Journal Logo


Spontaneous Eye Blink Rate Connects Missing Link between Aerobic Fitness and Cognition


Author Information
Medicine & Science in Sports & Exercise: July 2021 - Volume 53 - Issue 7 - p 1425-1433
doi: 10.1249/MSS.0000000000002590


The greater the habitual physical activity, the higher the aerobic fitness. A physically active lifestyle, as indicated by aerobic fitness, improves cognitive and mental health across the life span. Many studies, including our own, have shown that aerobic fitness is positively associated with prefrontal executive function via frontal functional and/or structural change (1–4). The neuromodulatory mechanisms underlying this linkage between aerobic fitness and executive function, however, remain unrevealed.

The brain dopaminergic system is a potential candidate for the neuromodulatory mechanisms underlying the missing link between aerobic fitness and executive function because dopaminergic function is associated not only with prefrontal executive function (5,6) but also with motivated behavior including physical activity (7,8). In addition, the dopaminergic function is functionally changed by habitual physical activity (9–12). Accumulating evidence reveals that the brain dopaminergic system originating in the ventral tegmental area and substantia nigra is linked with executive functions via function mainly in the prefrontal cortex (PFC) and striatum, which is to say the frontostriatal network (5,6,13). Recent rodent studies have shown that dopaminergic activity facilitates voluntary physical activity (7,8) and voluntary physical activity increases dopamine concentration in the PFC (10), which indicates that the dopaminergic function underlies the neural basis of greater physical activity and fitness. Some of the latest human studies using positron emission tomography (PET) support the findings in animal studies. Cross-sectional studies have suggested that between-person differences in habitual physical activities and aerobic fitness were positively correlated with striatal D2-like receptor availability in older adults (11,14); however, it remains unclear whether dopamine mediates the association between physical activity, as indicated by aerobic fitness and prefrontal executive function, probably because of the methodological difficulties.

Spontaneous eye blink rate (sEBR) could be helpful in addressing the link between dopamine, fitness, and executive function since over 30 yr of studies have shown that sEBR can be used as a noninvasive brain dopaminergic system indicator (15–18). Several animal and human pharmacological studies have revealed that dopamine agonists increase sEBR whereas dopamine antagonists decrease sEBR through the involvement of basal ganglia and the spinal trigeminal complex (18–21). An increase in sEBR induced by amphetamine correlates with the change in the levels of endogenous dopamine in the human striatum as measured by PET (22). Moreover, sEBR is altered in clinical conditions associated with dopaminergic system dysfunctions (e.g., Parkinson’s disease, schizophrenia, attention deficit/hyperactivity disorder [ADHD], etc.) (15,18,20), and dopamine precursor administration partially restores lowered sEBR in Parkinson’s disease (20). In addition, there is much evidence that shows the functional relationship between sEBR and dopamine-related cognitive function, including prefrontal executive function (19,23–25). On the other hand, a single bout of vigorous aerobic exercise has been shown to increase sEBR in adolescent boys with ADHD (26). Therefore, sEBR enables the examination of the mediation effect of dopamine on the effect of physical activity on prefrontal executive function, in noninvasive and stress-free experiments.

Functional near-infrared spectroscopy (fNIRS), a noninvasive optical neuroimaging tool, has often been applied to explore the possible involvement of the PFC in the association between aerobic fitness and executive function. In our study, we used fNIRS combined with the color–word Stroop task (CWST) to examine the effects of aerobic fitness and acute exercise on inhibitory control of a core component of executive function (1,27–31) and found that the left dorsolateral PFC (l-DLPFC), one of the essential subregions working with CWST performance (32,33), was active as a neuronal substrate by which acute exercise improves CWST performance in young adults (27–30). These results together with further evidence showing that dopamine regulates task-related neural efficiency (NE) via increased signal-to-noise ratio in task-relevant brain regions (34,35) lead us to postulate that NE in the l-DLPFC is involved in the association between aerobic fitness and CWST performance.

Here we sought to examine the hypothesis that sEBR mediates the association between aerobic fitness and CWST performance. Moreover, we assessed NE in the l-DLPFC using fNIRS to explore the involvement of efficient prefrontal activation with the mediation effect of sEBR.



Thirty-five healthy, young, Japanese males (age, 18–24 yr old) participated in this study. All participants were right-handed and nonsmokers. The choice to include male participants only was based on previous reports indicating sex differences for sEBR (15). All participants were psychiatrically normal (screened using the Beck Depression Inventory II, score less than 20). No participant reported a history of respiratory disease, circulatory disease, or neurological disease or had an illness requiring medical care. All participants had normal or corrected-to-normal vision and normal color vision. Written informed consent was obtained from all participants. This study was approved by the Institutional Ethics Committee of the University of Tsukuba and was in accordance with the latest version of the Declaration of Helsinki. Table 1 depicts the demographic data of participants.

TABLE 1 - Participant demographics data.
Measure Mean ± SD
Sample size (n) 35
Age (yr) 20.9 ± 1.8
Height (cm) 173.2 ± 4.5
Weight (kg) 64.3 ± 8.2
BMI (kg·m−2) 21.4 ± 2.6
Graded exercise test
 V˙O2peak (mL·kg−1⋅min−1) 44.7 ± 7.0
 HRpeak (bpm) 179.5 ± 9.3
 WRpeak (W) 240.7 ± 34.4
 RPEpeak (score) 19.5 ± 0.9
sEBR (per min) 20.5 ± 8.7
CWST performance
 Neutral accuracy (%) 97.7 ± 4.9
 Incongruent accuracy (%) 94.6 ± 7.8
 Stroop interference accuracy (%) −3.1 ± 8.0
 Neutral RT (ms) 713.7 ± 136.8
 Incongruent RT (ms) 883.0 ± 192.4
 Stroop interference RT (ms) 169.2 ± 94.9
 Neutral IES 730.3 ± 132.8
 Incongruent IES 935.5 ± 206.4
 Stroop interference IES 205.1 ± 116.6
BDI II (score) 7.5 ± 4.8
Values are presented as mean ± SD.
BDI, Beck Depression Inventory; BMI, body mass index; CWST, color-word Stroop task; HR, heart rate; IES, inverse efficiency score; RPE, ratings of perceived exertion; RT, reaction time; sEBR, spontaneous eye blink rate; V˙O2peak, peak oxygen uptake; WR, work rate.

Experimental procedure

For each participant, testing occurred on three separate days. On the first day, they completed demographic/health questionnaires and performed a graded exercise test to assess individual aerobic fitness levels. On the second day, their sEBR was recorded. On the third day, at least 48 h after the fitness assessment, all participants performed the CWST. On the test day, they refrained from any exercise and the consumption of caffeine and alcohol before the experiment to control for outside factors that could affect cardiovascular and neurocognitive function.

Cardiorespiratory aerobic fitness assessment

The individual aerobic fitness levels were determined using a graded exercise test with a recumbent ergometer (Corival Recumbent, Lode, The Netherlands). Peak oxygen uptake (V˙O2peak), the gold standard measurement of aerobic fitness, was determined by continuously measuring oxygen uptake during an incremental test to exhaustion. After warming up for 3 min at 30 W, the workload increased by 20 W·min−1 constantly and continuously until the maximal effort was reached. The pedal rotation speed was kept at 60 rpm. Exhaled gas was analyzed using a gas analyzer (Aeromonitor AE-310S; Minato Medical Science, Osaka, Japan). Heart rate (HR) was measured throughout the assessment. Participants were asked to indicate their subjective feeling about the exercise intensity using the Borg RPE (15-point scale: 6, no exertion; 20, maximal exertion). All participants exercised until they could no longer maintain a pedal rotation speed of 60 rpm. V˙O2peak was determined when at least two of the following criteria were satisfied: 1) the respiratory exchange ratio (R) exceeded 1.10, 2) achievement of 90% of age-predicted peak HR (220 − age), and 3) an RPE of 18 or more (36,37).


An event-related version of the CWST was used as in our previous studies (1,27–31). Two rows of letters were presented on the screen, and participants were asked to respond to whether the color of the letters in the top row corresponded to the color name shown in the bottom row (Fig. 1). Participants pressed either “yes” or “no” buttons with their right or left forefingers to respond. Each experimental session consisted of 30 trials, including 10 neutral, 10 congruent, and 10 incongruent trials presented in random order. For the neutral condition, the top row contained groups of X’s (XXXX) printed in blue, yellow, red, or green and the bottom row contained the words BLUE, YELLOW, RED, or GREEN printed in black. For the congruent condition, the top row contained the words BLUE, YELLOW, RED, or GREEN printed in a color congruent with that of the bottom row. For incongruent conditions, the color word in the top row was printed in an incongruent color to produce interference between the color word and the color name. The correct answer ratio assigned to yes and no was 50%. The bottom row was presented 100 ms after the top row to achieve sequential visual attention (27). Each trial was separated by an interstimulus interval showing a fixation cross for 9–13 s to avoid prediction of the timing of the following trial (27). The stimulus was presented on the screen for 2 s. All words were written in Japanese. Mean reaction time (RT) and accuracy rate were recorded for each condition, and then the RT was adjusted for accuracy (inverse efficiency score (IES) = RT [ms]/accuracy [%] × 100). This combined value of RT and accuracy rate was used to maximize the power of analyses by reducing individual differences in the speed–accuracy trade-off as in several recent neuroimaging studies (38,39). In addition, the single index was suitable for calculating the NE score, which will be explained below. The IES difference between incongruent and neutral conditions (IES of incongruent condition – IES of neutral condition) was used as a measure of Stroop interference. Cortical hemodynamic changes in the l-DLPFC were monitored with fNIRS while participants performed the CWST.

Examples of the Japanese-version CWST. Stimuli of single trials for the neutral, congruent, and incongruent trials of the CWST are presented. Translations into English are denoted in parentheses.

sEBR measurements

Participants sat in front of an A4 size white poster, located 70 cm from the participant. The participants were asked to look at a fixation cross on the center of the poster at rest. sEBR (per minute) was recorded for 5 min using a camcorder (120 frames per second, 2560 × 1440; Hero 7; GoPro, Inc., San Mateo, CA) set below the poster. sEBR was independently assessed from the recordings by two researchers (25), and the average of their two scores was adopted. Correlations between researchers exceeded r = 0.99. We previously confirmed the strong correlation with a measurement using a vertical electrooculogram (VEOG) recording method (n = 7, r > 0.99) (see Document, Supplemental Digital Content 1, Spontaneous eye blink rate counted by vertical electro-oculogram, The individual sEBR was calculated by dividing the total number of eye blinks during the 5-min measurement interval by 5. All sEBR data were collected by 6:00 pm because sEBR can be less stable at night (15).

fNIRS measurements

We used a multichannel fNIRS optical topography system (ETG-7000; Hitachi Medical Corporation, Tokyo, Japan) set with two wavelengths of near-infrared light (785 and 830 nm). The composition of the fNIRS probe using two sets of 4 × 4 multichannel probe holders and their placement followed the same procedure as in our previous studies (1,27–31). We adopted a virtual registration method to register fNIRS data to Montreal Neurological Institute (MNI) standard brain space (40). With this method, a virtual probe holder is placed on the scalp: the holder’s deformation is simulated, and thus probes and channels can be registered onto reference brains in the magnetic resonance imaging (MRI) database (41). A statistical analysis of the MNI coordinate values for the fNIRS channels enabled us to produce the most precise estimates of the locations of given channels for the group of participants and the spatial variability associated with the estimations (42). Lastly, the estimated locations were given anatomical labels using a MATLAB function that reads anatomical labeling information coded in a macroanatomical brain atlas (43).

fNIRS data analysis

In this study, we focused on the l-DLPFC as region of interest (ROI) because previous neuroimaging studies and electrical brain stimulation studies identified that the l-DLPFC is one of the vital regions for executive function (13), especially for CWST performance (30,32,33).

Individual timeline data for each channel’s oxygenated hemoglobin (oxy-Hb) signal were preprocessed using a band-pass filter (high-pass filter, 0.04 Hz; low-pass filter, 0.3 Hz). Channelwise and subjectwise contrasts were obtained from the preprocessed time series data by calculating the intertrial mean of differences between the oxy-Hb signals of the baseline (0–2 s before trial onset) and the peak (4–11 s after trial onset) periods (Fig. 2A). Based on a method widely used in anatomical labeling systems, LBPA40 (43), four adjacent channels were combined to form the l-DLPFC region (channels 13, 14, 16, and 17) (Fig. 2B).

A, Cortical activation patterns during the color–word matching Stroop task (CWST) in the current study. This graph depicts the timelines for oxy-Hb signals in response to the neutral trial (broken line) and the incongruent trial (solid line) in the left dorsolateral prefrontal cortex (l-DLPFC). B, The spatial profiles of fNIRS channels and ROI segmentation used in the current study; which were introduced from the previous studies (40,41). Channel numbers and FT7 and FT8 in the international 10–20 EEG standard positions are denoted above the corresponding locations. The channels enclosed by the black broken lines were defined as the l-DLPFC, and their data were integrated for further analyses.

Calculation for NE

Curtin et al. (44,45) proposed a novel metric of NE, which is a combination of task-related cortical activation measures and cognitive performance measures. According to this metric, high NE is achieved when cognitive performance is high and the corresponding cortical activation is low, and vice versa. The NE score was calculated using a composite score from the task performance (i.e., Stroop interference IES) and cortical activation (i.e., the change oxy-Hb measured using fNIRS) (44,45). Cortical activation and Stroop interference IES metrics were converted into z-scored measurements according to group, then NE was calculated using the distance of the point from the zero-efficiency line and assessed as a measurement (i.e., where normalized Stroop task performance = normalized l-DLPFC activation). The equations below depict the NE calculation (see equations [1] [2]). With respect to Stroop task performance, we reversed the positive and negative because a lower Stroop interference represents better performance.

NE=cognitive performancecortical activation/2
NE=ZStroop task performanceZoxyHbchange inlDLPFC/2

Statistical analysis

First, correlation analyses were performed to examine the associations between aerobic fitness (V˙O2peak), sEBR, Stroop interference IES, and l-DLPFC activation controlling for potential confounding variables (age and body mass index [BMI]) (2,25). Then we applied a mediation analyses to examine the mediation effect of sEBR (mediator variable) on the relationship between V˙O2peak (independent variable) and executive function (dependent variable) by using a multiple regression approach (46) and the nonparametric bootstrapping procedure, as in previous studies (1). The confounding variables (i.e., age and BMI) were controlled as covariates. According to the moderation model by Baron and Kenny (46), analysis should be conducted in four steps to test the mediation effect. In this study, the following analyses were performed. We tested 1) whether V˙O2peak (independent variable) was associated with Stroop interference IES (dependent variable), 2) whether V˙O2peak (independent variable) was associated with sEBR (mediator variable), 3) whether sEBR (mediator variable) was associated with Stroop interference IES (dependent variable), and 4) whether the prospective mediation effect was significant when the association between V˙O2peak and Stroop interference IES became significantly weaker (partial mediation) or insignificant (full mediation) after the inclusion of sEBR. To examine the significance of the mediation effect, we used the bootstrapping method recommended for relatively small sample sizes because bootstrapping is a nonparametric test that does not require the assumption of normality. For the bootstrapping method, thousands of samples are taken from a given data set, and the indirect effects for each resample are estimated. These estimations are used to examine mediation directly. Indirect effects can be estimated by subtracting the independent variable’s direct effects on the dependent variable after controlling for the role of the proposed mediators from the total effect of the independent variable on the dependent variable without controlling for the proposed mediators. The macro then uses a bootstrapping procedure in which the data are resampled 5000 times and asymmetric confidence intervals (CI) are generated to examine the significance of the indirect effect. The indirect effect of an independent variable on a dependent variable through a mediator variable is significant if the CI values do not overlap with zero. Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 26 (SPSS Inc., Chicago, IL). As for mediation analysis, we used the PROCESS macro for SPSS (Hayes model 4). The sample size was determined from the effect size of multiple regression analysis in mediation analyses. From our previous studies, a large effect size (f2 > 0.35) was predicted in this study (1). According to Green (47), a sample size of 35 is required in the case of four predictor variables (i.e., two explanatory variables and two control variables). Statistical significance was set a priori at P < 0.05.


Verification of CWST performance and l-DLPFC activation related to Stroop interference

We examined whether the Stroop interference effect (i.e., incongruent vs neutral) could be observed in this study. For task performance, paired t-test showed significant Stroop interference in accuracy rate (t34 = 2.34, P < 0.05), RT (t34 = 10.55, P < 0.001), and IES (t34 = 10.41, P < 0.001). Next, we examined whether the l-DLPFC was activated by Stroop interference. For cortical activation, the increased oxy-Hb change levels in the l-DLPFC as Stroop interference were not significant (one-sample t-test, t34 = 1.75, P = 0.09), but a positive effect (Cohen’s d = 0.30) comparable with previous studies (1,27–31) was observed. These results support that there was a Stroop interference effect on both behavioral performance and l-DLPFC activation in this study.

The association between aerobic fitness, sEBR, and CWST

Partial correlation analyses controlling for age and BMI revealed significant correlations of higher V˙O2peak and higher sEBR to less Stroop interference IES (pr = −0.35, P = 0.04; pr = −0.46, P = 0.007, respectively) (Fig. 3A, B). In addition, V˙O2peak was positively correlated with sEBR (pr = 0.38, P = 0.03, adjusted for age and BMI) (Fig. 3C). Next, we conducted the mediation analysis. Figure 4 and Table 2 show results of mediation analyses with age and BMI as covariates. Higher V˙O2peak was significantly associated with less Stroop interference IES (Fig. 4, path c) and higher sEBR (Fig. 4, path a). When sEBR was added as the predictor, it was negatively significantly associated with Stroop interference IES (Fig. 4, path b), and the significant association between V˙O2peak and Stroop interference IES diminished and became not significant (Fig. 4, path c′). This mediation effect was further examined using nonparametric bootstrapping procedures. The indirect effect (path a*b) was −0.131, and the 95% bootstrap CI of the indirect effect (path a*b) did not contain zero (95% CI = −0.296 to −0.004), indicating significant mediation. These results suggest that the relationship between aerobic fitness levels and Stroop interference IES is mediated by sEBR.

A, Relationship between aerobic fitness (V˙O2peak) and Stroop interference IES. B, Relationship between sEBR and Stroop interference IES. C, Relationship between aerobic fitness (V˙O2peak) and sEBR.
Mediation model. sEBR as a mediator of the effect of aerobic fitness (V˙O2peak) on Stroop interference. Path a, effect of V˙O2peak on sEBR; path b, effect of mediator on sEBR; path c, total effect of V˙O2peak on Stroop interference; path c′, direct effect of V˙O2peak on Stroop interference through a mediator. Age and BMI were entered as covariates for all paths. β indicates standardized regression coefficient. *P < 0.05.
TABLE 2 - Results of the mediation analysis.
R 2 B SE B β t Value ΔF
IV to MV (Path a)
 Model 0.189
 V˙O2peak 0.468 0.205* 0.374 2.278
 Age −1.448 0.831 −0.294 −1.742
 BMI 0.516 0.566 0.152 0.913
Total effect of IV on DV (path c)
 Model 0.315**
 V˙O2peak −5.337 2.533* −0.318 −2.107
 Age 27.890 10.258* 0.422 2.719
 BMI −18.935 6.981* −0.416 −2.712
Direct effect of MV on DV (path b) and IV on DV (path c′)
 Model 0.414** 0.099* 5.077
 V˙O2peak −3.142 2.573 0.187 −1.221
 sEBR −4.694 2.084* −0.350 −2.253
 Age 21.090 10.105* 0.319 2.087
 BMI −16.511 6.651* −0.363 −2.483
*P < 0.05.
**P < 0.01.
R2, coefficient of determination; B, unstandardized regression coefficient; SE B, standard error.

Association with NE of l-DLPFC

In addition, we tested the association with NE of the l-DLPFC. First, a significant correlation between l-DLPFC activation and Stroop interreference IES was confirmed (pr = −0.37, P = 0.04, adjusted for age and BMI). Next, to test the association with NE, the NE score was calculated with l-DLPFC activation and Stroop interference IES. Figure 5A shows the results of the efficiency graph for the relationship between l-DLPFC activation and Stroop task performance. sEBR was positively correlated with NE (pr = 0.36, P = 0.04, adjusted for age and BMI) (Fig. 5B), but V˙O2peak was not (pr = 0.10, P = 0.59, adjusted for age and BMI).

A, Efficiency graph for normalized l-DLPFC activation vs normalized Stroop task performance. B, Relationship between sEBR and NE. NE = (normalized Stroop task performance − normalized l-DLPFC activation)/


Although aerobic fitness is associated with prefrontal executive function, the exact neuromodulatory mechanisms behind the connection remain uncovered. Here we aimed at examining the hypothesis that sEBR, a possible marker reflecting the brain’s dopaminergic system, explains the association between aerobic fitness and prefrontal executive function. The mediation analyses revealed that the association between higher aerobic fitness and higher CWST performance was mediated by higher sEBR. In addition, sEBR was positively associated with the NE of the l-DLPFC. These results suggest that higher aerobic fitness has a provocative link with higher prefrontal executive function mediated in part by brain dopaminergic regulation.

As expected, we first found significant correlations among higher aerobic fitness, less Stroop interference, and higher sEBR (Fig. 3). This led us to perform a mediation analysis to determine whether sEBR, as a noninvasive and reliable marker for dopaminergic function in the recent human neuroscience research (15–17,19), mediates the relationship between higher V˙O2peak and less Stroop interference IES (higher executive function) (Fig. 4). So far, several studies have shown that higher aerobic fitness is correlated with higher prefrontal executive function (1,3,4) and that higher dopaminergic function and sEBR are correlated with higher prefrontal executive function (5,23,24), respectively. No study, however, has yet to investigate the positive connection between aerobic fitness and sEBR, and sEBR and executive function. Thus, evidence that sEBR significantly mediates the association between aerobic fitness and Stroop interference was first observed in the present study.

Although previous studies tended to focus on changes in neural activation and/or structure as the mediation candidate for the association between aerobic fitness and executive function (1–4), there has been no evidence for the mediation effect of neuronal substrates such as neurotransmitters and neuromodulators. A more recent study examined the relationship between aerobic fitness, working memory, and D2/3 receptor availability using PET in older adults. They reported a positive association between aerobic fitness and D2/3 receptor availability (11). Considering previous reports that sEBR is reflected with D2-like receptor density and dopamine concentration in the striatum (17,22), our sEBR findings are partially in agreement. In addition, animal studies have also reported that dopamine drives the motivation of physical activity (7,8) and that long-term exercise functionally changes the dopamine receptor in the striatum (12) and elevates levels of endogenous dopamine concentration in the PFC (10). In summary, the mediation effects of sEBR on the missing link observed in this study suggest that dopamine, which drives high physical activity and enhances executive function, has an essential neuromodulatory role in linking aerobic fitness and cognition.

The l-DLPFC, one of the dopaminergic circuits (6,13), is an essential brain region for Stroop task performance (32,33). As with our previous methods, which classified activated channels into the lateral PFC subregion according to the MNI space–based anatomical labeling (1,27–31,43), we used fNIRS to measure the activation of the l-DLPFC. As expected, the Stroop task performance positively correlated with task-related l-DLPFC activation, corresponding to previous neuroimaging studies (32). These results indicate that higher task-related cortical activation leads to higher executive function. Interestingly, the NE score calculated based on task performance and task-related l-DLPFC activation was positively correlated with sEBR (Fig. 5). This result suggests that higher sEBR is associated with higher NE of the l-DLPFC (i.e., performance is high, and the corresponding l-DLPFC activation is relatively low) rather than l-DLPFC activation itself. This insight corresponds to the well-accepted idea that higher dopaminergic function leads to NE in the l-DLPFC, which is to say that the signal-to-noise ratio increases (34,35). As DLPFC efficiency is one feature of the dopaminergic system, the current results suggest that sEBR may be working as a dopaminergic function. Indeed, dopamine-related psychiatric disorders, such as ADHD and Parkinson’s disease, have shown inefficient PFC activation compared with healthy controls (35,48). Moreover, a previous study on older adults with high physical fitness, including aerobic fitness, revealed a high executive function and efficient task-related DLPFC activation compared with the low fitness group (49). Thus, it is tempting to conclude that dopamine-associated NE in the l-DLPFC partially mediates the association between aerobic fitness and executive function.

It is possible that the functional connectivity of the l-DLPFC with other brain regions (e.g., striatum and anterior cingulate cortex) with dopaminergic pathways works for leading to higher performance via efficient l-DLPFC activation (13). There are dominant brain dopaminergic systems: the nigrostriatal pathway and the mesolimbic/cortical pathway. These pathways work in conjunction with each other, are integrated as frontostriatal networks, and play a key role in efficient executive processing in the PFC (6,13). Considering rodent studies reporting that exercise training changes the function of both the nigrostriatal and mesolimbic/cortical pathways (10,12), aerobic fitness enhancement might positively affect prefrontal executive performance via a high function of frontostriatal networks consisting of these dopaminergic pathways. To explore this in detail, the functional connectivity between l-DLPFC and other brain regions via the dopaminergic pathways should be investigated using high-resolution functional MRI and PET.

In terms of methodology, we tested our hypothesis on the basis of the premise that sEBR is a noninvasive measure that reflects dopaminergic activity. Although decades of evidence is supportive of an association between the sEBR and the dopaminergic system and that sEBR can be used as a reliable marker (15–19), the specific neural circuits modulating sEBR remain an open question (50,51). Our data that a higher sEBR is linked with higher l-DLPFC efficiency support that sEBR works as a dopaminergic function because l-DLPFC efficiency is a potential feature of dopaminergic regulation (34,35). To date, cortical–basal ganglia and brainstem modulations related to the dopaminergic activation have been considered the main candidates for blink generation systems (18,20,21). Taken together with the previous data that acute exercise might increase sEBR (26), it can be assumed from the present results that fitness enhancement induced by physical activity might affect these blink generation systems via dopaminergic regulation.

The current findings that suggest the importance of the dopaminergic system, affecting motivation, on the fitness–cognition link may open a new avenue for developing an effective exercise regimen. Physical inactivity, which impairs both lifestyle-related diseases and mental health, is now a global issue. Perhaps, physical inactivity and related psychosomatic disorders may be caused by dopaminergic dysfunction (7,8). Although a regular exercise regimen is at the leading solution, exercise conditions focusing on dopaminergic function may better enhance the beneficial effects of exercise and adherence. Already, some evidence is beginning to emerge (e.g., motivational mood while exercising boosts the beneficial effect on executive function with l-DLPFC activation [30]), and our current study supports such future neuroscience research in sports and exercise.


The present study included only male participants because the female cycle phase may influence sEBR and dopaminergic activity (15). In addition, future studies should clarify whether the relationships observed in the healthy young adults in this study can also be observed in other populations, especially those with a higher sEBR than healthy controls, such as schizophrenic patients (18). Finally, the potential involvement of the noradrenergic system as another possible neuromodulatory system cannot be ruled out (52). Thus, further studies examining the interactive involvement of brain monoamine neurotransmitters such as dopamine and noradrenaline may be helpful for a more holistic understanding of the link between aerobic fitness and executive function.


This study provides the first evidence that sEBR mediates the association between aerobic fitness and executive function. In addition, sEBR is positively associated with l-DLPFC efficiency. This supports the hypothesis that brain dopaminergic function with prefrontal NE works to connect a missing link between executive function and aerobic fitness. These findings provide a novel insight into the neuromodulatory mechanisms underlying the influence of aerobic fitness on prefrontal executive function.

This work was supported in part by the Japan Society for the Promotion of Science (JSPS) Grant 16H06405 (H. S.), 18H04081 (H. S.), and 20 J20893 (R. K.) and the Japan Science and Technology Agency (JST) Grant JPMJMI19D5 (H. S.). The authors thank members of the Laboratory of Exercise Biochemistry and Neuroendocrinology for their assistance with data collection. They express their gratitude to M. Noguchi (ELCS English Language Consultation, Japan) for helping with the manuscript.

The authors declare that they have no conflicts of interest. The results of the present study do not constitute endorsement by the American College of Sports Medicine and are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.


1. Hyodo K, Dan I, Kyutoku Y, et al. The association between aerobic fitness and cognitive function in older men mediated by frontal lateralization. Neuroimage. 2016;125:291–300.
2. Themanson JR, Pontifex MB, Hillman CH. Fitness and action monitoring: evidence for improved cognitive flexibility in young adults. Neuroscience. 2008;157(2):319–28.
3. Weinstein AM, Voss MW, Prakash RS, et al. The association between aerobic fitness and executive function is mediated by prefrontal cortex volume. Brain Behav Immun. 2012;26(5):811–9.
4. Ludyga S, Mücke M, Colledge FMA, Pühse U, Gerber M. A combined EEG-fNIRS study investigating mechanisms underlying the association between aerobic fitness and inhibitory control in young adults. Neuroscience. 2019;419:23–33.
5. Vernaleken I, Buchholz HG, Kumakura Y, et al. “Prefrontal” cognitive performance of healthy subjects positively correlates with cerebral FDOPA influx: an exploratory [18F]-fluoro-L- DOPA-PET investigation. Hum Brain Mapp. 2007;28(10):931–9.
6. Cools R. Chemistry of the adaptive mind: lessons from dopamine. Neuron. 2019;104(1):113–31.
7. Zhu X, Ottenheimer D, DiLeone RJ. Activity of D1/2 receptor expressing neurons in the nucleus accumbens regulates running, locomotion, and food intake. Front Behav Neurosci. 2016;10(66):1–10.
8. Garland T, Schutz H, Chappell MA, et al. The biological control of voluntary exercise, spontaneous physical activity and daily energy expenditure in relation to obesity: human and rodent perspectives. J Exp Biol. 2011;214(2):206–29.
9. Robertson CL, Ishibashi K, Chudzynski J, et al. Effect of exercise training on striatal dopamine D2/D3 receptors in methamphetamine users during behavioral treatment. Neuropsychopharmacology. 2016;41(6):1629–36.
10. Chen C, Nakagawa S, Kitaichi Y, et al. The role of medial prefrontal corticosterone and dopamine in the antidepressant-like effect of exercise. Psychoneuroendocrinology. 2016;69:1–9.
11. Jonasson LS, Nyberg L, Axelsson J, Kramer AF, Riklund K, Boraxbekk C-JJ. Higher striatal D2-receptor availability in aerobically fit older adults but non-selective intervention effects after aerobic versus resistance training. Neuroimage. 2019;202:116044.
12. Foley TE, Fleshner M. Neuroplasticity of dopamine circuits after exercise: implications for central fatigue. Neuromolecular Med. 2008;10(2):67–80.
13. Nagano-Saito A, Leyton M, Monchi O, Goldberg YK, He Y, Dagher A. Dopamine depletion impairs frontostriatal functional connectivity during a set-shifting task. J Neurosci. 2008;28(14):3697–706.
14. Köhncke Y, Papenberg G, Jonasson L, et al. Self-rated intensity of habitual physical activities is positively associated with dopamine D2/3 receptor availability and cognition. Neuroimage. 2018;181:605–16.
15. Jongkees BJ, Colzato LS. Spontaneous eye blink rate as predictor of dopamine-related cognitive function—a review. Neurosci Biobehav Rev. 2016;71:58–82.
16. Eckstein MK, Guerra-carrillo B, Miller AT, Bunge SA. Beyond eye gaze: what else can eyetracking reveal about cognition and cognitive development? Dev Cogn Neurosci. 2017;25:69–91.
17. Groman SM, James AS, Seu E, et al. In the blink of an eye: relating positive-feedback sensitivity to striatal dopamine D2-like receptors through blink rate. J Neurosci. 2014;34(43):14443–54.
18. Karson CN. Spontaneous eye-blink rates and dopaminergic systems. Brain. 1983;106(3):643–53.
19. Cavanagh JF, Masters SE, Bath K, Frank MJ. Conflict acts as an implicit cost in reinforcement learning. Nat Commun. 2014;5(5394):1–10.
20. Bologna M, Fasano A, Modugno N, Fabbrini G, Berardelli A. Effects of subthalamic nucleus deep brain stimulation and L-dopa on blinking in Parkinson’s disease. Exp Neurol. 2012;235(1):265–72.
21. Kaminer J, Powers AS, Horn KG, Hui C, Evinger C. Characterizing the spontaneous blink generator: an animal model. J Neurosci. 2011;31(31):11256–67.
22. Boileau I, Dagher A, Leyton M, et al. Modeling sensitization to stimulants in humans: an [11C]raclopride/positron emission tomography study in healthy men. Arch Gen Psychiatry. 2006;63:1386–95.
23. Zhang T, Mou D, Wang C, et al. Dopamine and executive function: increased spontaneous eye blink rates correlate with better set-shifting and inhibition, but poorer updating. Int J Psychophysiol. 2015;96(3):155–61.
24. Korponay C, Dentico D, Kral T, et al. Neurobiological correlates of impulsivity in healthy adults: lower prefrontal gray matter volume and spontaneous eye-blink rate but greater resting-state functional connectivity in basal ganglia-thalamo-cortical circuitry. Neuroimage. 2017;157:288–96.
25. Tharp IJ, Pickering AD. Individual differences in cognitive-flexibility: the influence of spontaneous eyeblink rate, trait psychoticism and working memory on attentional set-shifting. Brain Cogn. 2011;75(2):119–25.
26. Tantillo M, Kesick CM, Hynd GW, et al. The effects of exercise on children with attention-deficit hyperactivity disorder. Med Sci Sport Exerc. 2000;34(2):203–12.
27. Byun K, Hyodo K, Suwabe K, et al. Positive effect of acute mild exercise on executive function via arousal-related prefrontal activations: an fNIRS study. Neuroimage. 2014;98:336–45.
28. Yanagisawa H, Dan I, Tsuzuki D, et al. Acute moderate exercise elicits increased dorsolateral prefrontal activation and improves cognitive performance with Stroop test. Neuroimage. 2010;50(4):1702–10.
29. Kujach S, Byun K, Hyodo K, et al. A transferable high-intensity intermittent exercise improves executive performance in association with dorsolateral prefrontal activation in young adults. Neuroimage. 2018;169:117–25.
30. Suwabe K, Hyodo K, Fukuie T, et al. Positive mood while exercising influences beneficial effects of exercise with music on prefrontal executive function: a functional NIRS study. Neuroscience. 2021;454:61–71.
31. Ochi G, Yamada Y, Hyodo K, et al. Neural basis for reduced executive performance with hypoxic exercise. Neuroimage. 2018;171:75–83.
32. Macdonald AW. Dissociating the role of the dorsolateral prefrontal and anterior cingulate cortex in cognitive control. Science. 2000;288(9):1835–7.
33. Frings C, Brinkmann T, Friehs MA, Lipzig T. Single session tDCS over the left DLPFC disrupts interference processing. Brain Cogn. 2018;120:1–7.
34. Nyberg L, Andersson M, Kauppi K, et al. Age-related and genetic modulation of frontal cortex efficiency. J Cogn Neurosci. 2013;1–10.
35. Cools R, Stefanova E, Barker RA, Robbins TW, Owen AM. Dopaminergic modulation of high-level cognition in Parkinson’s disease: the role of the prefrontal cortex revealed by PET. Brain. 2002;125(3):584–94.
36. Howley ET, Bassett DR, Welch H. Criteria for maximal oxygen uptake: review and commentary. Med Sci Sports Exerc. 1995;27(9):1292–301.
37. Midgley AW, McNaughton LR, Polman R, Marchant D. Criteria for determination of maximal oxygen uptake: a brief critique and recommendations for future research. Sport Med. 2007;37(12):1019–28.
38. Kim K, Bohnen NI, Müller MLTM, Lustig C. Compensatory dopaminergic-cholinergic interactions in conflict processing: evidence from patients with Parkinson’s disease. Neuroimage. 2019;190:94–106.
39. Muhle-Karbe PS, Jiang J, Egner T. Causal evidence for learning-dependent frontal lobe contributions to cognitive control. J Neurosci. 2018;38(4):962–73.
40. Tsuzuki D, Jurcak V, Singh AK, Okamoto M, Watanabe E, Dan I. Virtual spatial registration of stand-alone fNIRS data to MNI space. Neuroimage. 2007;34(4):1506–18.
41. Okamoto M, Dan I. Automated cortical projection of head-surface locations for transcranial functional brain mapping. Neuroimage. 2005;26(26):18–28.
42. Singh AK, Okamoto M, Dan H, Jurcak V, Dan I. Spatial registration of multichannel multi-subject fNIRS data to MNI space without MRI. Neuroimage. 2005;27(4):842–51.
43. Shattuck DW, Mirza M, Adisetiyo V, et al. Construction of a 3D probabilistic atlas of human cortical structures. Neuroimage. 2008;39(3):1064–80.
44. Curtin A, Ayaz H. Chapter 22 - Neural efficiency metrics in neuroergonomics: theory and applications. In: Ayaz H, Dehaisn F, editors. Neuroergonomics: The Brain at Work and in Everyday Life.   Academic Press; 2019. pp. 133–40.
45. Curtin A, Ayaz H, Tang Y, Sun J, Wang J, Tong S. Enhancing neural efficiency of cognitive processing speed via training and neurostimulation: an fNIRS and TMS study. Neuroimage. 2019;198:73–82.
46. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research. Conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51(6):1173–82.
47. Green SB. How many subjects does it take to do a regression analysis. Multivariate Behav Res. 1991;26(3):499–510.
48. Bédard ACV, Newcorn JH, Clerkin SM, et al. Reduced prefrontal efficiency for visuospatial working memory in attention-deficit/hyperactivity disorder. J Am Acad Child Adolesc Psychiatry. 2014;53(9):1020–30.
49. Voelcker-Rehage C, Godde B, Staudinger UM. Physical and motor fitness are both related to cognition in old age. Eur J Neurosci. 2010;31(1):167–76.
50. Dang LC, Samanez-Larkin GR, Castrellon JJ, et al. Spontaneous eye blink rate (EBR) is uncorrelated with dopamine D2 receptor availability and unmodulated by dopamine agonism in healthy adults. Eneuro. 2017;4(5):e0211–7.
51. Sescousse G, Ligneul R, Van Holst RJ, et al. Spontaneous eye blink rate and dopamine synthesis capacity: preliminary evidence for an absence of positive correlation. Eur J Neurosci. 2018;47:1081–6.
52. Arnsten AFT. Catecholamine modulation of prefrontal cortical cognitive function. Trends Cogn Sci. 1998;2(11):436–47.


Supplemental Digital Content

Copyright © 2020 by the American College of Sports Medicine