activation of slow twitch units while the maximal M wave is the electrical counterpart of the activation of fast twitch units (Maffiuletti et al., 2000b). As is known the different types of motor units include related types of muscle fibres and corresponding thresholds, respectively (Moritani, 2003). Based on our results it is conceivable that RPMS with the used stimulation protocol affected muscle fibres which are part of the fast twitch units. According to this RPMS probably acted on muscle fibres involved in high threshold motor units which are essential for the generation of the maximal M wave. After the treatment these fibres likely possess a RPMS-induced decline in their responsiveness to the supramaximal stimulus. Therefore we suggest that the impairment of the maximal M wave is due to alterations in the peripheral structures below the point of stimulation which modified the integrity of neuromuscular propagation. In the future, RPMS could be an attractive alternative to EMS for rehabilitation and training applications. However, research is needed to further elucidate the effects.

It is concluded that the RPMS protocol probably did not affect spinal excitability but acted on the muscle fibres which are part of fast twitch units and mainly responsible for the generation of the maximal M wave. RPMS probably modified the integrity of neuromuscular propagation.