Enhanced skeletal muscle contractile function and corticospinal excitability precede strength and architectural adaptations during lower-limb resistance training

Evidence suggests that initial training-induced increases in muscle strength are primarily the result of neural adaptation (Komi et al. 1978; Moritani and deVries 1979; Narici et al. 1989; Reeves et al. 2005). Early electromyography (EMG) studies demonstrated an association between increased strength and gross efferent output (Hakkinen and Komi 1983; Aagaard 2003). However, strength training adaptation varies greatly between individuals (Dankel et al. 2020; Roberts et al. 2018) with a cascade of neural, mechanical, and architectural adaptations taking place (Maughan et al. 1983; Higbie et al. 1996; Hubal et al. 2005; Erskine et al. 2010; Blazevich et al. 2008). Now, thanks to technological advancements, specific areas of the neuromuscular pathway can be investigated in isolation, shedding greater light on these initial adaptations. Specifically, individual motor unit (MU) properties (Van Cutsem et al. 1998; Vila-Chã et al. 2010), and parameters of the corticospinal tract and primary motor cortex (M1) (Carroll et al. 2002; Selvanayagam et al. 2011) have been studied. Collectively, these findings suggest that specific locations and responses underpin the early adaptations that typically occur within the first ~ 6 weeks after commencing strength training (Wilson et al. 2019). Observed responses include increased excitability (Goodwill et al. 2012; Weier et al. 2012) and decreased inhibition (Latella et al. 2012) of the corticospinal pathway, as assessed using transcranial magnetic stimulation (TMS) (Kidgell et al. 2017). Such alterations serve to increase motor neuron output and increase force production (i.e., strength) (Kidgell et al. 2017; Siddique et al. 2020). Although, these findings represent one of several areas of potential neural adaptation to resistance training they do not confirm alterations to MU behaviour. Therefore, further investigation, incorporating simultaneous assessment of different regions within the neuromuscular network would provide a more complete understanding of the neural contribution to strength adaptation.

Increased agonist muscle activation resulting from resistance training (Pucci et al. 2006; Jenkins et al. 2017) has been attributed to altered MU behaviour (Van Cutsem et al. 1998); specifically, increased firing rate following periods of resistance training has been detected by intramuscular EMG (Vila-Chã et al. 2010) and by decomposing surface EMG (dEMG) (Del Vecchio et al. 2019). However, conflicting findings have shown unaltered MU firing rate (Rich and Cafarelli 2000; Pucci et al. 2006; Sterczala et al. 2020), reaffirming some of the uncertainty around strength training adaptation. While EMG provides a useful, non-invasive method to assess MU adaptations, it does not allow us to make direct inferences about adaptations within the spinal excitatory/inhibitory networks. Therefore, concurrent measurements of MU discharge properties alongside the associated corticospinal networks of excitation and inhibition could provide clearer insight into the precise nature of the early adaptations to resistance training. To date, similar approaches have been performed in isolation, but as yet, simultaneous measurements across the breadth of the neuromuscular network have not been conducted as part of a robust training intervention paradigm.

We previously demonstrated that tensiomyography- (TMG) derived measurement of muscle belly radial displacement (Dm), which describes the magnitude of muscle deformation in response to percutaneous electrical stimulation, is inversely related to changes in muscle architecture (muscle thickness and fibre pennation angle) (Wilson et al. 2019). When assessing muscle atrophy during bed-rest, Šimunič et al. (2019) demonstrated an increase in Dm prior to any observable change in muscle thickness or pennation angle. This evidence suggests that the time-course of contractile mechanics and architectural changes are not exclusively linked. Although increases in pennation angle and muscle thickness have been shown to contribute to later increases in muscle strength (after > 4-weeks training) (Aagaard et al. 2001; Blazevich et al. 2003), it is not known whether contractile properties would be modified prior to these hypertrophy-orientated architectural adaptations. With early increases in strength primarily attributed to neural adaptations, it may be possible that early alterations in contractile properties are influenced by altered excitation–contraction (E–C) coupling (Calderón et al. 2014). Previously, TMG has been used to infer alterations in E–C coupling following exercise-induced muscle damage (EIMD) (Hunter et al. 2012); where increased contraction time (Tc), the time taken for the muscle belly to reach peak displacement, was associated with secondary EIMD markers. This research demonstrated the capability of TMG to assess early changes in contractile mechanics. To further our understanding of resistance training adaptation it would be beneficial to investigate the time-course of alterations in contractile mechanics in relation to other areas within the neuromuscular pathway. This exploration will provide improved clarity around the underpinning central and peripheral mechanisms responsible for strength adaptation. We anticipate that this research will expand on the growing body of work which challenges some long-held beliefs regarding the neural contributions to resistance training adaptation (Pearcey et al. 2021).

The aim of this study was to track the time-course of changes in muscle contractile properties with relation to muscle architecture and neuromuscular adaptation across a 6-week resistance training programme. By assessing the time-course of changes throughout the neuromuscular pathway we aimed to further elucidate the mechanisms responsible for increasing strength shortly after commencing a new programme of strength training. We hypothesised that over the course of the 6-week resistance training intervention increased strength would be accompanied by reduced Tc and Dm of agonist muscles, indicating improved contractile mechanics and that these contractile adaptations would occur prior to any discernible change to muscle architecture. We additionally hypothesised that increased strength during the initial stages of training would be accompanied by an increase in corticospinal excitability, reduced corticospinal inhibition, and increased MU firing rate.

Three participants withdrew from the study; 1 from the training group for failing to complete > 90% of training sessions; 2 from the control group, respectively for sustaining an injury unrelated to the study and for failing to pass the TMS screening, leading to a final sample size of 40 participants; 22 in the training group (10 males/12 females; 173.48 ± 5.20 cm; 74.01 ± 13.13 kg; 19.14 ± 1.30 years) and 18 in the control group (10 males/8 females; 175.52 ± 7.64 cm; 70.92 ± 12.73 kg; 20.61 ± 0.94 years).

In the present study we aimed to determine the time-course of neuromuscular adaptations to a resistance training programme among novice participants. Training took place over a 6-week dynamic resistance training period and participants demonstrated increased 5-RM BS supporting the effectiveness of the adopted training programme. A decrease in VL and RF Dm was observed after 2-weeks training and prior to any increase in isometric strength, neural adaptation, or change in muscle architecture. After 4-weeks of training, an increase in maximal quadriceps strength was observed, and accompanied by, increased corticospinal excitability; however, no changes in VA or MU MFR within the quadriceps were seen. Furthermore, an additional increase in maximal quadriceps strength was observed from Wk 4 to Post-intervention, alongside increases in muscle thickness and pennation angle. All effect sizes associated with major findings are presented in Table 6.

Training-induced increases in 5-RM BS strength may be subject to a technique-learning effect; while any perceived strength gain based solely on specific adaptation to a repeatedly trained action should be considered with care, we caution against dismissing initial gains due to potential learning effect, as this effect could still be considered a relevant adaptation (Dankel et al. 2017). Nonetheless, increases in MVC, which we observed at Wk 4 and then again Post-intervention (Table 1), represent an unbiased measure of increased quadriceps strength. In this study, we observed increased quadriceps muscle contractile function and viscoelasticity of the muscle–tendon complex (Evetovich et al. 1997), as shown by decreased VL and RF Dm (Fig. 3A, B). Decreased quadriceps Dm before altered muscle architecture suggests a separate mechanistic change may be responsible for increased muscle tone. A recent study (Šimunič et al. 2019) demonstrated similar decreases in Dm prior to changes in muscle architecture during a recovery intervention following atrophy, also supporting the notion of mechanisms other than architectural adaptation being responsible for altered contractile properties. We previously reported that impaired E-C coupling has been represented by reduced Dm and increased Tc accompanying increased limb girth following EIMD (Hunter et al. 2012). However in the current study, no accompanying change in force output (Table 2) or Tc (Fig. 3C, D) was observed, indicating E-C coupling was not impaired. The present reduction in Dm was ~ 10% less than that shown by Hunter et al. (2012), indicating the less severe form of EIMD that commonly occurs in early stages of resistance training (Damas et al. 2015). A plausible explanation for reduced Dm is the alteration of intra-muscular tissue fluid content (Kasuga 2015) known to occur with EIMD in the early stages of resistance training (Chen et al. 2012). Such a non-invasive marker of contractile function could be useful for practitioners to gain objective insight into the efficacy of training interventions in their early stages. Future work should look to directly measure intra-muscular fluid content changes following resistance training concurrently with changes in contractile properties, to provide precise mechanistic understanding. High reliability of both Dm and Tc has been reported previously, using the same protocol that we have here adopted (Martín-Rodríguez et al. 2017) and the inter-individual spread in Dm recorded in this study is also similar to that reported elsewhere for lower-limb muscles (Llurda-Almuzara et al. 2020). Therefore, while we can be confident that the values at the upper and lower extremes of our range are not the result of measurement error, the significance of relatively high/low Dm within the context of our study is uncertain. This uncertainty arises because TMG measures skeletal muscle contractile mechanics in vivo, meaning that criterion referenced validity is inherently difficult to quantify (Macgregor et al. 2018). That we observed minimal relationships between changes in strength and Dm in both muscles suggests that the variance in Dm is not meaningful. TMG is not without limitations, we have previously raised questions around the external validity of the technique in relation to sports performance (Macgregor et al. 2018). While our current study, along with previous research (Wilson et al. 2019), has revealed associations between TMG-derived parameters and muscle function, it is still evident that the level of force evoked by electrical stimulation using TMG is < 10% of MVC (Ditroilo et al. 2011; Šimunič et al. 2011), which may—at least in part—explain the minimal relationships that we have observed between changes in Dm and strength.

Initial increases in muscular strength after 4-weeks of training were accompanied by neural adaptation in the form of increased MEP amplitude (Fig. 4A), partially confirming our secondary hypothesis. Increased MEP amplitude represents an increase in corticospinal excitability which includes the excitability of M1 and the efficiency of descending volley transmission through the spinal cord and into the muscles (Di Lazzaro et al. 2004). Similar results have previously been shown using dynamic and isometric training interventions, whereby increased corticospinal excitability has accounted for early strength gain (Griffin and Cafarelli 2007; Leung et al. 2013; Mason et al. 2017), although others have observed no change in MEP amplitude (Kidgell and Pearce 2010). Along with the findings of a recent meta-analysis (Siddique et al. 2020) the present study indicates that early-increases in muscular strength can be attributed to improved efficacy of neural transmission along the descending corticospinal tract. Interestingly, whilst corticospinal excitability remained increased (compared to baseline) at Post-intervention assessment, there was no further increase beyond Wk 4 (Fig. 4A). There are limited data regarding adaptations in corticospinal parameters after the initial increases presently observed as, by and large, previous studies have employed 3–5 week training interventions (Kidgell et al. 2017; Siddique et al. 2020). However, present data would suggest corticospinal excitability may not increase further after the initial observed change but rather, remain elevated after 4 weeks of training. We are not the first to report very early elevation in excitation, with little subsequent increase; Mason et al. (2020) reported greater area under the MEP recruitment curve following only one training session, with no further changes. We might speculate that the greater and more distal muscle mass involved in the present study (quadriceps vs. wrist flexors) may explain the delayed plateau in excitability response. Future work investigating the time-course of corticospinal excitability should look to determine the nature of this ‘ceiling effect’ and the influences of new training stimuli, including over a longer training duration to incorporate substantial muscle hypertrophy, and among a variety of muscle groups. The retention of the adaptations we have here observed could additionally be examined by prolonging measurements following cessation of the training stimulus (i.e., a detraining period). Since participants in this study were previously unaccustomed to resistance exercise, our findings are specific to novice exercisers. In this population we have here observed an early plateau in neuromuscular adaptation (i.e., within 6-weeks); whether a similar response could be expected among previously resistance trained individuals—for example, when starting a novel training programme or recommencing training after a period of interruption—is unclear. The principle of muscle memory associated with myonuclear domain size that has been reported in non-human mammalian muscle (Bruusgaard et al. 2012; Gundersen et al. 2018) has not been translated to humans (Psilander et al. 2019) but should not be dismissed as a possible mechanism for augmented adaptation to resistance exercise among non-novices. Therefore, future research should seek to build on our current findings by exploring early responses to resistance exercise in previously trained individuals.

At odds with recent meta-analyses (Kidgell et al. 2017; Siddique et al. 2020), and our own secondary hypothesis, was a lack in change of corticospinal inhibition (Fig. 4B). Recently, Ansdell et al. (2020) also demonstrated no change in cSP duration when assessing short-term training adaptations; however, these authors also saw no change in MEP amplitude unlike the present data. Therefore, it is possible the presently observed results are not due to alterations in M1 but rather somewhere else along the corticospinal tract. However, as TMS MEPs are unable to precisely differentiate between intra-cortical mechanisms (Brownstein et al. 2018) this cannot be confirmed. It should also be acknowledged that whilst shown to be a reliable method of analysis (Damron et al. 2008), cSP measurement does involve practitioner discretion as to when discernible EMG signal re-commences following the silent period; holding a potential for variance. Future work should explore techniques that allow such differentiation to be made within the corticospinal tract, such as stimulation of the cervico-medullary junction to determine efficacy of corticospinal-motor neuronal synapses (Nuzzo et al. 2016). Similarly, despite increases in voluntary strength (5-RM and MVC), we observed no change in VA. Trezise and Blazevich (2019) previously reported increased VA following a longer training programme (10 weeks) of similar frequency to our own, but it is noteworthy that this change showed no relationship with improved isometric strength. While VA may or may not be seen to increase following training, other neuromuscular adaptations are more clearly related to improving strength capability.

In order to gain comprehensive and concurrent insight into early-resistance training adaptations we employed sEMG decomposition to explore MU discharge property adaptations (Rich and Cafarelli 2000; Kamen and Knight 2004). No change in MU MFR, measured at 60% of MVC, was observed at any time point in either group (Table 5). The present MFR data is at odds with a recent study by Del Vecchio et al. (2019) who observed increased MFR in tibialis anterior after 4-weeks isometric training. Disparity between these findings may be due to methodological differences such as the training-specific tests in which MFR was assessed, and the specific equipment (and subsequent algorithms) used to collect MFR data. Another methodological difference was that Del Vecchio et al. (2019) pooled MFR data from multiple contractions, while we did not. Pooling MFR data from multiple contractions has the potential to alter numbers of low-threshold (faster firing) MUs (Van Cutsem et al. 1998) and inadvertently influence MFR. Indeed, MU MFR analysed on a per-contraction basis (as adopted here) suggests MFR is not altered Post-training (Beck et al. 2011; Stock and Thompson 2014; Sterczala et al. 2020). Furthermore, changes in adipose tissue thickness may arbitrarily alter MU properties derived from sEMG due to spatial filtering (Petrofsky 2008); to add a further measure of control to the present study, we measured adipose tissue thickness at the VL and RF electrode sites, with both remaining unchanged over the duration of testing (Table 4). It is possible that resistance training presently altered the recruitment thresholds of MUs, or the degree of MU hypertrophy (Sterczala et al. 2020); but this cannot be confirmed as it was not possible to measure relationships between MU property-recruitment thresholds. Therefore, in line with a previous suggestion (Contessa et al. 2016), future studies should look to these relationships when investigating training induced changes in MU behaviour to obtain greater clarity along the recruitment threshold spectrum.

Later increases in MVC strength (Wk 4 to Post) were accompanied by architectural adaptations in both VL and RF; namely increased pennation angle and muscle thickness (Table 2). Previously, increases in muscle thickness and pennation angle have been demonstrated following similar timeframes of resistance training (Blazevich et al. 2003; DeFreitas et al. 2011; Damas et al. 2015), and have been shown to be contributory to increased capacity for maximal force production (Aagaard et al. 2001; Campbell et al. 2013). It may be beneficial for future studies in this area to also include global measures of body composition to track changes in muscle mass. We observed that VL muscle thickness increased prior to RF (Table 2); this inter-muscle difference possibly resulting from differing stimuli during training (Floyd 2014). Previously, differences in hypertrophic response have been observed between VL and RF (Mangine et al. 2018), with differences in joint articulation involved in the exercises used in the present intervention suggested as explanatory; despite both muscles being controlled by the same innervation point (Page et al. 2019). Despite this inter-muscle difference in adaptation, our findings support the initial hypothesis of muscle architecture enhancements accounting for strength gain in the latter stage of training, as VL muscle thickness also increased from Wk 4 to Post-intervention. Interestingly, VL muscle thickness increased concurrently with corticospinal excitability after 4-weeks training, suggesting that neural and architectural adaptations may not be mutually exclusive. Additionally, corticospinal excitability remained elevated in the presence of further quadriceps architectural adaptation (Wk 4-Post), supporting the notion that a cumulative effect of neural and architectural adaptations account for strength gains. Despite our participants being unaccustomed to resistance exercise, all were healthy and otherwise physically active, as well as being young adults. We know that from the age of ~ 50 years, skeletal muscle mass is progressively lost at a rate of 1–2% per year (Baumgartner et al. 1998; Lauretani et al. 2003) and aged muscle is characterized by smaller muscle fibre diameter, more-varied fibre size, and reduced number of muscle fibres—specifically type II fibres (Frontera et al. 2000; McPhee et al. 2018; Tintignac et al. 2015). In older age, loss of muscle function is greater and more rapid than would be expected from the reduction in muscle mass alone, due in part to motor unit denervation and neuromuscular junction degeneration (McPhee et al. 2018; Mosole et al. 2014). Therefore, we caution against applying our present findings to older adult populations; similarly, further research is required to understand the early responses to novel resistance exercise among sedentary and/or unhealthy population groups.

As previously stated, our study was not designed to compare within-group response differences between male and female participants. Nonetheless, while the present sample (20 males/20 females) may be suitable to infer general conclusions regarding novice resistance exercisers, further research is needed to explore whether inter-sex differences may exist. It is well-understood that male and female responses to resistance exercise differ; men typically maintain ~ 10 kg greater muscle mass than women irrespective of overall body mass (Rossetti et al. 2017), so despite similar relative hypertrophy following resistance training (Abe et al. 2000; Hubal et al. 2005), in absolute terms men can gain up to twice as much muscle mass compared to women (Ivey et al. 2000). Inter-sex differences in TMG-derived properties have been less thoroughly investigated; among healthy individuals, women have been observed to present lower Dm than men in lumbar region (Lohr et al. 2020) and lower-limb musculature (Kusumoto et al. 2023). On the other hand, strong, resistance-trained women displayed greater Dm in lower-limb musculature compared to similarly well-trained men (Herring et al. 2021). Interestingly, Kojić et al. (2021) observed similar decreases in Dm among men and women following a 7-week resistance training intervention, but notably there were no inter-sex differences in Dm before training commenced among those individuals. Taken together, we might speculate that the existing evidence suggests the potential for TMG to distinguish between male and female musculature among healthy, non-resistance trained individuals, but the adaptive response to resistance training appears unlikely to differ between men and women.

We are the first to demonstrate that increased quadriceps muscle contractile function and corticospinal excitability can be observed in the early stages of training prior to increased strength and architectural adaptation. However, precise mechanistic understanding of this change in skeletal muscle contractile mechanics remains to be determined. The integrated assessment approach employed in the present study supports the consensus that early strength is attributed to changes in the neural physiology, however further work is needed to precisely determine the location of adaptation. The cumulative effects of neural and physiological adaptations observed here may provide practitioners with some greater clarity of the time-course aspect of training-induced adaptations. This knowledge may inform the efficacy of training, rehabilitation monitoring and planning considerations.