Neuromuscular Fatigue to Power Loading Using a Weight-Stack Device Fitted With or Without Additional Rubber Band Resistance

Abstract

Peltonen, H, Walker, S, Häkkinen, K, and Avela, J. Neuromuscular fatigue to power loading using a weight-stack device fitted with or without additional rubber band resistance. J Strength Cond Res 28(7): 1802–1811, 2014—The addition of a rubber band (RB) to traditional weight-based resistance may be beneficial regarding the deceleration phase common during explosive power actions with light weights. This would lead to greater force and muscle activity during the latter part of the movement, but this may influence the distribution of neuromuscular fatigue. Fifteen men performed 2 knee extension power loadings (5 × 5 × 40% of 1 repetition maximum [1RM]) using a typical weight-stack device without (WS) and with the addition of a RB (WS + RB) in a crossover design. Concentric contractions were performed as fast as possible, and analyzed for torque, velocity, power, and vastus lateralis and medialis surface electromyography (EMG) parameters over 20° segments (range of motion, 60–180°) from throughout the loading. Additionally, maximum isometric knee extension was analyzed preloading and postloading for torque and EMG parameters, and for voluntary activation level using the twitch interpolation technique. Peak concentric power occurred at 100–120° on both devices but was greater during WS + RB (p ≤ 0.05). Electromyographic amplitude was greater in WS at 80–120°, p ≤ 0.05–0.01. Loading-induced reductions in peak concentric power (−12 ± 8 vs. −4 ± 10 kW, p ≤ 0.05) and voluntary activation level (−5.5 ± 7 vs. −1.1 ± 5%, p ≤ 0.05) were greater when using WS + RB. However, a greater reduction in root mean square of electromyography over 0–100 milliseconds was observed after WS (−31 ± 22 vs. 10 ± 26%, p < 0.05) accompanied by a large reduction in torque over 0–100 milliseconds. It seems that activation strategies, and consequently neuromuscular fatigue, during power loading could be modified through the addition of RB resistance.

Introduction

Explosive strength training (i.e., power training) with weight-based resistance can provide the same type of inertia as in common sport performances. It seems logical that the transfer effect of inertia-based strength gains to sport performance may be efficient; however, this should be supported by empirical studies (¹²). During explosive power actions with light weights, the resistance is the greatest at the beginning of the movement (the acceleration phase) and the need for maximal muscular activity decreases when the movement proceeds (³⁴) with concomitant reduced force and power (³⁵).

Previous studies (^27,35) have shown that acute effects of power resistance exercise primarily reduce rapid force production and muscle activity rather than maximal force output. One possibility for the “targeted” effect on the initial contraction capabilities is that, when performing explosive actions with weight-based resistance, the movement is performed for as little as 50% of the whole range of motion under a muscle activation strategy which aims to accelerate the action (³²). Consequently, it could be speculated that neuromuscular fatigue, and also long-term adaptation to training, would be specific to the onset of force production rather than maximum force production (^1,8,16).

For populations such as athletes or elderly people, it may be beneficial to improve both rapid and maximum force production, 1 possibility to extend the resistance, and muscle activity over the entire range of motion is to combine the load with a rubber band (RB) (^10,22,41). The effects of RBs govern the inertial properties of weight-based resistance (¹²) creating a steadier and greater total resistance throughout the concentric phase of the movement, but also adding workload during the eccentric phase (¹⁰). Therefore, the additional RB may change the fatigue distribution in the neuromuscular system and, thus, maximum force production may also be affected by acute loading. This modification in resistance mechanics may eventually lead to training-induced improvements in both explosive and maximal force capabilities during power training. Additionally, as improved maximal strength is associated with enhanced rapid force production (^1,3,4), explosive actions with RBs may be advantageous in some situations compared with traditional resistance exercise techniques.

In previous studies, RBs have been used typically in multijoint extension exercises, such as squats or bench press (^{20,22,36,37,41}). These exercises use the muscle groups' increasing capability of force production throughout the movement (²⁶) and are commonly used in practice. Single-joint exercises, however, offer a more controlled experimental set-up in which to investigate neuromuscular fatigue. One such study (¹³) showed that the number of repetitions was significantly reduced when adding RBs to the weight-stack resistance. Unfortunately, this and other single-joint studies (^2,29) did not seek to identify the cause of fatigue between these 2 resistance modes. Thus, the purpose of this study was to examine the acute neuromuscular responses during and after a typical power loading protocol using bilateral weight-stack resistance knee extensor device with and without additional RB resistance. This study used a combination of surface electromyography (EMG) and electrical stimulation to determine central and peripheral fatigue, as well as kinetic and kinematic data. It was hypothesized that power loading with additional RBs would induce similar reductions in rapid force production and muscle activity, but additionally, would lead to impaired maximal force production and muscle activation (i.e., central fatigue).

Methods

Experimental Approach to the Problem

The experimental design included a familiarization session and 2 loading protocols (Figure 1) using a weight-stack knee extension device with (WS + RB) or without RB (WS) resistance (Figure 2). Usually, power training is performed by multijoint movements; however, this unique single-joint study design made use of acute neuromuscular responses induced by voluntary contractions and electrical stimulation to determine loading-induced fatigue during repetitions and fatigue distribution between central and peripheral components during a typical power loading protocol.

Figure 1: Measurement and loading protocol. MVC = maximal isometric voluntary contraction.

Figure 2: Knee extension device and the attachment of the rubber band.

Subjects

Fifteen healthy young men (28 ± 5 years, 180.8 ± 3.6 cm, 78.7 ± 9.8 kg) volunteered as subjects. The subjects had no regular strength training background, but they were all physically active habitually performing low-intensity endurance-based physical activity (e.g., jogging and cross-country skiing) for a few hours per week. Full details about possible risks or discomfort were given to the subjects, and they signed the informed consent. The study was performed in accordance with the Declaration of Helsinki 1975, and protocol was accepted by the Ethics Committee of the University of Jyväskylä.

Familiarization

In the familiarization session, surface EMG placements were measured and marked by indelible ink tattoos and submaximal muscle, and nerve stimulation was performed to familiarize the subjects with these testing procedures. The subjects practiced maximum isometric knee extension trials, and they performed a few repetitions on the bilateral weight stack device (David 200; David Health Solution Ltd., Helsinki, Finland) in a seated position, and the device was set according to the anatomical dimensions of the subject. The knee and hip angles on the isometric condition were 107 and 110°, respectively. After that 1 repetition maximum (1RM), load without RB resistance was determined. The resistance provided by WS was also variable because the weight-stack system used a cam wheel in the lever arm mechanism. The range of the motion for the knee joint was from 60 to 180°, and hip joint angle was set to 110° throughout the movement. After the familiarization session, subjects recovered at least 4 days before the first loading. The order of loading was randomized, and the recovery time between the different loadings was at least 1 week. The subjects had their own constant test start time during the day, and they avoided strenuous physical activities for 48 hours before each loading.

Subjects were instructed to consume 0.5 L water 1 hour before they arrived to loading, as hydration status has been shown to affect strength levels (²³). In both loading sessions, subjects performed a warm-up that consisted of 6 self-paced bilateral knee extension repetitions at 40% of 1RM on the loading device. The loading protocol itself consisted of 5 × 5 × 40% of 1RM load with a 3-minute recovery between sets. Each concentric repetition was performed as fast as possible, and the ankle pad was stopped after it was “kicked” before the subject was performing controlled eccentric part of repetition. The load (40% 1RM) was calculated from the 1RM of the familiarization session. In the combined resistance loading, an RB [circumference: 2,080 mm, width: 30 mm, thickness: 5 mm, and color: blue; details McMaster et al. (³⁰)] was attached to the weight stack (Figure 2), with the same load that was used without the RB (40% 1RM). The RB was tightened to begin resisting the movement after 70° knee joint angle. The RB, which was used in this study, was chosen from commercial available training bands based on its stiffness qualities. For power training, the chosen RB is probably one of the most suitable to govern the inertial properties of weight stack. For that reason, it was not our intention to match the mechanical loading between the 2 conditions after the 70° knee joint angle when the RB resisted the movement.

Measurements

The premeasurements (pre) consisted of withdrawal of blood samples for blood lactate analyses, unilateral maximal isometric voluntary contraction (MVC), resting twitch (RT) torque, and superimposed twitch (SIT) torque during MVC at 107° knee joint angle. In addition, maximal compound mass action potentials (M wave) were measured at a standing rest position. Maximal isometric voluntary contraction, M wave, SIT and RT torques, and blood lactate were measured again immediately after the loadings (post) (Figure 1). The loading device (David 200) was modified with a locking system and force sensors to allow assessment of these tests in isometric condition. Torque, knee joint angle, and EMG activities were measured throughout the loading repetitions. The analyses for the loadings were done from the second repetition of the first set and the fifth repetition from the fifth set. To analyze the concentric actions, the range of motion (60–180°) was divided into six 20° sectors.

Electrical Stimulation

Muscle Stimulations

Two pairs of carbon film muscle stimulation electrodes (V-Trodes; Mettler Electronics Corp., Anaheim, CA, USA; diameter 70 mm) were placed on to the mid and distal portion of the quadriceps muscle group (right leg). Stimulation electrode pairs were galvanically separated, and the skin under the electrodes was shaved and cleaned.

Unilateral isometric knee extension torque responses because of the stimulation during resting condition were determined at the 107° knee joint angle on the loading device. The current was increased progressively in 20 mA steps between stimulations when the torque response was higher than that of the previous stimulation. When the maximal torque response was reached, 50% of the stimulation current was added. This supramaximal stimulus (150%) was used for all subsequent stimulations. The electrical stimulator (Digitimer Stimulator Model DS7AH; Digitimer Ltd., Hertfordshire, United Kingdom) produced single rectangular pulses (1 milliseconds, 400 V).

Resting stimulations were performed 2 times with 1 minute between the twitches. Resting twitch torque, rate of twitch torque development (RTTD), and half-relaxation time (1/2RT) were analyzed from each twitch. The SIT stimulation protocol included RTs before and after MVC, and 1 twitch was delivered during the plateau phase of MVC. The subjects were instructed to increase their torque progressively toward the maximum, and they were able to reach MVC within 5 seconds. Voluntary activation level was calculated according to the formula by Harridge et al. (¹⁸):

where, Pts is the difference between the voluntary torque and twitch torque from the SIT, and Pt is the RT torque after MVC.

Nerve Stimulations

The same constant current stimulator (Digitimer) was used to stimulate the femoral nerve beneath the inguinal ligament. The cathode (1 cm diameter) was attached into the femoral triangle at the placement that gave the strongest response to a weak stimulation current. This location was marked on the skin for replacement. The anode was attached on the greater trochanter. Monophasic rectangular pulse wave (1 milliseconds single-pulse, 400 V) was delivered to evoke maximal compound mass action potentials (M wave) in vastus lateralis (VL) muscle. M-wave amplitude was followed after every 10 mA current stages until there were plateaus in the M-wave amplitude (again 50% of the stimulation current was added for the measurements). Latency time, peak-to-peak amplitude, and duration of M wave were measured and analyzed in a standing rest position before and immediately after the loadings.

Electromyography

Electromyography of the VL and vastus medialis (VM) muscles during knee extension was recorded (NeuroLog 824; Digitimer Ltd.) from the right leg using bipolar surface electrodes (10 mm pickup area and 20 mm inter-electrode distance). The electrodes were placed according to SENIAM guidelines (¹⁹). The placements of the electrodes were tattooed with small ink dots before skin preparation. The tattoos ensured that electrodes were on the same location of the muscle during both loadings. Eletromyographic signals were sampled at 2,000 Hz, and they were pre-amplified (500 gain) at a bandwidth of 10–500 Hz for later analysis. Signals passed through an analog-to-digital (A/D) board converter (Power 1401) to a computer using Signal 2.16 software (Cambridge Electronic Design, Cambridge, United Kingdom). Electromyography signal data were bandpass filtered (20–350 Hz) before analysis and were transformed to root mean square (rms) form before being normalized to isometric MVC EMGrms values.

Electromyography activities of the VL and VM muscles were averaged and combined. Rectus femoris muscle activity was not measured because of the muscle stimulation electrodes. Maximal EMG activity was analyzed from the plateau phase of isometric MVC over a 1,000 milliseconds time window, and EMG activity during rapid torque production was analyzed from initial 100 milliseconds of isometric MVC (EMG100 milliseconds). The averaged median frequency of the VL and the VM muscles were determined from the isometric MVC before and after loading.

Torque and Angle

The weight-stack resistance device was equipped with in-built force and angle sensors allowing evaluation of concentric actions. Calibrations of all sensors were accomplished before the beginning of the study. Force and angle signals were sampled at 2,000 Hz, and signals were low-pass filtered (force 20 Hz and angle 75 Hz). From these parameters, mean angular velocity and mean power values during single concentric repetitions were analyzed for each of six 20° knee joint angles. Mean power was calculated as the mean angular velocity multiplied by the mean torque for each sector separately. During isometric knee extension trials, the plateau of isometric MVC was used to determine mean maximal torque with a 1,000 milliseconds time window. The rate of torque development during the greatest 10ms (RTD) and the initial 100ms (T100ms) were analyzed from the fastest MVC trial before and after loading.

Blood Collection

Blood samples were collected into 20 μL capillary tubes and mixed with 1 ml hemolyzing solution. Automatic blood lactate analysis (Biosen; EKF Diagnostic, Madgeburg, Germany) was performed after testing according to the manufacturer's instructions.

Statistical Analyses

The results are presented as mean and SDs in the text. All dependent variables were assessed by using a 2-way analysis of variance with repeated measures (2 loadings × 6 angle sectors or 2 loadings × 2 time points). When a significant F value was observed, Bonferroni post hoc procedures were performed to locate the pair-wise differences. The relative changes (before vs. after loading) were analyzed with a paired 2-tailed Student's t-test between resistance modes. Statistical significance was accepted at p ≤ 0.05.

Reliability values for the measurements used were at acceptable levels; isometric torque is 0.981 and 3.4%, EMGrms is 0.918 and 7.2%, EMG median frequency is 0.957 and 6.8%, maximum twitch torque is 0.994 and 1.3%, maximum RTTD is 0.997 and 3.2%, and calculated voluntary activation is 0.732 and 1.9%.

Results

Dynamic Measurements

Unfatigued Condition

During the second repetition of the first set, average resistance was 57 ± 26% (p < 0.001) higher and average power was 59 ± 26% (p < 0.001) higher, but average velocity was 7 ± 5% (p < 0.001) lower in WS + RB resistance than WS resistance (Table 1).

Table 1: The average values of the entire dynamic contraction from the second repetition of the first set.^*

During the first 20° (from 60 to 80°) knee joint angles, the torque-angle relationship between resistance modes was similar, but it increased significantly at all knee joint angles from 80 to 180° (Figure 3A) on the WS + RB compared with WS resistance. According to these differences, power was significantly greater at all knee joint angles from 80 to 180° (Figure 3C) on the WS + RB resistance, even though the angular velocities were significantly higher on the WS resistance after 100° knee extension angle (Figure 3B).

Figure 3: A) Torque-angle curves, (B) velocity-angle curves, (C) power-angle curves, and (D) muscle activity-angle curves of dynamic knee extensions during unfatigued condition from the second repetitions of the first set and after the loading from the fifth repetitions of the fifth sets. *p ≤ 0.05, **p < 0.01, ***p < 0.001, due to loading.

p < 0.01–0.001, between resistance modes during the second repetitions of the first set in the unfatigue condition.

In the beginning of the loading, EMGrms activity reached peak values (206 ± 74% and 165 ± 72%; from preloading isometric EMG) significantly earlier and at smaller knee angles on the WS (343 ± 81 milliseconds, 80–100°) than on the WS + RB resistance (474 ± 44 milliseconds, 120–140°). Significant decrements were observed in the EMGrms activity of the quadriceps muscles during 80–120° knee angles on the WS + RB compared with that of WS resistance (Figure 3D).

Fatigued Condition

During the fifth repetition of the fifth set, decreases in torque production occurred throughout 80–120° of varying magnitudes from −4 ± 2% at 120° to −3 ± 3% at 80° (p < 0.001 and p ≤ 0.05) knee joint angles on the WS + RB resistance, whereas torque did not decrease significantly on the WS resistance (Figure 3A). However, the angular velocity decreased significantly during loading between 120and 160° on the WS and 80–160° knee joint angles on the WS + RB resistance (Figure 3B).

Power reduced −11 ± 10% (p ≤ 0.05) between 120 and 140° on the WS resistance and decreases in power production were in range from −8 ± 3% to −17 ± 6% (p < 0.01–0.001) between 80 and 160° knee joint angles during loadings on the WS + RB resistance (Figure 3C). At the knee angles that produced peak concentric power (i.e., 100–120°), loading-induced reductions were greater (p ≤ 0.05) after WS + RB (−11.6 ± 8 kW) compared with WS (−4 ± 10 kW). Root mean square of electromyographic activity of the dynamic contractions did not change significantly during loadings on both resistance modes.

Isometric Measurements

Voluntary Contraction

Maximal isometric force decreased after loading by −6 ± 4% (p < 0.01) on the WS and by −7 ± 9% (p ≤ 0.05) on the WS + RB; however, maximal EMG activities during MVC did not change significantly after either resistance mode. The reductions in T100 ms were −27 ± 26% (p ≤ 0.05) and −16 ± 10% (p < 0.01) on the WS and on the WS + RB resistance modes, respectively, but reduction of EMGrms was significant only on WS (−31 ± 22% p < 0.01). The RTD decreased similarly and significantly after both resistance modes (−26 ± 17%, WS and −28 ± 16%, WS + RB p < 0.001) (Figure 4). Moreover, loadings did not lead to significant changes in median frequencies at postloading EMG activities.

Figure 4: Relative torque and root mean square of electromyography (EMGrms) characteristics of explosive produced maximal isometric voluntary contraction (MVC) before and after loading. *p ≤ 0.05, **p < 0.01, ***p < 0.001.

Muscle Stimulation

Voluntary activation level declined significantly (−5 ± 7%, p ≤ 0.05) only after the WS + RB resistance loading. Furthermore, changes in voluntary activation level were significantly different (p ≤ 0.05) between the different resistance modes after loading (Table 2). Resting twitch and RTTD were significantly enhanced and 1/2RT decreased after loading on both resistance modes. In addition, RT and RTTD were significantly higher after loading on the WS + RB resistance mode than after the WS resistance mode.

Table 2: Characteristics of muscle stimulation before and after loading and their relative changes.^*

Nerve Stimulation

No statistical differences in latency times, peak-to-peak amplitudes, or durations in maximal M waves between pre and postloading measurements or resistance modes were observed (Table 3).

Table 3: Absolute values of M wave and median frequency before and after loading.^*

Blood Lactate

Blood lactate concentrations increased from 1.8 ± 0.6 to 3.4 ± 1.3 mmol·L⁻¹ and from 1.5 ± 0.3 to 4.3 ± 0.7 mmol·L⁻¹ (p < 0.001), on the WS and WS + RB resistances, respectively, with a statistical difference (p < 0.01) between the resistance modes.

Discussion

During the present knee extension power loadings, the main findings were: (a) concentric peak power of both resistance modes occurred during 100–120° knee joint angle, (b) the magnitude of loading-induced reduction in concentric peak power was greater after the WS + RB resistance than WS resistance, (c) during isometric contraction after loading, the reduction in voluntary activation level was significantly greater after the WS + RB resistance than WS resistance; however, (d) greater reductions in EMG and force during the first 100 milliseconds after WS were observed.

The resistance of the WS + RB mode was higher than that of the WS mode after the extension of the knee joint beyond 80°. This was expected because of the additional resistance of the RB. In parallel, concentric power was greater after 80° knee joint angle and angular velocity was lower from 100° knee joint angle in the WS + RB resistance mode compared with WS mode.

Muscular power is used as one of the main parameters in the measurement of explosive athletic performance, and several studies have used power or its components (torque and angular velocity) to indicate fatigue because of different kinds of power loadings (e.g., ^9,25). Typical power loading (e.g., 5 × 5 × 40% 1RM) using WS caused decreases in angular velocity between 120 and 160° knee joint angle, which led to decreased power at 120–140° joint angles in the end of loading. As may have been expected, due to greater resistance and total work during WS + RB, larger responses were observed in concentric torque, angular velocity, and power. During WS + RB, decreased angular velocity and power were observed between 80 and 160° knee joint angles in the end of loading. Additionally, concentric torque decreased between 80 and 120° joint angles, which was not observed during loading using WS. The close pattern between decreases in angular velocity and power during both loadings suggests that decreased power production was largely dependent on the muscles' ability to contract at high velocities because of the moderate loads that were used (²⁴).

Previous studies observed that the reductions in isometric force production during the first 100 milliseconds were greater than maximal force (^27,35). Loading using WS appeared to specifically affect both torque production and muscle activity during the first 100 milliseconds to a greater extent than WS + RB. It is possible that the greater muscle activity observed during the initial phase of the dynamic movement led to the greater decreases in EMG 100 milliseconds compared with WS + RB. One possible explanation for the lower resistance, higher muscle activity conflict could be differences in anticipation of the movement velocity between the resistance modes. The protocol of this study included a pause before concentric movement, which should eliminate differences in the utilization of the stretch shortening cycle e.g., the effect of elastic energy (⁷) or Ia afferent activity. Newton et al. (³²) observed a similar behavior of the activation pattern in their study, when comparing explosive bench press actions with and without bar throwing. Finally, the lack of central fatigue on the WS resistance, assessed by voluntary activation level, is in-line with previous high volume explosive jumping studies (^11,38). In other words, pure WS resistance would appear to be more specific to ballistic-type power training.

In contrast, the additional RB increased resistance progressively throughout the concentric phase of the explosive movement. The band, thus, prevented excessive momentum and maintained the load contact throughout the knee joint extension (¹²). Importantly, an explosive action is necessary to build momentum and allow full range of motion, as there is considerably higher resistance at larger knee joint angles. This maximal and continuing need for torque production created a situation where central fatigue could be observed as the magnitude of fatigue resembled maximal strength loading (^35,40). Beelen et al. (⁵) speculated that failure of voluntary drive might be influenced more by high force production with slow contraction velocities than repeated fast contractions. This view was also supported by Newham et al. (³¹), although fatigue might also be dependent on the duration of contraction time (⁶). Velocity and duration of contraction are naturally linked together; however, these observations are in-line with the present results. In this study, the reduction of voluntary activation level was not because of any changes in the neuromuscular propagation, since the present power loadings did not induce significant changes in the maximal M-wave properties and median frequency, similarly to the study by Linnamo et al. (²⁸).

It is fair to assume that the greater metabolite accumulation, observed by significantly higher, but still low, blood lactate concentration, is a consequence of increased workload during loading using WS + RB (¹⁴). For example, performing 15 sets of 1RM did not lead to increased blood lactate concentration on the same WS device (^40,35). However, the metabolic by-products accumulated during power loading were not sufficient to cause significant fatigue in contractile properties of the muscle, as neither resistance mode showed altered passive twitch characteristics. On the contrary, RT torque and RTTD were enhanced after loading on the WS + RB resistance. Possible causes of this observation could be a consequence of increased muscle temperature or increased muscle-tendon complex stiffness because of the higher workload (²¹). Alternatively, these enhancements may possibly indicate greater postactivation potentiation after power loading with the WS + RB resistance mode (¹⁷).

Some methodological considerations from this study should be noted. First, examining voluntary activation level at MVC may not be the optimal method to measure all aspects of neural fatigue during explosive power loading (^39,15), as the greatest change in muscle activation has been observed to occur during the first 100 milliseconds of the contraction (²⁷), especially as observed during the WS loading in the present study. Additionally, it should be noted that the magnitude of central fatigue, observed in the present study was lower than that observed in our previous study using the same WS device and loading protocol (³⁵). These minor, but meaningful differences might be explained by inter-individual variations between these 2 studies. Second, rapid single-joint resistance training movement includes a risk of injury, if the mass (e.g., via ankle pad) is fixed to the limb. However, in this study, subject's limb was free to move backward during the concentric phase; and therefore, it is not likely that the momentum of the mass would have affected the involved structures of joint and muscles considerably during the deceleration phase (³³).

Practical Applications

This study showed clear differences in neuromuscular fatigue due to the addition of a RB during power loading. After WS + RB, MVC was decreased, and this was accompanied by decreased voluntary activation level, indicating central fatigue. Conversely, muscle activation was greater during the initial part of the movement (60–120°) during WS leading to greater fatigue over the first 100 milliseconds. These results may represent different muscle activation strategies during these 2 power loadings, and it seems that there is specific loading-induced fatigue. Using a single-joint exercise as a surrogate for more common power exercises, it may be suggested that if practitioners/athletes/trainers wish to improve both rapid and maximal force production (and associated muscle activation) then the addition of RBs to free weights may be beneficial. Alternatively, if the goal is to improve solely rapid force production and muscle activation then free weights alone may offer the best training paradigm.

Acknowledgments

This work was funded in part by personal grants to Heikki Peltonen from the Ellen and Artturi Nyyssönen Foundation and the Rector of the University of Jyväskylä, Finland. All authors declare no conflicts of interest. The results of this study do not constitute endorsement of the product by the authors or the National Strength and Conditioning Association.