Benefit of human moderate running boosting mood and executive function coinciding with bilateral prefrontal activation

Running, compared to pedaling is a whole-body locomotive movement that may confer more mental health via strongly stimulating brains, although running impacts on mental health but their underlying brain mechanisms have yet to be determined; since almost the mechanistic studies have been done with pedaling. We thus aimed at determining the acute effect of a single bout of running at moderate-intensity, the most popular condition, on mood and executive function as well as their neural substrates in the prefrontal cortex (PFC). Twenty-six healthy participants completed both a 10-min running session on a treadmill at 50%\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\dot{\text{V}}\text{O}}_{{{\text{2peak}}}}$$\end{document}V˙O2peak and a resting control session in randomized order. Executive function was assessed using the Stroop interference time from the color-word matching Stroop task (CWST) and mood was assessed using the Two-Dimensional Mood Scale, before and after both sessions. Prefrontal hemodynamic changes while performing the CWST were investigated using functional near-infrared spectroscopy. Running resulted in significant enhanced arousal and pleasure level compared to control. Running also caused significant greater reduction of Stroop interference time and increase in Oxy-Hb signals in bilateral PFCs. Besides, we found a significant association among pleasure level, Stroop interference reaction time, and the left dorsolateral PFCs: important brain loci for inhibitory control and mood regulation. To our knowledge, an acute moderate-intensity running has the beneficial of inducing a positive mood and enhancing executive function coinciding with cortical activation in the prefrontal subregions involved in inhibitory control and mood regulation. These results together with previous findings with pedaling imply the specificity of moderate running benefits promoting both cognition and pleasant mood.


Discussion
Here we used fNIRS to test the hypothesis that acute running stimulates the PFC, which results in increasing positive mood and executive function. We found that a single bout of moderate-intensity running at 50%VO 2peak , the most popular running condition, enhanced mood not only through arousal but also through pleasure and also influenced executive function, as evidenced in the shortening of Stroop interference RT, while activating the cortical areas associated with cognition and mood regulation. These results clearly support the hypothesis that acute moderate running benefits mood and executive function coincident with cortical activation in the prefrontal subregions involved in mood regulation. Referred to our previous evidence that suggested single benefit of moderate pedaling exercise in enhancing cognition without reporting pleasant mood change basedon the TDMS, the current study could demonstrate dual benefits of moderate running exercise in increasing cognition and pleasure level, which may influence the brain in regulating stress, leading to mental health 4,5,43,46-48 . Running apparently enhanced arousal levels compared to a resting control, and this is consistent with our previous studies 4,5 . Moreover, we also revealed that 10 min of moderate-intensity running benefited pleasure levels, which also agrees with previous evidence reporting increased positive affection after a single bout of moderate-intensity running 49,50 . Based on these findings, running may be considered an exercise mode that benefits mood, which is an important factor for exercise adherence 51 . Another psychological response observed here is an observed RPE that was lower than the reference range 45 . Theoretically, RPE is a recognized marker of intensity and homeostatic disturbance during exercise as well as an effective way to monitor physical effort 52,53 . In regulating RPE, we propose that the prefrontal cortex may be involved in the interpretation of physiological signals, which are integrated between internal and external factors present in exercise environments to determine exercise tolerance 52,54 . Therefore, running seems to have the potential to influence prefrontal regions to diminish perceived physical effort (lower RPE) that presumably benefits exercise performance. However, this account remains somewhat speculative.
The behavioral response investigated using the CWST revealed longer RT with higher ER in the incongruent condition compared to the neutral condition, demonstrating that the Stroop interference effect could be observed across all sessions. The Stroop interference time was subsequently determined, and the results support our hypothesis that running shortened Stroop interference time compared to the resting control [3][4][5] . Moreover, significant coincidence between enhanced mood (increased arousal and pleasure levels) and shortened Stroop interference time was revealed. Together, these results allow us to postulate that a running-induced positive mood could enhance behavioral Stroop performance resulting in increased executive function, although the directionality of this relationship cannot be directly inferred from our data 34,37 .
The prefrontal cortex -the neural substrate underpinning the performance of the CWST -revealed significant increases of Oxy-Hb in response to Stroop interference for all ROIs, showing that the Stroop interference effect could be observed as in our previous studies 4,5 . The effect of running on prefrontal activation was subsequently determined, and the results showed that running led to significant increases in activation in all ROIs except the l-VLPFC. The l-DLPFC, which plays an important role in the implementation of cognitive control, had increase of Oxy-Hb response to Stroop interference, which corresponds to the results of previous studies including our own 4,5,30,55,56 . Besides, the l-DLPFC was reported to implicate in mood regulation in mood disorders, demonstrating significant coincidence between running-induced positive mood and l-DLPFC activation as we hypothesized [19][20][21] . The right DLPFC, which is involved in conflict resolution (faster reaction time after incongruent trials) and Stroop accuracy, was also showed significant activation 31,57,58 . The right-VLPFC, which is associated with mood and emotion regulation, was observed to have significant activation coinciding with increased arousal and pleasure levels [19][20][21]59 . In addition, the FPA, a brain region which generally activates with the DLPFC to response to tasks involving manipulation and monitoring, such as planning for action, was observed to have significant both sides activation 60 . It is considered that running-induced bilateral prefrontal activation, differing from the previous studies using pedaling exercise that revealed prominent l-DLPFC activation (Fig. 5) [3][4][5]43,61 . Taken together, these results support our hypothesis that an acute bout of moderate-intensity running elicits mood improvement and enhances executive function coinciding with prefrontal subregion activations involved in mood regulation.
The neural mechanisms for running-elicited cortical activation have remained unclear. Running requires topdown feedforward control responses to multi-modal sensory information that partially involve the prefrontal cortex to execute coordinated movement and balance [13][14][15]17,18 . Specific features of running, such as foot strike, may benefit the brain activation by enhancing blood flow velocity in the middle cerebral artery 24 . However, the mechanism of this finding still unclear. Besides, a mechanical force from vertical head acceleration, may induce serotonin receptor internalization in the prefrontal cortex, influencing emotion and cognitive control based-on animal study 25 . With respect to the theoretical concept of acute exercise-enhanced cognition, it is considered that running has the potential to induce neural activation by activating the RAS to enhance arousal levels, and by projecting neuronal signals to the prefrontal cortex 32,33,[37][38][39] . This neuronal response is mostly modulated by three neurotransmitter systems: the noradrenergic, dopaminergic and serotoninergic systems, which originate from the brainstem and project profusely throughout the forebrain, playing a role in cognition and mood regulation [34][35][36][37][38][39] .
That insufficient physical activity leads to physical and mental illness has been recognized as a global issue. It is important to demonstrate a minimal effective exercise that will benefit both mental and physical health. We have found that even a 10-min single bout of moderate-intensity running can boost both mood and executive function; in particular, broad prefrontal activation as well as enhanced pleasure levels are important for exercise www.nature.com/scientificreports/ adherence. These fruitful findings are meaningful in terms of health promotion because running is considered a practical exercise and is one of the most popular recreation activities worldwide.
Here we investigated only the effect of running, but not a pedaling because of the following methodological limitations. First, to gain statistical robustness through the process of neuroimaging analysis using fNIRS, we must limit session numbers. Second, we have already found the exact findings on pedaling using the same experimental protocol for ensuring the specificity of running effects 4,5,30 . Indeed, we successfully revealed the dual effects of running boosting mood and executive function, different from those pedaling (Fig. 5).

Conclusions
The current study reveals that a 10-min single-bout of moderate-intensity running elicits a positive mood and increased executive function by enhancing arousal levels coincidentally with activation in prefrontal subregions involved in mood regulation. We provide evidence for the neural substrates behind mood enhancement and increased executive function induced by acute moderate-intensity running. To this end, these finding are valuable in supporting moderate running effect on mental health since running is an easily accessible form of exercise requiring minimal equipment and sport structure. This should shed light on the specificity of running among a varieties of exercise prescriptions promoting mental health.

Materials and methods
Participants. Twenty-six young adults participated in this study. All participants were Japanese native speakers, right-hand dominant, with normal, or corrected-to-normal vision and normal color vision. No participant reported a history of neurological or psychiatric disorders, or had a disease requiring medical care. The characteristics of participants are presented in Table 2. This study was approved by the Institutional Ethics Com-  Experimental procedure. The experimental procedure consisted of three major steps. First, we measured maximal oxygen uptake ( V O 2peak ) to determine the appropriate individual intensity of moderate exercise, which is 50%VO 2peak based on the definition of exercise intensity of the American College of Sports Medicine 45 . Second, we investigated non-cortical physiological signals induced by running in 6 participants, such as skin blood flow, to specify proper fNIRS measurement times after running in order to eliminate possible contamination of the fNIRS signal 62 . We found that all physiological variables fully returned to the baseline within 15 min. Therefore, 15 min after running, we started fNIRS measurements to avoid possible signal contamination ( Fig. 6a  and b). Details of the two preliminary steps are provided in Supplementary C. Finally, we determined the effect of a single bout of moderate-intensity running on mood and executive function. Two sessions, control (CON) and running (RUN), were randomly conducted with a counterbalance measure design on different days. In the RUN session, fNIRS was used to measure prefrontal hemodynamic changes while performing the CWST before and 15 min after running to avoid non-cortical physiological signal contamination. Participants ran for 10 min on a treadmill at a personalized speed. Mood state was assessed before and after running. In the CON session, participants rested for 10 min instead of running (Fig. 6c).
Psychological measures. Psychological mood state was measured before and after running using the Two-Dimensional Mood Scale (TDMS) 63 . The TDMS is a momentary mood scale that consists of eight words describing arousal and pleasure states (energetic, lively, lethargic, listless, relaxed, calm, irritated and nervous). The participants were asked to indicate how they were feeling for each mood-expressing word using a 6-point Likert scale that ranged from 0 = Not at all to 5 = Extremely. Then, arousal and pleasure levels were calculated.
Behavioral measures. Executive function was investigated using an event-related version of the CWST 3,4,43,64 . The test appeared on the screen with 2 rows of letters (Fig. 7a). The participants were instructed to decide whether the color of the letters in upper row corresponded to the color name in the lower row. They were also asked to place their index fingers on "yes" and "no" buttons and respond to the test by pressing the appropri- www.nature.com/scientificreports/ ate button as fast as they could. Subsequently, reaction time (RT) and error rate (ER) were calculated. The CWST consisted of three conditions: neutral, congruent and incongruent. For the neutral condition, the upper row was presented as a row of X's (XXXX) printed in red, green, blue or yellow, and the lower row presented the word 'RED' , 'GREEN' , 'BLUE' or 'YELLOW' printed in black. For the congruent condition, the upper row contained the word 'RED' , 'GREEN' , 'BLUE' or 'YELLOW' printed in the congruent color (e.g., 'RED' was printed in red), and the lower row presented the same words as in the lower row of the neutral condition. For the incongruent condition, the upper row presented the color word printed in an incongruent color to produce interference between the color word and the color name (e.g., 'RED' was printed in green), and the lower row presented the same words as in the lower row of the neutral and congruent conditions (Fig. 7b). Each experimental session consisted of 30 trials, made up of 10 neutral, 10 congruent and 10 incongruent trials, which appeared in random order. The upper row was presented 100 ms before the lower row in order to shift visual attention. The trial remained on the screen until the participant responded or for 2 s, whichever was shorter. Then, a fixation cross appeared on the screen as an inter-stimulus interval for 10-12 s to avoid prediction of the timing of the subsequent trial. Stroop interference, an index of executive function in the prefrontal cortex, was calculated as the difference in reaction time between the incongruent and neutral conditions 30 . All words were written in Japanese. The participants performed three practice sessions to ensure that they understood and were familiarized with the CWST well before starting the experiment.
fNIRS data acquisition. Cortical hemodynamic changes were monitored using the multichannel fNIRS optical topography system ETG-7000 (Hitachi Medical Corporation, Japan) and adopting two wavelengths of near-infrared light (785 and 830 nm). The optical data from fNIRS was analyzed based on the modified Beer-Lambert Law as previously described 65,66 . With this method, we were able to calculate signals reflecting the oxygenated hemoglobin (oxy-Hb) concentration changes in a unit of millimolar-millimeter (mM•mm) 66 . The sampling rate was set at 10 Hz. For fNIRS probe placement, two sets of 4 × 4 multichannel probe holders, which consist of 8 illuminative and 8 detective probes arranged alternately at an inter-probe distance of 3 cm resulting in 24 channels (CH) per set, were placed to cover both lateral prefrontal activation foci as in previous studies 3,4 . The left probe holder was placed such that probe 5 (between CH4 and CH11) was placed over FT7, with the www.nature.com/scientificreports/ medial edge of the probe column parallel to the medial line. Likewise, the right probe holder was symmetrically placed on the left hemisphere (Fig. 7c). We adopted a virtual registration method to register fNIRS data to Montreal Neurological Institute (MNI) standard brain space 67,68 . With this method, we were able to place a virtual probe holder on the scalp by stimulating the holder's deformation and by registering probes and channels onto a reference brain in the magnetic resonance image (MRI) database 69,70 . We probabilistically estimated the MNI coordinate values for fNIRS in order to obtain the most likely microanatomical prediction for locations of the given channels as well as the spatial variability of the estimation 71 . Finally, a MATLAB function was used to label the estimated locations in a macro-anatomical brain atlas 72 .
Analysis of fNIRS data. Prefrontal Oxy-Hb changes that occurred while performing the CWST were calculated as shown in our previous studies 3,4,[40][41][42][43] . Individual timeline data for each channel were preprocessed with a band-pass filter using a high-pass filter (0.04 Hz) to remove baseline drift and a low-pass filter (0.30 Hz) to screen out heartbeat pulsations. Then, channel-wise and subject-wise contrasts were calculated by the intertrial mean of differences between peak (4-11 s after trial onset) and baseline (0-2 s before trial onset) periods. The contrasts obtained were subsequently subjected to a second level of random effects group analysis. This study adopted LBPA40, a widely used method among anatomical labeling systems, to combine 3-4 neighboring channels to form each region of interest (ROI) 72 . The regions included the DLPFC (l-DLPFC: channels 13, 14, 16 and 17; r-DLPFC: channels 35, 36, 39 and 40), the VLPFC (l-VLPFC: channels 2, 6 and 9; r-VLPFC: channels 26, 29 and 33) and the FPA (l-FPA: channels 3, 7 and 10; r-FPA: channels 25, 28 and 32). LBPA40 is considered valid because optical properties of neighboring channels are known to be similar 73 . However, with this method, optical properties in different ROIs can cause systematic bias during statistical analysis. Therefore, we limited analyses to ROI-wise and used a false discovery rate (FDR) to control the low proportion of false positives 74 .
Statistical analysis. Psychological mood state was subjected to repeated-measures two-way ANOVA with session (CON, RUN) and time (pre, post) as within-subject factors. Bonferroni's post hoc test was subsequently processed after a significant F value was observed. Regarding behavioral Stroop performance, RT and ER were first subjected to repeated-measures three-way ANOVA with condition (neutral, incongruent), session (CON, RUN) and time (pre, post) as within-subject factors to examine whether the general tendencies for the Stroop task could be reproduced in all conditions. Then, the effect of running on Stroop task performance was analyzed by repeated-measures two-way ANOVA with session (CON, RUN) and time (pre, post) as within-subject factors followed by Bonferroni's post hoc test. For the fNIRS data, pre-sessions of cortical activation for CON and RUN sessions, which were free from any effect of running, were first examined to determine whether the Stroop interference effect (the difference in oxy-Hb between incongruent and neutral conditions) could be observed. The Stroop interference levels for the CON and RUN sessions were averaged and served as substrates for a ROI-wise analysis with FDR for controlling the low proportion of false positives. Only significant ROIs for Stroop interference were subsequently analyzed for the effect of running on prefrontal activation using repeated-measures twoway ANOVA with session (CON, RUN) and time (pre, post) as within-subject factors followed by FDR correction. This study adopted the McNemar test to assess the association among running-influenced mood change, running-effected behavioral Stroop performance and running-elicited cortical activation in a binomial manner.

Data availability
All data that support the findings of this study are available from the corresponding author by request with no restrictions.