Modulations of corticospinal excitability following rapid ankle dorsiflexion in skill- and endurance-trained athletes

Corticospinal plasticity, the ability of the brain to modify neuronal connections, is essential for learning, motor control, improved memory, and recovery from brain injury (Kantak et al. 2012). It can be modified by conditions such as visual, auditory, and proprioception information through adapting the neural connections (Pascual-Leone et al. 2005). Motor training is a process of acquiring information from external sources and accomplishing movement, which is intrinsically associated with neuroplasticity. Long-term sports training improves corticospinal plasticity for motor learning (Hötting and Röder 2013; Singh et al. 2016). Meanwhile, different categories of training such as endurance training and skill training seem to modify the neural system differently (Schlaffke et al. 2014).

Endurance training aims to increase the capacity of continuous motor output by repeating the same movement sequence and, therefore, increasing the efficiency of the movement (Barnes and Kilding 2015). Endurance training increases cognition and neuroplasticity in several brain regions such as the cerebellum, hippocampus, and cerebral cortex via different mechanisms of global affection, such as altered blood volume in the brain and lactate induces elevation of neural growth factors and, but does not alter specific motor map organization or synapse number (synaptogenesis), which is produced by motor learning (Thomas et al. 2012; Taubert et al. 2015). At the spinal level, the excitability of the motor neurons is known to adapt after long-term endurance training (Koceja et al. 2004). Following rapid toe movement, well-trained swimmers demonstrated higher spinal excitability than non-trained individuals (Ogawa et al. 2009), and similar results were later presented in endurance runners (Ogawa et al. 2012). In general, long-term endurance training results in enhanced spinal excitability and provides increased blood flow, and oxygen delivery through angiogenesis, to brain regions, but appears not to participate directly in the modulation of synaptic number or topology (Churchill et al. 2002; Taubert et al. 2015; Chen et al. 2019).

On the other hand, skill training is defined as the acquisition and subsequent refinement of novel movement sequences (Adkins et al. 2006). According to neuroimaging studies, when learning a new specific exercise (e.g., dancing, gymnastics) that triggers motor skill learning processes, greater neural networks are activated within the brain area of the focused task compared with simple movements (e.g., grasping and moving small objects) (Papale and Hooks 2018; Ungerleider et al. 2002). A transcranial magnetic stimulation (TMS)-based experiment showed that skill-trained athletes (dancing, gymnastics, and figure skating) have higher capacity for corticospinal plasticity of the test-relevant muscle (soleus) compared to endurance-trained athletes (cross-country skiing, orienteering), (Kumpulainen et al. 2015). During a single motor skill learning session, there is an increase in corticospinal excitability in the area controlling the corresponding limb (Suzuki et al. 2012), which may be related to decreasing cortical inhibitory neurotransmission (Kolasinski et al. 2019). From previous studies, both single session of skill training seems to modify cortical behavior, and long-term skill training has been shown to result in changes of corticospinal excitability and different cortical responsiveness versus endurance training (Suzuki et al. 2012; Perez et al. 2004).

One method to investigate the effects of interventions on corticospinal plasticity is the stretch reflex test, where neural responses are recorded by surface electromyography (EMG) (Hagbarth 1967). Such a method can be combined with stimulation methods, such as TMS, to determine the contribution of different parts of the neural system to the reflex response (Budini et al. 2017). This stretch reflex contraction occurs naturally in locomotion (i.e., stretch-shortening cycle), but the exposure to and fatigue induced by these actions during training (Avela and Komi 1998) could be hypothesized to be different between skill- and endurance athletes leading to differential corticospinal responses. For a healthy human, an imposed dorsiflexion of the ankle joint leads to a series of clear responses in the EMG of the stretched muscles. The main response, with an onset latency at 40‒50 ms, is called the short-latency stretch reflex (SLR) and is mediated by a monosynaptic reflex loop (Fellows et al. 1993; Lee and Tatton 1975). The classic view is that SLR is “purely” under spinal control, whereas only the long-latency reflex (LLR ~ 90 ms) (Dietz et al. 1984) can be influenced by cortex behavior (Petersen et al. 1998) based on latencies likely for a transcortical loop (Evarts 1973). This view is supported by TMS-evoked MEP responses not affecting SLR or medium-latency reflex (MLR), but being facilitated at LLR₂ (after ~ 120 ms) (Taube et al. 2008). Based on the movements performed during training, e.g., endurance athletes performing repetitive stretch–shortening cycles while skill athletes perform regularly changing movement patterns, it is reasonable to assume that skill- and endurance athletes differ in their corticospinal control of movement and that this may become apparent in different phases following muscle stretch.

The aim of the current study was to explore the contribution of/and the underlying corticospinal mechanisms mediating motoneuronal responses to stretch reflex of skill- and endurance-trained athletes by recording MEPs in the soleus muscle. Both skill and endurance training are known to lead to neuronal adaptation and mechanisms of neuronal modulation in short-term sports training has been explored (Kolasinski et al. 2019), but there are contentions about how the different types of long-term training affects neuroplasticity. Whether muscle stretch influences corticospinal facilitation/inhibition differently in endurance- and skill-trained athletes and, thus the mechanism(s) behind natural movement remain unknown. It was hypothesized that endurance-trained athletes would show more prominent modulation at SLR, while skill-trained athletes would show higher modulation after SLR (SLR + 120 ms).

There were no differences between groups in training years (endurance group: 11 ± 3 years; skill group: 13 ± 3 years, P = 0.330), rMT (endurance group: 54 ± 7% stimulator output; skill group: 47 ± 8% stimulator output, P = 0.055) or MVC (endurance group: 297 ± 67 Nm; skill group: 227 ± 86 Nm, P = 0.140).

This study investigated changes in corticospinal excitability at different latencies relative to rapid dorsiflexion between skill- and endurance-athlete groups. As planned, the passive ankle dorsiflexion led to a stretch reflex in the soleus muscle, which is an important part of proprioceptive processing and results in afferent feedback to both spinal and supraspinal centers. It was hypothesized that the endurance-trained athletes would show more prominent corticospinal modulations at SLR, while skill-trained athletes would show higher modulation during the period where a supraspinal reaction to the movement is prominent (SLR + 120 ms). In line with the hypothesis, the present study showed higher MEPs at p120 in the skill group. However, in opposition to the hypothesis, the endurance group did not demonstrate more prominent corticospinal modulation at SLR. Finally, MVC was strongly correlated with resting MEPs in the skill group, which was not the case in the endurance group. On the other hand, a strong correlation between stretch reflex and MEPs was observed at SLR in the endurance group but not in the skill group.

In the present study, MEPs at p120 were higher than at Onset only in the skill group. p120 took place 120 ms after SLR, which was approximately at the latency of the second long-latency reflex (LLR₂) reported by Taube et al. (2008). The increased MEPs at LLR₂ indicated modulation of corticospinal excitability, while reduced H-reflex at the same time point suggested that this modulation was cortical in nature (Taube et al. 2008). At this phase, there is sufficient time to allow different pathways, including cortical and spinal, to contribute to the recorded MEPs’ facilitation and inhibition. Greater MEPs in the skill group 120 ms after SLR suggests that they have a greater or more long-lasting facilitation of corticospinal excitability than endurance-trained athletes after rapid ankle dorsiflexion, even in a passive condition. One important suggestion on the mechanisms of motor learning-induced cortical plasticity is that synaptic connections at the cortical level are modified through LTP (Friedman and Donoghue 2000). LTP is a prerequisite for synaptogenesis and, thus, skill training has, indeed, been shown to lead to synaptogenesis of the motor cortex (Adkins et al. 2006). PAS, which is an artificial intervention pairing electrical stimulation of somatosensory nerves and TMS of the corresponding area of the motor cortex, can produce LTP-like plasticity in the synapse. The amount of PAS-induced LTP-like plasticity increase depends on the number of active synapses. Therefore, a PAS intervention has been used as a measure of corticospinal plasticity (Lazzaro et al. 2009). In a study by Kumpulainen et al. (2015), PAS induced increased MEP in skill athletes, but not in endurance athletes or untrained adults, revealing higher corticospinal plasticity and greater synaptogenesis at the cortical level of skill-trained athletes. Previous findings suggest that skill training results in increasing adaptability of corticospinal plasticity and, thus, skill-trained athletes may preferentially rely more on cortical sources for voluntary movement as was the case in the present study at p120 and a120 condition.

It is important to note that there was no evidence of voluntary muscular activity prior to the stretch or after the stretch reflex response had abated in the passive trials (i.e., the muscle was silent). We speculate that this was a ‘priming’ mechanism in the skill group to modulate top-down responses by motor programs stored in the central nervous system after the rapid perturbation (Pierrot-Deseilligny & Burke, 2005). Possible explanations are as follows: first, for skill-trained athletes, there are more voluntary movement changes in training and competition, which need to be controlled by the motor cortex, cerebella, or somatosensory association cortex (Kurtzer 2015; Suminski et al. 2007). Second, central control has been exposited in the processing of, e.g., expected postural response (Horak et al. 1989). Thus, following the perturbation, the skill athletes may have been ‘primed’ for a voluntary response after the rapid ankle movement. In support of this contention, a strong positive correlation between MVC and resting MEPs was observed in the skill group only. MVC force is dependent on recruitment of motor units and the force-producing capacity of muscle fibers. A higher MEP value is related to higher excitability of motor cortical output cells and motor neurons during voluntary contraction (Taylor et al. 2002). Therefore, for skill-trained athletes, corticospinal excitability plays an important part in voluntary movement and is possibly observed in our enhanced p120 MEP_AVG during the phase where supraspinal reaction would be possible as a cortical adaptation to a top-down strategy in response to rapid ankle dorsiflexion.

Weaker, but supporting, evidence for greater reliance on cortical involvement in skill athletes was found in ICF. During all conditions, average ICF values were higher in the skill than endurance group and a significant main effect for the group was observed, indicating that ICF was higher. Nevertheless, there were no (pairwise) statistical between-group differences at any condition, which dilutes confidence in making such inferences. The cortical mechanisms of ICF are not fully clear. Increased ICF is known to be strongly influenced by decreased GABAergic inhibition or a separate increase in glutamatergic facilitation (McGinley et al. 2010; Ziemann 2003). Since the two receptors of GABAergic inhibition, GABA_A and GABA_B, influence SICI and LICI, respectively (McDonnell et al. 2006; Kujirai et al. 1993), and that no differences in SICI or LICI were observed, it may be that potentially higher ICF in skill athletes was due to glutamatergic facilitation. While this is speculative, glutamatergic facilitation is one of the important molecular mechanisms for LTP, and as such, a greater ICF would support Kumpulainen et al. (2015) findings that skill athletes have higher corticospinal plasticity than endurance athletes.

Even though LICI showed facilitation (i.e. > 100% of test MEP) in the present study, this has been previously shown to occur when 50 ms inter-stimulus interval is employed in the assessment of LICI (Valls-Solé et al. 1992; Di Lazzaro et al. 2002). This presumably occurs because of increased postsynaptic excitability elicited by the conditioning stimulus or stimulus-induced activity in subcortical regions (Bolden et al. 2017; Valls-Solé et al. 1992; Di Lazzaro et al. 2002). However, the present study is not able to determine the precise mechanisms underpinning this finding.

At SLR, SR/ MEP_AVGs ratio was used to normalize MEPs with stretch reflex responses to reveal whether TMS has an additive effect on EMG amplitude, which was expected to demonstrate a between-group difference in the present study. However, although the ratio in each group (endurance group = 1.8, skill group: = 1.3) was raised (i.e. , > 1), no significant between-group differences were observed. In a previous study between endurance athletes and a non-trained group, endurance runners demonstrated higher monosynaptic reflex excitability by enhanced stretch reflex response, which highlights enhanced modulation of spinal excitability after long-term endurance training (Ogawa et al. 2012). In the present study, there were also no between-group differences observed for SICI or LICI at SLR. These findings imply that, in the non-motor control task (i.e., resting muscle), corticospinal modulation did not affect SLR differently between training groups.

On the other hand, SLR MEP_AVG showed a strong correlation with stretch reflex in the endurance group only. The monosynaptic spinal loop likely contributes more to corticospinal excitability than the supraspinal loop in the fast response phase after rapid ankle dorsiflexion. Possible reasons for the relationship in endurance athletes are as follows: training that includes repetitive stretch–shortening cycle actions was the predominant training form of eight out of this study’s ten endurance athletes (cross-country skiing, long-distance running, and triathlon). It is widely known that training-induced muscle and spinal motor neuron adaptation occurs from such a stimulus (Churchill et al. 2002; Taubert et al. 2015; Avela and Komi 1998). In rapid ankle movement, i.e., stance and swing phase in running, fast spinal loop modulation helps to keep balance during body oscillations since modulations via cortical processes would be too slow (Tahayori and Koceja 2012). In agreement with this hypothesis, a study performed using the H-reflex method revealed that after a long period of typical endurance training (well-trained swimmers in the study), athletes demonstrated greater spinal excitability (i.e., increased H-reflex) than non-trained individuals (Ogawa et al. 2009). However, the present study was unable to determine whether spinal excitability is indeed higher in endurance training athletes compared with skill training athletes through direct comparisons. There were some individuals in the endurance group seemingly specifically adapted for high excitability post-stretch from our results, and perhaps the lack of between-group differences may have been due to high within-group variance and low sample size.

Some study limitations should be considered. Due to COVID-19 and quarantine policies of the laboratory, we were only able to recruit and complete testing for 20 participants. Therefore, this sample size may not have been sufficient to determine between-group differences considering the large variabilities of MEP amplitudes when responding to rapid ankle dorsiflexion. Although it is not possible to perform a priori sample size estimation for novel measurements, the convention within the field is that typical sample sizes per group are approximately 15 (e.g., Kumpulainen et al. 2015; Wächli et al. 2017). Therefore, only the clearest differences in corticospinal plasticity between skill- and endurance athletes may have reached the level of statistical significance in the present study. This may explain, for example, significant main effects for IC, but no significant differences when post hoc tests were performed to ascertain specific differences between conditions. It is also possible that different loading patterns may induce differences in corticospinal plasticity between sports. For example, the two swimmers included in the present study may have added variance to the results given that their sport does not include stretch–shortening cycle actions in the triceps surae muscles through repetitive ground contact as in running. We suggest that sport and training characteristics should be considered when recruiting athletes as participants if the research involves corticospinal responses during motor tasks. Finally, information regarding the exact phase of the menstrual cycle was not collected in the present study. It is currently debatable whether testing during different menstrual cycle phase would influence the data (Ansdell et al. 2019; El-Sayes et al. 2019), but it may be pertinent to consider in the future.

This study observed a similar pattern of corticospinal modulation, as revealed by MEP_AVG, in long-term trained endurance- and skill athletes during and following rapid ankle dorsiflexion. However, corticospinal excitability (MEP_AVG) was enhanced 120 ms after muscle stretch in skill-trained athletes, together suggesting a ‘priming’ of corticospinal excitability during the supraspinal reaction phase. Our skill-trained athletes demonstrated a positive relationship between MEP amplitude and MVC, supporting the view that some reliance on corticospinal excitability for voluntary action is particularly important for skill athletes.