Using cluster and rest redistribution set structures as alternatives to resistance training prescription method based on velocity loss thresholds

Introduction

The beneficial effects of resistance training (RT) on muscular strength, power, speed, endurance, balance, coordination, and hypertrophy are widely recognised (Suchomel, Nimphius & Stone, 2016; Suchomel et al., 2018). In addition, RT is often used for injury prevention and rehabilitation purposes and has an important role in general well-being due to numerous beneficial effects on health and quality of life (O’Connor, Herring & Caravalho, 2010; Feigenbaum & Pollock, 1999). However, adaptations to RT may differ depending upon several factors such as the choice and order of the exercises, training intensity, volume, frequency, rest interval duration, set structure, and velocity of the repetitions (Schoenfeld, Ogborn & Krieger, 2017; Grgic et al., 2018a, 2018b; Pareja-Blanco et al., 2014; Jukic et al., 2020, 2021). Of these, training volume and intensity have received the majority of attention in the literature (Schoenfeld, Ogborn & Krieger, 2017; Schoenfeld et al., 2017). On that note, exercise intensity during RT has traditionally been determined as the load lifted relative to the one-repetition maximum (%1RM) while set volume has frequently been prescribed according to a theoretical maximum number of repetitions per set that can be performed against a given %1RM until reaching muscular failure (Bompa & Haff, 2009). Although this way of prescribing training loads and volume is relatively simple and practical and can be managed with relative ease with large groups of people, it does not account for physiological and psychological stressors that can affect an individual’s day-to-day performance (Mann, Ivey & Sayers, 2015). For instance, maximal strength can fluctuate daily when an individual is fatigued or significantly increase within a few weeks due to training adaptation (Padulo et al., 2012). Furthermore, the number of repetitions that can be completed against a given %1RM has been shown to be both individual- and exercise-specific (Richens & Cleather, 2014; Sánchez-Medina & González-Badillo, 2011). To combat these issues, the velocity-based training (VBT) approach to RT has been shown as a highly effective and reliable methodology for RT monitoring and prescription (González-Badillo & Sánchez-Medina, 2010; Sánchez-Medina & González-Badillo, 2011).

Among other advantages, VBT allows for the objective assessment of the neuromuscular fatigue incurred during a set by monitoring the velocity loss (VL) of the repetitions (Sánchez-Medina & González-Badillo, 2011). This is because performing multiple repetitions in a consecutive fashion without rest between repetitions (i.e., traditional sets [TS]) inevitably results in fatigue accumulation, especially when large number of repetitions are performed, and the loads used (Jukic et al., 2020). In addition, monitoring VL reached in each set serves as a precise method of quantifying the level of effort (i.e., repetitions performed with respect to the maximum number that can be completed) (Sánchez-Medina & González-Badillo, 2011). Since level of effort is an important factor in determining the acute responses (Pareja-Blanco et al., 2018) and subsequent adaptations to RT (Pareja-Blanco et al., 2017), many studies using VBT approaches to RT have examined the effects of training with different magnitudes of VL on maximal strength, hypertrophy, muscle endurance and performance of athletic tasks. In brief, higher VL (>20%) maximised hypertrophic adaptations but resulted in a significant reduction in the IIX muscle fiber phenotype (Pareja-Blanco et al., 2017, 2020) whereas lower VL (≤20%) resulted in similar improvements in muscle strength and endurance as well as performance of sport-specific tasks such as vertical jumping and sprinting (Pareja-Blanco et al., 2017). However, Pareja-Blanco et al. (2020) recently showed that here might be an upper and lower VL threshold that should be prescribed during RT to induce optimal training adaptations. It was concluded that moderate VL thresholds (i.e., VL10 and VL20) should be chosen to optimise adaptations to RT since VL0 seemed to induce levels of fatigue that were too low to maximise adaptations, whereas VL40 did not promote further muscle strength and hypertrophy compared to VL20.

Despite the benefits of monitoring repetition velocity during RT, doing so requires the use of valid and reliable velocity measuring devices (e.g., linear position transducers). While the cost of these devices has decreased in recent years, they are still not affordable to many. Even if velocity measuring devices are available, their use with groups of people can be an impractical task (e.g., team sports, group sessions). In that regard, more heuristic approaches to controlling movement velocity during RT could prove beneficial. Cluster sets (CS), where intra-set rest periods are implemented in addition to the pre-existing inter-set rest periods, are a straight-forward method to reduce fatigue while allowing higher velocity repetitions to be performed during RT (Jukic et al., 2020). However, CS extend total training time relative to TS which might not always be feasible from a practical standpoint. A more time efficient alternative to CS is to simply redistribute the total rest time of TS structures to include shorter and more frequent rest intervals. Just like CS, this set structure, known as rest redistribution (RR), has been shown to be effective—albeit to a lesser extent—in alleviating acute mechanical, metabolic, and perceptual markers of fatigue during RT (Jukic et al., 2020), and thus effective at inducing positive training adaptations (Jukic et al., 2021). With these points in mind, it is possible that both CS and RR set structures could be used as an a-priori alternative to commonly used VL thresholds. Indeed, Tufano et al. (2017) observed that 12-s inter-repetition rest periods allowed for 36 consecutive back squat repetitions to be performed without dropping below VL20 while examining the effects of CS and RR on mechanical performance whereas Jukic & Tufano (2019) recently showed that RR allowed almost all repetitions (~17.5 out of 18) in a clean pull exercise to be performed with less than 20% VL regardless of the load used across three sets. However, these studies used either CS or RR protocols in isolation, did not have a TS protocol, and used only lower body exercises in their protocols. Since both CS and RR could affect RT performance differentially (Jukic et al., 2020, 2021) with these effects often being exercise specific (Latella et al., 2019), it would be beneficial to examine both CS and RR as an alternative to VL thresholds (e.g., VL20) across multiple sets and exercises within the same study.

To shed light on whether CS and RR could be used as an a-priori alternative to VL, we examined the effects of TS, CS and RR on (a) VL reached during back squat (SQ) and bench press (BP) exercises across multiple sets; and (b) the number of instances at which VL20 was not exceeded. Based on previous research (Jukic & Tufano, 2019; Jukic et al., 2020, 2021), we hypothesised that: (a) lower VL will be experienced during CS and RR compared to TS during both SQ and BP across all sets (CS < RR < TS); and (b) Both CS and RR would allow for the greatest number of repetitions to be performed above the VL20 threshold (CS > RR > TS) and could potentially serve as an a-priori alternative to VL20 thresholds.

Materials and Methods

Results

For descriptive purposes, raw data with medians and interquartile intervals as well as their distribution is presented in Fig. 1. Results from the linear mixed-effects model (interactions and main effects) are described in the text, whereas estimated marginal means with 95% confidence intervals and pairwise comparisons are illustrated in Fig. 2. The final linear mixed-effects model on the velocity loss experienced during TS, CS and RR revealed a non-significant interaction between the set structure, set number, and exercise (F = 0.793; p = 0.530). Similarly, exercise × set number and set structure × set number interactions did not reach statistical significance (F = 0.693–0.772; p = 0.500–0.544). However, set structure × exercise interaction was significant (F = 9.846; p = 0.0005). In addition, significant main effects of set structure (F = 97.811; p < 0.0001; TS > CS > RR), exercise (F = 281.153; p < 0.0001; BP > SQ) and set number (F = 7.770; p = 0.0005; Set 3 > Set 2 > Set 1) were observed for the amount of VL reached.

The final generalised linear mixed-effects model for the number of instances participants were able to stay above the 20% VL threshold revealed no significant set structure × exercise × set number, exercise × set number, set structure × exercise, and set structure × set number interactions (χ² = 0.015–4.417; p = 0.959–0.110). However, significant main effects of set structure (χ² = 45.356; p < 0.0001; RR > CS > TS), exercise (χ² = 162.987; p < 0.0001; SQ > BP), and set number (χ² = 14.813; p = 0.0006; Set 1 > Set 2 > Set 3) were found for staying above the 20% VL threshold. Specific comparisons (odds ratios) are presented in Table 1. Predicted probabilities for staying above the 20% VL threshold broken down by the set structure, exercise and set number are illustrated in Fig. 3.

Discussion

The purpose of this study was to examine the effects of CS, RR, and TS on neuromuscular performance and to determine the viability of using CS and RR as alternatives to prescribing VL20% thresholds across multiple BP and SQ sets. The main findings of this study revealed that (a) the inclusion of 30 s intra-set intervals during CS, and creating shorter but more frequent rest periods during RR prevented the VL experienced during TS with RR being the most effective set structure for this purpose; (b) these effects were more pronounced during BP than SQ (i.e., the magnitude of VL was lower); (c) both CS and RR allowed the vast majority of repetitions to be performed above VL20%; and (d) velocity was better maintained above VL20% during the BP compared to SQ during CS and RR, with RR also being more effective than CS for maintaining velocity during the BP exercise. Taken together, data from the current study show that CS and RR are effective at reducing the overall fatigue-induced decrease in velocity compared with TS while also allowing the vast majority of repetitions to be completed above VL20%—though the magnitude of these effects are likely to be exercise-specific.

As expected, significantly lower VL was observed during both CS and RR when compared to TS, regardless of the exercise and the set number—a finding that is in agreement with the literature (Jukic et al., 2020). However, our first hypothesis was only partially supported since, although both CS and RR were effective at reducing VL experienced during TS, RR was more effective than CS. The way in which CS and RR set structures were constructed could potentially explain the greater effects of RR over CS on VL. Namely, RR set structure consisted of 9 sets of 2 repetitions with 45 s of inter-set rest whereas CS consisted of 3 min of rest after every set of 6 repetitions which was interspersed by 30 s of rest after every 2 repetitions. Due to this setup, inter-set of RR (45 s) can be aligned and compared with the intra-set rest of CS (30 s). In this regard, longer rest periods after every 2 repetitions during RR than CS might have allowed better energy substrate restoration, while only 30 s of rest during CS caused some residual fatigue to be carried over from the previous set of 2 repetitions, subsequently resulting in greater VL compared to RR (see Supplemental Files S1 and S2 for visualisation). However, although plausible, this is speculative since energy substrate restoration was not measured in the present study. Regardless of the mechanism behind this finding, it must be noted that RR set structures are also more efficient as they do not extend total training time like CS, but still attenuate neuromuscular fatigue (and seemingly to a greater extent than CS, at least in the present study). This is particularly important for practitioners who are often time constrained. Furthermore, the magnitude of the effect of CS and RR on VL, when compared to TS, was lower for the SQ than the BP exercise since much lower VL was experienced during TS, and subsequently CS and RR, in the SQ than the BP exercise. Therefore, when considering the findings from the present study, findings from a recent meta-analysis (Latella et al., 2019), and distinct neuromuscular and perceptual fatigue profiles reported in the literature for upper- and lower-body muscles (Vernillo et al., 2018; Mayo, Iglesias-Soler & Fernández-Del-Olmo, 2014), the effects of CS and RR on the overall neuromuscular performance might be exercise specific.

Although reducing neuromuscular fatigue, as quantified by VL incurred in a training set during resistance training, is a very beneficial feature of CS and RR set structures, the VL experienced might still be unnecessarily high (i.e., higher than 25%) which could result in sub-maximal training adaptations (Pareja-Blanco et al., 2017, 2020). In this regard, a significant body of literature now suggests terminating training sets after VL20–25% has been reached. The rationale behind this suggestion is based on the findings from longitudinal studies which collectively show how VL20–25% is the most efficient VL threshold as it induces hypertrophic adaptations (Pareja-Blanco et al., 2020), allows for positive neuromuscular adaptations (Pareja-Blanco et al., 2020), preserves the expression of the fast twitch muscle fibres with considerably lower training volume being performed compared to higher VL thresholds (Pareja-Blanco et al., 2017). However, valid and reliable measuring devices are needed to prescribe sets using VL which are still not affordable to many, and therefore making this training methodology inaccessible to many. The present study shows the viability of both CS and RR to potentially serve as cost-free alternatives to VL20% resistance training prescription. Specifically, the participants were able to perform the vast majority of repetitions during both CS and RR with less than 20% VL compared to TS. Therefore, CS and RR were effective not only at allowing for higher absolute velocities during the whole training session (average velocity across sets), but also at minimising the degree of fatigue (i.e., velocity loss) experienced at the end of each training set. The present results agree with the findings from the only two other studies—to our knowledge—that have taken a similar approach to assess the ability of CS and RR to maintain velocity above certain thresholds (Tufano et al., 2017; Jukic & Tufano, 2019). However, those studies used either CS or RR protocols in isolation, did not have a TS protocol, or used only lower body exercises. Therefore, the present study expands on previous findings demonstrating that the effects of CS and RR, compared to TS, were more pronounced during the BP exercise, with RR also being more effective than CS in maintaining repetition velocity above VL20%. As stated, the greater effects of RR compared to CS could be attributed to how these set structures were constructed (i.e., longer more frequent rest periods during RR). However, the greater benefits of both CS and RR during the BP could likely be explained by the smaller muscle groups involved in this exercise—which results in more localised fatigue—compared with the SQ where the fatigue is distributed among a greater amount of muscle mass. The relative position of the sticking region in these exercises may also explain these VL differences, as the SQ allows more time/distance for force generation after such region. This probably also explains why overall VL was greater in the BP compared to SQ exercise. In turn, participants managed to maintain repetition velocity in the SQ exercise even during TS (~21%), although their velocity maintenance was still considerably higher during CS and RR (~10%). In this regard, it can be concluded that CS and RR will provide greater benefits for the BP exercise due to the greater localised fatigue that is typically experienced in this compared to the SQ exercise, but also that CS and RR will allow for the extremely high velocity maintenance in the SQ exercise. Thus, practitioners should be aware that the same VL threshold (or the use of the same CS or RR setup) may produce divergent training adaptations when used for different exercises. With that said, the use of CS and RR during resistance training can be of a great practical value as it can produce similar outcomes as prescribing VL20% (Pareja-Blanco et al., 2017, 2020) while avoiding the financial and potential logistical constraints associated with the use of velocity tracking devices.

Conclusions

The use of CS and RR attenuated VL experienced during TS across multiple sets of both BP and SQ exercises. In addition, the use of CS and RR allowed for the vast majority of the repetitions to be performed above VL20%. These beneficial effects of CS and RR were greater in the BP exercise likely due to the smaller muscle groups involved in this exercise, thus resulting in more localised fatigue accumulation compared to the SQ exercise. While this suggests that the magnitude of the effect might be exercise-specific, velocity maintenance in the SQ exercise is still expected to be considerably greater during CS and RR than TS. Importantly, for the purpose of maximising mechanical outputs in the shortest time period, RR emerged as the most efficient set structure among the three since it allowed for the lowest VL among the three set structures, and greatest number of repetitions to be performed above VL20% without extending the total training time—which was the case during CS. Collectively, while both CS and RR are effective tools for reducing the overall fatigue experienced during resistance training, they also allow repetition velocity to be maintained above VL20%. Therefore, the beneficial effects of terminating training sets after VL20% demonstrated in prior literature can be achieved by using CS and RR, while avoiding the financial and logistical constraints associated with the use of velocity tracking devices.

Supplemental Information