| | Reliability of techniques to assess human neuromuscular function in vivoReceived 2 September 2005; received in revised form 21 November 2005; accepted 28 November 2005. published online 23 January 2006. Abstract The purpose of this study was to comprehensively evaluate the reliability of a large number of commonly utilized experimental tests of in vivo human neuromuscular function separated by 4-weeks. Numerous electrophysiological parameters (i.e., voluntary and evoked electromyogram [EMG] signals), contractile properties (i.e., evoked forces and rates of force development and relaxation), muscle morphology (i.e., MRI-derived cross-sectional area [CSA]) and performance tasks (i.e., steadiness and time to task failure) were assessed from the plantarflexor muscle group in 17 subjects before and following 4-weeks where they maintained their normal lifestyle. The reliability of the measured variables had wide-ranging levels of consistency, with coefficient of variations (CV) ranging from ∼2% to 20%, and intraclass correlation coefficients (ICC) between 0.53 and 0.99. Overall, we observed moderate to high-levels of reliability in the vast majority of the variables we assessed (24 out of the 29 had ICC >0.70 and CV<15%). The variables demonstrating the highest reliability were: CSA (ICC=0.93–0.98), strength (ICC=0.97), an index of nerve conduction velocity (ICC=0.95), and H-reflex amplitude (ICC=0.93). Conversely, the variables demonstrating the lowest reliability were: the amplitude of voluntary EMG signal (ICC=0.53–0.88), and the time to task failure of a sustained submaximal contraction (ICC=0.64). Additionally, relatively little systematic bias (calculated through the limits of agreement) was observed in these measures over the repeat sessions. In conclusion, while the reliability differed between the various measures, in general it was rather high even when the testing sessions are separated by a relatively long duration of time. 1. Introduction  Great advances in methodological assessment techniques over the last few decades have led to the ability to non-invasively assess the function of the human neuromuscular system. Methods such as magnetic resonance imaging (MRI), electromyography (EMG), and electrically evoked muscle contractions permit non-invasive determinations of a wide range of both neurophysiologic and skeletal muscle variables. Utilizing these techniques allows for one to study various phenomena in vivo, perturb the system minimally or not at all, and allow repeated measurements of the same site. As a result, there have been significant advances in our understanding of how the nervous and muscular systems function during physical activity as well as disease states. Additionally, over the past few years there has been a burgeoning interest in non-invasive techniques to assess neuromuscular function to the extent that entire meetings have been devoted to this sole topic [17]. Through the use of a combination of voluntary and electrically evoked muscle contractions coupled with recordings of electrical (EMG) and mechanical (force) responses one can non-invasively investigate the complex physiological processes involved in muscle activation, including both neural (central) and contractile (peripheral) properties. The uses of these techniques are widespread, and particularly valuable in determining where a decrement or enhancement in voluntary force (muscle strength) is spatially localized following an acute stimulus (i.e., fatigue, drug administration) [35], [59], chronic changes in physical activity (i.e., bedrest, exercise) [1], [20], or pathological conditions (i.e., amyotrophic lateral sclerosis) [38], [39]. For example, with respect to central activation, the interpolated twitch technique (involving superimposition of supramaximal electrical stimulation during a maximal voluntary contraction) allows for the quantification of muscle inactivation due to lack of central drive to the muscle [7]. Additionally, evaluation of the H-reflex response allows for the investigation of changes in transmission in spinal pathways, which can be used to gain insight into the excitability of the spinal reflex as well as nerve conduction velocity in the reflex loop [53], [64]. At the level of the sarcolemma, peripheral activation can be assessed via changes in the evoked compound muscle fiber action potential (CMAP), which reflects the temporal and spatial summation of the motor unit and muscle fiber action potentials which are influenced by various factors (i.e., muscle fiber conduction velocity, number of activated motor units, intracellular action potential characteristics) [37]. Moreover, through the evaluation of evoked forces, insight into the contractile properties of skeletal muscle can be obtained. For example, by comparing peak forces obtained at varying stimulation frequencies insight into the excitation–contraction (E–C) coupling processes can be garnered, as changes in the low-to-high frequency ratio is indicative of alterations in E–C coupling [21], [34]. Furthermore, by studying the rates of evoked force development and relaxation, information on intracellular Ca++ transients, muscle fiber type, and cross-bridge function can be obtained [11], [26], [29], [30], [58], [65], [66], [67], [68], [69], [70], [71]. The combination of the aforementioned, coupled with imaging techniques such as magnetic resonance imaging (MRI) which can evaluate muscle morphology [54], [55], allows for a broad appraisal of the human neuromuscular system, can serve to assess the site of adaptations, and thus can be scientifically and clinically valuable. Despite the widespread use of these techniques, there are few studies on the long-term stability of these measures. For example, the vast majority of the previous reliability studies have been conducted with testing sessions separated by ∼1-week or less [18], [44], [45], [51]. This relatively short duration between reliability assessment sessions seems problematic when considered in the context of studies investigating interventions are generally longer in duration. This is especially true for techniques such as surface EMG, where the reliability is highly dependent on the electrode location, which may be biased when evaluated serially over short time frames (days to ∼1-week) due to the shaved and abraded skin guiding the repeat electrode placement. Additionally, while there are presently a large number of studies investigating the reliability of the aforementioned variables, these papers typically only report on one or a few parameters [13], [28], [31], [40], [44], [45]. Thus, it is difficult and time-consuming for a clinician or researcher to gather and synthesize this data. Therefore, the purpose of this study was to comprehensively evaluate the reliability of a large number of commonly utilized non-invasive experimental tests of in vivo human neuromuscular function separated by a relatively long-duration of time (4-weeks). 2. Materials and methods 2.1. General overview of the experimental design The neuromuscular function of the left plantarflexor muscle group was assessed before and after 4-weeks where subjects were asked to maintain their normal physical activity patterns. The plantarflexors were chosen for several reasons, namely the ability to easily evoke action potentials (i.e., H-reflex), and the common usage of this muscle group in research studies [1], [6], [20], [36], [42], [61]. A variety of specific variables that are used to evaluate both neurological and contractile function were measured. These variables were tested sequentially in all subjects at each testing session, and are described in detail below. Relative reliability was assessed for each dependent variable using the calculation of intraclass correlation coefficients (ICC), and absolute reliability was assessed via coefficient of variation (CV), standard error of the mean (SEM) and limits of agreement (LOA). 2.2. Subjects Seventeen healthy subjects participated in the study (Table 1). Subjects were excluded if they had a history of cardiovascular or neuromuscular disease as assessed by a medical history questionnaire, or if we were unable to obtain maximal evoked properties with certainty (i.e., compound muscle fiber action potential or twitch force). Prior to testing we met with subjects for an orientation meeting where the procedures and tests were explained. They were asked to refrain from strenuous physical activity 24-h prior to the testing session, and avoid caffeine the day of the testing session. For each individual, testing sessions were conducted at the same time of day to control for variations due to circadian rhythms, and calf skin temperature was assessed (SKT100C temperature module, MP150, BioPac Systems Inc., Goleta, CA). If differences in temperature were greater than ∼1°C upon reporting for the second testing session, the limb was appropriately adjusted using warm or cool compresses. During the 4-weeks that separated the testing sessions subjects were asked to maintain their normal physical patterns, and during three days of this period wore a physical activity monitor (AMP-331 Dynastream Industries, Alberta, CA) to determine their level of physical activity (Table 1). The local ethics committee approved the experimental protocol, and all subjects provided written informed consent prior to testing. | | |  | Sex | N | Age (years) | Height (cm) | Weight (kg) | Physical activity (steps/day) | |
|---|
| Female | 12 | 21.0±0.73 | 161.0±1.6 | 65.1±3.1 | 10,547+1151 | | | Male | 5 | 20.6±1.9 | 177.8±3.1 | 63.0±4.0 | 8671±584 | | | Total | 17 | 20.9±0.72 | 165.9±2.4 | 64.4±2.4 | 9912±911 | | | | |
2.3. Electrical recordings Surface EMG signals were recorded from the soleus (SOL), medial gastrocnemius (MG), lateral gastrocnemius (LG) and tibialis anterior (TA) muscles using similar techniques previously described by our laboratory [14], [15]. Briefly, the skin was shaved, abraded and then cleaned with alcohol to minimize skin impedance. EMG signals were recorded with bipolar surface electrodes (Ag–AgCl, potential sensitive area of 22-mm, 40-mm center-to-center interelectrode distance; Kendall Medi-Trace 200, Chicopee, MA). The SOL electrodes were placed parallel to the muscle fibers on the lateral aspect clearly below the belly of the gastrocnemius muscles. The MG and LG electrodes were placed parallel to the muscle fibers, just distal from the knee and ∼2cm medial or lateral to the midline. The TA electrodes were placed parallel to and just lateral to the medial shaft of the tibia, at ∼1/4 the distance between the knee and ankle. These electrode locations were based on recommendations from Cram and Kasman [16]. After the placement of the electrodes during the first testing assessment, anatomical locations were noted by taking distance measurements from the popliteal fossa (for the SOL, MG and LG), and the tibial tuberosity (for the TA). These measurements were utilized for electrode placement during the second assessment. A reference electrode was placed on the anterior superior iliac spine. The EMG signals were amplified 500-times, band-pass filtered between 10 and 500Hz, and sampled at 2500Hz unless otherwise stated (MP150, BioPac Systems Inc., Goleta, CA). The EMG signal was saved for subsequent analysis. 2.4. Electrical stimulation Evoked responses were elicited by transcutaneous stimulation of the tibial nerve by use of a cathode located in the popliteal fossa (Ag–AgCl, 36-mm diameter; Kendall Medi-Trace 200), and an anode placed on the lower-third of the posterior thigh (Ag–AgCl, 48-mm diameter, Kendall Medi-Trace 530). The optimum site for stimulation was first located by a hand-held stimulation probe. The stimulus consisted of a 1-ms square pulse (Grass S88 stimulator coupled with a Grass SIU5 stimulus isolation unit, Astro-Med Inc., West Warwick, RI). 2.5. Mechanical recordings To measure plantarflexion force, subjects were seated in a custom-modified plantarflexion dynamometer (Parabody 826, LifeFitness, Schiller Park, IL), which allowed for strict control of hip, knee and ankle joint angles. The left leg was positioned in the dynamometer with the hip, knee and ankle joint angles secured at 90°. Isometric force was measured by a force transducer attached to a lever arm that the subjects plantarflexed against (MLP-300-T, Transducer Techniques, Temecula, CA), amplified and recorded at 625-Hz using a 16-bit data acquisition card (MP150, BioPac Systems Inc., Goleta, CA). The exerted force was displayed on a 43-cm computer monitor (AppleVision, Apple Computer Inc., Cupertino, CA, USA) located 1-m directly in front of the subject. 2.6. Magnetic resonance imaging Serial axial plane magnetic resonance imaging scans were acquired from the lower leg using a 1.5T Phillips Intera whole body scanner (Philips Medical Systems, Bothell, WA). These procedures were similar to those previously described [54], [55]. Briefly, 10-mm-thick transaxial images (2122-ms repetition time, 10.1-mm slice-to-slice interval) were obtained of the lower left leg. During acquisition the subjects lay supine with the feet secured to a custom-designed footplate which maintained the ankle joint angle at 90°. 3. Variables assessed Note. Due to the large number of variables assessed, it is beyond the scope of this paper to address their physiological meaningfulness. Rather, the reader is referred to a number of review and original research articles devoted to this topic [1], [2], [5], [7], [22], [23], [24], [27], [33], [47], [49], [53], [58], [59], [60], [65], [66]. 3.1. Muscle strength To determine muscle strength subjects performed a minimum of 4 maximal voluntary isometric contractions (MVC) with a 1–2min rest period between each contraction. Subjects gradually increased force production over the first second, and then exerted a maximum effort for ∼3–4s. If subjects continually recorded more force with increasing trials, or if the trials were not within 5% of each other, additional trials were performed until a plateau was reached. During testing strong verbal encouragement was provided by the investigators. The trial consisting of the highest value was considered the MVC force and used in subsequent analyses. 3.2. Voluntary electromyogram The interference EMG signals recorded during the MVC trial were analyzed surrounding peak force (1024-point epoch; ∼400-ms). The amplitude of the signals was quantified by calculating the root-mean-squared EMG (RMS EMG) of the SOL, MG, LG and TA muscles. To evaluate muscle coactivity, defined as the activity of antagonistic muscles (i.e., RMS EMG of the TA during plantarflexion) relative to the agonistic muscles activity (i.e., RMS EMG of the SOL, MG and LG during plantarflexion), a ratio was calculated (RMS EMG of the TA: Summed RMS EMG of the SOL, MG and LG). Additionally, the median frequencies (FMED) of the agonistic muscles were calculated as previously described by our laboratory [15]. 3.3. Muscle cross-sectional area The MR images were analyzed using the National Institutes of Health ImageJ software (http://rsb.info.nih.gov/ij/). Muscle anatomical cross-sectional area (CSA) was calculated for the SOL, MG and LG (Fig. 1). This calculation was based on an average CSA over five-slices obtained from the mid-belly of the muscle. Extreme care was taken to insure that the slices chosen for analysis were of the same anatomical location at both time points (accomplished by matching the slices for anatomical features [i.e., fascial characteristics]). 3.4. Evoked muscle forces Supramaximal electrical stimulation was utilized to evoke twitch, doublet (10-ms interpulse interval) and post-activation potentiated (PAP) doublet force (Fig. 2). The PAP doublet was administered ∼2-s after the cessation of a 5-s MVC (Fig. 2). Peak net force for the respective conditions were calculated. 3.5. Rates of evoked force development and relaxation The rates of evoked force development (+df/dt) and relaxation (−df/dt) were calculated for the resting doublet force response (Fig. 3). For the rate of force development, force–time curves displayed a biphasic waveform with an initially faster rate of development through the first-half in comparison with the second-half (Fig. 3). Thus, we calculated the rate of force development for the early phase (10–40% of peak force), late phase (60–90% of peak force) and throughout the entire contraction (10–90% of peak force). Additionally, the rate of force relaxation was calculated between 90% and 50% of peak force. 3.7. Compound muscle fiber action potential amplitude and duration While subjects were lying prone on an examination table (see H-reflex excitability below for more details), a single, supramaximal electrical stimulation pulse elicited the soleus compound muscle fiber action potential (CMAP, also commonly referred to as Mmax) (Fig. 4A). The peak-to-peak amplitude of the CMAP was calculated, as well as the duration of the first negative peak (zero-to-zero crossing). The sampling rate was increased to 200kHz for this measure to increase the resolution of calculating the CMAP duration. 3.8. Specific force In vivo specific force (force per unit area) was estimated by dividing the evoked doublet force by the MRI-derived anatomical muscle CSA (N/cm2). 3.9. H-reflex excitability The Hoffman reflex (H-reflex) is a spinal reflex response resultant from submaximal electrical stimulation of sensory nerve fibers which project back on and excite the α-motoneuron pool to create an action potential in the innervated skeletal muscles. As such, it is a commonly utilized tool to assess the excitability of spinal α-motoneurons, while also reflecting transmission efficiency (i.e., presynaptic inhibition) in Ia afferent synapses [1]. The excitability of the H-reflex response in the soleus muscle was assessed while the subjects’ lay prone on an examination table. During this assessment subject posture was strictly controlled as the head was maintained in medial alignment with the torso, and the arms rested comfortably by their side. A cylindrical cushion was placed under the talus to maintain an ankle joint angle of 45°. The H-reflex was assessed using techniques previously described [1]. In brief, electrical stimulation intensity was adjusted and continuously monitored to evoke a muscle action potential (M-wave) with a peak-to-peak amplitude equal to 20±2.5% of the CMAP. The corresponding peak-to-peak amplitude of the H-wave was calculated and expressed as a percentage relative to the CMAP (H20:CMAP) (Fig. 4B). Averages of eight trials were used in this calculation. 3.10. Peripheral nerve conduction and M-wave latency Indexes of nerve conduction were estimated from the latency responses in the H-reflex recordings (described above in H-reflex Excitability). These measures were calculated as: (1) the time difference between the M- and H-waves (reflex latency), and (2) the time interval between stimulus onset and the M-wave (M-wave latency) (Fig. 4B). The exact M- and H-wave onsets were identified by closely examining the signals (via magnification) and identifying where the EMG signal began to change from baseline. The reflex latency represents the time required for Ia conduction from the stimulus site (politeal fossa) to the spinal cord, synaptic delay at the motorneuron and then back down the motorneuron to the stimulus site. This measure is a commonly used as an index of nerve conduction velocity [61], [64]. The M-wave latency represents the time required for peripheral nerve conduction from the stimulus site through the branching motor axons and across the neuromuscular junction. Again, to increase the resolution of the time computations, EMG sampling rate was increased to 200kHz. 3.11. Rate of voluntary force development While sitting in the plantarflexion dynamometer subjects were asked to perform an explosive contraction by pushing up with their calf muscles as fast possible. The rate of voluntary force development (V+df/dt) was calculated between 5% and 40% of MVC force. 3.12. Isometric steadiness The ability to perform a steady submaximal contraction was assessed by subjects attempting to match a target force (displayed as a single target line on the computer screen) equal to 25% MVC for ∼10-s. Fluctuations in force were quantified by calculating the CV in force over the last-8-s of the contraction. 3.13. Time to task failure The time to task failure of a sustained, submaximal contraction (maintained at 20% of MVC by matching a target line) was measured as previously described [14]. 3.14. Statistical analysis Test–retest reliability was analyzed using coefficient of variation (CV), limits of agreement (LOA) and intraclass correlation coefficients (ICC) (two-way random effects model single measure reliability). The CV was calculated to represent intrasubject variation between the two testing sessions. This was performed by calculating the CV for each subject, and then reporting the mean CV for the respective dependent variables. For example, if a subject displayed an M-wave latency of 6.15ms on visit 1 and 6.49 on visit 2, we calculated their CV as: standard deviation of the two visits divided by the mean of the 2 visits times 100. So, for this example the CV=(0.240416/6.32)×100=3.80%. The ICC (2,1) is a two-way random effects model with single measure reliability in which variance over the repeated sessions is considered. SPSS (SPSS Inc., Version 10.0, Chicago, IL) was used to calculate the ICC. Due to the ceiling effect associated with the measure of central activation the ICC was not calculated for this variable. In addition to the ICC, a ‘relative reliability’ statistic which assesses the reproducibility of measurement relative to a sample of repeated measurements [4], we also chose to assess ‘absolute reliability’ (the degree in which the repeated measures vary). To fully understand the absolute stability of a measure it is important to understand the contribution of the main components of measurement error. In general, measurement error is broken into two classes: systematic bias and random error. Systematic bias represents the orderly changes in a measure over time, such as a learning effect; whereas random error is the result of biological or mechanical variation [4]. Since several of our measures required voluntary task performance (i.e., steadiness, rate of voluntary force development) the potential for systematic bias exists. Therefore, we utilized the limits of agreement (LOA) method (a measure of absolute reliability), which is a statistical technique that is useful in partitioning out systematic bias vs. random error [4], [10]. In doing this, Bland–Altman plots were generated for each variable and analyzed for the presence of heteroscedasticity (Fig. 5) [10]. Heteroscedasticity is when the residuals are not equally distributed throughout the range of scores of the dependent variables (Fig. 5A), whereby homoscedasticity is when the residuals are approximately equal for all dependent variable scores (Fig. 5B). This was determined by examining the correlation (R2) between the absolute differences and the mean values. R2 values between 0 and 0.1 were considered homoscedastic (no relation between error and the size of the measured variable) and systematic bias and random error were then calculated [4]. R2 values greater than 0.1 were heteroscedastic (amount of random error increases as the measured values increases) and the ratio LOA were then calculated [4]. The LOA ratio is calculated using the following equation: LOA ratio=[(SDdiffs/AVGmeans)×1.96]×100. Where SDdiffs is the standard deviation of all of the difference scores (visit 2−visit 1 calculated for each subject), AVGmeans is the average of all of the mean scores (mean of visits 1 and 2 for each subject), and the factor of 1.96 represents the inclusion of 95% of observations of the difference score. The LOA ratio is interpreted as “any two tests will differ due to measurement error by no more than X% either in the positive of negative direction” [4]. We should caution the interpretation of our LOA findings based on our relatively small sample size impacting the standard deviation of the sample and resulting in a wider LOA. 5. Discussion Overall, we observed moderate to high-levels of reliability in the vast majority of the in vivo measures of neuromuscular function that we assessed (24 out of the 29 assessed variables had ICC>0.70 and CV<15%). In general, the MRI-derived CSA, muscle strength and nerve conduction measures were the most highly reproducible, while the amplitude variables derived from the voluntary EMG signal during an MVC were the least reproducible. Our observation of high reliability in MVC force (ICC=0.97) is similar to that reported by Gabriel et al., who reported an ICC of 0.94 for elbow flexion strength when tested two weeks apart [25]. It has been suggested that when a longer duration of time separates testing sessions, the reliability of isometric strength is compromised [41]. Our findings do not support this previous observation as we observed excellent reliability in assessing PF MVC force even when testing sessions were separated by a relatively long-time period (Table 2, Table 3). Despite the observation of stability in PF strength, the surface EMG signal recorded from these contractions displayed a much higher degree of variability. The RMS EMG calculations were particularly variable (Table 2, Table 3), with the LG, MG, TA and Coactivity Ratio being less stable than the SOL. While some investigators have reported a considerably higher degree of reliability of this parameter [44], studies with testing sessions separated by several weeks tend to report lower reliability [25], [41]. The frequency domain data appeared to be more stable over time when compared to the time domain data, as attested to by all of our calculated measures of reliability (i.e., reduced CV) (Table 2, Table 3). This finding is consistent with previous reports, as our ICC’s ranged between 0.71 and 0.88 which is comparable to those reported by others (ICC range of 0.65–0.93) [25], [41], [44]. It is interesting to note that our observed median frequency values are somewhat higher than those previously reported [9], [56]. It is difficult to say why this is, but it could be related to differences in the biomechanical positioning during testing, as our setup probably resulted in the muscle fibers being in a shorter position (knee joint was flexed in our study vs. extended in the others), which has been shown to result in higher spectral frequencies when compared to a lengthened muscle [46]. There are a number of factors that influence the interference electromyogram signal during voluntary contractions which collectively account for the observed variation. It is possible that our relatively large electrode size and interelectrode distance resulted in cross-talk, and thus affected the reliability of the measures. For further discussion of influences on the voluntary EMG signal the reader is referred to an article by Farina et al. [24]. A higher degree of reliability was observed for the electrically evoked EMG signal characteristics versus that of the voluntary signal. Here, the CMAP displayed high relative reliability for both the amplitude and duration (ICC>0.8, CV<10%) (Table 2). Additionally, there was no evidence of systematic bias present (Table 3). Previous authors have reported similarly high reliability for CMAP amplitude over time [12], [13], [19], [43], but to our knowledge reliability of the CMAP duration has not been previously reported. EMG data recorded during the H-reflex responses (H20:CMAP, NCV and M-wave latency) also showed high reproducibility (Table 2, Table 3). Our measure of H-reflex excitability (H20:CMAP) displayed an ICC identical to that reported by Hopkins et al. (ICC=0.93), who evaluated the SOL response over 5-consecutive days [32]. Similar findings have also been observed for the H-reflex in the flexor carpi radialis [13], and vastus medialis [52] although much lower reliability has been reported by others [3]. Our estimate of peripheral nerve conduction velocity, evaluated through the latency times in the H-reflex responses, also demonstrated high-levels of reliability, with a very low CV of only 1.3%, which is consistent with the high stability of this measure reported by Troni et al. [64]. With regard to the evoked peak forces, we observed a heteroscedastic response for all variables (twitch, doublet and PAP force), with a large amount of variability in the between subjects responses. In terms of relative reliability, the twitch and doublet forces exhibited slightly higher reproducibility than the PAP force, which is in agreement with the recent report of Morton et al. [50]. It is probable that this is attributable to slight timing differences of the delivered PAP stimulus. Our peak twitch force displayed a lower ICC (0.80) than that reported for the elbow flexor muscles (ICC=0.91) [63]. However, our CV (∼11%) is similar to those reported for the dorsiflexor when the ankle joint angle is positioned at an intermediate range (∼6–12%) [40]. The reliability of rate of doublet force development was similar to that observed for rate of force relaxation (ICC’s ∼0.80, CV’s ∼10%; Table 2), and the reliability of the +df/dt force–time curves does not appear to vary depending on the portion of the curve selected (i.e., the initial portion vs. the later portion). The rate of voluntary force development had a slightly higher ICC (0.92) than that of the evoked force, but a similar mean CV (12%). Our reliability findings for the rate of voluntary force development is higher than that previously reported for the PF (ICC=0.63) [62], although it is comparable to that for elbow extension and flexion (ICC=0.87 and 0.81, respectively) [48]. The assessment of central activation failure using the interpolated twitch method revealed that none of the study participants were able to fully activate (defined as central activation=100%) their muscles on both testing sessions, however, four subjects did achieve full activation during the first visit and two during the second. While the relative reliability of this measure was rather good (CV=3.9%), evaluating the reliability of central activation using the LOA method indicates that a small amount of systematic bias (−1.53%) was present between the two sessions. This finding indicates that over the repeat sessions, there was a tendency for an orderly decrease in central activation. However, it should be noted that this slight decrease was not statistically significant between the two time points when a pairwise comparison was made (p=0.49). Morton et al. observed a similar degree of variation (CV=3.4%) in central activation of the knee extensors tested 7-days apart [50]. Additionally, this degree of variation is congruent with central activation assessments made via transcranial magnetic stimulation [63]. In assessing the variability of fatigue (time to task failure) and neuromotor performance task (steadiness) we observed wide ranging reliability. The isometric steadiness was highly reproducible, while the time to task failure of a sustained, submaximal contraction was more varied (Table 2, Table 3). Our ICC finding for the time to task failure was very similar to that observed for sustained knee extensor contractions at 20% MVC separated by 1-week (ICC=0.64 and 0.68, respectively) [57]. Additionally, we did observe some systematic bias (−12s) in this measure. This systematic bias and lower degree of reliability (when considered in comparison to the other voluntary tasks that we assessed such as steadiness and strength) is not surprising, as it has been suggested that practice of a sustained isometric contraction enables some, but not others, to prolong their endurance time due to adaptive changes in the neuromuscular activation pattern during the fatiguing contractions [33]. However, we must caution our findings using the LOA methodology based on our small sample size which can be problematic due to single or a few outliers skewing the data and creating high standard deviations. It should be noted that the subjects used in this study, were young, physically active individuals with a low body mass index, of which the majority (∼70%) were female. Therefore, it may be inappropriate to extrapolate these findings to other populations where the physical characteristics may directly influence many of these variables. For example, it is well known that subcutaneous adipose tissue acts as a low-pass filter on the recorded surface EMG signal [8], thus in populations where this may vary (i.e., obesity) the reliability of these respective parameters may be different. Additionally, caution should be made in inferring our findings for the plantarflexors to other muscle groups, as it is probable that they would vary between muscles [7]. Lastly, it should be noted that the reliability of the measured neurophysiologic or contractile parameters in this study does not indicate whether they are useful and/or appropriate in the diagnosis of neuromuscular disorders. In summary, we have presented the relative and absolute reliability of a large number of different non-invasive methods and their subsequent variables commonly used to asses the neuromuscular function of humans. While the reliability differed between the various measures, in general the reliability is rather high even when the testing sessions are separated by a long duration of time. Knowledge of this variation is critically important as it will allow clinicians to delineate between the expected random variation in these measures versus that of a ‘real change’ due to a disease state or intervention. Additionally, these findings are valuable to scientists as they will aid in calculating sample sizes for studies and comparing, and improving upon, the precision of measurements using different methods and/or equipment. Acknowledgment This study was funded in part by a National Aeronautics and Space Administration (NASA) Grant (NGT5-50446). References [1]. [1]Aagaard P, Simonsen EB, Andersen JL, Magnusson P, Dyhre-Poulsen P. Increased rate of force development and neural drive of human skeletal muscle following resistance training. J Appl Physiol. 2002;93(4):1318–1326. [2]. [2]Aagaard P, Simonsen EB, Andersen JL, Magnusson SP, Dyhre-Poulsen P. Neural adaptation to resistance training: changes in evoked V-wave and H-reflex responses. J Appl Physiol. 2002;92:2309–2318. [3]. [3]Ali A, Sabbahi MA. Test–retest reliability of the soleus H-reflex in three different positions. Electromyogr Clin Neurophysiol. 2001;41(4):209–214. [4]. [4]Atkinson G, Nevill AM. Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Med. 1998;26(4):217–238. MEDLINE |
CrossRef
[5]. [5]Basmaijian JV, DeLuca CL. Muscle alive – their functions revealed by electromyography. Baltimore: Williams and Wilkins; 1985;. [6]. [6]Behm DG, St-Pierre DM. Fatigue characteristics following ankle fractures. Med Sci Sports Exerc. 1997;29(9):1115–1123. MEDLINE |
CrossRef
[7]. [7]Behm DG, Whittle J, Button D, Power K. Intermuscle differences in activation. Muscle Nerve. 2002;25(2):236–243.
CrossRef
[8]. [8]Bilodeau M, Cincera M, Gervais S, et al. Changes in the electromyographic spectrum power distribution caused by a progressive increase in the force level. Eur J Appl Physiol Occup Physiol. 1995;71(2–3):113–123. MEDLINE |
CrossRef
[9]. [9]Bilodeau M, Goulet C, Nadeau S, Arsenault AB, Gravel D. Comparison of the EMG power spectrum of the human soleus and gastrocnemius muscles. Eur J Appl Physiol Occup Physiol. 1994;68(5):395–401. MEDLINE |
CrossRef
[10]. [10]Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1(8476):307–310.
CrossRef
[11]. [11]Booth J, McKenna MJ, Ruell PA, et al. Impaired calcium pump function does not slow relaxation in human skeletal muscle after prolonged exercise. J Appl Physiol. 1997;83(2):511–521. [12]. [12]Bril V, Ellison R, Ngo M, et al. Electrophysiological monitoring in clinical trials. Roche Neuropathy Study Group. Muscle Nerve. 1998;21(11):1368–1373.
CrossRef
[13]. [13]Christie AD, Inglis JG, Boucher JP, Gabriel DA. Reliability of the FCR H-reflex. J Clin Neurophysiol. 2005;22(3):204–209. MEDLINE [14]. [14]Clark BC, Collier SR, Manini TM, Ploutz-Snyder LL. Sex differences in muscle fatigability and activation patterns of the human quadriceps femoris. Eur J Appl Physiol. 2005;94(1–2):196–206. MEDLINE |
CrossRef
[15]. [15]Clark BC, Manini TM, The DJ, Doldo NA, Ploutz-Snyder LL. Gender differences in skeletal muscle fatigability are related to contraction type and EMG spectral compression. J Appl Physiol. 2003;94(6):2263–2272. [16]. [16]Cram JR, Kasman GS. Introduction to surface electromyography. Gaithersburg: Aspen; 1998;. [17]. [17]Damon B, McCully K. Co-Chairs. Workshop on investigation of human muscle function in vivo, Nashville, TN, USA; 2005. [18]. [18]Dankaerts W, O’Sullivan PB, Burnett AF, Straker LM, Danneels LA. Reliability of EMG measurements for trunk muscles during maximal and sub-maximal voluntary isometric contractions in healthy controls and CLBP patients. J Electromyogr Kinesiol. 2004;14(3):333–342. Abstract | Full Text |
Full-Text PDF (195 KB)
|
CrossRef
[19]. [19]de Carvalho M, Lopes A, Scotto M, Swash M. Reproducibility of neurophysiological and myometric measurement in the ulnar nerve-abductor digiti minimi system. Muscle Nerve. 2001;24(10):1391–1395.
CrossRef
[20]. [20]Duchateau J. Bed rest induces neural and contractile adaptations in triceps surae. Med Sci Sports Exerc. 1995;27(12):1581–1589. MEDLINE [21]. [21]Edwards RH, Hill DK, Jones DA, Merton PA. Fatigue of long duration in human skeletal muscle after exercise. J Physiol. 1977;272:769–778. MEDLINE [22]. [22]Enoka RM, Christou EA, Hunter SK, et al. Mechanisms that contribute to differences in motor performance between young and old adults. J Electromyogr Kinesiol. 2003;13:1–12. Abstract | Full Text |
Full-Text PDF (253 KB)
|
CrossRef
[23]. [23]Enoka RM, Fuglevand AJ. Motor unit physiology: some unresolved issues. Muscle Nerve. 2001;24(1):4–17.
CrossRef
[24]. [24]Farina D, Merletti R, Enoka RM. The extraction of neural strategies from the surface EMG. J Appl Physiol. 2004;96(4):1486–1495. [25]. [25]Gabriel DA, Basford JR, An KN. Neural adaptations to fatigue: implications for muscle strength and training. Med Sci Sports Exerc. 2001;33(8):1354–1360. MEDLINE |
CrossRef
[26]. [26]Gollnick PD, Korge P, Karpakka J, Saltin B. Elongation of skeletal muscle relaxation during exercise is linked to reduced calcium uptake by the sarcoplasmic reticulum in man. Acta Physiol Scand. 1991;142(1):135–136. MEDLINE |
CrossRef
[27]. [27]Hagg GM. Interpretation of EMG spectral alterations alteration indexes at sustained contraction. J Appl Physiol. 1992;73(4):1211–1217. [28]. [28]Hanchard NC, Williamson M, Caley RW, Cooper RG. Electrical stimulation of human tibialis anterior: (A) contractile properties are stable over a range of submaximal voltages; (B) high- and low-frequency fatigue are inducible and reliably assessable at submaximal voltages. Clin Rehabil. 1998;12(5):413–427. MEDLINE |
CrossRef
[29]. [29]Harridge SD, Bottinelli R, Canepari M, et al. Whole-muscle and single-fibre contractile properties and myosin heavy chain isoforms in humans. Pflugers Arch. 1996;432(5):913–920. MEDLINE |
CrossRef
[30]. [30]Hill CA, Thompson MW, Ruell PA, Thom JM, White MJ. Sarcoplasmic reticulum function and muscle contractile character following fatiguing exercise in humans. J Physiol. 2001;531(Pt 3):871–878. MEDLINE |
CrossRef
[31]. [31]Holmback AM, Askaner K, Holtas S, Downham D, Lexell J. Assessment of contractile and noncontractile components in human skeletal muscle by magnetic resonance imaging. Muscle Nerve. 2002;25(2):251–258.
CrossRef
[32]. [32]Hopkins JT, Ingersoll CD, Cordova ML, Edwards JE. Intrasession and intersession reliability of the soleus H-reflex in supine and standing positions. Electromyogr Clin Neurophysiol. 2000;40(2):89–94. [33]. [33]Hunter SK, Enoka RM. Changes in muscle activation can prolong the endurance time of a submaximal isometric contraction in humans. J Appl Physiol. 2003;94(1):108–118. [34]. [34]Jones DA. High- and low-frequency fatigue revisited. Acta Physiol Scand. 1996;156(3):265–270. MEDLINE [35]. [35]Kalmar JM, Cafarelli E. Effects of caffeine on neuromuscular function. J Appl Physiol. 1999;87(2):801–808. [36]. [36]Kawakami Y, Akima H, Kubo K, et al. Changes in muscle size, architecture, and neural activation after 20 days of bed rest with and without resistance exercise. Eur J Appl Physiol. 2001;84:7–12. MEDLINE |
CrossRef
[37]. [37]Keenan KG, Farina D, Merletti R, Enoka RM. Influence of motor unit properties on the size of the simulated evoked surface EMG potential. Exp Brain Res [in press], doi:10.1007/S00221-005-0126-7. [38]. [38]Kent-Braun JA, Ng AV. Specific strength and voluntary muscle activation in young and elderly women and men. J Appl Physiol. 1999;87(1):22–29. [39]. [39]Kent-Braun JA, Walker CH, Weiner MW, Miller RG. Functional significance of upper and lower motor neuron impairment in amyotrophic lateral sclerosis. Muscle Nerve. 1998;21:762–768.
CrossRef
[40]. [40]Koh TJ, Herzog W. Evaluation of voluntary and elicited dorsiflexor torque–angle relationships. J Appl Physiol. 1995;79(6):2007–2013. [41]. [41]Kollmitzer J, Ebenbichler GR, Kopf A. Reliability of surface electromyographic measurements. Clin Neurophysiol. 1999;110(4):725–734. Abstract | Full Text |
Full-Text PDF (253 KB)
|
CrossRef
[42]. [42]Koryak Y. Contractile properties of the human triceps surae muscle during simulated weightlessness. Eur J Appl Physiol. 1995;70:344–350. [43]. [43]Kwon O, Lee KW. Reproducibility of statistical motor unit number estimates in amyotrophic lateral sclerosis: comparisons between size- and number-weighted modifications. Muscle Nerve. 2004;29(2):211–217.
CrossRef
[44]. [44]Larsson B, Karlsson S, Eriksson M, Gerdle B. Test–retest reliability of EMG and peak torque during repetitive maximum concentric knee extensions. J Electromyogr Kinesiol. 2003;13(3):281–287. Abstract | Full Text |
Full-Text PDF (120 KB)
|
CrossRef
[45]. [45]Mathur S, Eng JJ, MacIntyre DL. Reliability of surface EMG during sustained contractions of the quadriceps. J Electromyogr Kinesiol. 2005;15(1):102–110. Abstract | Full Text |
Full-Text PDF (333 KB)
|
CrossRef
[46]. [46]Merletti R, Lo Conte LR, Cisari C, Massazza U. Effect of ankle joint position on electrically evoked surface myoelectric signals of the tibialis anterior muscle. Arch Phys Med Rehabil. 1993;74(5):501–506. MEDLINE |
CrossRef
[47]. [47]Milner-Brown HS, Miller RG. Muscle membrane excitation and impulse propagation velocity are reduced during muscle fatigue. Muscle Nerve. 1986;9(4):367–374.
CrossRef
[48]. [48]Mirkov DM, Nedeljkovic A, Milanovic S, Jaric S. Muscle strength testing: evaluation of tests of explosive force production. Eur J Appl Physiol. 2004;91(2–3):147–154. MEDLINE |
CrossRef
[49]. [49]Moritani T, Yoshitake Y. 1998 ISEK Congress Keynote Lecture: The use of electromyography in applied physiology. International Society of Electrophysiology and Kinesiology. J Electromyogr Kinesiol. 1998;8(6):363–381. Abstract | Full Text |
Full-Text PDF (691 KB)
|
CrossRef
[50]. [50]Morton JP, Atkinson G, Maclaren DP, et al. Reliability of maximal muscle force and voluntary activation as markers of exercise-induced muscle damage. Eur J Appl Physiol. 2005;. [51]. [51]Palmieri RM, Hoffman MA, Ingersoll CD. Intersession reliability for H-reflex measurements arising from the soleus, peroneal, and tibialis anterior musculature. Int J Neurosci. 2002;112(7):841–850. MEDLINE |
CrossRef
[52]. [52]Palmieri RM, Ingersoll CD. Intersession reliability of a protocol to assess reflex activation history in the vastus medialis. Int J Neurosci. 2005;115(5):735–740. MEDLINE |
CrossRef
[53]. [53]Pierrot-Deseilligny E, Mazevet D. The monosynaptic reflex: a tool to investigate motor control in humans. Interest and limits. Neurophysiol Clin. 2000;30(2):67–80. MEDLINE |
CrossRef
[54]. [54]Ploutz LL, Tesch PA, Biro RL, Dudley GA. Effect of resistance training on muscle use during exercise. J Appl Physiol. 1994;76:1675–1681. [55]. [55]Ploutz-Snyder LL, Tesch PA, Crittenden DJ, Dudley GA. Effect of unweighting on skeletal muscle use during exercise. J Appl Physiol. 1995;79(1):168–175. [56]. [56]Ricard MD, Ugrinowitsch C, Parcell AC, et al. Effects of rate of force development on EMG amplitude and frequency. Int J Sports Med. 2005;26(1):66–70. MEDLINE |
CrossRef
[57]. [57]Rochette L, Hunter SK, Place N, Lepers R. Activation varies among the knee extensor muscles during a submaximal fatiguing contraction in the seated and supine postures. J Appl Physiol. 2003;95(4):1515–1522. [58]. [58]Rome LC, Lindstedt SL. The quest for speed: muscles built for high-frequency contractions. News Physiol Sci. 1998;13:261–268. [59]. [59]Russ DW, Kent-Braun JA. Sex differences in human skeletal muscle fatigue are eliminated under ischemic conditions. J Appl Physiol. 2003;94:2414–2422. [60]. [60]Sale D. Postactivation potentiation: role in performance. Br J Sports Med. 2004;38(4):386–387.
CrossRef
[61]. [61]Scaglioni G, Ferri A, Minetti AE, et al. Plantar flexor activation capacity and H reflex in older adults: adaptations to strength training. J Appl Physiol. 2002;92(6):2292–2302. [62]. [62]Sleivert GG, Wenger HA. Reliability of measuring isometric and isokinetic peak torque, rate of torque development integrated electromyography and tibial nerve conduction velocity. Arch Phys Med Rehabil. 1994;75(12):1315–1321. MEDLINE [63]. [63]Todd G, Taylor JL, Gandevia SC. Reproducible measurement of voluntary activation of human elbow flexors with motor cortical stimulation. J Appl Physiol. 2004;97(1):236–242. [64]. [64]Troni W, Cantello R, Rainero E. The use of the H reflex in serial evaluation of nerve conduction velocity. Electroencephalogr Clin Neurophysiol. 1983;55(1):82–90. MEDLINE |
CrossRef
[65]. [65]Westerblad H, Allen DG. The contribution of [Ca2+]i to the slowing of relaxation in fatigued single fibres from mouse skeletal muscle. J Physiol. 1993;468:729–740. MEDLINE [66]. [66]Westerblad H, Allen DG. The influence of intracellular pH on contraction, relaxation and [Ca2+]i in intact single fibres from mouse muscle. J Physiol. 1993;466:611–628. MEDLINE [67]. [67]Westerblad H, Allen DG. Relaxation, [Ca2+]i and [Mg2+]i during prolonged tetanic stimulation of intact, single fibres from mouse skeletal muscle. J Physiol. 1994;480(Pt 1):31–43. [68]. [68]Westerblad H, Allen DG. The role of sarcoplasmic reticulum in relaxation of mouse muscle; effects of 2,5-di(tert-butyl)-1,4-benzohydroquinone. J Physiol. 1994;474(2):291–301. MEDLINE [69]. [69]Westerblad H, Allen DG. Mechanisms underlying changes of tetanic [Ca2+]i and force in skeletal muscle. Acta Physiol Scand. 1996;156(3):407–416. MEDLINE [70]. [70]Westerblad H, Allen DG. Slowing of relaxation and [Ca2+]i during prolonged tetanic stimulation of single fibres from Xenopus skeletal muscle. J Physiol. 1996;492(Pt 3):723–736. [71]. [71]Westerblad H, Lannergren J, Allen DG. Slowed relaxation in fatigued skeletal muscle fibers of Xenopus and Mouse. Contribution of [Ca2+]i and cross-bridges. J Gen Physiol. 1997;109(3):385–399. MEDLINE |
CrossRef
 Brian Clark received his BS degree in Biology from Western Carolina University in 1998, an M.S. in Exercise Physiology (2001), and a Ph.D in the field of Neuromuscular Physiology (2006) from Syracuse University. He is currently an Assistant Professor of Physiology in the Department of Biomedical Sciences of Ohio University. His research interests include identifying adaptations in neurological and contractile properties following prolonged periods of alterations in physical activity level (i.e., disuse), the mechanisms of human muscle fatigue, and clinical neuromuscular pathophysiology.  Summer Cook received a B.S. degree in Exercise Physiology at East Stroudsburg University in 1999, an M.S. in Exercise Science and is currently working on her Ph.D. in Exercise Physiology at Syracuse University. She is a member of the American College of Sports Medicine (ACSM) and is a candidate for the Student Representative on the Board of Trustees of the ACSM. Her research interests include aging and in vivo investigation of neural and skeletal muscle function.  Lori Ploutz-Snyder received her B.S. from the Honors Tutorial College at Ohio University, M.S. from the Department of Zoological and Biomedical Sciences, and Ph.D. from the Department of Biological Sciences at Ohio University. Subsequently, she completed postdoctoral work studying muscle physiology and muscle functional magnetic resonance imagining in the Departments of Physiology and Radiology at Michigan State University. She is currently Chair and Associate Professor in the Department of Exercise Science at Syracuse University where she also directs the Musculoskeletal Research Laboratory. Her research interests focus on skeletal muscle, MRI, and physical function in the aging. Musculoskeletal Research Laboratory, Department of Exercise Science, Syracuse University, 820 Comstock Avenue, Room 201, Syracuse, NY 13244, USA  Corresponding author. Tel.: +1 315 443 1411; fax: +1 315 443 9375.
PII: S1050-6411(05)00151-3 doi:10.1016/j.jelekin.2005.11.008 © 2006 Published by Elsevier Inc. | |
|
FULL TEXT00151-3/journalimage?img=PIIS1050641105001513.gr1.lrg.gif&fig=fig1&kwhquery=&issn=1050-6411&ishighres=false&allhighres=false&free=yes','journalimage');)
00151-3/journalimage?img=PIIS1050641105001513.gr2.lrg.gif&fig=fig2&kwhquery=&issn=1050-6411&ishighres=false&allhighres=false&free=yes','journalimage');)
00151-3/journalimage?img=PIIS1050641105001513.gr3.lrg.gif&fig=fig3&kwhquery=&issn=1050-6411&ishighres=false&allhighres=false&free=yes','journalimage');)
00151-3/journalimage?img=PIIS1050641105001513.gr4.lrg.gif&fig=fig4&kwhquery=&issn=1050-6411&ishighres=false&allhighres=false&free=yes','journalimage');)
00151-3/journalimage?img=PIIS1050641105001513.gr5.lrg.gif&fig=fig5&kwhquery=&issn=1050-6411&ishighres=false&allhighres=false&free=yes','journalimage');)