tDCS over the left prefrontal cortex does not affect time-trial self-paced cycling performance: Evidence from oscillatory brain activity and power output

Transcranial direct current stimulation (tDCS) is a non-invasive brain stimulation technique, which is increasing in popularity amongst sport scientists. The rationale of using tDCS as a tool in sport is that stimulating brain areas related to exercise performance would make athletes to boost their physical performance. Here, in a pre-registered, (https://osf.io/rf95j/), we tested the hypothesis that tDCS over the left dorsolateral prefrontal cortex (DLPFC) influences performance in a 20’ time-trial self-paced exercise and electroencephalographic (EEG) oscillatory brain activity. 36 trained male cyclists completed a 20’ time-trial self-paced exercise in three separate sessions, corresponding to three stimulation conditions: anodal, cathodal and sham. tDCS was administered before each test during 20’ at a current intensity of 2.0 mA. In each session, power output, heart rate, rate of perceived exertion (RPE) and EEG (at baseline and during exercise) was measured. We found that neither anodal, cathodal nor sham improved performance, affected heart rate, RPE or EEG activity. Our data suggest that the effects of tDCS on endurance performance should be taken with caution, with the brain region stimulated being an important factor.


Introduction 30
Self-paced exercise refers to a physical activity in which the effort needs to be evenly distributed and 31 monitored in order to complete the task without reaching premature exhaustion 1 . Performance in self-32 paced exercise is undoubtedly related to the functioning of peripheral body systems, such as the 33 muscles, heart, lungs etc., as well as the brain. In this respect, self-pacing during exercise is a 34 challenging cognitive task 2 , as it requires constant control and monitoring of internal (e.g., heart rate) 35 and external inputs (e.g., a bump on the road while cycling), while maintaining the goals of the task 36 (e.g. completing a set distance as fast as possible). In other words, self-paced exercise can be regarded 37 as an executive task, with high demands of self-control, goal-monitoring and inhibition 3 .

38
Research in cognitive neuroscience has long pointed to the prefrontal cortex as a key brain area 39 involved in executive processing 4 . Interestingly, the few neuroimaging studies testing participants 40 while exercising have shown activation of the prefrontal cortex, together with the expected sensorio-41 motor recruitment 5 , which reinforces the hypothesis of the crucial role of executive processing on 42 self-paced exercise. The rationale of the present study was therefore that stimulation of the prefrontal 43 cortex via transcranial direct current will affect self-paced exercise performance.

44
Transcranial direct current stimulation (tDCS) is a non-invasive electrical brain stimulation technique 45 that is able to induce cortical changes by depolarizing (anodal) or hyperpolarizing (cathodal) a 46 neuron's resting membrane potential 6  and mood in elite athletes of different sport modalities (n = 10) after ten days of anodal stimulation 63 over the left dorsolateral prefrontal cortex, which, according the authors, may contribute to 64 4 performance gains, greater well-being and faster recovery. However, due to the lack of a control 65 condition (Borducchi et al.) and small sample sizes in their studies (like in almost every previous study 66 on tDCS and sport performance), the above results should be considered with caution.

67
The present (pre-registered, https://osf.io/rf95j/) research is novel as it is the first to directly test the 68 hypothesis that stimulation of the prefrontal cortex would affect performance in a 20' time-trial self-69 paced exercise bout in trained male cyclists. More precisely, we expected that activation via anodal 70 stimulation would improve performance, whilst inhibition of the prefrontal cortex via cathodal 71 stimulation would impair performance (compared to a sham condition). The indexes of physical 72 performance were the power output during exercise and the RPE after the self-paced exercise.

73
Additionally, we asked participants to perform an executive task 14 after the exercise. The purpose was 74 to test the hypothesis that any change on physical performance produced by the tDCS over the 75 prefrontal cortex would modulate the subsequent (known 15 ) effect of exercise on inhibitory control.

76
This is in line with the idea of a bi-directional relationship between exercise, brain and cognition 15 , 77 i.e., brain and cognitive functioning influences exercise performance and vice versa. Brain electrical 78 activity was measured at rest, during exercise, and during the cognitive task by recording 79 electroencephalography (EEG) in order to examine the effects of tDCS at brain level. Even though the

Exercise performance 88
The average power output during the time trial self-paced exercise was not significantly different 89 (F(2,34) = 0.31, p > 0.05) between conditions (see Fig. 1

91
The heart rate signal for three participants was lost during the 20' time-trial self-paced exercise, 92 consequently they were removed from the subsequent analysis (n= 33). The average heart rate during

Electrical brain activity (EEG) 100
Due to excessive noise in the EEG signal, five participants were not included in the EEG analysis (n= 101 31). The analysis of tonic spectral power (see Fig. 2) did not provide any significant difference (all ps 102 > 0.05) between conditions (anodal, cathodal and sham), and for each period of time (baseline-pre; 103 baseline-post, warm-up, self-paced exercise and recovery).

104
The event-related spectral perturbation (stimulus-locked) analysis in the flanker task (see Fig. 3) did 105 not reveal any main effect of condition for the congruent or incongruent trial (both ps > 0.05).

Executive task 107
A main effect of stimulus was reported in the flanker task, with participants being less accurate (M=

113
To the best of our knowledge, this is the first study testing the influence of prefrontal cortex tDCS' 114 stimulation on self-paced exercise and brain activity during exercise. The main finding of this study 115 was that 20' anodal or cathodal tDCS' stimulation (relative to sham) over the left dorsolateral 116 prefrontal cortex did not affect exercise performance or brain electrical activity. Moreover, neither 117 sRPE, EEG or cognitive performance were affected by the stimulation.

118
Our findings indicated that anodal or cathodal tDCS applied over the left dorsolateral prefrontal cortex 119 before exercise did not modulate exercise performance during a 20' time-trial self-paced exercise. This 120 finding contrast the results of the two previous studies testing the effect of tDCS over the same brain    2) The selection of the electrode montage. To date, M1 is the most recurrent stimulated area. It is 128 suggested that M1 stimulation is likely to increase power output during exercise or to reduce RPE for 129 a given force or power 8 . It could then be argued that tDCS would only affect physical performance 130 when applied to M1. However, there is a previous study targeting the left prefrontal cortex finding 131 significant results 12 ; and another stimulating the temporal cortex, 10 , which may jeopardize the

148
The hypothesis that anodal would increase EEG amplitude was not confirmed in the present study.

149
After the 20' stimulation, the EEG spectral power was similar across all condition for each period of 150 time. This null effect could be again explained by the low intensity of the stimulation. Vöröslakos et 151 al. 20 found that currents between 4-6 mA should be delivered to modulate EEG amplitude.

152
The rationale of including the flanker task after the cycling self-paced exercise was that any change in

176
The method and planned analyses of this study were pre-registered on the Open Science Framework 177 (OSF). This was done on June 29, 2017, and can be found at https://osf.io/rf95j/. The raw files can be 178 found in OSF.

179
Previous research has shown that the effect of tDCS on exercise performance is small-medium.

180
Additionally, we considered that a medium effect would be appropriate in terms of the potential future 181 practical application of the findings from this type of research to elite cyclists. Therefore, according to   limitation of our study. Nevertheless, that was motivated for two reasons: 1) our participants were 197 experienced cyclists used to performing self-paced exercise, and given their expertise, the purpose was 198 that of familiarize them with the laboratory setting testing procedure; 2) we were aware that the 10' 199 test was performed after the maximal incremental exercise test and participants were already fatigued. were freely able to change gearing and cadence throughout. Participants were aware of the elapsed 218 time, but they did not have feedback on performance (wattage and heart rate) during, or after the self-219 paced exercise. Heart rate was measured continuously throughout the protocol (V800, Polar Electro, 220 Finland). Immediately after exercise, we asked the participant to rate their session RPE (sRPE) 24 .

221
Finally, participants completed a 5' cool-down and the executive task. The interval between the 222 different sessions was at least 48h to allow the full recovery and to minimize carryover effects.

265
All analyses were completed using statistical nonparametric permutation tests with a Monte Carlo 266 approach. These tests do not make any assumption of the underlying data distribution, are unbiased, 267 and as efficient and powerful as parametric statistics. When statistical significance (p < 0.05) was 268 found, values were corrected by the false discovery rate method. The effect of experimental condition 269 (anodal, cathodal, sham) on self-paced exercise power output, heart rate and RPE were analysed using 270 a within-subject design condition.

271
Spectral power was analysed using a within-participants' design with the factor of stimulation (anodal, 272 cathodal, sham). Each period (Baseline, Warming Up, Exercise, Cooling Down) was tested separately 273 for significance. In the absence of strong a priori hypotheses over the frequency range and channels 274 which tDCS may induce a change, we use a stepwise, cluster-based, non-parametric permutation test

276
For the cognitive task, we analysed the event-related spectral perturbation main effects of stimulation 277 (anodal, cathodal, sham) for each stimulus (congruent, incongruent) by applying the cluster-based 278 approach 30 . In order to reduce the possibility that the type II error rate was inflated by multiple 279 comparisons correction, we set an a priori criteria of collapsing data into four frequency bands: Theta