Edited by: Stephane Perrey, Université de Montpellier, France
Reviewed by: Concepción Padilla, University of Cambridge, United Kingdom; Keita Kamijo, Waseda University, Japan
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Acute exercise consistently benefits both emotion and cognition, particularly cognitive control. We evaluated acute endurance exercise influences on emotion, domain-general cognitive control and the cognitive control of emotion, specifically cognitive reappraisal. Thirty-six endurance runners, defined as running at least 30 miles per week with one weekly run of at least 9 miles (21 female, age 18–30 years) participated. In a repeated measures design, participants walked at 57% age-adjusted maximum heart rate (HRmax; range 51%–63%) and ran at 70% HRmax (range 64%–76%) for 90 min on two separate days. Participants completed measures of emotional state and the Stroop test of domain-general cognitive control before, every 30 min during and 30 min after exercise. Participants also completed a cognitive reappraisal task (CRT) after exercise. Functional near-infrared spectroscopy (fNIRS) tracked changes in oxygenated and deoxygenated hemoglobin (O2Hb and dHb) levels in the prefrontal cortex (PFC). Results suggest that even at relatively moderate intensities, endurance athletes benefit emotionally from running both during and after exercise and task-related PFC oxygenation reductions do not appear to hinder prefrontal-dependent cognitive control.
Aerobic exercise is thought to influence emotion and cognitive control in a dose-dependent manner (Dietrich,
Among the few studies to have evaluated endurance exercise effects on emotion and domain-general cognitive control, endurance exercise ranging from a 1 h to a 6-day run benefited multiple aspects of emotion, including lowering tension, anger and depression and elevating happiness and calmness (Markoff et al.,
Neuroimaging may help to clarify exercise-dependent changes in frontal regions of the brain implicated in cognitive control. Functional near-infrared spectroscopy (fNIRS) measures changes in oxygenated and deoxygenated hemoglobin (O2Hb and dHb) from superficial layers of cortex (Cui et al.,
Given the evidence that exercise influences domain-general cognitive control, as well as associated changes in PFC oxygenation, it follows that exercise may influence more specific aspects of cognitive control, such as the cognitive control of emotion, i.e., emotion regulation. Emotion regulation refers to cognitive processes that enable individuals to regulate their own emotions, through both conscious and non-conscious processes, by increasing or decreasing the experience of negative or positive emotions (Gross,
Given the increasing popularity of endurance running events (Douglas and Fuehrer,
Domain-general cognitive control was measured via the Stroop test of selective attention and response inhibition (Stroop,
A secondary objective was to assess whether endurance exercise influences emotion. Given that acute exercise tends to enhance positive emotion in physically fit individuals (Reed and Ones,
Thirty-six individuals (21 female, 15 male; age 18–30 years) participated for monetary compensation of $150 USD (see Table
Sample characteristics (
Average | SD | Minimum | Maximum | |
---|---|---|---|---|
Age | 23.4 | 3.6 | 18 | 30 |
Weekly running total distance | 36.9 | 9.3 | 30 | 65 |
Miles (kilometers) | (59.4) | (15.0) | (48.3) | (104.6) |
Weekly long run | 11.8 | 3.3 | 9 | 26 |
Miles (kilometers) | (19.0) | (5.3) | (14.5) | (41.8) |
BMI | 22.2 | 3.6 | 17.6 | 38.2 |
Godin total leisure time | 80.5 | 24.6 | 42 | 125 |
Beck depression inventory | 2.2 | 2.3 | 0 | 9 |
Perceived stress scale | 17.0 | 6.4 | 6 | 37 |
State-trait anxiety inventory—trait | 33.6 | 5.3 | 23 | 50 |
Reappraisal | 5.2 | 0.7 | 4.0 | 6.8 |
Suppression | 3.5 | 1.3 | 1.0 | 5.8 |
The experiment used a repeated measures design, with Exercise Intensity (Walk, Run) as the within-participants factor. Sample size estimation was based on effect sizes from Ando et al. (
The Godin Leisure Time Questionnaire (Godin and Shephard,
Distribution of Beck Depression Inventory (BDI), State-Trait Anxiety Inventory -Trait (STAI-Trait) and Perceived Stress Scale (PSS) scores (
<10 | 10–18 | 19–29 | 30–63 | |
BDI | ||||
20–30 | 30–40 | 40–50 | >51 | |
STAI | ||||
0–10 | 11–20 | 21–30 | 31–40 | |
PSS |
The feeling scale (FS) is a one-item inventory measuring the extent to which participants feel pleasant or unpleasant and ranges from “very good” (+5) to “very bad” (−5; Hardy and Rejeski,
The felt arousal scale (FAS) is a one-item inventory measuring feelings of arousal and ranges from “low arousal” (1) to “high arousal” (6; Svebak and Murgatroyd,
The rating of perceived exertion (RPE) is a commonly used one-item self-report measure of perceived physical exertion and ranges from “no exertion at all” (6) to “maximal exertion” (20; Borg,
The Stroop test (Stroop,
Dependent measures included accuracy and response time (on accurate trials only). “Stroop interference” was calculated by the difference in response time between incongruent and congruent trials, with lower scores reflecting enhanced performance.
The CRT involves viewing a series of negative pictures from the International Affective Picture System (IAPS; Lang et al.,
Participants completed an open-ended response questionnaire, including the two items, “What strategies did you use to maintain?” and “What strategies did you use to decrease?” Open-ended responses were coded according to Opitz et al. (
fNIRS was performed using the NIRSport (NIRx Medical Technologies, LLC, Glen Head, NY, USA). This continuous-wave fNIRS system consisted of eight light sources and eight detectors, with a 3 cm source-detector separation comprising 21 channels, all across the dorsal and anterior PFC (see Figure
Functional near-infrared spectroscopy (fNIRS) Probe Setup (red dots = sources, blue = detectors). Figure previously published in Giles et al. (
Participants completed three sessions: one practice session and two experimental sessions. During the practice session, interested participants first completed the consent and were screened for eligibility. If participants met inclusion and exclusion criteria described above, they completed the Godin Leisure Time Questionnaire, BDI, PSS, Emotion Regulation Questionnaire, and trait subscale of the State-Trait Anxiety Inventory. They then completed practice trials of the Stroop test while walking and running.
During each of the two experimental sessions, participants first donned the fNIRS and HR monitors (Polar model RS800CX) and completed a baseline set of FS and FAS (verbal report), and Stroop test (keyboard press). They warmed up by walking for 5 min at 2.5 miles per hour (MPH), i.e., 4.0 kilometers per hour (KPH). Participants then walked at 57% age-adjusted maximum HR (HRmax; range 51%–63%) or incrementally increased the speed until running at 70% HRmax (range 64%–76%), after which they adjusted their speeds to remain with the prescribed HR zone, for 90 min. The Walk and Run conditions were completed in counterbalanced order. They were chosen based on the American College of Sports Medicine’s classification of exercise intensity as light and moderate, respectively (Garber et al.,
Schematic representation of one study session. The experiment consisted of two experimental sessions, during which participants either Walked or Ran for 90 min. Abbreviations include the feeling scale (FS), felt arousal scale (FAS), rated perceived exertion scale (RPE) and cognitive reappraisal task (CRT).
Exercise sessions were separated by at least 1 week to minimize injury risk and practice effects on the tasks. To reduce diurnal variation in cognitive and physical performance, within-participant test sessions were scheduled for approximately the same time of day (± 1 h). To reduce the influence of hydration status on cognitive and physical performance, participants were asked to consume
Stroop test response time and accuracy were analyzed using repeated measures ANOVA with Intensity (Run, Walk), Time (Pre-Exercise, Minute 30, Minute 60, Minute 90, Post-Exercise) and Congruency (Congruent, Incongruent) as fixed factors. For Stroop results analyzed separately for before vs. after exercise and within exercise, see
FS, FAS and RPE data were analyzed using the aligned rank transform (ART) for nonparametric factorial data analysis (Wobbrock et al.,
An effect was deemed statistically significant if the probability of its occurrence by chance was
The NIRStar acquisition software (NIRx Medical Technologies, LLC, Glen Head, NY, USA) was used to record fNIRS data and to evaluate its signal-to-noise ratio. The nirsLAB data analysis package (NIRx Medical Technologies, LLC, Glen Head, NY, USA) was used for all subsequent calculations. Raw data for all channels were visually inspected, spike artifacts were removed, and faulty channels were removed from subsequent analyses (an average ± SD of 1.24 ± 1.85 channels per task iteration for the Stroop test, and 0.21 ± 0.63 for the CRT). All channels were band-pass filtered, with low cutoff frequency = 0.01 Hz and high cutoff frequency = 0.1 Hz. The modified Beer–Lambert law was used to compute estimates of changes in O2Hb, dHb and tHb (i.e., O2Hb + dHb) levels from the frequency-filtered data, using the 30–60 s time period before each task iteration as the baseline (Sassaroli and Fantini,
The Statistical Parametric Mapping (SPM) utilities incorporated into nirsLAB were used to determine event-related changes in O2Hb, dHb and tHb during the Stroop and CRT. SPM employs the general linear model (GLM) to identify O2Hb, dHb and tHb hemodynamic brain responses with reference to experimental factors. Level-1 analyses (SPM 1) assess differences on a within-participant basis and were used to generate parameter estimates (β weights) for each channel and factor, i.e., exercise Intensity, Time and Congruency for the Stroop test and exercise Intensity, picture Valence and regulation Instruction for the CRT.
Stroop test O2Hb, dHb and tHb β weights for each channel were analyzed using repeated measures ANOVAs with exercise Intensity (Run, Walk), Time (Pre-Exercise, Minute 30, Minute 60, Minute 90, Post-Exercise) and Congruency (Congruent, Incongruent) as within-participants factors. Similarly, cognitive reappraisal test β weights for each channel were analyzed using repeated measures ANOVAs with exercise Intensity (Run, Walk) and: (1) preparatory period: picture Valence (Negative, Neutral); or (2) regulation period: picture Valence/Instruction (Negative/Decrease, Negative/Maintain, Neutral/Maintain) as within-participants factors. Alpha levels were Bonferroni corrected for multiple comparisons (α = 0.05 divided by 21 channels resulted in α = 0.0024) (Kopton and Kenning,
Emotional valence was more positive in the Run than in the Walk condition,
Feeling scale (FS) and felt arousal scale (FAS) means, standard error of the means (SEM), medians, and interquartile ranges (ICQ) for each Exercise and Time (
Feeling scale | Felt arousal scale | ||||||||
---|---|---|---|---|---|---|---|---|---|
Mean | SEM | Median | IQR | Mean | SEM | Median | IQR | ||
10 Min | Run | 2.7 | 0.3 | 3 | 3 | 1.8 | 0.1 | 2 | 1 |
Pre-exercise | Walk | 2.4 | 0.3 | 3 | 3 | 1.8 | 0.2 | 1 | 2 |
Min 30 | Run | 3.1 | 0.3 | 3 | 2 | 2.8 | 0.1 | 3 | 1 |
Walk | 2.8 | 0.3 | 3 | 2 | 2.2 | 0.1 | 4 | 1 | |
Min 60 | Run | 3.3 | 0.3 | 3 | 2 | 3.0 | 0.2 | 3 | 2 |
Walk | 2.7 | 0.3 | 3 | 1 | 2.4 | 0.2 | 2 | 1 | |
Min 90 | Run | 3.4 | 0.4 | 3 | 2 | 3.3 | 0.2 | 3 | 1 |
Walk | 2.6 | 0.3 | 3 | 1 | 2.5 | 0.2 | 2 | 1 | |
30 Min | Run | 1.9 | 0.3 | 2 | 2 | 2.3 | 0.2 | 2 | 2 |
Post-exercise | Walk | 1.6 | 0.3 | 2 | 3 | 2.0 | 0.2 | 2 | 2 |
Emotional arousal was higher in the Run than in the Walk condition,
Perceived exertion was higher during the Run than Walk condition,
HR was higher during the Run than Walk condition,
Consistent with classic findings (Stroop,
Stroop response times (seconds) and accuracy means (SEM) for each Exercise and Time (
Run | Walk | ||||||||
---|---|---|---|---|---|---|---|---|---|
Congruent | Incongruent | Congruent | Incongruent | ||||||
Response time (s) | 10 Min pre-exercise | 1.16 | (0.03) | 1.17 | (0.03) | 1.14 | (0.03) | 1.16 | (0.03) |
30 Min | 1.10 | (0.03) | 1.12 | (0.03) | 1.13 | (0.03) | 1.17 | (0.03) | |
60 Min | 1.10 | (0.03) | 1.11 | (0.03) | 1.11 | (0.03) | 1.13 | (0.03) | |
90 Min | 1.07 | (0.03) | 1.10 | (0.03) | 1.11 | (0.03) | 1.12 | (0.03) | |
30 Min post-exercise | 1.13 | (0.03) | 1.15 | (0.03) | 1.18 | (0.03) | 1.20 | (0.03) | |
Accuracy | 10 Min pre-exercise | 0.98 | (0.01) | 0.98 | (0.01) | 0.97 | (0.01) | 0.98 | (0.00) |
30 Min | 0.98 | (0.01) | 0.99 | (0.00) | 0.98 | (0.01) | 0.98 | (0.01) | |
60 Min | 0.98 | (0.01) | 0.99 | (0.00) | 0.98 | (0.01) | 0.98 | (0.01) | |
90 Min | 0.99 | (0.01) | 0.99 | (0.01) | 0.97 | (0.01) | 0.99 | (0.01) | |
30 Min post-exercise | 0.99 | (0.01) | 0.98 | (0.01) | 0.98 | (0.01) | 0.97 | (0.01) |
We hypothesized that Stroop performance would be higher during the first hour of the Run than Walk, and then the trend would reverse. In partial support of our hypothesis, analysis of response times revealed an Intensity by Time interaction,
We hypothesized that Stroop-evoked O2Hb would increase more during the Run than the Walk during the first hour of exercise, after which point O2Hb would decrease during the Run and increase during the Walk. A main effect of exercise Intensity showed that O2Hb was lower in response to the Stroop test during the Run than Walk across channels 4, 9, 12 and 18 (
Channel 12 oxygenated hemoglobin (O2Hb), total hemoglobin (tHb) and deoxygenated hemoglobin (dHb) concentration (mean ± standard error of the mean (SEM), in μM) for each Exercise and Time during the Stroop test (
Analysis of ratings of unpleasant emotion suggested that the pictures induced negative emotion as intended and that the sample exhibited successful reappraisal, in that ratings of negative emotion were higher for Negative/Maintain trials (mean ± standard error of the means (SEM) = 5.16 ± 0.14) than for Negative/Decrease trials (mean ± SEM = 4.80 ± 0.14), which in turn was higher than for Neutral/Maintain trials (mean ± SEM = 1.32 ± 0.04),
Descriptive statistics from the open-ended portion of the post-experiment questionnaire are presented in Table
Uninstructed emotion regulation strategies within the cognitive reappraisal task means (SEM) for each Exercise (
Overall | Run | Walk | ||||
---|---|---|---|---|---|---|
CR present | 0.81 | (0.04) | 0.83 | (0.06) | 0.78 | (0.07) |
Non-CR present | 0.29 | (0.05) | 0.17 | (0.06) | 0.42 | (0.08) |
Multiple strategies | 0.11 | (0.04) | 0.03 | (0.03) | 0.19 | (0.07) |
To determine whether participants’ use of uninstructed emotion regulation strategies influenced the extent to which they rated the pictures as unpleasant, analysis of reappraisal success was repeated using only participants who employed cognitive reappraisal as the sole emotion regulation strategy (
Reappraisal success means (SEM) within participants who adhered to the cognitive reappraisal instruction only (
Cortical activation during the CRT was analyzed separately for the: (1) “preparatory period” consisting of the first 4 s of each trial in which participants viewed the Negative and Neutral pictures, i.e.,
Channel 4 O2Hb, tHb and dHb β weight means (SEM) during the CRT preparatory period (upon picture presentation
During the regulation period, a main effect of exercise picture Valence/Instruction showed that O2Hb was higher in response to Negative pictures, both when instructed to Decrease and Maintain, than Neutral pictures across channels 4 and 16 (
Channel 4 O2Hb, tHb and dHb β weight means (SEM) during the CRT regulation period (
The present experiment evaluated the influence of endurance exercise (i.e., Run relative to Walk) on emotion, domain-general cognitive control, the cognitive control of emotion, and associated changes in PFC oxygenation. Endurance exercise improved response times and reduced Stroop test-evoked PFC O2Hb relative to the Walk, with no concurrent change in selective attention. Further, the Run enhanced cognitive reappraisal adherence and success.
Participants completed measures of emotional state before, every 30 min during, and after the 90-min Run or Walk. Participants generally felt positive throughout both the Run and Walk and the Run augmented positive emotion and arousal more so than the Walk (see Table
Participants also completed the Stroop test of selective attention before, every 30 min during, and after the 90-min Run or Walk. Participants exhibited shorter response times during the run than during the walk (see Table
Stroop interference did not differ between the two exercise conditions. Thus, we did not support the second premise of the RAH model, in that exercise did not impair higher-order cognitive function. Despite the prevalence of the RAH model, there is little consensus on if and how exercise influences non-motor cognitive processes. While a number of studies have found that exercise influences cognitive control (e.g., Dietrich and Sparling,
PFC O2Hb and tHb increased in response to the Stroop test before exercise and decreased during exercise, while dHb did not change (see Figure
Participants completed the CRT following their 90-min run and walk, which involved cognitively reappraising and maintaining their emotional responses toward negative and neutral pictures. Participants better adhered to the instructed emotion regulation strategies following the run, as they tended to use more uninstructed strategies following the walk (see Table
Changes in PFC oxygenation during the CRT were limited to those driven by picture valence, with no effects of exercise intensity or emotion regulation instruction (see Figures
The present findings suggest that moderate-intensity endurance exercise enhances positive emotion and emotion regulation success using cognitive reappraisal. We did not find evidence of endurance exercise influence on cognitive control. However, four primary limitations curtail our ability to generalize the findings to exercise as a whole. First, participants’ average exertion ratings of “somewhat hard” during the run suggest that despite the relatively long duration, they did not reach an intensity akin to the ventilatory threshold or respiratory compensation threshold at which emotional responses and cognitive control would decline. Given that changes in emotion and cognitive control may occur more as a function of exercise intensity than duration (Kilpatrick et al.,
Second, participants in the present study were free from psychological disorders. However, depression, anxiety and stress influence cognitive control and the cognitive control of emotions (Castaneda et al.,
Third, the Stroop test in its present iteration was perhaps not sensitive to changes that may occur as a function of exercise. Previous studies have utilized more challenging versions of the Stroop test (Yanagisawa et al.,
Finally, near-surface blood flow (e.g., skin blood flow) has been shown to increase during exercise, and changes to skin blood flow may correlate with changes in cerebral O2Hb (Miyazawa et al.,
The present experiment suggests that endurance exercise akin to 90 min moderate-intensity running increases positive emotion during exercise, and the cognitive control of emotion using reappraisal after exercise. Further, although endurance exercise reduced PFC activation, domain-general cognitive control remained stable. Thus, even at relatively moderate intensities, endurance athletes benefit emotionally from running both during and after exercise, and task-related PFC oxygenation reductions do not appear to hinder prefrontal-dependent cognitive control. However, it remains unclear whether the same effects would persist at higher exercise intensities, akin to athletes’ high intensity training and race experience, or in response to more challenging cognitive control tasks. Thus, the findings add to the growing body of literature demonstrating complex relationships between endurance exercise, emotion, and emotion regulation, and future research should expand upon the conditions under which endurance exercise may enhance or impair emotion and cognitive control.
The datasets for this manuscript are not publicly available because the funding agency does not permit public release of data products. Requests to access the datasets should be directed to Grace Giles at
GG, ME, TB, HU, CM, HT and RK contributed to study concept. GG completed data preparation and wrote the first draft of the manuscript. GG, MD, TB, HU, HG and RB contributed to data analysis. All authors contributed to manuscript revision, read and approved the submitted version.
Drs. Barbour and Graber are affiliated with Photon Migration Technologies, Corp, which is the parent company of the manufacturer, NIRx Medical Technologies, LLC, of the NIRS device used in this study. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
This study was effected as part of the doctoral dissertation, “Exercise, Emotion, and Executive Control” at Tufts University (Giles et al.,
The Supplementary Material for this article can be found online at: