| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Reports |
Y Laufer, PT, DSc, is Faculty Member, Physical Therapy Department, University of Haifa, Faculty of Social Welfare and Health Studies, Eshchol Building, Room 910, Haifa, Israel 31905
M Elboim, PT, MSc, is a PhD candidate in the Physical Therapy Department, University of Haifa
Address all correspondence to Dr Laufer at: yocheved{at}research.haifa.ac.il
Submitted January 2, 2008;
Accepted June 9, 2008
 | Abstract |
|---|
Objective: The objective of this study was to compare maximally induced strength (force-producing capacity) of contractions, muscle fatigue, and discomfort associated with an LPC and with 3 KACs differing in frequency and duration of burst modulation.
Design: This was a repeated-measures trial, with randomized order of current presentation.
Setting: The study was conducted in the physical therapy laboratory at the University of Haifa.
Subjects: Twenty-six volunteers without impairments, with a mean age of 27.4 years (SD=5.0, range=21–45), participated.
Intervention: All currents were applied in separate sessions to the wrist extensors of each subject. Currents consisted of an LPC with a 50-Hz pulse frequency and 3 KACs with a 2.5-kHz carrier frequency, including the "Russian current" (RC) burst modulated at 50 Hz with 25 cycles per burst and 2 currents burst modulated at 20 or 50 Hz with 10 cycles per burst.
Measurement: The maximal electrically induced isometric force, the force integral of 21 electrically induced consecutive contractions, and the degree of discomfort were recorded.
Results: Force of contraction was not affected by type of current. The LPC was least fatiguing, and the RC was most fatiguing, with the 2 other KACs having an intermediate effect. Degree of discomfort was higher with the KAC modulated at 20 Hz.
Conclusions: When comfort, strength, and fatigue are considered jointly, the LPC is advantageous. Electrically induced fatigue is affected by the number of cycles per second, rather than the number of bursts per second.
| Introduction |
|---|
 TopAbstract Introduction MethodResultsDiscussionConclusionReferences |
|---|
It has been demonstrated that skeletal muscles fatigue more rapidly during repetitive stimulation than during volitional contractions.10,11 Two factors contribute to this effect: motor unit recruitment order and activation frequency. Motor unit recruitment order in voluntary contractions begins with fatigue-resistant muscle fibers working at a motor unit firing rate of no more than 30 Hz when a contraction is initiated and at an even lower rate as the contraction is sustained.12,13 In contrast, the recruitment order during NMES is either random14 or begins with the fast-fatigable muscle fibers.15 Furthermore, in order to obtain a contraction of maximum force via NMES, it is important to use higher-than-normal stimulation frequencies in the range of 50 to 80 Hz.16 Yet, the higher the stimulation frequency, the more rapid the decline in force production capabilities.17–19 Thus, in determining the optimal stimulation characteristics for muscle rehabilitation, it is important to select parameters that evoke strong muscle contraction with minimal discomfort and minimal fatigue.
Two types of stimuli are commonly used in clinical practice: low-frequency pulsed currents (LPCs) and kilohertz-frequency alternating currents (KACs).20 The LPC stimulators typically deliver either monophasic or biphasic pulses at frequencies ranging between 1 and 150 Hz. The KAC stimulators deliver a biphasic, symmetrical waveform (typically sinusoidal) at frequencies ranging between 1 and 10 kHz. The rationale for using currents in the kilohertz-frequency range is related to the role of the skin as a capacitive barrier to the flow of current. Thus, with the increased current frequency, the skin may offer progressively lower impedance, allowing less electrical energy to dissipate peripherally and more electrical energy to penetrate to the muscle.21,22
However, as nerves cannot respond at kilohertz frequencies23 and the muscle fatigues rapidly with increased frequency,17 KACs are typically modulated to produce bursts at low frequencies ranging between 1 and 150 Hz, with varying burst-interburst ratios (burst duty cycles).20 In accordance with a phenomenon first described by Gildemeister,24 it is assumed that the successive pulses within each burst will summate to produce depolarization at subthreshold levels, with the total number of action potentials (APs) determined by the number of bursts. This phenomenon recently has been substantiated in studies demonstrating that the threshold voltage decreases in accordance with an increase in the number of alternating current cycles per burst (as determined by the burst duration).25,26 The accepted explanation for this phenomenon is that when the nerve fiber membrane undergoes subthreshold polarization and successive pulses occur during the refractory period before the changes in ion concentrations have time to recover, these changes can summate to produce an AP.
The most well-known KAC used clinically for muscle strengthening is "Russian current" (RC), which is a 2.5-kHz sinusoidal alternating current applied at a burst frequency of 50 Hz and a burst duty cycle of 1:1.27 This current gained popularity following its introduction in 1977 by the Russian scientist Kots, who claimed that it produced muscle strength gains of up to 40% in athletes (see review by Ward and Shkuratova27). Yet, despite the popularity of the RC, few studies have examined its efficacy in comparison with LPCs. In a recently published study, Ward et al28 determined that, although the classical RC stimulation elicited significantly less discomfort than a monophasic pulsed current of equal pulse duration and frequency, the torque produced by the RC was significantly lower. Moreover, in a previous study,29 not only did monophasic and biphasic waveforms generate contractions with greater torque than the RC, but both waveforms also were significantly less fatiguing.
It has been suggested that although low-intensity motor threshold stimulation of burst-modulated KAC may result in subthreshold summation leading to a single AP per burst,24,25 perhaps at the higher stimulation intensities that are used to elicit strong muscle contractions, the summation of just a few cycles in each burst may be sufficient to produce an AP. Thus, in the RC, where the 10-millisecond burst duration contains 25 sinusoidal cycles, it is possible that each burst elicits several APs, leading to a nerve firing rate of some multiple of 50 Hz, which, in turn, results in the observed accelerated muscle fatigue.29–32
To the best of our knowledge, no one study has compared the effect of the RC and a low-frequency current with respect to strength of contraction, degree of fatigue, and degree of discomfort. Moreover, the present study was designed to examine whether the previously noted strong fatiguing effect of the RC is indeed a function of the number of cycles per burst. To this end, we compared an LPC with biphasic, symmetrical square pulses delivered at 50 Hz with 3 current settings, all utilizing a 2.5-kHz carrier frequency. One setting involved the classical RC with a burst frequency of 50 Hz and a burst duty cycle of 1:1, generating a total of 1,250 cycles per second. The other 2 current settings involved either a 20- or 50-Hz burst frequency at a burst duty cycle of 1:4, each generating a total of 500 cycles per second (Tab. 1). Phase duration was identical in all 4 currents (200 microsecond). Our hypothesis was that the degree of fatigue would be determined by the total number of cycles per second, rather than by the number of bursts per second, with the current setting involving a higher number of cycles per second and inducing greater fatigue.
|
| Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
Testing Procedure
Each subject participated in 4 separate 30- to 45-minute testing sessions conducted by the same researcher. The time that elapsed between sessions ranged from 48 hours to 7 days. Delayed-onset muscle soreness in the stimulation area was cause for postponing repeated testing; however, none of the subjects reported such soreness beyond 3 days. The effect of a current with one particular stimulus setting on contraction strength and fatigue of the wrist extensors of the nondominant hand was tested in each session. The order of stimulating currents was randomized between subjects using computer-generated random numbers. The participants remained blinded to the stimulus parameters used throughout the experiment. The stimulus characteristics used for testing are provided in Table 1. The 932 Myomed unit,* which is a constant-current clinical stimulator, was used to elicit all of the electrically induced contractions.
Each session included the following stages: setup, determination of the maximal voluntary isometric contraction (MVIC), habituation to the stimulus, determination of the maximal electrically induced contraction (MEIC), extended voluntary contraction, and a fatigue test (a summary of the sequence of events is presented in Tab. 2). At the conclusion of the final testing session, subjects indicated the degree of discomfort associated with each of the currents.
|
to the area of stimulation. The proximal electrode was placed 1 cm distal to the lateral epicondyle on the line between the humeroradial joint and the inferior radioulnar joint. The distal electrode was placed midway on the same line so that it lay over the extensor digitorum muscle. The subjects then were stabilized in a seated position with the forearm fully supported, while the shoulder and elbow were both maintained at approximately 60 degrees of flexion. The subjects’ forearm was fixated with Velcro straps to a specially designed splint, with the wrist at the end of the splint, allowing for gravity-assisted wrist flexion. A specially designed strap was attached with Velcro around metacarpal bones 2 through 5, parallel to the wrist joint, with the thumb web space determining the distance between the strap and the wrist joint. One end of a high-tension wire was connected to the wrist strap at the palmar surface of the hand in line with the third metacarpal, and the other end was connected to a model SM250 QMA force transducer.
The force transducer was maintained at 90 degrees in relation to the palmar surface of the hand. The length of the tension wire was adjusted so that when the wrist extensors contracted, the hand would rise until the wrist angle reached 0 degrees of wrist extension, at which time a pull was exerted on the transducer. Thus, the apparatus provided measurement of isometric wrist extension contraction, which allowed for computerized online reading of the contraction force. The QMA software package provided summary data of force and force-integral outputs at a 36-Hz data acquisition rate.
MVIC test.
Following one practice trial, the subjects were requested to "raise your hand as hard as you can" for 3 consecutive contractions and to relax completely between contractions. The duration of each contraction was 3 seconds, and the rest period between trials was 60 seconds.
Habituation to stimulus.
After a short intermission, subjects were habituated to the electrical current to be tested in that particular session. To do so, the current amplitude was gradually increased until subjects indicated to the tester that their limit of tolerance had been reached. Although the actual intensity of the current was not disclosed to the subjects, they could observe the force detected by the force transducer on the computer screen. The process was repeated for 2 to 4 trials until no additional current intensity could be tolerated.
MEIC test.
Following a 5-minute rest period, the stimulator was programmed to deliver 3 consecutive stimulations with the maximally tolerated intensity determined in the previous stage. Stimulation time consisted of 3 seconds of ramp-up time, 3 seconds of on time, and 90 seconds of rest time. The subjects were instructed to fully relax during the electrically stimulated contractions and to "let the machine work for you." It should be noted, however, that some of the subjects could tolerate current intensities strong enough to induce a co-contraction of the wrist flexor muscles, thus reducing the recorded force of the wrist extensors. As force of contraction was monitored throughout the testing session, whenever a co-contraction occurred, it was observed that at a certain point the recorded MEIC started to decrease as intensity was increased. In these cases, the current intensity was limited to ensure maximal recorded force of the wrist extensors.
Extended voluntary contraction.
Following a 5-minute rest period, the subjects were requested to perform a 7-second-long maximal voluntary contraction. The force integral, as determined by the area under the force curve (AUC), was used to determine test-retest reliability of this measure and to normalize the force integral of the electrically induced contractions during the fatigue test.
Fatigue test.
The fatigue component of the protocol followed a 5-minute rest period. Utilizing each subject's maximally tolerated stimulus, the electrically induced fatigue test consisted of a series of 21 seven-second-long electrical stimulations (with a 3-second ramp-up time), separated by 3-second-long rest periods. Subjects were asked to relax during the electrically stimulated muscle contractions. As the muscle fatigued, stimulation at a constant intensity did not consistently elicit a contraction of sufficient force to raise the hand to 0 degrees of dorsiflexion and to exert a pull on the force transducer. However, due to occasional recovery of the muscle during subsequent stimulations, the entire series of 21 contractions was completed in all tests.
Degree of discomfort.
At the conclusion of the last session, the 4 currents used in the study were reintroduced consecutively in the same order initially presented in the 4 sessions. The intensity of each current was increased gradually until subjects reported that maximum tolerable stimulation has been reached. At this point, the subjects were asked to rate, on a 10-point visual analog scale (VAS), the degree of discomfort associated with the applied current.
Data Analysis
Intraclass correlation coefficients (ICCs) for the voluntary force measured with the strain gauge were calculated using the means of the 3 voluntary contractions performed consecutively at the beginning of each assessment session. The ICC of the force-time integral was calculated using the AUC of the single, 7-second-long voluntary contraction performed prior to the fatigue test in each of the 4 assessment sessions. The ICCs were determined by the variance components derived from the appropriate random-effects analysis-of-variance (ANOVA) models, using data from all 4 sessions of data collection.
The force of the electrically induced contractions produced by each waveform was determined by using the mean of the 3 MEICs, expressed as a percentage of the mean of the 3 MVICs. The degree of discomfort was assessed with the VAS score. Mixed-model ANOVAs, with the waveform as the independent variable repeated across subjects (which served as a random factor) and followed by Tukey-Kramer analysis, were used to compare force of contraction and degree of discomfort across the different waveforms.
A similar analysis was used to determine the following 2 variables used to describe the effect of current waveform on fatigue: (1) percentage of elicited contractions and (2) number to 
50%. For the percentage of elicited contractions, the number of stimulations that elicited a contraction with measurable force was counted for each fatigue trial and was expressed as a percentage of 21, which was the total number of possible contractions. The number to 50% refers to the number of contractions elicited until the force of contraction was 50% of the initial contraction for the first time.
Additional analysis of the effect of stimulation on muscle fatigue assessed the effect of stimulation on the force-time integral of the consecutive contractions, using 2 variables. First was the AUC expressed as a percentage of the MVIC (AUC %MVIC). For this variable, the AUC of each 7-second-long contraction during the fatigue test was expressed as a percentage of the AUC of the 7-second-long voluntary contraction performed prior to the fatigue test. Second was the AUC expressed as a percentage of the initial contraction (AUC %100). Because the change in the AUC can be affected not only by the rate of decrease in force production but also by the strength of the initial contraction, the AUC (normalized to the voluntary AUC) was thus normalized to the initial contraction produced by the specific waveform. In this analysis, the AUC of the first contraction for each waveform was considered as 100%, and consecutive contractions were expressed as a percentage of the initial contraction. A regression model was constructed to evaluate the differences in fatigue across waveforms over the 21 contractions. Because the fatigue versus contraction number profile had the appearance of a reciprocal relationship, the AUC for each contraction (expressed as %MVIC in one analysis and as %100 in the second analysis) was converted to 1/(AUC + 1) before conducting the analysis. A random-effects model was used to accommodate the repeated measures from each subject. Tukey-Kramer analysis was used for comparison across different waveforms. The analysis was performed using SAS version 6.09.
| Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
Strength of Maximal Electrically Induced Contractions
The group means and standard deviations of the MEICs, expressed as a percentage of the MVIC for the 4 currents, are presented in Table 3. The ANOVA showed no significant differences in strength among the contractions induced by the different currents.
|
|
50%. The ANOVA indicated a significant effect of current for this variable (F3,75=6.00, P=.001). Tukey-Kramer post hoc tests indicated that the number to 50% was significantly lower for the 50-Hz 1:1 current than for either the LPC or the 20-Hz 1:4 current settings (P<.0005 and P=.04, respectively). The observed difference between the 50-Hz 1:1 and 1:4 currents did not reach significance (P=.06). Figures 2 and 3 present the group means of AUC %MVIC and AUC %100 per current, respectively. The ANOVA indicated a significant effect of current on both AUC %MVIC and AUC %100 (F3,2151=14.29, P=.0001 and F3,2151=8.66, P=.0001, respectively), as well as a significant effect of contraction number on both variables (F3,2151=1958.48, P=.0001 and F3,2151=119.31, P=.0001, respectively). No interaction effect was present for either variable. Although the absence of an interaction effect indicates no difference in the rate of change between contractions (the slope), the significant effect of current indicates a difference in the magnitude of fatigue value between currents. Table 4 presents a summary of the Tukey-Kramer tests comparing the degree of fatigue induced by the different waveforms.
|
|
|
| Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
Our study results demonstrated that stimulation with an LPC resulted in significantly less fatigue than stimulation with a 2.5-kHz KAC, irrespective of its burst frequency and burst duty cycle. This held true when the force integral of each contraction was expressed as a percentage of the MVIC, as well as when it was normalized to the first contraction. Moreover, no difference in the degree of fatigue was demonstrated between the 2 KACs, which differed only in their burst frequencies (20 Hz versus 50 Hz) and not in their total number of cycles per second. At the same time, the comparison between the 2 currents with an equal burst frequency (50 Hz) demonstrated significantly greater fatigue when the burst duration was longer and included 25 cycles per burst (the RC), as opposed to the shorter burst duration with 10 cycles per burst (50-Hz 1:4 current). Thus, the rate of fatigue induced by the KAC was clearly related to the total number of pulses delivered, rather than to the number of bursts.
In a study using a similar fatiguing protocol, Laufer et al29 previously demonstrated that the RC induced accelerated fatigue in comparison with the LPC at 50 Hz. However, there is one fundamental difference in the fatiguing protocols used in these 2 studies. In the earlier study, muscle stimulation was terminated when it failed to induce a contraction strong enough to raise the limb against gravity, whereas stimulation was maintained until the entire series of 21 consecutive trials was completed in the present study. As demonstrated in Figure 1, successive stimulations with the LPC resulted in a gradual decline in the force-integral values, whereas all KACs induced great fluctuations in the force integrals of successive contractions. This example clearly demonstrates that although the 3 currents with a 2.5-kHz carrier frequency induced a dramatic decrease in the force integrals of successive contractions, the muscle often was able to recover very quickly at least some of its ability to elicit a detectable force.
Thus, in the presented example, even when fatigue completely inhibited detectable muscle contraction in the 2 currents with a total of 500 cycles per second (50-Hz 1:4 and 20-Hz 1:4 currents), it took 1 to 2 contractions for the muscle to recover at least partially so as to elicit a measurable contraction. Moreover, even with the RC, which was dramatically more fatiguing, the muscle recovered to some extent after 9 stimulations that were unsuccessful in eliciting a contraction. This recovery of contractile force was demonstrated repeatedly by the majority of the subjects.
The fatigue pattern observed following stimulation with the KACs in the present study is in keeping with a phenomenon known as "high-frequency fatigue."18 This form of fatigue is characterized by an extensive loss of force at high frequencies of stimulation and a rapid recovery when stimulation frequency is reduced. It is believed to be related to a failure of the AP to propagate within the muscle fibers.35 Although high-frequency fatigue produces dramatic loss of force, it is not considered a normal mechanism of fatigue during sustained voluntary contraction.9 Therefore, it seems that no benefit would be gained by training a muscle with a highly fatiguing NMES protocol, which has no relevance to the force-generating and force-maintaining characteristics of skeletal muscles.
The higher rate of fatigue induced by the RC, in comparison with the other 2 KACs, and the characteristic high-frequency fatigue of these currents support our hypothesis that the total number of APs is not determined solely by the number of bursts per second. As has been postulated previously,29–32 the results strongly suggest that at intensities eliciting strong muscle contractions, it may be sufficient for just a few of the cycles within each burst to summate and produce an AP, resulting in multiple APs per burst. For example, the fatigue curves in the present study can be explained if we assume the occurrence of one AP per 5 cycles within each burst. Thus, the 50-Hz 1:4 current with 10 cycles per burst and 50 bursts per second had the same total number of APs as the 20-Hz 1:4 current with 25 cycles per burst and only 20 bursts per second (a total of 100 APs). By the same calculation, the RC, with its relatively long burst duration (25 cycles) and a 50-Hz burst frequency, may have produced a total of 250 APs. To the best of our knowledge, the effects of burst duration and frequency on fatigue have not been examined previously. However, previous studies32,36 have demonstrated that longer burst durations elicit weaker muscle contractions. Given the dramatic fatiguing effect observed in the present study, it is possible that the weaker contractions were due to a decline in force occurring during the contractions as a result of the bursts of longer duration.
The present study determined no difference in the degree of discomfort associated with the classical RC (50-Hz 1:1 current) and the low-frequency current. Previous research findings regarding degree of discomfort associated with KAC and LPC are inconsistent. Similar to our results, Lyons et al37 reported no significant difference in discomfort between a 2.5-kHz alternating current and a low-frequency current. In another study,28 significantly more discomfort was associated with a low-frequency current than with an RC. These contradictory results are probably related to the modes used for reporting discomfort. In the first study, as in the present study, subjects were asked to rate the degree of discomfort associated with the maximal current tolerated using a VAS, whereas the latter study tallied the number of times that the subjects reported an unpleasant feeling during the ramp-up to maximum tolerable stimulation.
Ward and colleagues26,38,39 determined that in the current frequency range between 1 and 10 kHz, discrimination between pain and motor stimulation is optimal around 10 kHz. This held true regardless of whether the applied current was burst modulated or single sine-wave stimulation. Thus, although threshold voltages decrease with the increase in the number of alternating current cycles per burst,25 to the extent that a separation between motor and pain thresholds is a predictor of comfort, their results imply that the important factor affecting comfort is the phase duration of a single sine wave, rather than the number of cycles per burst. This conclusion is further supported by the results of the present study, which showed no difference in the degree of discomfort between the 2 currents with equal carrier and burst frequencies (50-Hz 1:1 and 50-Hz 1:4 currents) but a different number of cycles per burst (25 versus 10, respectively).
The only current in the present study that was perceived as less comfortable than either the LPC or the RC was the current burst modulated at 20 Hz (20-Hz 1:4 current). In fact, although the total number of cycles per second in the 20-Hz 1:4 and the 50-Hz 1:4 currents was the same, most participants perceived the 20-Hz 1:4 current as a significantly different "type" of current. Unlike the currents involving stimulation at 50 Hz, it was perceived more as a "pulsating" current, as each of the 20 bursts could be separately distinguished, much like an LPC with a pulse frequency of 20 Hz. Thus, the degree of discomfort seems here to be more likely the result of the number of bursts per second—whether involving a single pulse, as in the LPC, or multiple cycles, as in the KACs.
Several limitations need to be considered in interpreting our results. The measured force of the extensor muscles in the present study was limited in some subjects by a co-contraction of the wrist flexor muscles. This may have obscured the true, electrically induced maximal force elicited by the various currents. The effect of a co-contraction has not been previously referred to as a confounding factor in studies examining the strength of maximally induced contractions, probably because the majority of those studies focused on the much larger quadriceps femoris muscle.29,36,37 However, it is possible that better control of the co-contraction would have been achieved in the present study had we used smaller electrodes.
The on/off ratio used in the present study (10 seconds on/3 seconds off) is not the ratio typically used during muscle strengthening protocols, where the ratio used ensures much longer rest periods between contractions.20 Therefore, it is possible that the dramatic, electrically induced fatigue observed in our study will not be as major a factor in the clinical use of KACs for muscle strengthening. Yet, a recent study37 comparing the muscle torque induced with an LPC and a KAC, using the typical 10 seconds on/50 seconds off ratio, demonstrated a significantly lower torque integral following 10 contractions with the KAC. Thus, the often-used 1:5 duty cycle may not be sufficient to offset the strong fatiguing effect of the RC.
The greater discomfort associated with the 20-Hz 1:4 current was somewhat surprising because the subjects were asked to rate the degree of discomfort associated with each current when stimulation was set at maximum tolerable intensity. Presumably, at this stimulation intensity, degree of discomfort should have been identical in all currents. Nevertheless, subjects frequently indicated that the "pulsating nature" of the KAC modulated at 20 Hz, as opposed to the more "continuous" sensation of the other currents, rendered it less pleasant. Although rating the degree of discomfort at a fixed current intensity, rather than at the maximally tolerated intensity, may have provided us with a more absolute indication of the relative discomfort of each current, we believe that because during clinical application of NMES one normally increases intensity to maximum tolerance, the information obtained in the present study is more clinically relevant.
The findings of the present study have important clinical implications. Although both LPC and KACs can be used to induce contractions of sufficient force, which could be applied clinically for strengthening purposes, burst-modulated currents at kilohertz frequencies induce high-frequency fatigue, which inhibits the ability of the muscle to sustain forceful contractions. As these currents offer no advantages in terms of either peak force or degree of discomfort, we recommend, based on the present findings, that clinicians applying NMES preferentially choose an LPC. Future studies should determine to what degree the application of an LPC at higher frequencies than the 50 Hz examined here induces high-frequency fatigue and what on/off duty cycle ratio is best suited to offset this phenomenon. Such studies are necessary not only with subjects who are healthy but also with patient populations, as muscle atrophy and weakness may affect the muscles’ susceptibility to fatigue. As the present study was limited to a single application of each current, future studies are clearly necessary to examine the accumulative training effect of the various currents following repeated sessions of NMES.
| Conclusion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
| Footnotes |
|---|
The study was approved by the Institutional Review Board of the University of Haifa.
An abstract and oral presentation of this research were given at the 58th Conference of the Israeli Physical Medicine and Rehabilitation Society; Israel; November 29, 2007.
* Enraf-Nonius, Delft, the Netherlands. 
Velcro USA Inc, 406 Brown Ave, Manchester, NH 03103.
QMA Systems, PO Box 170, Gainesville, GA 30503.
SAS Institute Inc, PO Box 8000, Cary, NC 27513.
| References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConclusionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  A. R Ward Electrical Stimulation Using Kilohertz-Frequency Alternating Current Physical Therapy, February 1, 2009; 89(2): 181 - 190. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
