To our knowledge, this is the first study that comprehensively examined the reliability and magnitude of work completed across VL thresholds during RT while also exploring the effects of several factors potentially influencing the use of VL thresholds in practice. The main findings of this study were 1) the agreement of the number of repetitions performed across VL thresholds between two testing sessions was not acceptable, regardless of the load used; 2) the agreement of the velocities associated with the first instance when a given VL threshold was exceeded between two testing sessions was not acceptable, regardless of the load used; 3) the number of repetitions performed across VL thresholds was affected by sex, load, and emotional stability, whereas the magnitude of the velocity associated with the first instance a given VL threshold was exceeded was affected by sex, testing session, and load; 4) whether the same VL threshold was reached by the same person in two consecutive days was affected by load, whereas whether someone would experience a 50% VL in a set depended upon the load used and emotional stability of the person; and 5) whether someone would perform multiple repetitions within a given VL threshold was affected by sex, load, training history, strength, and emotional stability. Considering the above, it can be posited that 1) VL thresholds cannot be used to reliably control the amount of work completed during RT, but rather as a general variable quantifying the amount of fatigue experienced by an individual; and 2) VL thresholds should be highly individualised when used in practice since load, sex, training status, history, and psychological traits can all influence, at least to some extent, the variability in responses to different VL thresholds.
The agreement of the actual volume of work completed until reaching different VL thresholds between two consecutive testing sessions (72 hours apart) was not acceptable since the limits of agreement constantly sat outside the ± 2 repetitions margin, regardless of the load used. For instance, with 70% of 1RM, the upper CI of the upper LoA (estimate = 5.47) and lower CI of the lower LoA (estimate = -5.36) were higher and lower than 5 repetitions, respectively. This effectively means that, for the same threshold, individuals can be expected to perform anywhere between less than 5 and more than 5 repetitions on two consecutive testing sessions in a controlled environment (i.e., laboratory settings). While the extent to which this difference in work would affect longitudinal adaptations is not currently quantified, it is likely that the effect is profound. This contention is supported by the discrepancy in the literature regarding the effectiveness of various VL thresholds for inducing training adaptations. For instance, lower VL thresholds were equally effective and more beneficial for muscle strength and power adaptations, whereas higher VL thresholds were superior for muscle hypertrophy compared to lower ones [25]. However, when the volume of work was matched between different VL threshold groups, no differences were observed in muscle strength, hypertrophy or even gains in performance of athletic tasks [12, 13]. In addition, individuals might not reach the same VL threshold on two consecutive training sessions, as found in the present study (Fig. 5b), which compromises the ability of this method to control RT volume. This would only be exacerbated outside of controlled lab environments due to the complex interplay of different levels of fatigue, motivation, prior injury, and general readiness to train across individuals in common practical settings. Finally, the agreement of the velocities associated with the first instance when a given VL threshold was exceeded between two testing sessions was also not acceptable, regardless of the load used. Therefore, VL thresholds do not seem to be a reliable option for monitoring and prescribing RT volume.
The number of repetitions performed across VL thresholds was affected by sex, load and emotional stability, whereas the velocity associated with the first instance a given VL threshold was exceeded was affected by sex, load, and day. In this regard, both the number of repetitions and repetition velocity decreased as the load increased across VL thresholds. Although this finding may seem logical, it implies that the same amount of work cannot be completed just by selecting a given VL threshold, as the work completed is also load dependent. Thus, monitoring and performance evaluation (e.g., pre to post training cycle) should be done across the loading spectrum, while comparing performances for each load individually. Furthermore, regardless of the load used, females completed more repetitions than males across VL thresholds, while males performed repetitions at higher velocities. These findings align with the reported physiological and neuromuscular differences, often resulting in differences in exercise performance and recovery, between men and women.[26] Specifically, on average, men are stronger, and women are less fatigable; able to sustain force at the same relative intensity for a longer period [27, 28]. These sex differences in strength and fatigability were previously attributed to variation in muscle phenotype[29] insofar that women have smaller muscle fibres than men [29] and a higher proportion of type I fibres relative to type II [30], with greater muscle capillarisation [31], blood flow during exercise [32], and with distinct glycolytic and oxidative capacities [33, 34]. While relative strength did not influence the number of repetitions performed or repetition velocity across VL thresholds in the present study, sex differences in fatiguability could explain why females did more work, on average, compared to males across VL thresholds, and sex differences in the proportion of type I fibers relative to type II fibers, could explain why males performed repetitions at higher velocities, on average, across VL thresholds. Finally, individuals with higher levels of emotional stability tended to perform more repetitions across VL thresholds, on average. In general, high scorers on neuroticism – the trait on the opposite spectrum of emotional stability – tend towards heightened perception of fatigue symptoms [17]. Indeed, elevated neuroticism, by definition, is related to more frequent and intense experience of negative emotional states and elevated neuroticism is positively associated with fatigue [17, 16]. Furthermore, individuals with higher levels of neuroticism feel exhausted more frequently and report more severe fatigue than those with lower neuroticism.[35] It may be that individuals higher in emotional stability have in fact stronger fatigue tolerance (rather than propensity to experience less fatigue), thus enabling them to complete a greater amount of work across VL thresholds. Importantly, conscientiousness did not affect the number of repetitions performed or repetition velocity across VL thresholds in the present study, although this personality trait has been previously linked with fatigue [16]. Conscientiousness generally describes people who are task oriented. Since all participants who completed all the procedures were verbally encouraged and acknowledged for their efforts during the study, it may be that this was enough for people of all levels of conscientiousness to feel as though the task was successfully completed, thus preventing any influential effects of this personality trait that might have occurred if the experimental situation was less structured without clear instructions, encouragement and acknowledgments for completing the task (i.e., the testing session). The amount of VL experienced during RT has been repeatedly shown to be a useful, non-invasive indicator of the acute metabolic stress, hormonal response and mechanical fatigue during RT, and as such could be used to quantify fatigue induced by the training set [7, 8]. However, it seems that this monitoring practice could be drastically improved by considering individual trainee characteristics and training conditions since sex, the choice of load, and emotional stability all affected responses to different VL thresholds during RT.
There are several considerations when implementing velocity-based approaches to RT. For instance, the repetition-velocity relationship might not be linear for all individuals [36], and sets can be terminated after one or two repetitions exceed a predetermined threshold [37]. It seems logical that these considerations also apply to using VL thresholds during RT. It may be that whether individuals 1) can perform multiple repetitions within a single VL threshold; and 2) have the capacity to experience a 50% VL in a set are additional factors worth considering when implementing VL thresholds in practice, since both will likely affect the repetition-velocity relationship, as well as when sets are terminated. Indeed, whether individuals can do multiple repetitions within a single VL threshold is further influenced by sex, load, strength, emotional stability, and prior training practices. In this regard, females, stronger individuals, individuals higher in emotional stability, and those who generally perform higher repetitions in training (> 12) and leave 2–4 repetitions in reserve, are more likely to perform multiple repetitions within a single VL threshold. The effects of sex may be attributed to differences in fatiguability and muscle phenotype between males and females, whereas those with higher emotional stability had greater fatigue tolerance, allowing them to complete more work within the same VL threshold. It is not surprising that stronger individuals, and those who perform higher number of repetitions during their own training were more likely to perform multiple repetitions within a single VL threshold. Namely, stronger individuals would generally possess greater neural drive [38], myofibrillar cross-sectional area [39], and superior intermuscular coordination [40], all likely enhancing the ability to complete more work. In addition, generally performing a high number of repetitions during RT increased the odds of doing multiple repetitions within a single VL threshold likely due to these participants having greater muscle endurance (i.e., training specificity principle). Interestingly, those who usually left 2–4, compared to 0–2 reps in reserve in their own training were more likely to perform multiple repetitions within a single VL threshold. Furthermore, only load and emotional stability explained the ability of the individuals to experience a 50% VL during RT. Indeed, experiencing 50% VL with higher loads is unlikely given that repetition velocity sharply decreases as the load increases, thus not giving the opportunity to reach high VL. However, individuals higher in emotional stability were more likely to experience high VL during RT, probably due to their ability to cope with fatigue, as previously alluded to. At least some of these findings could also be explained by the fact that individuals have their own “pattern” of experiencing VL (Supplementary file V). For instance, some people experience a gradual decline in velocity as the number of repetitions increase, whereas others maintain high velocity in the beginning and then experience a sudden drop in velocity, while others experience an early drop in velocity and then maintain velocity near the end of a set. The exploratory correlations between the number of repetitions performed until reaching 10, 20, 30, and 40% VL and the maximum number of repetitions completed in a set across loads support the concept that individuals have unique VL patterns. If there was indeed a general pattern of VL, one would assume very high correlations (r > 0.9) between the number of repetitions performed until reaching any of the VL thresholds and the maximum number of repetitions performed in that set (until failure). However, this was not the case, especially not with the loads (e.g., 70–80% 1RM) and VL thresholds (0–20%) typically used to optimise neuromuscular adaptations during RT, as the correlations never surpassed r = 0.63 (Supplementary File VI). While this insight is entirely exploratory, it seems reasonable to hypothesise that individuals possess their own patterns of VL, which limits the generalisability of specific VL threshold stimuli across individuals. Considering all the above, it seems prudent to consider the choice of load and individual trainee characteristics (i.e., training status, history, and psychological traits) when implementing VL thresholds in practice as each affects the variability of responses.
The present study took a unique approach to examine a range of factors related to the use of VL thresholds during RT, and there are several considerations when interpreting the data. Firstly, the reliability and magnitude of the work completed until reaching different VL thresholds may not necessarily transfer to other exercises. However, the effects of trainee characteristics are likely applicable to a variety of exercises. Secondly, while attempts were made to represent a broad range of participants (e.g., training experience, strength levels, sex etc.) these findings may not generalise to sedentary populations since the participants had at least 6 months of RT experience. This at least partially explains why training experience did not influence any of the outcomes in the present study. Thirdly, since not every set in the present study was performed in a completely fresh state, some residual fatigue may have affected performance in subsequent sets and perhaps between day agreement for the amount of work completed across VL thresholds. However, 72 hours of rest between sessions and long rest intervals between sets, with participants performing sets with loads in a descending order should have ensured that the potential effects of fatigue were minimised. Fourthly, while efforts were made to balance the numbers of female and male participants, the number of females in the present study was considerably lower compared to males (due to COVID-related issues with recruitment). However, the females had a wide range of strength levels, training experience, and different training practices, improving this sample’s generalisability. Nevertheless, future studies should aim to balance the number of male and female participants when making comparisons, where possible. Finally, the present study could not determine variability in metabolic, neuromuscular, and biomechanical responses to different VL thresholds. Therefore, future studies should quantify the extent that the demonstrated variability in the amount of work completed until reaching different VL thresholds affects the variability in acute metabolic, neuromuscular, and biomechanical responses to VL thresholds, as well as in longitudinal adaptations.