Participants

Eighteen healthy young males (age: 22.6 ± 5.7 years old; height: 180 ± 5.7 cm; weight: 74.5 ± 7.8 kg) gave their written informed consent to participate in the present study. Participants reported a regular practice of Parkour, at a wide range of expertise. Our sample size calculation was based on changes in motor performance before and after a tDCS session during an explosive task (jumps) [27]. Considering a significance level of 5%, a power of 90%, and a standard deviation for paired differences of 6.5 (calculated from standard deviations with an hypothesis of a covariance of measures of 50%), 15 jumpers are needed to meet the objectives of the study. Sample size calculation was performed on PASS 13 Power Analysis and Sample Size Software (2014) [28]. Their total training volume (in hours), estimated with their experience (in years) and training frequency (in hours/week) has been evaluated and ranged from 156 to 10,608 hours (mean: 4343 ± 3508). Data were collected from May 2019 to May 2021, and accessed May 2021. Research protocol was approved by the regional ethic committee (CPP-Est-IV) under the number 18/47, registered on ClinicalTrials.org (NCT03937115), and conducted in accordance with the last version of the Declaration of Helsinki.

Experimental set-up and procedures

This study was a double-blinded, randomized, sham-controlled trial to determine the effects of tDCS applied over M1 or dlPFC on static and dynamic balance of Parkour practitioners. The protocol was carried out in 3 experimental sessions separated by at least 48h. The following conditions were performed randomly–anodal tDCS over right M1, anodal tDCS over dlPFC and sham stimulation. The randomization was done at the visit level, meaning that the order of the sessions was randomly assigned for each participant. To ensure the blinding of the type of tDCS applied toward participants, participants were told that any of the three experimental conditions could be a placebo session, independently of the electrode placement. Participants were asked after each session to report if they felt the stimulation was active or placebo. Most subjects (55.6%) thought they had received active stimulation during the placebo session. No difference has been reported in adverse events experienced (tingling, itching, etc.) between the placebo and the active sessions. During each session, subjects performed postural tasks on a force plate before and after tDCS intervention. They executed a bipedal and unipedal static posture before and after tDCS conditions. Similar tests were performed in PRE and POST measurements, but the testing order was randomized across conditions and participants.

Assessment of static posture

Static posture was appraised by means of two experimental conditions–bipedal and unipedal stance, performed on a force plate (Kistler Instrument Corp., Winterthur, Switzerland). The force plate allowed continuous recording of CoP displacements (sampling frequency: 1000 Hz) in both medio-lateral and postero-anterior axis. Bipedal stance was assessed by asking participants to stand still on the force plate, feet shoulder-width apart and arms relaxed. Unipedal stance was evaluated by asking participants to stand still on the left foot by bending the knee of the right leg. Contact with the supporting leg was not allowed. Foot placement and spacing was measured during the first session to be reproduced in the two others. In the two postural conditions, 30 seconds recording were performed, during which participants were asked to keep quiet, eyes open. The last 20 seconds were taken for analysis.

Transcranial direct current stimulation

tDCS was delivered by a neurostimulation device (StarStim®, Neuroelectrics©, Barcelona, Spain) adapted to double-blind procedure. It was transmitted by two saline-soaked synthetic sponge electrodes (Sponstim®, 25cm²) placed on the scalp or shoulder. Every participant agreed to undergo three 20-minute sessions of tDCS (two actives and one sham), separated by 48 hours. Electrodes placement was carried out using the international EEG 10/20 system. The experiment included three stimulation sequences: (i) anode facing the dlPFC (F3 position) and cathode facing the right supra-orbital region (AF8 position), (ii) anode over the dlPFC (F3) and cathode facing the right supra-orbital region (AF8), which is related to sham condition, and (iii) anode facing the right M1 (FC2 position) and cathode placed on the left shoulder. Sequence order was randomized by computer beforehand. Participants were told that any of the three conditions, even including different electrode montages, could be a SHAM condition. During active stimulation conditions, a 2 mA current was generated. The sham condition corresponds to a gradual increase in the intensity of the current during the first 30 seconds up to 2 mA (ramp-up), as recommended by previous literature (Gandiga et al., 2006; Moreira et al., 2021; Swann et al., 2015)

Data analysis

First of all, no missing data has to be reported in the whole data set. In static conditions, i.e. bipedal and unipedal upright standing, CoP characteristics were analyzed taking the 20 seconds trial as a whole. The force-plate represents a two-dimensional plane in which the x-axis represents the postero-anterior axis and the y axis the medio-lateral axis, since participants were always oriented in the same direction. The total sway path (length of CoP displacement) has been determined as the total distance covered by the CoP during the 20 seconds of interest. The area of CoP displacements was also analyzed as the area of the ellipse that includes 90% of CoP points for 20 seconds. Postero-anterior and medio-lateral amplitudes were determined as the maximal amplitude (maximal value–minimal value) covered by the CoP in the ellipse area in the x-axis and y-axis, respectively.

The relationship between these CoP variables and training volume of the participants (in hours, see “participants” section) were evaluated by determining the coefficient of correlation (Pearson’s correlation) to test the effect of experience on postural control.

The relative changes in CoP variables with the different tDCS intervention were determined in percentage by the following formula: [(POST-PRE)/PRE]*100. These relative changes plotted against training volume or initial performance (averaged PRE value).

Statistical analysis

All data are presented as the mean ± standard deviation (S.D.). Each test (unipedal, bipedal) was considered separately. The normality of the data sets was verified by the Shapiro-Wilk test, and variance homogeneity by the Levene test. Outcome variables tested were, in both bipedal and unipedal standing: CoP length (mm), CoP area (mm²), antero-posterior amplitude of CoP displacement (mm), medio-lateral amplitude of CoP displacement (mm). Two-way repeated measures ANOVA were used to assess differences between the 3 conditions with factors “time” (pre, post) and “tDCS intervention” (M1, dlPFC, SHAM). A one-way repeated measures ANOVA was performed on the relative changes, with factor “tDCS intervention” (M1, dlPFC, SHAM). Post-hoc tests were performed with Bonferroni test. Relationships between the several variables were assessed through Pearson’s coefficient of correlation. To account for session order effect (since tDCS conditions were performed randomly), two-way repeated measures ANOVA were also used to assess differences between the 3 conditions in their chronological order with factors “time” (pre, post) and “number of the session” (first, second, third). Effects size are indicated as the partial eta square (ƞ²p), Statistical analysis was performed using STATISTICA (10.0 version, Statsoft, Tulsa, Okhlaoma, USA). The level of statistical significance was set at P < 0.05.

Neurophysiological mechanisms

It has been shown that anodal stimulation of dlPFC could lead to an improvement of postural control, in younger healthy subject [14]. The result of the present study did not meet the initial hypothesis since no effect has been found in parkour practitioners. A possible explanation of this lack of results could probably be linked to the convergence of two factors: 1) parkour athletes present already high balance abilities 2) the dlPFC is not the most involved area in posture. Indeed, traceurs demonstrated strong abilities to maintain posture at least in the tested tasks [17], making the room for improvement far less important than in recreationally active participants. Therefore, although the dlPFC do have a role in managing standing balance [15], its role remains minor as compared to other brain region. The dlPFC is known to play a role in motor planning [29], and its stimulation was shown to be effective in modulating accurate pointing performances in parkour practitioners [12]. From the present result, regarding posture it seems that parkour athletes would need a stronger stimulus.

On the contrary and in line with previous results on healthy young recreationally active participants [4,6,30,31], anodal M1 stimulation had a positive effect on balance control. M1 stimulation would be associated with neurophysiological changes over several hours [32] corresponding to an increase in the motor region excitability which would prevent decline in downward neural control. tDCS would also be able to modulate the motor cortical transmitter system by decreasing the activity of the gamma-aminobutyric acid (GABA)-ergic system [33], whose primary function is to inhibit nerve impulses. tDCS applied to M1 would therefore strengthen postural control by its ability to request more neural networks of the sensorimotor system [7,34]. In addition, this trial adopted an extracephalic montage, which maximizes the effect of tDCS over the anode and can imply the modulation of a larger neural path [12], including subcortical structures involved in postural control[34]. Thus, the neural activation of muscles engaged in maintaining balance would be more efficient, which could justify the reduction of CoP displacements for static stance. As previously stated, traceurs showed a greater cortico-spinal excitability and a different spinal/cortical activation pattern as compared to untrained individuals [20]. Indeed, traceurs exhibit greater corticospinal excitability of lower limb muscles, but lower spinal excitability, as compared to untrained participants. tDCS intervention on M1, but not on dlPFC, was shown to modulate cortical descending command and spinal excitability of traceurs [23]. The particularly low level of soleus spinal excitability of traceurs, even as compared to other sport populations such as gymnasts [22], may make them particularly prone to respond to M1 tDCS intervention. Indeed, since soleus muscle is one of the most important effectors of static balance, this would explain the neurophysiological mechanisms behind M1 tDCS efficiency, particularly in this population.

However, M1 stimulation did not produce comparable results from bipedal to unipedal balance. Even if the decrease of both CoP length and area was significant for both bipedal and unipedal conditions, this was not the case for medio-lateral (ML) displacements, for which the PRE-POST difference was significantly higher in unipedal condition only. More precisely, unipedal stance showed an increase in the ML movements after stimulation, while those made in the postero-anterior (PA) axis were reduced. This result was not expected but could finally attest for a change in postural control strategy [35]. Unipedal balance requiring offsetting movements, decreasing sway on the PA axis could lead to a strong increase of ML oscillations. Unipedal balance involves tightening the calf muscles–sural triceps including gastrocnemii and soleus muscles–to reduce PA displacements or contracting the foot soles and ankles stabilizing muscles, such as the peroneus longus, to limit movements along the ML axis. These two coping strategies do not appear during the bipedal balance.

Considered the reference posture in humans, standing position offers better stability given that the support polygon in which we are involved is larger. Moreover, our results clearly showed that the amplitude of displacements on both axes was less important during bipedal condition (Fig 1E and 1F). These findings agree with a previous research, where larger PA movements were detected when keeping posture became less easy [36]. Bipedal posture may then not be sufficiently sensitive to small variations such as those induced by tDCS, particularly in well-trained athletes. Unipedal standing seemed to represent a more complex task, allowing to discriminate the effect of tDCS. Indeed, for instance the ellipse of CoP displacement was nearly 5 times wider in unipedal stance than in bipedal stance. Stimulating M1 had then a wide field of action to enhance the characteristics of CoP in unilateral task (Fig 2F). Motor abilities and particularly postural strategies can become more effective with training and practice [37]. Therefore, even with a good initial performance in balance, tDCS was efficient to modulate unipedal stance in well trained athletes. This possibly means that tDCS similarly affected uni- and bi-pedal strategies.

tDCS and targeted population

tDCS’s impact varied depending on the level of expertise of the participants. The systematic review of Brachman et al., 2017 concluded that balance training was an effective tool in improving postural control [38]. Holding balance is therefore at the center of training which justifies that the most experienced traceurs showed a lower CoP length and area. Parkour practitioners already demonstrated a significant better postural control performance than sedentary people during landing [21]. The effects of physical activity on postural control depends on the time of practice [39]. Traceurs, therefore, represent a population on which brain stimulation is effective on postural balance through the stimulation of pre-existing neural networks. However, a ceiling effect can be reached for more experienced practitioners. When carefully looking at individual data, it can be observed that for some individuals the CoP length did not decrease after M1 tDCS. Therefore, a limit can be established between those who showed a significant downward modulation of CoP length after tDCS (n = 15/18 for bipedal stance and n = 12/18 for unipedal stance) for and those who did not (n = 3/18 for bipedal stance and n = 6/18 for unipedal stance). According to the training volume of these participants, the limit could be established between 6000 hours (for unipedal stance) and 8000 hours of training (for bipedal stance). In fact, the more the subject is trained, the less room there is for improvement. Finally, it can also be noticed that a great inter and intra-subject variability was reported regarding corticospinal excitability modulation following tDCS, which can be linked with a lack of knowledge concerning the current transfer location [40]. Besides participants’ expertise in the applied tasks, the effects of tDCS may also be affected by other intrinsic factors such as individual anatomic variations [41].

Conclusion

To summarize, the results of our study suggest that offline M1 stimulation with an extracephalic montage was suitable to improve postural control in Parkour practitioners, contrary to dlPFC stimulation. Among other intrinsic factors, we found that the effects of tDCS were correlated with participants’ expertise in Parkour, as expressed by their total training volume. Then, although tDCS seems to be effective in modulating postural control in trained individuals, this technique may have limitations in the most highly trained of them.