In recent years, the use of prolonged (i.e., from a few min to 60 min) local vibration (LV) has emerged as a new rehabilitation method (Alashram et al., 2019; Wang et al., 2020; Avvantaggiato et al., 2021). LV applied to a muscle or its tendon produces repetitive changes in muscle length, inducing in turn muscle spindle Ia afferents discharge (Burke et al., 1976a) that project at both spinal and cortical levels, with the potential for both acute and chronic neuromuscular adaptations (Souron et al., 2017a). While most of the studies on effects of LV have been obtained while applying LV to relaxed muscles (Noma et al., 2012; Tavernese et al., 2013), there exist alternative studies within the literature that applied LV on contracted muscles (Marconi et al., 2011; Celletti et al., 2017), considering that contraction may increase Ia afferents discharge (and likely optimize LV effects) through increased α-γ co-activation (Burke et al., 1976b). In a recent systematic review including 12 studies investigating the potential for LV to induce some motor conditioning, it was accordingly suggested superimposing LV to muscle contraction may be more effective (Fattorini et al., 2021). Yet, there is a paucity of studies comparing LV applied on relaxed vs. contracted muscles (Filippi et al., 2009; Brunetti et al., 2012) so that cumulative evidence is needed. One way to question the potential benefit of superimposing LV to contraction is to investigate the well-known LV-induced acute changes in central nervous system excitability.

For instance, a decrease in spinal loop excitability (i.e., depression in Hoffmann reflex amplitude) has been largely reported after prolonged LV for lower limb muscles, i.e., mainly the soleus muscle (Ekblom and Thorstensson, 2011; Souron et al., 2019), but also for upper limb muscles, i.e., flexor carpi radialis (FCR) muscle (Nito et al., 2021), even when LV was applied during contraction (Rocchi et al., 2018). H-reflex is commonly used as a proxy of motoneuronal excitability but it can be influenced by different mechanisms as presynaptic inhibition or homosynaptic post-activation depression, limiting its interpretation (Misiaszek, 2003; Knikou, 2008).

In recent studies we used corticospinal tract electrical stimulation to directly assess motoneuron excitability after prolonged LV applied either to the muscle belly or its tendon (Souron et al., 2017b; Souron et al., 2019; Kennouche et al., 2021). This is possible because stimulation of the descending tracts directly activates large diameter axons which are not subjected to presynaptic inhibition (Taylor, 2006). After prolonged LV, we suggested a decrease in motoneuronal excitability as evidenced by a decrease in amplitude of evoked responses to corticospinal tract electrical stimulation in the quadriceps (Souron et al., 2017b; Kennouche et al., 2021) and soleus (Souron et al., 2019) muscles. This was confirmed for the FCR muscle, with evoked responses to cervico-medullary electrical stimulation (i.e., cervico-medullary evoked potentials, CMEPs) being decreased after 6 and 20 min of LV (Nito et al., 2021), with no study investigating the influence of contraction on the modulation of motoneuronal excitability after prolonged LV.

Besides changes in motoneuronal excitability, LV may also alter corticospinal excitability which can be investigated through recording of motor evoked potentials (MEPs) evoked by transcranial magnetic stimulation (TMS). For instance, TMS is a well-known non-invasive method to investigate use-dependent plasticity in corticospinal excitability following a variety of interventions (Moscatelli et al., 2021). Yet, previous studies did not show any MEP modulation immediately after prolonged LV to the Achilles tendon (Lapole et al., 2012) or to the FCR (Steyvers et al., 2004) despite greater MEP amplitude 30–60 min after the end of LV exposure. Conversely, it was reported an increase in MEP amplitude immediately after 10 min of LV superimposed to finger contraction and this was still persistent 30 min after vibration cessation (Christova et al., 2010), yet it was conversely reported a decrease in MEP amplitude 30 min after a LV exposure superimposed to wrist flexion (Marconi et al., 2008). Because MEP amplitude depends on the whole corticospinal pathway (Devanne et al., 1997), results regarding corticospinal excitability changes after LV are difficult to interpret because of confounding effects prolonged LV could have at both supraspinal (i.e. cortical excitability) and spinal (i.e. motoneuronal excitability) levels (Souron et al., 2017a). Comparing changes in CMEP and MEP responses may allow to distinguish neural changes origin between supraspinal and motoneuronal levels, the MEP/CMEP ratio being a proxy of cortical excitability (i.e., an increase in the MEP/CMEP ratio would suggest an increase in cortical excitability) (Taylor et al., 2002; McNeil et al., 2013; Pearcey et al., 2014; Brownstein et al., 2021). Using such ratio, we previously suggested increased cortical excitability after prolonged LV to the quadriceps muscle (Souron et al., 2017b; Kennouche et al., 2021).

Altogether, while greater LV effects could be hypothesised to occur when combined to contraction (i.e. due to increased Ia afferent discharge), studies from the literature do not allow to clearly determine whether applying LV on relaxed or contracted muscles would lead to similar acute adaptations within the central nervous system. The aim of our study was therefore to investigate the acute changes of FCR motoneuronal and cortical excitability following prolonged LV superimposed or not to voluntary wrist flexion. To this end, we recorded CMEP and MEP responses and we further calculated MEP/CMEP ratios. We hypothesized that LV would decrease CMEP and increase MEP/CMEP, with superimposed LV to wrist flexion leading to greater effects.

For the vibration condition (VIB_RELAXED), LV at a fixed frequency of 100 Hz (VCAR0044-0075-00, SUPT Motion) was applied to the muscle belly of the relaxed right FCR (Figure 2). The application lasted 10 min and was repeated three times with an interval of 1 min as described in previous studies (Marconi et al., 2008; Tavernese et al., 2013; Celletti et al., 2017).

For the contraction condition (CONTRACT), participants were asked to voluntarily perform isometric wrist flexion at 10% MVC (thanks to a visual feedback of the force trace). Participants had to contract for 15 s, then relax for 15 s, and this was repeated for 10 min three times with an interval of 1 min in-between. The “contract/relax” rhythm was provided to the participant by an audio soundtrack.

For the vibration + contraction condition (VIB_CONTRACT), a combination of LV and muscle contraction was performed. Participants had to contract and relax as presented above for 10 min three times and LV was applied continuously on the right FCR during the 10 min periods.

For the control condition (CONTROL), participants arm was installed in the ergometer with the right FCR fully relaxed for 30 min.

The aim of the present study was to investigate the influence of 30 min of LV superimposed or not to voluntary wrist flexion on acute changes in motoneuronal and corticospinal excitability. To this purpose, we recorded CMEP and MEP responses of similar size during a low-intensity contraction and we further calculated MEP/CMEP ratios. Immediately after LV applied both during a relaxed state and during wrist flexion, we observed a decrease in CMEP responses, suggesting a decrease in motoneuronal excitability, with no changes after wrist flexion only. MEP amplitude were not modulated suggesting a greater cortical excitability after the two LV conditions (i.e., an increase in MEP/CMEP ratio).

From a methodological point of view, all responses recorded in the present study were evoked during a low-intensity contraction by considering a constant level of electromyographic activity. This allowed us to maintain a stable muscle activation across timepoints (Brownstein et al., 2021). Similarly, we adjusted stimulations intensities at baseline trying to target the same peak-to-peak amplitude for all the evoked responses (i.e., CMEP and MEP) (Pearcey et al., 2014; Kennouche et al., 2021). This contrasts with one of our previous studies where the size of the responses to TMS and corticospinal tract electrical stimulation were not matched, leading us to likely test different proportions of the motoneuron pool (Souron et al., 2017b). We acknowledge that we were not able in the present study to systematically match the size of the evoked responses, as demonstrated by MEP/CMEP ratios far from the expected value of 1 for this ratio at baseline (i.e., PRE) for some participants, likely because of responses variability and difficulties in evoking large CMEPs in some participants.

As expected, CMEPs responses remained similar between PRE and POST measurements in the CONTROL condition, so they were in the CONTRACT condition (i.e., 10% MVC). This latter result is in agreement with previous findings of unchanged responses to corticospinal tract stimulation after repeated low-intensity contraction (i.e., 5% MVC) (Petersen et al., 2003). After prolonged FCR LV (i.e., VIB_RELAXED condition), we observed a 32% decrease in CMEP amplitude, as also previously reported on upper limb (Nito et al., 2021) and lower limb (Souron et al., 2017b; Souron et al., 2019; Kennouche et al., 2021) muscles, suggesting decreased motoneuronal excitability as a consequence of repeated Ia afferents inputs onto to alpha-motoneurons during LV (Souron et al., 2017b; Kennouche et al., 2021; Nito et al., 2021). While we initially hypothesized greater CMEP depression in the VIB_CONTRACT than VIB_RELAXED condition, the 41% decrease in CMEP amplitude in VIB_CONTRACT was not different than the decrease observed in VIB_RELAXED.

Despite a decrease in CMEPs amplitude after LV, we reported no immediate vibration-induced changes in MEPs amplitude regardless of muscle state. As corticospinal excitability depends on both cortical and spinal levels, MEP amplitude should be interpreted in light of changes at the spinal level (McNeil et al., 2013). In our study, the increase in MEP/CMEP ratios after LV applied on relaxed and contracted muscles suggests an increase in cortical excitability. For instance, MEP responses were not altered after prolonged LV despite the decrease in motoneuronal excitability (i.e., decreased CMEP responses), likely masking an increase in cortical excitability (i.e., increased MEP/CMEP ratios). Similar findings were previously reported by Kennouche et al. (2022) and Souron et al. (2017) on the quadriceps muscle. One first potential hypothesis to explain the observed LV-induced increased cortical excitability could be an increase in descending drive during the low-intensity contraction. For instance, this could have been necessary to reach the target EMG activity despite the LV-induced decrease in motoneuronal excitability. However, this would not explain previous findings of increased resting corticospinal excitability 30–60 min after prolonged LV (Steyvers et al., 2004; Lapole et al., 2012). Therefore, an alternative hypothesis to explain previous findings as well as present results would be cortical excitability priming (i.e., cortical excitability modulation) after prolonged LV, likely through cortical projections of Ia afferents (Mima et al., 1997). This may rely on topographically and functionally specific reciprocal connections between primary somatosensory cortex and primary motor cortex, as suggested in a TMS study by increased excitability of the sensorimotor pathways for some responders to LV (Lapole and Tindel, 2015).

Because a greater discharge of Ia afferents is suggested to occur during LV superimposed to muscle contraction (i.e., VIB_CONTRACT) than LV applied on relaxed muscle (i.e., VIB_RELAXED) (Burke et al., 1976b), we initially hypothesized greater CMEP decrease and MEP/CMEP ratio increase after VIB_CONTRACT than VIB_RELAXED. Yet, changes were not different between the two muscle states. This would agree with previous findings where a single session of peripheral nerve electrical stimulation increased MEP amplitude when applied at rest but not when superimposed to isometric voluntary contraction, likely as a result of gating of sensory inputs during isometric contraction (Saito et al., 2014). For instance, gating of sensory inputs is known to occur during voluntary contraction (Rushton et al., 1981; Kakigi et al., 1995), through inhibitory circuits located at both spinal and supra-spinal levels (Seki and Fetz, 2012; Lei et al., 2018). We may therefore speculate that the expected greater Ia afferent discharge in VIB_RELAXED did not lead to greater effects because of similar gating of afferent inputs.

Alternatively, the lack of differences between VIB_CONTRACT and VIB_RELAXED could be the result of an inability of VIB_CONTRACT to actually increase Ia afferents discharge when compared to VIB_RELAXED. In the original study of Burke et al. (1976) using microneurography, it was actually demonstrated increased Ia afferents discharge during LV, and this LV-induced increased discharge was even greater when LV was superimposed to an isometric voluntary contraction. Yet, the latter result was only observed for Ia afferents that were responding submaximally during the relaxed state. For those Ia afferents that were already responding one-to-one to LV, voluntary contraction usually had no influence on discharge rate (Burke et al., 1976b). Because the LV characteristics we used in the present study are thought to be already optimal for Ia afferent stimulation (Vedel and Roll, 1982), this may explain why prolonged LV led to similar acute changes in motoneuronal and cortical excitability whether it was applied on the relaxed or contracted FCR.

Yet, when considering studies that investigated the long-term effects of three consecutive days of 30-min LV to the quadriceps in elderly women (Filippi et al., 2009) and female volleyball players (Brunetti et al., 2012), increased leg power (Filippi et al., 2009; Brunetti et al., 2012), decreased knee joint laxity (Brunetti et al., 2012), as well as increased vertical jump (Filippi et al., 2009) were reported to be greater when LV was superimposed to voluntary knee extension than when LV was applied alone. It should be however acknowledged that none of the two aforementioned studies included a group performing contraction only. Moreover, LV was applied while the leg was extended so that the quadriceps muscle was in a shortened length. This may have led to reduced responsiveness of muscle spindles to vibration (Burke et al., 1978; Souron et al., 2018), so that performing a contraction actually increased Ia afferents discharge (Eklund et al., 1964; Eklund and Hagbarth, 1966). Further studies could usefully explore how combining different muscle length and muscle state could influence LV effects.